ML20205J773

Coordination of Safety Research for the Babcock and Wilcox Integral System Test Program

ML20205J773
Person / Time
Issue date:	03/31/1987
From:	Sursock J, Matt Young NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
To:
References
NUREG-1163, NUDOCS 8704010410
Download: ML20205J773 (233)

Text

_ _ - . . -- - _ -

! NUREG-1163 1

i l Coordination of Safety Research

! for the Babcock and Wilcox

! Integral System Test Program 5

i i

l U.S. Nuclear Regulatory

! Commission r

M. W. Young J. P. Sursock i

pa "8 N *

%q{*

,,t .

2, i

f I

i

! E?M8,ing07oasi  !

j lib 3 R pyg i

1.-.--_..-_-.. . - - , - . . . _ - _ _ _ . _ _ _ _ _ . _ _ _ . - , . . , . - - - - - . . . _ - - . . _ . - _ , . _ _ . . , . _ _ . . . _ . . . _ . . _ . , - - , . _ - - - _ _ _ - - - _ -

NOTICE Availability of Reference Materials Cited in NRC Publications Most documents cited in NRC publications will be available from one of the following sources:

1. The NRC Public Document Ronm,1717 H Street, N.W.

Washington, DC 20555

2. The Superintendent of Documents, U.S. Government Printing Office, Post Office Box 37082, Washington, DC 20013-7082

3. The National Technical Information Service, Springfield, VA 22161 Although the 'isting that follows represents the majority of documents cited in NRC publications, it is not intended to be exhaustive.

Referenced documents available for inspection and copying for a fee from the NRC Public Docu-ment Room include NRC correspondence and internal NRC memoranda; NRC Office of Inspection and Enforcement bulletins, circulars, information notices, inspection and investigation notices; Licensee Event Reports; vendor reports and correspondence; Commission papers; and applicant and licensee documents and correspondence.

The following documents in the NUREG series are available for purchase from the GPO Sales Program: formal NRC staff and contractor reports, NRC-sponsored conference proceedings, and NRC booklets and brochures. Also available are Regulatory Guides, NRC regulations in the Code of Federal Regulations, ar.d Nuclear Regulatory Commission issuances.

Documents available from the National Technical Information Service include NUREG series reports and technical reports prepared by other federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to the Nuclear Regulatory Commission.

Documents available from public and special technical libraries include all open literature items, such as books, journal and periodical articles, and transactions. Federal Register notices, federal and state legislation, and congressional reports can usually be obtained from these libraries.

Documents such as theses, dissertations, foreign reports and translations, and non-NRC conference proceedings are available for purchase from the organization sponsoring the publication cited.

Single copies of NRC draft reports are available free, to the extent of supply, upon written request to the Division of Technical information and Document Control, U.S. Nuclear Regulatory Com-mission, Washington, DC 20555.

Copies of industry codes and standards used in a substantive manner in the NRC regulatory process j are maintained at the NRC Library, 7920 Norfolk Avenue, Bethesda, Maryland, and are available i there for reference use by the public. Codes and standards are usually copyrighted and may be purchased from the originating organization or, if they are American National Standards, from the American National Standards Institute,1430 Broadway, New York, NY 10018.  :

1 I

i

NUREG-1163 EPRI NP-4353-SR R2 Coordination of Safety Research l for the Babcock and Wilcox Integral System Test Program Manuscript Completed: January 1987 Date Published: March 1987 M. W. Young, U.S. Nuclear Regulatory Commission J. P. Sursock, E'octric Power Research Institute Division of Reactor System Safety Offico of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Wcshington, D.C. 20555

...y, 7

l l

l l

'b ABSTRACT 1

1 j This report describes the MIST facility and all the IST support projects sponsored i by the USNRC and by EPRI. These support projects have been deemed to play an j essential role in helping resolve issues raised by MIST scaling compromises. Each i

j support project is described in detail and application of the expected data to I

resolution of issues is discussed.

The combined effort of MIST and seven other support projects will resolve

! virtually all questions addressed by the IST program.

I i

1 I

l 4

I

)

4 i

1 f

)

i i

y d

l i i

j l

1 iff 4

4 l

1

TABLE OF CONTENTS Section Page 1 STATEMENT OF PURPOSE 1-1 1.1 Background 1-1 1.2 Purpose of Report 1-4 2 REPORT STRUCTURE 2-1 3 INTEGRAL TEST FACILITIES 3-1 3.1 MIST Program (Undefined Tests) 3-3 3.1.1 Facility Description 3-3 3.1.2 Scaling Approach 3-25 3.1.3 Scaling Compromises Investigated 3-34 3.1.4 Test Procedure: General Approach and Techniques 3-34 3.1.5 Appilcation of Data 3-36 3.2 EPRI Integral Test facility at SRI International (SRI-2) 3-46 3.2.1 Facility Description 3-46 3.2.2 Scaling Approach fnr the Facility 3-58 3.2.3 Scaling Compromises Investigated by SRI-2 3-61 3.2.4 Test Procedure 3-62 3.2.5 Application of Data 3-62 3.3 University of Maryland at College Park (UMCP) 3-78 3.3.1 Facility Description 3-78 3.3.2 Scaling Approach 3-89

' 3.3.3 Scaling Compromises Investiqated by BiCP 3-103 3.3.4 Test Procedure 3-106 3.3.5 Application of Data 3-118 I

Y

. . _ - - . - _ - - _ . _ - .. . - _ = - . _ . - . . . - - ._ . . . . . - . _ -

j i

Section Pa.ge i

4 SEPARATE EFFECTS TEST FACILITIES 4-1 j

4.1 Argonne National Laboratory--Flow Regimes 4-1

]

4.1.1 Facility Description 4-1 l l 4.1.2 Scaling Approach 4-6 4.1.3 List of Issues Addressed by the Project 4-18

! 4.1.4 Test Procedure 4-10 4.1.5 Application of Data 4-19

! 4.2 SAIC--Large Pipe Flow Regimes 4-22 i

4.2.1 Facility Description 4-23 l 4.2.2 Scaling Approach 4-27 j 4.2.3 Scaling Compromises Investigated 4-28 r

4.2.4 Test Procedure 4-28 i 4.2.5 Application of Data 4-29  ;

4.3 SAIC--MIST Auxiliary feed Study 4 41 j 4.3.1 Facility Description 4 41 4.3.2 Scaling Approach 4 44 4.3.3 Issues Addressed by the Project 4-48

])! 4.3.4 Test Procedure 4-48 4.3.5 Application of Data 4-48 j 5 ANALYTICAL MODELS AND CODES 5-1

! 5.1 TRAC-PF1/M001 COMPUTER CODE 5-1  :

5.1.1 TRAC Description 5-1 1

5.1.2 TRAC Scalin9 5-4 i 5.1.3 Scaling Compromises investigated 5-5 i 5.1.4 Analysis Procedure 5-6 j 5.1.5 Application of Code Results 5-7 I

5.2 TETRATECil/EPRI Two-Phase Flow Puup Model 5-7

)i i 5.2.1 Tetra Tech /EPRI Analytical Two-Phase Pump Model S-8 Description  ;

5.2.2 Aoplication of Model 5 14 ,

)

i 6 SU'4:iARY AND CONCLUSION g-j

RECOMMENDATIONS 7 REFERENCES 7-1 i

(

i APPENDIX A A-1

) 1 i APPENDIX B 0-1 1

i l

4 i

vi u . _ _ _ _ . _ _ - _ _ _ _ _ _ . _ _ _

LIST OF FIGURES Figure g 3-1 Reactor Coolant System Arrangement 3-4 3-2 Reactor Vessel Arrangement 3-7 3-3 RV Upper Plenum Vertical Section 3-9 3-4 Top Plenum Cover Plate 3-10 3-5 Upper Plenum Cylinder 3-11 3-6 MIST Reactor Vessel Concept 3-12 3-7 Upper and Lower Downcomer 3-14 3-8 MIST RVVV Circuit 3-16 3-9 19-Tube Once Through Steam Generator 3-17 3-10 Comparison of 19-Tube and Prototype OTSG Tube Support Plates 3-18 3-11 OTSG Tube Support Plate Locations 3-19 3-12 SRI-2 Reactor Coolant System Arrangement 3-47 3-13 SRI-2 Reactor Vessel Details 3-4E 3-14 SRI-2 Vent Valve Design 3-50 3-15 SRI-2 Upper Vessel Components 3-51 3-16 SRI-2 Reactor Vessel Baffle Plates 3-52 3-17 SRI-2 Reactor Vessel Upper Shell 3-53 3-18 SRI-2 Reactor Vessel Lower Assembly 3-54 3-19 SRI-2 Steam Generator Plenums 3-55 3-20 SRI-2 Loop Instrumentation 3-57 3-21 Flow Regime Transition 3-66 3-22 Predicted Flow Regimes in Vertical Hot Leg for SRI-2 and Plant 3-69 3-23 Transfer Function Diagrams 3-72 3-24 UMCP Reactor Coolant System Arrangement 3-79 3-25 UMCP Reactor Vessel and Downcomer 3-80 ,

1 3-26 UMCP Reactor Vessel Vent Valve 3-81 i 3-27 Design Details of 34CP Candy Cane View Port 3-83 3-28 Design Details of WiCP Hot Leg View Port 3-84 3-29 Design Details of WiCP Cold Leg View Port 3-85 3-30 Spacing of Tubes in UMCP Steam Generator 3-87 3-31 UMCP Loop Instrumentation Location 3-90 3-32 Predicted Plant Pressurizer Water Level 3-108 vii i

l Figure pg 3-33 Integrated Mass Flow (Break Flow & HPI Flow) 3-109 3-34 Predicted Pressure of Vessel as Function of Time 3-110 3-35 Normalized Pressure as Function of Normalized Liquid Inventory 3-119 4-1 ANL Basic HLUB Loop Design 4-2 i 4-2 ANL Facility Straight inlet Section 4-4

,i

4-3 ANL Facility Partial Reactor Vessel and Hot Leg Horizontal 4-5 i Section Simulation 4-4 Typical Loop With Once-Through Steam Generator 4-16 i

4-5 EPRI/SAIC Test Facility (4-inch pipe) 4-24 4-6 EPRI/SAIC Test Facility (12-inch pipe) 4-25 4-7 Flow Pattern Map for Vertical Flow 4-31

) 4-8 Flow Pattern Map for Vertical Tubes 4-32 4-9 Flow Pattern Map for Vertical Tubes 5.0 cm dia., Air-Water 4-33 l

l 4-10 Predicted Bubbly to Slug Flow Transition Lines for Plant 4-37 vs. Experiment 4-11 Predicted Churn / Slug to Annular Flow Regime Transition 4-38 Lines for Plant vs. Experiment j 4-12 SAIC Auxiliary Feedwater Test Section Schematic 4-42 i 4-13 SAIC Auxiliary Feedwater Conceptual Loop Design 4-43 i 4-14 Flooding Prediction at the Tube Support Plate 4-45 5-1 One-Dimensional Control Volume Method for Rotating Machines 5-9 5-2 Comparison of Present Theory for Homologous Heads with CREARE 5-13 4

and B&W Data for v/aN between 0.8 and 1.2 1 5-3 Comparison of the Head between the Theory and Experimental 5-15 j Data for a Mixed-Flow Pump i 5-4 Comparison of the Head between the Theory and Experimental 5-16 Data for a Radial-Flow Pump

)

i I

l l

f i

i viii l

1 1

. _ . . . _ - . _ _ _ _ _ . ~ - . . . _ _ _ _ _ _ - . _ _ . - - . _ _ . . _ . _ . _ . ~ _ - _ . . - . -.__-_,_- . _ , . _ - .

LIST OF TABLES i

i 8

Table Page i 1-1 Support Projects Coordination Matrix to Address 1-3 l

Issues Raised by MIST Scaling Compromises

, 3-1 MIST Atypicalities 3-2 3-2 Post-58LOCA Events 3-25 3-3 MIST Documents on Facility Design and Scale 3-32 1 3-4 Potential Use of MIST to Address MIST Atypicalities 3-35 3-5 MIST Test Matrix: Test Group 30, Mapping 3-38

]

3-6 Scaling Criteria for SRI-2 3-60 l 3-7 SRI-2 Test Procedures 3-63 j 3-8 UMCP Instrumentation Location 3-91 3-9 Dominant Dimensionless Groups for Each Scaling Compromise (UMCP) 3-104 4-1 Summary of Scaling Analysis for SAIC Aux Feed Tests 4-47 4

i j

l 1

l 3

i ix

)

n . - . . , -- - - - . - , , , , . w .- ---. - - - - - - - - - - - - - - - , , - , -,,,,--,--,e .w . -.--en , - - - - - - - - , - , . . - > - - - - - - , - - - ,

i j

4 ACKNOWLEDGMENTS f Several authors have participated in the preparation of this report:

l Jim Gloudemans for MIST (Sec. 3.1) i J-P. Sursock and Bob Kiang for SRI-2 (Sec. 3.2)

Y. Y. Hsu and Gary Pertmer for UNCP (Sec. 3.3)

M. Ishii for ANL facility (Sec. 4.1) i Ab Hashemi for SAI facilities (Sec. 4.2 & 4.3)

) Thad Knight for TRAC Code (Sec. 5.1)

J.H. Kim for TETRATECH/EPRI i

Their contributions are greatly appreciated.  !

i i

J l

i i t i

1 f

4 i

i i

j t 1

i t

i i

X 1

i NOMENCLATURE--ANL/ SRI-2 A Nondimensional area a Flow area a

s Wall cross sectional area B Biot number f C

p Fluid heat capacity C pg Solid heat capacity d Hydraulic diameter F Friction number f Friction factor g Gravity h Heat transfer coefficient j j Total volumetric flux kg Thermal conductivity of solid K Orifice coefficient t Axial length t

e Equivalent length for distributed losses t

h Length of hot fluid section L Hondimensional length N Phase change number pch N

sub Subcooling number N

Fr Froude number N

d Orift flux number Ng Frictionnumber(two-phase)

N o

Orificenumber(two-phase) p Pressure q Heat generation density in solid Q3 Heat source number

! R Richardson number St Stanton number t Time i fluid temperature Tg Solid temperature T Saturation temperature at T Characteristic time ratio i

u Velocity (liquid)

U Nondimensional velocity l

xi i

i

Vjg Drift velocity V Volume x Quality Greek Symbols a Vold fraction as Solid thermal diffusivity a Thermal expansion coefficient a Conduction thickness aH gg Latent heat aH sub Subcooling op Pressure drop

., ao Density difference ou Viscosity difference u Viscosity of liquid t Wetted (heated) perimeter o Censity of liquid og Density of solid Subscripts 1 ith section o Referenceconstant(heatedsection) s Solid h Hot c Cold R Model to prototype ratio g Vapor

()m Model

()p Prototype i

I a

Nih

NOMENCLATURE.-SAIC l

D Tube diameter 9 Gravitational acceleration Entranceregionlength(lengthofchurnflowregion)

IE Ug3 Gas superficial velocity UL3 Liquid superficial velocity Ug Meanvelocity(Ut3 + Ugs) a Void fraction  !

! og Gas density oL Liquid density vt Liquid kinematic viscosity o Surface tension f

I r

i I i 1

xiil 1

i NOMENCLATURE--MARYLAND l

A Area c Specific heat e Internal energy f Friction flow loss F force g Gravitational acceleration h Height or length, convective heat transfer coefficient k Entrance and minor flow loss or thermal conductivity K* As defined by equation (3-55) t Effective thermal height L Length of component m Mass m Mass flow rate M Total Mass p Pressure it Heat transfer rate Q Volume flow rate t Time i Temperature or time period u Velocity v Specific volume V Velocity Potential head or elevation a Volumetric expansion coefficient y Specific Weight u Viscosliy o Density t ilme constant t xlv

Section 1 STATEMENT OF PURPOSE l

l 5

\

?i e

~

j /;

, ,< 1

[

t 1.1 BACXcA000 . -

~

The Integral Systiml Test (IST) program wss initiated in JJIy 1983 as a rtsult.of [

extensive

• discussions among the NRC, LPkt, the Babcock & kMicou Ownersi Group - -  !

(BWOG), and Babcock & Wticox. The major v!ements pf this program are: l l

The OT!S program (8W raised loop scaled factitty) e ,  ;

,; .F 1 e- The MIST program (8W lower loop scaled fact 1'Ity) ..-

e The.8WOG code assessment pregram f

e Support projects sponsored by EPRI or NRC.

The MIST program is a key componer.t of this effort, and by far the most expen-

f sive. its objective is to design, build, and operate a scaled version cf a ,

f typical BW reactor coolant system design at Fototypical pressures and temper. j atures. The purpose of this work is to generat$ high quality emperimental data to ' l be used later for assessing tnerral. hydraulic safety computer codes. ihe Sr.ecific' ,t issues addressed in the program are identified 'and discussed in the' f taal Test

~

i AdvisoryGroup(TAC) report (1). Briefly, they address! ..

f

'l e Decay heat removal by single- and two r#ase natuul circulation <

l e Consequences of interruptions of eatun;1'ctreulation. ,

e Decay heat removal by boiler.cond m ser ude , j

',i e 1.ongtermcoolinjb.v'<wansofnaturalcirculat'.one,rotherwta.e(e.g.,

f FeedandBleed) ,

ConscencrJ of steam generate- tube ru,$ture e ,

, i

/

l e Consequences of loop-to-loop osc111ations,

)

The MIST scaling and design approachtas are based on the so called " volume scaling" concept: The elevations of major corptinents and of flow paths are duplicated, the volume ratios and power densttlet are preserved, tne'6ydraulle resistances are ,

simulated,andthe;bermodynamicistdtevarlablesareprototypical. A suhl fra tor 6

11 '

y A t

- - _- - - h . -

of 1/017 was cdopted for MIST. A detailed description of the MIST final design and the sca)ing rationale are provided in the MIST facility specification (eport(2).

\

As for any complex thermal-hydraulic system, ideal scaling--if at all possible--is well beyond the reach of available resources and time constraints. Thus, design

compromises had to be introduced (e.g., atypical downcomer and reactor vessel vent valves). These design Lorrpromises are a potential source of distortion of some physical phenomena (e.g., variation of flow regimes in the hot legs) which in turn l

could lead to atypical transient behavior (e.g., premature interruption of natural circulation). It soon became clear that a good assessment of the impact of these

' design compromices on transient behavior is essential. It also became apparent that weU -coordinated support projects, sponsored separately by NRC and EPRI, l could be brought to bear heavily on these questions.

I Two " facility coordination" meetings were held. The objectives of these meetings 4

were to establish:

a. What are the relevant support projects?

b. What issues / compromises can they address, and what is the optimal allocation of available resources?

c. How can the results of these projects (data or calculations) be used to resolve the questions associated with the MIST design compromises?

These tectings were very successful in achieving a high level of coordination.

ihe essence of these meetings is summarized in.lable 1-1. The MIST design compromises and their expected impact on key issues are listed in the first column. The support projects are listed across the table. The " matrix" indicates what tacility is best suited to study the effects of a particular compromise /

. distortion. A cursory look indicates that most issues are covered by the support projepts. The cross-comparison of results from the different projects is clearly expected to be a most effective and productive way to resolve the remaining isnes.

l 1-2

Table 1-1 SUPPORT PROJECT COORDINATION MATRIX TO ADDRESS ISSUES RAISED BY MIST SCALING COMPROMISES SUPPORT PROJECTS 1 2 3 4 5 6 7 8 9 10 o>  %

.!! - & .2

$ 0 .  ? o MIST SCALING COOR0MISES a, I r E 7 ^ $ 3: $ e 5 5 Af fecte<f issuefs) 5 C k - b d S a - - M M be W 54 5 s E $ $ E E Ae E fe A. (5) nc Flow 4 Density Fields +

(6) DC Tangantial Resistance +

(7) RVVV Simulation

3. Re-establish Natural Circulation X X

4. Long. Term Cooling X X

5. Interloop Interactions X(b) X(b) X(b) X(b) X(b)

P. (1) Hot leg Separation +

(12) Hot leg Flow Regime

1. Interruption of Natural Circulation X X X X X

3. Re-establish Natural Circulatios X X X X W.

C. (3) AFW Multidinensionality

?. Establish RCM 0 0 X 0

6. Combined Pri/Sec Blowdown A SGTR a 0; X 0(c)

O. (4) TCP 26 Characteristics

3. Re-establish Natural Circulation:

Degradation X Cavitation E. (2) Si Metal Mass

5. Interloop Interactions X F. (8) Piping Metal Mass

3. Re-establish Natural Circulation X co@ined X

4. Long. Term Cooling X combi ned X G. (10) Secondary Side

6. Combined Blowdown 4 SGTR , O combi ned 0 H. (11) One-Phase. 2-T Stratified Flow

3. Re-establish Natural Circilation A 0,

5. Interloop Interactions A 0 Legend: a = Change required; X = Applicable; O = Potential for resolving issues exists. Additional technical detall is needed to complete assessment.

(s) Need interf acility test to Itak results free three f acilities.

(b) JPS and TRAC Stability Analysis to link data from various facilities and prototype.

(ci Dartmouth Analysis -(sponsored by EPRI).

~.

1-3

Two letters issued jointly by NRC and EPRI were transmitted to the IST program participants (reproduced in Appendix A). The first letter requests the partici-pants (i.e., the project managers of the support facilities and analytical programs) to provide a document describing their work and how it can be applied to address the issues raised by MIST scaling compromises. It was also decided that a companion to this report would be issued to address, in greater detail, each integral facility scaling philosophy, technique, and evaluation and recommend a common basis for comparing these facility results. The second letter requested additional information from the integral test facility projects.

1.2 PURPOSE OF REPORT The purposes of this report are:

a. To describe the various support projects participating in this concerted effort.

b. To discuss how the results of individual projects can be applied to the resolution of the licensing issues as identified in the TAG report (1).

c. To demonstrate the powerful synergetic effects of a complex coordination effort between NRC and EPRI sponsored projects.

1-4

Section 2 REPORT STRUCTURE

The report is divided into three main chapters. The first chapter deals with integral test facilities (i.e., MIST, SRI-2, and UMCP). The second chapter describes the Separate Effect Test facilities (i.e., SAI and ANL flow regime and flow separation studies in the hot leg; SAI aux. feed. distribution study). The third chapter covers the contribution of codes (TRAC) and analytical models (e.g.,

I l

two-phase flow pump).

)

The chapters are divided into sections. Each section addresses an individual project and its contributions. The section is, in turn, divided into three or five subsections:

o Project description e Scaling approach (for experimental facilities) e Compromises / distortions investigated by the project e Test procedures (for experimental facilities) e Application of data / calculations.

Finally, a chapter entitled " Summary and Conclusions" presents the " integrated" viewpoint and the benefits derived from this coordination effort.

l i

2-1

~ - - - - - - - -

_______________l I

i l

i l

Section 3 INTEGRAL TEST FACILITIES I

i

INTRODUCTION The overall integral system test (IST) program includes three experimental inte-gral facilities which are sponsored under three separate contract entities. The Babcock & Wilcox (B&W) is under contract to construct and operate the Multiloop Integral System Test facility (MIST) which is a part of a jointly funded program supported by the United States Nuclear Regulatory Commission (NRC), B&W, the B&W Owners Group (B&WOG), and the Electric Power Research Institute (EPRI). The NRC is separately sponsoring the University of Maryland to build a reduced scale, lower budget integral facility relative to MIST. EPRI is providing sole sponsor-ship for SRI International to construct an integral facility which also is reduced scale yet unique and operates on a more modest budget.

This section will focus only on the three integral facilities and how these programs can address the MIST scaling compromises as identified in Table 3-1.

l l

l 3-1 l

l l ,,_ __ . _ _ _

Table 3-1 MIST ATYPICALITIES

1. Hot Leg Separation

2. S. G. Metal Mass

3. AFW Multidimensional i 4. RCP Two-Phase Characteristics

5. Downcomer Flow & Density Field

6. Downcomer Tangential Resistance

7. RVVV Simulation

8. Piping Metal Mass

9. Low Pressure Injection

10. Secondary Side

11. One-Phase, 2-T Strat. Flow

12. Hot Leg Flow Regime l

1

)

1 1

3-2

l 3.1 MIST PROGRAM (UNDEFINED TESTS) 3.1.1 Facility Description The MIST facility is an integral 2x4 test loop which consists of a reactor vessel, downcomer, four reactor vessel vent valves (RVVV), two hot legs, two 19-tube once-through steam generators (OTSGs), four cold legs, and four reactor coolant pumps.

Figure 3-1 illustrates the overall arrangement of the MIST reactor coolant system and identifies the diameters and schedules of the pipe sections and components.

The reactor coolant system is approximately 75' tall and 11' across (including the pressurizer). As indicated on Figure 3-1, the downcomer is external to the reactor vessel and centrally located in the MIST configuration. The four cold legs (two frcm each loop) enter the downcomer around its circumference at 90' spacings. The reactor coolant pumps are located at the highest cold leg eleva-tion, one per cold leg. T.e two 19-tube OTSGs are full-length subsections of their plant counterparts and are located to preserve similarity in flow path and length for each loop. The two hot leg takeoffs from the reactor vessel are spaced and angled to maintain equal flow lengths. The horizontal length of the hot leg takeoffs has been minimized to preserve flow typicality. The pipe length from the reactor vessel nozzle to the hot leg U-bend is approximately 45'. A pipe section is included between the hot leg U-bend and the upper tube sheet of each OTSG to simulate the prototypical ekvation and volume occupied by the steam generator upper primary plenum. Details of each loop component are described below.

HOT LEGS The two hot leg pipes will be identical except for the surge line connection on one of the hot leg pipes at an elevation of 27.25'. The hot legs will be fabri-cated from sections of 2-1/2" Schedule 80 stainless steel and Inconel 600 pipe.

Inconel 600 will be used upstream of the hot leg U-bend and 304 stainless steel will be used downstream. Both hot legs will be guard heated, using four separate control zones, to minimize piping heat losses.

The hot legs originate at the reactor vessel nozzle (elevation = 21.25'), extend  ;

0.40' horizontally, and then turn upward. The radius of curvature of the upward '

bend is 1.61'. The hot legs extend vertically upward to an elevation of 64.95' where the HLUBs start. The hot leg U-bend radius of curvature is 1.61', and the spillover vlevation is 66.46'. Downstream of the HLUBs, at an elevation of 53',

3" x 5" dif fusers make the transition from the hot legs to the steam generator upper p %nums. In addition to the penetration required for instrumentation, a  !

1 3-3

p .R- ELEV. 66*-5-1/2" ELEV. 52*-1-3/8" LOWER (SECONDARY) SIDE- - -

OF UPPER TUBE SHEET

! I i i O

e HOT LEG g 2-1/2" SCH 80 SECTION "A-A" LOOKING DOWN AT ELEV. 28'-0" ELEV. 24'-2" RVV VALVE ELEV. 27*-3" SPILLOVER SURGE UNE

[ TO HOT EG ELEV 25*-9" fl f 8

'! Il OVER Lp j , PRESSURIZER ELEV. 21*-3" u n - ELEV.17*-11-7/16"

, GB \

STEAM GENERATOR, _! ,

\ __ ELEV.15*-4-3/8" N PRESSURIZER REACTOR VESSEL H 80 6" SCH 120% '

ELEV. 4'-813/16" l li COLD LEG SOTTOM ACTIVE FUEL

,M A"2" SCH 80 ELEV.0*-0"

\ l UPPER (SECONDARY) SIDE OF LOWER TUBESHEET i ..

N _

1 Ir  ! IL .

ELEV. -O'-5-11/16" ]~ ll kt

.It i f:

e l- -

ELEV. -7*-41/16" Figure 3-1. Reactor Coolant System Arrangement 3-4

high point vent connection is provided at the apex of each HLUB. Viewports are provided, at approximately the midelevation and at the top of the HLUB, for obser-vation of fluid conditions.

COLD LEGS Four identical cold legs will be used in the MIST facility. The cold legs will be fabricated using 2" Schedule 80 stainless steel pipe. In one of the four cold legs, leak sites will be installed upstream and downstream of the reactor coolant pump site to provide simulated cold leg suction and cold leg discharge breaks.

Each cold leg will be guard heated, using three control zones, to minimize heat loss.

4 Two cold legs will originate at the outlet of each steam generator (-6.77') and drop vertically to the spill under elevation (-7.34'). The cold legs run hori-zontally for 2.42' and then turn upward. The cold legs extend vertically upward to the reactor coolant pump interface (24.75'). The reactor coolant pumps will be installed at the apex of the cold leg pipes so that the spillover elevation of 25.75' is maintained. Downstream of the coolant pumps, the cold legs are sloped (58.6') downward to the downcomer intersection (21.25').

An HPI injection nozzle will be included in each cold leg pipe downstream of the reactor coolant pump at an elevation of 23.02'. In one cold leg, a connection will be provided at 24.75', downstream of the reactor coolant pump, for the pressurizer spray line.

REACTOR VESSEL The reactor vessel will be a cylindrical stainless steel pressure vessel which contains a simulated core of 45 fuel rod size electric immersion heaters and four guide tubes. The mechanical design and arrangement of the reactor vessel must:

e Comply with applicable ASME boiler and pressure vessel codes.

e Provide a minimum leakage seal system for the heater / lower head pressure boundary.

e Provide a square subchannel for the core with a minimum of excess water volume and metal mass.

3-5

l The reactor vessel shell will be fabricated from 6" Schedule 120, 304 stainless steel pipe, as shown in Figure 3-2. The lower vessel flange will be a special design to reduce the diameter and thickness (and therefore metal mass). All open-ings in the vessel will be reinforced in compliance with the applicable ASME code. The applicable codes are:

1. ASME Section VIII, Division 1 "Unfired Pressure Vessels"

2. ASME Section I, " Power Boilers" l

The core simulation in the reactor vessel will consist of 45 electrically heated l filament-type heater rods. Each electrically heated rod will be geometrically similar to a PWR fuel rod and will be powered to simulate the fuel rod surface heat flux during small break loss-of-coolant accident (SBLOCA) transients.

< Each heater rod will have an outside diameter of 0.430 inches and will be 18 feet long. This length consists of 144 inches of heated length beginning at one end, and 72 inches of unheated length at the opposite end.

l l

Each of the heater rods will be capable of operating at 70 kW, which corresponds to a 100% power simulation. However, for the MIST program, each heater rod will be operated over a range of 0 to 7 kW which corresponds to O to 10% power simula-tion. At 7 kW operation, the average surface heat flux is 36.1 watts /in2. An axial cosine profile will be obtained over the full control range of 0 to 7 kW.

The model core will use full-length heater rods of prototypical geometry (0.430" rod outside diameter, 0.568" pitch between rods, square array). The number of mooel heater rods is the number of plant fuel rods divided by the model power scale factor:

N = 177x208 = 45 817 A 7x7 model array with four unheated rods (simulating control rod guide tubes) is selected; the four unheated rods matched the number in the plant again scaled as  ;

power: 1 l

177x17

= 3.7

" unheated

• 817 i

1 I

3-6

I l

REACTOR VESSEL SUPPORT ELEV. 27*.3" i

.l l

1 7-REACTOR VESSEL VENT

• VALVE OUTLET ELEV. 24*-2" ,

'f

\-h N ',' ..

lj- \ \

HEATER RODS (45)

,, 6" SCH 120 PlPE S A-312 TP304 UNHEATED GUIDE TUBES (4) 3 lI

.l HOT LEG CONNECTION SECTION "A-A" ELEV. 21*-3" ,g ,

b lll .

li h; ^

'l 1:i *

.. .,ELANGE REQUIRED FOR UPPER PLENUM SHROUD ACCESS TO UPPER HEATER CONNECTION l . . ..P ,

i TOP OF ACTIVE FUEL

1. ELEV.16*-B-13/16"

. .s j .i

. HEATERS i l i

"A"I '

"A" BOTTOM OF ACTIVE FUEL +I ELEV. 4*-813/16" N fI 1

1[ :f, r

I REACTOR VESSEL l'NLET -  !

ELEV. O' 511/16" #  !  ;

j,..

F i ,3 7J

-L l. J.

h.  !

hh SEAL ASSEMBLY ,_j Figure 3-2. Reactor Vessel Arrangement l l

3-7

The model core will be housed in a circular vessel of minimum inner diameter, the four vacant arc segments (between the square core and the circular vessel) will be inserted. The model core grid spacers will use prototypical (Inconel) grid straps.

The model reactor vessel lower plenum will encompass the downcomer-to-reactor vessel flowpath icwpoint, plus hardware for flow shaping to obtain a uniform radial flow profile at the core inlet. The lower plenum is to retain power-to-volume scaling.

The upper plenum and top plenums are designed to afford prototypical phase separa-tion upstream of the RVVV and top plenum. Circumferential symmetry of the hot leg and RVVV ports is preserved within the constraints imposed by the total loop arrangement. The upper plenum cylinder flow holes are sized and located to approximate prototypical relative hydraulic resistances and elevations. The upper plenum to top plenum separator approximates the plant open flow area. The inter-plenum separator is at the proper elevation, the top plenum height is shortened to preserve power-to-volume scaling.

The MIST upper plenum cylinder geometry cannot be power-scaled because the result-ing flow paths would have atypically large pressure drops and because the small flow areas could induce slug flow. The MIST upper plenum cylinder is designed to preserve the flow split between the direct path to the outlet nozzle and the path through the large holes, and to obtain prototypical moisture separation beyond the plenum cylinder holes.

The MIST plenum cylinder (Figures 3-3, 3-4, and 3-5) will have an outside diameter of 4.388 inches. The twenty-four 3 inch diameter holes opposite the outlet nozzle will be simulated by a single 1.7 inch diameter hole. The six 33-1/2 inch diameter holes will be simulated by two 3.97 inch diameter holes; the four 23-1/2 inch diameter holes will be simulated by two 2.276 inch diameter holes. The elevation of the bottom of each type of hole will be preserved in MIST.

Volume to power scaling of the upper plenum sets the elevation of the plenum cover plate at 8.156 feet above the end of the heated length of the core. The plenum cover plate will contain nine 0.7 inch diameter holes to simulate the hydraulic resistance between the top head region and the upper plenum. The elevation of the top of the reactor vessel will be 4.104 feet above the plenum cover plate based on volume-to-power scaling of the top head region. Figure 3-6 provides a comparison between model and plant reactor vessel elevations.

3-8

l l

l l

NINE 0.71 IN. ' '

DIA HOLES ELEY. 24.875' _ f_ .

PLENUM COVER PLATE

\ TM0 1 ELEV. 24.67" 3.97" TWO HOLES ELEV. 24.17'  ;

ELEV. 24.456' RVVV 50

- 4.388" -

d ..

ELEV. 21.25' . - 1.7" IN DIA.

, ., HOLE OPPOSITE HOT LEGS 2.624" - = 5.501" =

Figure 3-3. RV Upper Plenum Vertical Section 3-9 ,

=. - - - - _ . _.

1 l

1 4.388" -

l l I l I l U .71" I

h I I I ( l U

l nA t l l f l I

- 1 " --

Figure 3-4. Top Plenum Cover Plate i

I

)

I l

3-10 l

-. --- . _ ___ ._ -._ - - - _ - _ . - - - - . . . . - . _ ..__1

I 1

l l

1

/

/

ELEY. 24.762' '

f I l

l

. 6" 2.276" 3.97*

i I I \

ELEV. 24.456' ,

1 t

i 3.97" I

I i l 1.7"

~

ELEY. 21.25" /

\

i

}

- 1.7"

, l i

i i 1 5 I __ __ __ l Figure 3-5. Upper Plenum Cylinder i

i t

b i

l t 3-11 i

a

l 2

m eu t

34 .

asesse.

. W k~

ruec-2

' '/j 3

$ 29 - "

__ Plenum Internals l *N m e a er f =a =*".""

5 '

Are Not Shown l'

s 00 0 0 0.0 .

o

- 25 i

l

(- 3 M RVVV, 24'

• eseuen s.se l

I

"****** ~ ~ Zui,*,"" l;; - 21 - - 21 C 2

!) e

+

• - - 17 -- ""'

, & r . g _ . . _ . . . . -

l I s m lW - c 8 . s m

p ss I

u r

o

]

, p. .

p

• y mm j ,

u o J

y "

' o 1 w e j

l '

______b - _ _ __

s

$ -- 5 ' -5 O

h 41EIIIIIIIIIIIIU l I*

4 f

g . _1/2- ~1/2 , e g -I --

.e ... .

neacton vesset roa a in avet assemety stater gg g g Figure 3-6. MIST Reactor Vessel Concept i

i I

Downcomer. The downcomer is one of the most challenging components of the reactor coolant system to model due to the numerous and diverse phenomena which occur during different plant transients. The downcomer is the juncture between the inner flow path (through the reactor vessel vent valves) and the outer flow path (through the cold leg). It is the zone of recombination of inter cold leg flow and fluid density asymmetries. The downcomer along with the cold leg discharge piping define the region of interaction of cold fluid (high-pressure injection fluid or return loop flow at low flow rates) with reactor vessel vent valve steam. Effective modeling of plant fluid and thermal behavior requires that these and other multidimensional phenomena be considered; at least within the con-straints of the essentially one-dimensional design of the MIST facility.

The MIST downcomer uses two separate regions (see Figure 3-2 and 3-7) with differ-ent scaling emphasis in order to best preserve the important phenomena detailed in Ref. 2. The upper downcomer provides an annular mixing chamber for the loop, HPI, LPI, RVVV, and core flood steams. This annular concept allows loop-to-loop coupling for asymmetric loop behavior and provides limited geometric similarity to prototypical downcomer regions. The lower downcomer consists of a single pipe, power to volume scaled, which maintains the current liquid level and volume in the downcomer and core regions.

The MIST upper downcomer provides an annular mixing region 5.64 feet long. As noted in the Design Requirements section of Ref. 2, all critical spillover and juncture elevations are preserved. However, the upper two feet of nonflow volume of the upper downcomer were omitted to reduce excess volume resulting from Froude ,

scaling. The inner annular boundary is a 3 inch Schedule 160 pipe, with an 0.0. l of 3.5 inches. This pipe is welded into the top of the upper downcomer and capped at the bottom resulting in a voided region inside this pipe, further reducing I excess fluid volume. The outer annular boundary is an 8 inch Schedule 160 pipe with an I.D. of 6.813 inches. The resulting annulus, with a gap of 1.656 inches, is separated into four quadrants by baffle plates extending to within 0.25" of the outer annular boundary. One cold leg and one RVVV discharge into each quadrant at plant typical elevations. Two core flood nozzles, 180* opposed, enter the upper downcomer across two of the baffle pates at plant typical elevations. The baffles are rigidly attached to the inner pipe. The baffles will extend from the top of the upper downcomer to the transition to the lower downcomer, thus supporting any loop-to-loop flow or density asymmetries which may exist.

3-13

4 i u .j RVVV NOZZLES &

'l n

t UPPER DOWNCOMER U.

i xCOLD LEG NOZZLES

\Q-UPPER l' LOWER DOWNCOMER DOWNCOMER

. ...r...,.

{ )

9::., ~ .

- in,44_ 4 J ._

l ..../

)I e'

TOP VIEW lj h n Figure 3-7. Upper and Lower Downcomer 3-14

l l

' 1 The upper to lower downcomer transition will be accomplished by an 8 inch to 3 inch reducer, 7-1/2 inches long. Since the lower downcomer is fabricated from Schedule 80 pipe, the reducer must accommodate this transition also.

The MIST lower downcomer is made of 3 inch Schedule 80, 316 stainless steel in order to maintain correct liquid levels in the downcomer (relative to the reactor vessel) and to minimize metal mass.

The MIST lower downcomer will extend downward 19.71' to an elevation of -0.46' (referred to steam generator lower tube sheet). The downcomer flow will be split and returned to the lower plenum of the reactor vessel.

Reactor Vessel Vent Valve (RVVV). The MIST Reactor Vessel Vent Valve (RVVV) arrangement will simulate the prototypical RVVVs with a circuit from the reactor vessel upper plenum to the downcomer. All fittings and control valves used in the circuit will be of a gas tight construction to minimize loop leakage. The circuit will simulate the eight prototypical vent valves with four lines as illustrated in Figure 3-8.

A differential pressure measurement between the reactor vessel (at the RVVV eleva-tion) and the upper downcomer quadrant cooresponding to each MIST vent valve line will be used to control the condition (open or closed) of each control valve.

Each control valve will have a selector switch to allow manual remote operation or automatic control of the valve condition. In the automatic mode, the valves will open and close at preset differential pressures. The actuation differential pres-sures setpoints for each control valve will be continuously adjustable over the calibrated range of the differential pressure transmitter used. The RVVV circuit will typically be set to open in the range of about 1/8 to 1/4 psi. It will also be possible to actuate all four vent valves using one differential pressure measurement.

STEAM GENERATORS The OTSG is a single pass, counterflow, tube, and shell heat exchanger. It consists of 19 alloy 600 tubes with an outside diameter of 5/8 inch spaced on a triangular pitch of 7/8 inch. The tube bundle is enclosed in a hexagonal shell 3.935 inches across flats and is held in place by 16 carbon steel tube support I plates (TSPs) spaced at approximately 3 foot intervals. The distance between the secondary face of the lower and upper tube sheets is full length, approxi-mately 52'-1-3/8". The 19-tube OTSG is illustrated in Figure 3-9. The TSPs 3-15

I NMEACTOR VESSEL

• T \

r ,

\W

\ lpoWNCOMGR T~ / T W ptow RESTRICTORS O TROL VALVES Figure 3-8. MIST RVVV Circuit .

3-16

1 l

PRIMARY INLET If O

=

+625 3/8" m NUPPER TUBE SHEET STEAM OUTLET

+610-3 / 8" &

HIGH AFW =q INJECTION N TUBE SUPPORT

_

OO 1 0000 **

3.935" OOOOO "A" "A" OOO u OO %e se M

SECTION "A A"

=t

=

+71" &

LOW AFW INJECTION 0" p--

TOP OF LOWER LOWER TUBE SHEET l f

TU8E SHEET If i PRIMARY OUTLET i 1

Figure 3-9. 19-Tube Once Through Steam Generator i

i 3-17 l

19 TUBE OTSG TUBE SUPPORT PLATE (DRILLED) is a AREA #p h ] Oo g .

[ 0 1.3 / 1

'\l'QSkN2A - # Rg (IN.) R2(IN.) R3 (IN.) AREA (IN.)

\ - 0.32031 0.250 0.1907 0.051518 FULL SIZE OTSG TUBE SUPPORT PLATE (BROACHED)

I 2.' 4, .

SO #*0 0

-/ -

R As

( 0 ,[

3 R2 )

2 Rg (IN.) R2(IN.) AREA (IN.1 0.323 0.430 0.051866 0.320 0.430 0.053075 0.320 0.427 0.051445 0.323 0.427 0.050233 120' Figure 3-10. Comparison of 19-Tube and Prototypical OTSG Tube Support Plates 3-18 l

l

PRIMARY INLET 1P TUBE - --UPPER TUBESHEET SUPPORT d, / J -

\

l 48'3" _\_ _

45'6" ___

42'8" ___

39'9" ___

36'8" __,

33'10" _ _ . ,

30'10" __,

27'9" , _ _ _

24'-10" _ _ . .

21'8" , , , _ _

18'8" ___ i 15'10" ___

12'8" -

9 7" ___

s' 7" ___

3'5" ___

, , , _ _ O REFERENCE LINE

fj' 4 TOP OF LOWER TUBESHEET LOWER TU8ESHEET x

I f PRIMARY OUTLET Figure 3-11. OTSG Tube Support Plate Locations 3-19 l

i are 1-1/2" thick and are drilled in a manner which simulates the broached pattern of a full-size OTSG TSP. This is illustrated in Figure 3-10. The axial locations of the TSPs are defined on Figure 3-11.

In the OTSG, primary flow enters at the top, flows downward through the tubes, and exits at the bottom. The main feedwater enters the steam generator,at the bottom, boils on the outside of the tubes, and exits at the top. For MIST, both high and i low auxiliary feedwater (AFW) injection location will be available. The high AFW injection is located so that the distance from the nozzle to the highest TSP in the model is the same as in the Oconee III OTSGs. The low AFW injection is loca-ted at 5'-11". These AFW injection locations are shown in Figure 3-9. Addition-ally, auxiliary feedwater nozzles may be selected for either maximum or minimum wetting of the tubes in the MIST OTSGs.

j The configuration of the AFW nuzies for maximum and minimum wetting was deter-mined during the GERDA program using a plastic model of the 19-tube OTSG. The maximum wetting configuration consists of a 0.187 inch 10 tube with a 0.063 inch 00 cross insert located on three sides of the hexagonal shell. The minimum wetting configuration consists of a single 0.430 inch 10 tube located on one side of the hexagonal shell.

The pressure boundary between the primary and secondary circuits is established by carbon steel tube sheets placed at the top and bottom of the generator with a I distance of 52'-1-3/8" between the secondary faces of the tube sheets. Each tube i sheet is clad with Inconel on the primary side. The tubes pass through the drilled tube sheet and are welded at the primary face. The upper and lower tube  ;

sheets are 3 and 24 inches thick, respectively.

i The OTSG instrumentation locations are discussed in Ref. 2.

I PRESSURIZER The MIST pressurizer will be a 6" Schedule 120 stainless steel pipe approximately 11-1/2' high. The pressurizer main heaters will be selected to provide a scaled power of 2 kW. Additional heater capacity of 3 to 4 kW will be available for i facility operational purposes.

l l

The pressurizer surge line will be a 3/4" Schedule 160 stainless steel pipe. The

) elevation of the surge line to hot leg connection will be 27.25' and the surge i

l l 3-20 i

l l

line low point (top of pipe) will be at 15.36'. The elevation of the pressurizer l bottom will be at 17.95' to preserve the surge line sloop seal. It will be possi-ble to isolate the pressurizer by closing valves in the surge line. Also, a drain line will be provided.

l A pressurizer spray 11ne will be connected to the pressurizer at an elevation of 29.8'. The spray line will connect the pressurizer to the cold leg. The spray line will be isolatable by a remotely actuated valve. Two guard heater control zones will be used. The spray line also will be guard heated, but the controller will be set to maintain the spray line temperature equal to the cold leg temperature.

i A PORV connection will be provided at the top of the pressurizer. In addition, provisions for degassing the pressurizer (during system heat-up) and for providing a gas blanket in the pressurizer (used during loop-filling) will be included.

/

l REACTOR COOLANT PUMPS 4

{:

The MIST facility reactor coolant system will contain four reactor coolant pumps;

one in each of the four cold legs. Each MIST reactor coolant pump will have f scaled head-flow characteristics so that pump on-off operation (running, locked

}

rotor, pump bumps) during a SBLOCA can be simulated.

Seven design requirements and two additional design objectives for the MIST

.i reactor coolant pumps were established.

I First priority was assigned to single-phase pump performance characteristics. In single-phase saturated or subcooled water operation, the MIST reactor coolant pumps will obtain full plant prototypical head rise at scaled flowrates. Flow-q rates are scaled by 1/817, the same scale factor used in the MIST facility design.

Due to the scale factor used in the test loop design, scaled leak rates for j

SBLOCAs in MIST are very low in magnitude. Controlled leak discharge rates are on i the order of 0.005 lbm/sec for SBLOCA testing in MIST. The desire is to keep the system inventory leakage from the pumps small enough in magnitude to be small compared to the controlled leak rate. These considerations lead to the second f design requirement that system inventory leakage from the MIST reactor coolant

] pumps be less than 0.0001 lbm/sec per pump, which is essentially zero leakage.

1

)

3-21 a

v s

~ ,.

b The MIST facility pumps will be designed sd that,when not running, the pJmp .

impeller can be locked to prevent rotation (windmilling) in either the forward or reverse direction. The purpose of this third dasjgn requirement is to satisfy two plant simulation criteria. First, the full-sizt reactor coolant ptmps e1 equipped with an antireverse rotation mechanism to prevent windmilling in the reverse direction. Second, the MIST cold legs r;ust have prototypical head loss characteristics when the pumps'are unpowered. The reactor coolant pumps will be part of the cold leg resistance, and it will be much easier to ot'tain tha correct total head loss characteristics in each cold leg if the pump impeller'is fixed and the pump resistance is a constant. .

The fourth design requirement is that the locked rotor resist,anct-(differential pressure across pump with locked rotor) must be equal to the loned rotor resis-tance of the full-size plant reactor coolant pde s-(scaled), if the MIST'pemp resistance is less than plant typical, additional resistanca will be added in each cold leg to make up the difference. It 'is desirable that the prp resistance be $

less than plant typical so that the loop can be orificed in the, cold leg suction '

piping. Each cold leg orifice could then be used to obtain a fi m meaasarement.

The fifth design requirement is that the pumps be capable of controlled startups and coastdowns to simulate conditions during pump bump caeration. The full-size reactor coolant pumps are equipped ~with large flywheel; which le d to 1cng startup and coastdown periods. Operlitor action during a SBLOCA transient may include periodic " pump bumping" in which a pump is started, briefly operated, then shut 1) off and allowed to coast down. Tne purpose of the pump bump-is .t1 restart stalled ',

natural circulation flow in the primary coolant loop. To simulate' the effect of '

these " pump bumps," the startup and coastdown scaled flow versus time and suctica lift versus scaled flow should be obtained by the MIST' reactor coolant pumps.

The sixth requirement is for the pumps to b(. installed v4.rtic' ally with pump j suction taken from the bottom and' pump discharge exiting horizontally. Tris provides a prototypical fluid path in the vicinity of the pump 'mpeller y d best preserves the reactor coolant pump spillover elevation and geometry.

1 The seventh design requirement is that the pumps be designed to operate 'cu1-tinuously without damage under two-phase or single-chase steam flow corh tions.

A design objective for the MIST pump is that u.1 controlled heat loss he less than 1 kW from each pump. Desig'1 objectives are met as far as practical.

3.22

-m ,

4 Another design objective is that heat energy input to the Reactor Coolant System (due to impeller inefficiency and c.onduction heat transfer from the motor windings) should be close to the scaled heat energy input of the full-size reactor

^

coolant pumps.

EMERGENCY CORE COOLING SYSTEM 1

l ~~Thk emergency core cooling system (ECCS) will consist of high pressure injection (HPI) and core flood systems. These systems will provide volumetrically scaled injection flows and capacities which simulate their plant counterparts.

The MIST facility HPI system includes a single pump to supply a branched injection network from a deionized, deaerated water supply or from a water supply saturated with noncondensable gas. The branched injection network will consist of four lines drawing from a common header; each line will inject flow to one of the four cold legs in the loop on the discharge side of the reactor coolant pumps. The valves shown located in each line could be used to simulate different plant-

[ typical HPI configurations.

Each of the four injection lines will be sized and/or line restrictions added to obtain the plant-typical K-factor (irrecoverable pressure loss characteristics).

,, , The HPI injection nozzles will be sized to preserve the plant-typical ratio of injected fluid momentum to cold leg fluid momentum, i The head-flow characteristics of the plant HPI system will be simulated by supply-P ing the common header with total HPI flowrates volumetrically scaled from the plant. The common header will be supplied by a multipiston positive displacement 1 pump with a bypass flow loop. The control system for HPI flow will monitor loop pressure and adjust bypass flow to obtain the desired injection fl.owrates.

Two prototypical core flood tanks will be simulated in MIST by a single core flood l tank constructed of 6-inch Schedule 160 pipe, approximately 25 feet tall. Volu- l metrica11y scaling the single model tank to simulate two full-size plant tanks,

~

the model tank will have an internal volume of 3.55 cubic feet. The model core flood tank will be installed vertically with the bottom of the tank at the proto-

, typical elevation. The injection line from the core flood tank to the core flood nozzle will be sized and/or line restrictions added to obtain the plant-typical 7' K-factor (irrecoverable pressure loss characteristics). The core flood nozzle in i 3-23 j,

1 the reactor vessel downcomer will be sized to preserve the plant-typical ratio of core flood injection fluid momentum to downcomer fluid momentum.

The model core flood tank will initially contain 19.8 gallcns of deionized, deaerated water at' ambient temperature. The tank will be pressurized to 600 psig with nitrogen overpressure with the nitrogen supply isolated. A small core flood tank circulation pump will be used to draw fluid from the bottom of the tank and spray it back into the tank above the initial water level so that the core flood tank fluid is saturated with nitrogen gas.

\-

When the loop pressure is above 600 psig, the core flood tank is isolated from the. l loop by a control valve. A pressure switch will monitor the loop pressure; whent loop pressure drops below 600 psig, the pressure switch will signal the control valve to cpen and remain open. A check valve in the injection line between the control valve and the core flood nozzle will allow the injection of the core flood tank inventory to the 1000, but prevent flow from the loop to the core flood tank.

SECONDARY LOOP The secondary loop in the MIST facility design will be limited to providing the steam generators secondary inventory and those fluid boundary conditions which impact SBLOCA phenomena. This includes steam generator secondary level control, auxiliary feedwater, main feedwater, and controlled cooldown (050-100*F/hr) capacity.

The steam generators secondary inventory can be controlled automatically, or by the operator. Two modes of automatic inventory control will be available:

constant level control or a high-low setpoint control. In the constant level control mode, feedwater flowrate is continuously adjusted (automatically) to maintain a preset water level in the steam generators secondary side. In the high-low setpoint mode, two water level setpoints are preset in the controller.

When tre secondary level drops to or below the low setpoint, feedwater injection is actuated. Feedwater injection continues until the secondary level reaches the high setpoint. At this point, feedwater injection is stopped. When feedwater injection is halted, the secondary level begins dropping due to steaming. Feed-water injection is not retriitiated until the secondary level again reaches the low setpoint. i l

3-24

i l

During the phase of high-low setpoint control where injection flow is actuated to refill the secondary inventory to the high setpoint, feedwater flowrate can be controlled automatically to provide a preset constant rate of feed (constant rate of refill) or can be controlled automatically to provide scaled feedwater injec-tion flowrates as a function of steam generator pressure. The latter method can be used to simulate plant feedwater string head-flow characteristics.

Each feedwater injection circuit has a high-flow and a low-flow control valve.

The high-flow control valves' range is sufficient to provide flows simulating plant auxiliary feedwater system head-flow characteristics, as well as main feedwater operation up to 10% scaled core power. The low-flow control valve in each injection circuit is mainly for the loop operators use between tests and test initiation.

Steam flow fram each steam generator is controlled individually. Each steam generators steam flow control circuit also consists of a high-flow and a low-flow control valve. The high-flow control valve's range is sufficient for most testing requirements, and the low-flow control valve is mostly for the loop operators' use. Steam flow can be controlled automatically with a controller, programmed controller (for control cooldown), or controlled manually by the loop operator.

3.1.2 Scaling Approach Twenty-five post-SBLOCA events can be identified to describe a composite SBLOCA transient. A list of these events in their chronological order is provided in Table 3-2 and discussed in Ref. 2. System characteristics relevant to post-SBLOCA modeling include (1) elevation, (2) two-phase behavior, (3) volume, and (4) wall effects, (5) it recoverable pressure losses. These characteristics were ranked according to tneir importance for preserving simulation of tne events /pnenomena discussed above.

Elevation is considered of major significance in 9 of 25 events, particularly decoupling of primary-to-secondary heat transfer and events involving fluid buoyant forces. System two-phase behavior (voiding, phase separation, flooding, etc.) is of at least intermediate significance in more than two-thirds of the events. Fluid volume is ranked third of the five system characteristics. It is of particular significance in voiding and vapor compression events. Wall effects, notably steam condensation upon vapor compression, is ranked fourth. Last is irrecoverable pressure drop (fluid momentum loss) which nonetheless received one-third the score of the first-ranked event.

3-25

f Table 3-2 POST-SBLOCA EVENTS Event Chronological Number Event Initiation

1. Leak Flow 1

2. Pressurizer Draining 5

3. Primary Depressurization

4. Power and Flow Transient: Reactor and Reactor Coolant Pump Trips Transfer of Steam Generator Feed

5. Single-Phase Natural Circulation Intermittent Spillover Circulation

6. Hot Leg U-Bend Saturation and Voiding

, - 7. Reactor Vessel Upper Head Voiding

8. Reactor Vessel Vent Valve Actuation

9. Decoupling of Steam Generator, Steam Generator

Depressurization

10. Primary Repressurization

11. Leak-HPI (High Pressure Injection) Cooling i 12. Feed and Bleed (Primary) Cooling i 13. Downcomer and Cold leg Voiding and Condensation l 14. Asymmetric Conditions Among Cold Legs Boiler Condenser Mode

15. Steam Generator Condensation of Primary Steam

16. Primary Depressurization to CFT/LPI Pressures

17. Steam Generator Repressurization

]

Primary Refill

! 18. Compression of Primary Steam j 19. Venting of Primary Fluid 4 20. Subcooling of Primary Components

21. Spillover Circulation (Hot Leg U-Bend Refilled)

22. " Pump Bump" Cooldown i'

23. Controlled Steam Generator Depressurization and Primary 1 Cooldown i 24. Reinitiation of Natural Circulation j 25. Cooling of Idled Loop 4

I 3-26 .

1 l

i L_ -. - - - . . _

I I

Elevation. Elevation is ranked as the most significant model characteristic for the preservation of post-SBLOCA events, and thereby, for testing germane to the TAG issues. The ideal model maintains full plant elevations at every flow junc-tion, spillover elevation, and flowpath low point. As enumerated in Ref. 2, full elevation makes use of the existing model steam generator and its extensive testing and benchmarking. Full elevation facilitates the adaptation of model-verified codes to plant predictions. Most importantly, elevation differences govern certain post-SBLOCA events, viz., heat-transfer decoupling of the primary and secondary systems, and natural circulation flow reversals (note Ref. 2).

l lv Two-Phase Behavior. System two-phase behavior encompasses: hot leg U-bend voiding and phase separation; hot leg (horizontal and vertical) flow regimes, and lower region (reactor vessel upper head, downcomer, and/or cold leg discharge piping) voiding and condensation. Hot leg two-phase behavior is addressed at length in Ref. 2. Lower-region voiding and condensation are also addressed in their individual sections. Detailed discussion of two-phase behavior is deferred to these sections; only general observations are made herein.

Hot leg U-bend separation is nowhere correlated for larger pipe diameters, nor has it been tested at conditions approaching post-SBLOCA prototypical. No code can predict the details of even quasisteady HLUB two-phase flow interactions. This dearth of information is in contrast to the significance of HLUB two-phase perfor-mance. It governs the transient. If phase separation were not to occur after pump trip and during the reduction of system liquid inventory, the event sequence would be exceedingly mild. The primary would retain heat transfer coupling to the SG even with the collapsed HL level well below the elevation of the HLUB. The exposure of the two-phase primary mixture to condensation in the SG would be an excellent heat transfer mode. The primary would depressurize in response to SG secondary pressure control, ECCS injection would be augmented as primary system pressure was reduced, and plant control would be relatively straightforward.

There are indications and there exists a large school of thought that the course of the plant post-SBLOCA transient is radically different. The difference, of course, relates to HL and HLUB phenomena. It is thought that phase separation occurs after the HLUB fluid saturates, pumps are tripped, and core power decreases somewhat with posttrip decay. This separation, segregating vapor near the HLUB and liquid at the lower elevations, interrupts circulation when the level upstream of the HLUB falls below the elevation of the HLUB spillover. Heat transfer between the primary and the secondary is then impeded until the primary inventory 3-27 l

decreases sufficiently to bring the primary liquid-vapor interface into the SG heat transfer regions. Of course, this interruption event may engender primary repressurization, inhibit ECCS introduction, and so on.

Only one plant has experienced two-phase, reduced-inventory conditions. TMI-2 observations are not unambiguous, but they certe. inly indicate that phase separ-ation does ultimately occur. Observations from a small-scale integral system model of the B&W geometry also support the separation hypothesis; without forced flow and with decreasing system inventory, the HLUB experiences: liquid-only flow, bubbly flow, two-phase mixture flow with a developing and lowering liquid-vapor interface, and finally, vapor flow only (interrupting circulation). Without further belaboring the point, it seems obvious that separation will occur during a plant post-SBLOCA event, circulation will interrupt, and the model must also accommodate separation. These assertions are explored further in Ref. 2.

The small system /large system criterion of Dukler and Taitel (3) is pertinent to the occurrence of phase separation in the model. At model diameters less than their criterion for a vertical two-phase system, bubbly flow will not occur; hence separation is suppressed. This criterion then sets a lower limit (which is pres-sure dependent) for the model HL pipe diameter. At 100 psia, this minimum diameter is 1-3/4".

Volume. System volume scaling governs fluid transport time as well as the rate of change of fluid energy content. The ideal model system scales volume as power and retains volume-versus-elevation proportions. Coupled with power-scaled boundary a flowrates, the ideal system then loses inventory and level, and changes fluid properties, in entirely prototypical fashion.

In order to accommodate bubbly flow, the hot and cold legs diameters were over-sized compared to ideal volume scaling. As a result, the hot and cold legs volumes are three to four times larger than the ideal volume scaling. The other major primary loop component volumes are sized according to ideal volume scaling.

Wall Effects. System bounding containment effects include: thermal energy storage, containment conduction across a fluid thermal gradient (and hence inter-phase energy transfer), and heat losses to ambient through the containment.

The ideal model scales containment structural volume as contained fluid volume and avoids atypical metal appendages (massive instrument probes, flanges, conducting 3-28

~. - _ _ - _ _ _ _ _ _ _ _ _ _

l l

supports, and the like). Even with ideal metal scaling, fluid-to-metal coupling is atypically effective in reduced-size pipes, i.e., the convective resistance is reduced as the heat transfer surface area per contained fluid volume is increased. Moreover, required pipe wall thickness decreases roughly as the square root of the pipe diameter reduction, thus high pressure models inherently suffer this atypicality. The optimum model simply minimizes pipe wall metal while avoid-ing the massive appendages previously noted.

Heat losses to ambient are particularly relevant to post-SBLOCA studies. Scaled decay core power levels may be less than the model heat losses if only passive insulation is used; plant heat losses, on the other hand, are still much smaller than model losses to ambient. The optimum model thus must actively counter ambient losses. Guard heating (automatically controlling on the sign of the temperature gradient in the insulation just beyond the' pipe periphery) is well-suited for this purpose and has therefore been included in the MIST facility.

Wall effects during vapor compression nay be paramount, particularly in light of I

the excess model metal and increased fluid-to-metal coupling already noted.

i Because the optimum model still suffers these atypicalities, it is important to

identify and quantify the mechanisms of condensation.

In Ref. 2, the likely modes of condensation heat transfer are addressed: heat storage in the walls, direct steam-water interface heat transfer, and indirect interface heat transfer by axial conduction within the walls. Heat storage in the 1

walls is shown to be the predominant condensation mode; the other two modes are an order of magnitude smaller. (Fortunately, wall heat storage is readily calculated in either the model or the plant).

l Irrecoverable Pressure Drop. The ideal model preserves plant irrecoverable pres-sure drops (fluid momentum losses) at (power-) scaled flowrates. Furthermore, the 1 ideal model preserves component as well as system losses, and the loss fractions due to form (shock) and to friction.

Recognizing the recurring need for scaling compromises, additional guidelines may be discussed: with parallel flow branches, plant-typical hydraulic symmetry must be maintained. With branch points at the hot leg nozzles and at the steam generator outlets, it is desirable that the hydraulic resistances within each branch leg be maintained. Finally, with major inner and outer flow loops (RVVV versus loop, both the downcomer-reactor vessel), the resistance of each loop should be maintained.

3-29 1

Model piping diameters may be estimated to preserve frictional irrecoverable pres-sure drop at power-scaled mass flowrates (Ref. 2). Because of differences in component geometry between the plant and the model, this model pipe diameter may have to be modified to approximately conserve total (form plus friction) irre-coverable pressure drop. For example, the preservation of frictional pressure drop only indicates larger than area-scaled model pipes, thus the model form losses associated with contractions and expansions are less than those of the plant. This decrease of form loss must be accommodated by a corresponding increase of friction loss (using diameter slightly smaller than calculated by friction considerations only).

MIST System Specification. MIST is to be a 2x4 integral system model (2 hot legs and steam generators, 4 cold legs and reactor coolant pumps). It is to model the B&W-designed (177 fuel assembly) lowered-loop plants. Each major primary system component is to be represented. Primary boundary systems are to be included.

Secondary system modeling is to include the steam generator secondary side with appropriate bounde.ry conditions.

Model power scaling S(Q) is set by the ratio of the number of steam generator tubes in the plant to that in the model, 15520/19 = 817. Model power (q) is thus:

q = Q/S(Q) = 2700 MW/817 = 3.3 MW, or 33 kW per % full power.

The model is to be full elevation. Full-scale plant elevations are to be pre-served at every primary system flow junction, at each flow spillover point, and at each flowpath low point.

Model components are specified in two distinct fashions. Heat transfer compo-nents, viz. the core and steam generators, are modeled by replicating a moderate number of full-sized cells having prototypical geometry and of full length. Thus, these heat transfer corrponents have representative power density, surface heat flux, and fluid volume, flow area, and residence time.

The second category of model components encompasses the hot leg and cold leg piping, i.e., those components not amenable to modeling by extraction of proto-typical subsections.

3-30

The scaling implications of the HL and CL pipe diameters have been discussed in detail in Ref. 2. In summary, a hot leg diameter of 1.3" preserves fluid veloc-ities and power-to-volume scaling. A diameter of 2.5" preserves vapor Froude Number and 2.9" preserves frictional pressure drop; (because model form losses are less than those of the plant, at equivalent flowrates, total irrecoverable pres-sure drop can be preserved with excess frictional pressure drop).

A diameter of roughly 2" is needed to obtain a "large" system, one that will admit to bubbly flow, based on the vertical two-phase flow criterion of Dukler and Taitel. This "large"-system requirement must be met to ensure phase separation during the post-SBLOCA transient and hence the possibility of primary flow inter-ruption; it thus seems prudent to increase the HL diameter beyond 2" to allow for uncertainties in the correlated boundary.

Two candidate diameters are 3-in. Schedule 160 (2.626" id) and 2-1/2 in.

Schedule 80 (2.323" id). B&W has experience with the larger diameter; it does support bubbly flow and phase separation. The larger diameter also allows more residual loop pressure drop, which can be used to improve flow-orifice metering f sensitivity.

The smaller diameter is preferable based on the ratio of fluid volume to metal volume--its ratio is within 20% of that of the plant. It is an improvement over the larger diameter for power-to-volume and fluid velocity scaling (by 30%), but these scale factors are still 3 times plant typical.

The 2.3" diameter represents a somewhat tighter specification than the 2.6" diameter--closer to the larger system /small system boundary and less excess irrecoverable pressure drop margin, but improved power-to-volume and fluid veloc-ity (and hence transit time) scaling. It is thus the more desirable diameter of the two. The small system /large system boundary is 1-3/4" at 100 psia, decreasing to 0.7" at 2000 psia. The 2.3" diameter is at least 2.3/1.75 = 1-1/3 times this small-system limit. There appears to be ample room for correlation error, assur-ing that large system performance will be obtained. With this assurance, and considering the previously-listed advantages of the smaller pipe, the 2.3" (2-1/2 inch Schedule 80) hot leg pipe diameter is selected.

i MIST scaling is described in greater detail in Refs. 2 and 4. Additional scaling details are given in a series of B&W Calculation Packages, "32s," and are summar-ized in B&W "86" documents. The relevant MIST 32 and 86 documents are listed in 3-31 l

1

. Table 3-3 MIST DOCUMENTS ON FACILITY DESIGN AND SCALE Document Date of

Identification No. Release Title 86-1140704-00 8 JJn 83 MIST Core Design Information 86-1140704-01 2 Nov 83 MIST Core Design Information 86-1140704-02 6 Jun 84 MIST Core Design Information 86-1140704-03 2 Aug 84 MIST Core Design 32-1142134-00 22 Apr 83 MIST Facility Specifications, Elevations

. 32-1142134-01 29 Apr 83 MIST Facility Specifications, Elevations 86-1142135-00 22 Apr 83 MIST Facility Specifications, Elevations 86-1142135-01 7 Jun 83 MIST Facility Specifications, Elevations 86-1142551-00 20 May 83 MIST Facility Specifications, Fluid Volumes, and Metal Volumes 32-1142552-00 20 May 83 MIST Facility Specifications, Fluid Volumes, and Metal Volumes 32-1142593-00 20 Sep 83 Decay Power for Short Times After Infinite Operation 86-1142637-00 16 May 83 MIST Facility Specifications--System Characteristics 86-1142637-01 16 Jun 83 MIST Facility Specifications--System Characteristics 86-1142637-02 22 Jun 83 MIST Facility Specifications--System Characteristics 1

86-1143876-00 14 Jun 83 Preliminary Specification for Model RC Pump 1 32-1148431-00 14 Feb 84 Fission Power-MIST 32-1149122-00 19 Dec 83 MIST HPI Line Resistance 32-1149123-00 16 Mar 84 MIST CFT Design 32-1149123-01 9 May 84 MIST CFT Design 86-1149126-00 8 Feb 84 8EST Est. - Shutdown Margin 3-32 1

. _ - - - . . - . . . , . - . - , ~ .- .- . . _ . . . - - . . - . .-. . - . - . . - . . . . . - - -

l l

Table 3-3 (continued)

Document Date of Identification No. Release - Title 32-1149128-00 29 Feb 84 MIST Pressurizer Scaling 32-1149128-01 6 Jun 84 MIST Pressurizer Scaling 32-1149130-00 7 Mar 84 Justification for Deleting MIST Core Flow Distribution Tests 32-1149131-00 30 Aug 84 MIST CL Details i 86-1149139-00 30 Aug 84 MIST CL Details 32-1149138-00 8 May 84 MIST Reactor Vessel Scaling 32-1149140-00 31 May 84 MIST Upper Plenum Flow Splits 32-1149141-00 20 Jun 84 MIST Hydraulics Model 32-1149142-00 24 Jul 84 MIST Pressurizer Spray a P 32-1149143-00 23 Jul 84 MIST Conversion Factors & Pump Characteristics 86-1149144-00 24 Jul 84 MIST Conversion Factors & Pump Characteristics 32-1149145-00 1 Aug 84 MIST HPI 86-1149146-00 1 Aug 84 MIST HPI 32-1153978-00 (pending) (Codeinput) 86-1153979-00 18 Jan 85 PORV/ Safety Sizes 32-1153984-00 (pending) RVVV 86-1153985-00 (pending) RVVV 3-33

Table 3-3. Each document is available for copying and distribution, subject to Project Management Group approval.

3.1.3 Scaling Compromises Investigated The atypicalities which can be addressed by MIST are summarized in Table 3-4.

MIST was assigned to address 3 of its 12 atypicalities. Of these 3: MIST can well address AFW multidimensional effects in MIST; MIST is of limited usefulness regarding the SG Metal Mass atypicality (precluding substantial facility modifica-tions); and a secondary system modification is required to address its Secondary Side atypicality.

MIST can also be used to address two other of its atypicalities without modifica-tion, viz. hot leg separation and hot leg flow regime. MIST testing to address

] MIST atypicalities is generally of limited use regarding the direct application of this information to the code modeling effort. MIST testing is of similar benefit and applicability for all of the issues affected by each atypicality.

1 3.1.4 Test Procedure: General Approach and Techniques MIST test procedures are keyed to test types, which include: Debug, Character-  ;

ization, Mapping, and Composite. " Debug" tests are used to verify system and

, component operability; ad hoc tests are performed without formal test specifica-tions and with relatively limited test procedures.

I MIST characterization tests are performed to determine the behavior and impact of key subsystems. For example, RVW/Downcomer interactions are to be tested, as are the S.G. secondary control functions. Test conditions are selected to highlight j the interaction or subsystem being examined, rather than for plant-transient

typicality. Loop conditions thus range widely, e.g., from subcooled at atmo-spheric pressure to saturated at 2350 psia (the MIST PORV setpoint). Formal test specifications and procedures are developed for each of the characterization tests. Data taking employs the full set of instruments and both the high-speed

and low-speed data acquisition systems, as well as supplementary data recording techniques (strip charts and video tapes). The data scan frequency is keyed to l the rapidity and duration of the intaractions being characterized.

l MIST Mapping Tests (Test Group 30) examine the events which occur early in the j post-SBLOCA sequence, viz. saturation, two-phase circulation, intermittent circu-

! 1ation, flow interruption, and the BCM. Rather than being transient tests,

] 3-34 i

i

i 1

Table 3-4 i

THE POTENTIAL USE OF MIST TO ADDRESS THE MIST ATYPICALITIES i

MIST ATYPICALITIES Was MIST Can MIST address this Can the MI'T S information i assigned to atypicality? be applied directly to i (From Table 2 of Sursock/ Young address this the plant modeling effort?

i letter dated 25 Feb 85-- atypicality?

Appendix B)

Limited 4

Mod'n & Mod'n Yes No Yes Limited Rea'd No Yes Limited Rec'd

1. Hot Leg Separation X X X

2. S.G. Metal Mass X X X y 3. AfW Multidimensional X X X

4. RCP Two-Phase Characteristics X X X

5. Downcomer Flow & Density Field X X X

6. Downcomer Tangential Resistance X X X

7. RWV Simulation X X X

8. Piping Metal Mass X X X

9. Low Pressure Injection X X X

10. Secondary Side X X X

11. One-Phase, 2-T Strat. Flow X X X

12. Hot Leg Flow Regime X X X

i a

j i

pseudo-steady states are obtained by holding the system boundary conditions con-stant throughout each test; system total liquid inventory is decreased gradually I and is held constant when system interactions begin to change. The full set of i instruments and recording techniques are to be used.

I l The MIST Composite Tests are integral system transients, from reactor trip plus 1-1/2 minutes through postrefill cooldown. Tests will be conducted both para-f metrically (with limited simulation of operator actions, to enhance the under-I standing, comparison, and prediction of these tests) and with the more significant operator actions simulated. The loop is generally to be initialized in subcooled natural circulation at 3-1/2% of scaled full power plus losses to ambient. The j initial S.G. secondary control pressure will be constant at 1010 psia--this sets j the secondary saturation temperature (and T-COLD) at 545'F, and thus determines T-HOT (at approximately 590 to 600*F, depending on the S.G. thermal center eleva-

! tion). A test is begun by opening the specified leak and then actuating the appropriate boundary systems (power decay ramp, HPI, increased AFW) as the pres-surizer drains. This sequence obtains plant-similar conditions as the hot leg (upper elevation) fluid saturates. System thermodynamic conditions thus range from multiphase (subcooled, saturated, and superheated) fluid at high pressures to j predominantly subcooled at -200 psia. The full set of (critical) instruments are

! to be used. Recording methods will include the low-speed and high-speed Data Acquisition Systems, strip charts recordings of selected signals, and video tapes ,

j of the viewport displays.

i j 3.1.5 Application of Data (to Resolve the Issues or to Address the Atypicalities)

J

{ The application of MIST to the MIST atypicalities (and affected issues) is summar-

! ized in Table 3-4. Atypicalities are taken up individually in the following subsections. The description of each atypicality is generally excerpted from Part I of the MIST Facility Specification (Ref. 2).

i MIST Atypicality #1: Hot Leg Separation.

] Background. Separation will occur when the hot leg liquid inventory is j insufficient to support the cocurrent flow of two phase, i.e., when a liquid-l vapor interface develops in the hot leg.

1 i

l Leak fluid inventory loss rate was used to estimate the rate of change of hot i leg void fraction with fluid inventory. These were combined with an assumed l draining time, to provide an estimate of hot leg void fraction due to imposed j 3-36 i

liquid deficiency. The intersections of the j-based void fractions with those imposed by draining indicated a time at which HL phase separation occurs.

Also, the time (after trip and the start of draining) at which the j-based void fraction decreases below 25% signals the transition from slug to bubbly flow. These transition times (for the various assumed conditions, including maximum vapor superficial velocity) ranged from 3 minutes to almost 9 minutes after trip.

i Reactor vessel vent valve operation impacts these estimates. With increasing RVVV vapor flowrate, the j-based void fractions versus time decrease more rapidly. Bend-induced separation also has an impact. The HL U-bend volume above the spiller elevation is a small portion of the total het leg volume, viz. ~8% in the plant and -5% in the model (using a 1.6' bend radius).

Phase separation at the U-bend occurs as roughly 10% power, or less than one minute after reactor trip. Based on the relatively small volume of the upper U-bend, and the propensity of the bend to cause phase separation, it is expected that the draining of a relatively small amount of primary fluid will lead to the interruption of circulation of the plant and model hot leg U-bend.

Planned Tests. MIST Test Group 30 Mapping Tests, examines the interactions which occur early in the SBLOCA transient, including phase separation. Condi-tions are selected to highlight these interactions and to investigate their response to the major boundary conditions, viz. core power level, magnitude of HPI-leak cooling, RVVV performance, etc. These conditions are held constant throughout a mapping test; only primary system total liquid inventory is to be ,

varied during testing. Inventory is reduced in steps by imbalancing the leak l and HPI flow rates; it is then held constant at inventory plateaus to observe l system interactions at pseudo-steady state. This testing approach facilitates the code benchmarking of these results. The intertest conditions variations will identify those conditions having the greatest impact, and the range of conditions over which the system performance changes are most pronounced.

Code predictions at conditions of most interest are thereby facilitated.

Additional Tests. The planned Mapping Tests are to be equivalent to five Composite transients, or roughly 50 hours5.787037e-4 days <br />0.0139 hours <br />8.267196e-5 weeks <br />1.9025e-5 months <br /> of testing. It would be beneficial to increase this allocation in order to expand the range and combinations of conditions to be investigated. The draft Mapping Test matrix is shown in

, Table 3-5. It is expected that at least five of the nine test conditions listed in the matrix will be performed with the current allocation.

i j 3-37

, -. ~, , - -

r --v. - ,e , r --..4.-., --.. . . - - . , . - - - _ _ - . , _ . _ _ _ . . - _ . , , . ,

i

! .s Table 3-5 MIST TEST MATRIX: TEST GROUP 30, MAPPING Variable 1 2 3 4 5 I

Core Power Leak /HPI RVVV SG Level RCP Status

!  % Flowrate Status l (1bm/hr)

l. Settings 1 3-1/2 550 Auto 31.6' 0FF

! 2 1 275 Closed Band ON

! 3 -- 0 -- Minimum --

4 -- B1-CLD -- -- --

j

. Setting Tolerance: +0.1% balanced na 10.1 na 300000 1 1 1 1 1 300101 2 300202 2 300312 2 3 300403 2 300504 2 300604 3 300705 2 l 300802 4

• Notes:

i j 1. Increase the core power entries by the power required to offset losses to ambient.

2. Balance the leak and HPI flow rates. Discharge from the Oc-RV bottom.- If impractical, use the B1-CLD leak site and discard etting 300802. If neither j discharge site can be throttled, use a scaled 5 cm leak to approximate 550 lbm/hr and a scaled 2-1/2 cm 2 leak to approximate 275 lbm/hr. Control HP! to match the discharge flow rate.

3. The " Auto" RVVV settings are 0.125 and 0.04 psi to open and close, with independent valve operation. For the " closed" setting, manually close the valves.

! 4. SG 1evel: " Band" = band level control at 31.5 1 1.5'. " Minimum" = lowest ,

l 1evel while maintaining stable SG control. All use high-elevation AFW.

i

! 5. RCPs are either off or all on at full flow rate.

i i

i i

3-38 '

Appiteation. The Mapping Tests are directed to the generation of pseudo- '

steady state data during events which occur early in the post-SBLOCA sequence. Conditions are varied among the tests to obtain the sensitivity of events to major variations of boundary conditions. These results will be ideal for comparing observations to the code predictions of the early events (including hot leg separation). These comparisons, combined with the informa-tion from the separate-effects facilities, will enhance the ability of the codes to predict these early events in the plant.

MIST Atypicality #2: S.G. Metal Mass.

Atypicality. Steam generator wall effects include heat losses to ambient

] (primary to secondary to ambient) and energy storage in the steam generator j metal. Losses to ambient are countered by active guard heating of the entire j steam generator; these guard heaters act to maintain adiabaticity at the S.G.

outer metal surface.

The model S.G. uses prototypical tubes, tube support plates, and tube sheet j material and geometry, thus eliminating internal metal-effects atypicali-

! ties. The S.G. shell is overly thick in the scaled model S.G.; similarly the ratio of peripheral cells to unit (internal) cells is too large, this increased coupling of S.G. fluid to the shell through heat transfer exacer-bates the impact of excess shell metal in the model.

l The excess MIST S.G. metal mass will increase the ratio of stored energy to fluid energy. It will tend to dampen oscillatory events and to increase the j amount of energy removal required for cooldown.

i i

j Testing. Supplementary MIST tests could address the impact of S.G. metal on j system interactions or cooldown. The S.G. guard heater controls could be adjusted to vary the amount of energy stored in the metal (a guard-heaters-off

test is already included in the MIST test matrix). Assuming that the energy stored in the larger S.G. metal masses is more responsive to metal conduction

than to surface convection, variations of the metal stored energy may largely be constrained to increases from nominal. The transient events perceived to be most affected by metal stored energy effects would then be repeated with the altered metal energy content. The single metal thermocouple per S.G. tube f

sheet could be augmented to better indicate the metal stored energy.

3-39 4

_.._.________.________m,,._-,.._.__.____--.___,,..,.m_ _ _ _ _ . . - . _ _ . . . _ _ , _ - - , . _ , . _ _ _ . . , _ . - , ~ . , _ _ _ , . _ , . _ . , . , , - _ . . - . . - -

Application. The sensitivity of system response to metal stored energy could readily be compared to the code-predicted response. The required modeling detail would be determined. This code modeling information would be directly applicable to the plant modeling effort.

MIST Atypicality #3: AFW Multidimensionality.

Atypicality. Auxiliary feedwater effects refer primarily to the localized peripheral feed introduction and the resulting pronounced radial and circum-ferential heat transfer differences. Because three-dimensional effects do not adnit to geometric scaling, atypical AFW effects are likely in any scaled model. With full elevation in the steam generator, the tube bundle dimensions l and hence the impedence to horizontal (steam) flow between heat transfer l regions on the secondary side of the S.G. are reduced. (Bracketing methods of feed introduction--maximum wetting and minimum wetting--are available in MIST.)

The relatively few tubes in the model S.G. bring into close proximity the tube

! regions wetted upon AFW introduction and those which are unwetted. The impact

! of this dimensional atypicality can be addressed by using extremes of AFW wetting; the introduction of AFW through one relatively large nozzle wets only one or a few S.G. tubest at the other extreme. AFW injection using three dispersion nozzles wets virtually all the tubes (at the same AFW flowrate at which only a few tubes are wetted using the minimum-wetting configuration).

Tests using the extremes of tube wetting would give insight into tube wetting effects. They will not achieve plant similarity in this regard, at either extreme of tube wetting.

Testinq. Selected transients or portions of transients could be repeated using the maximum AFW wetting configuration rather than minimum wetting.

Candidate interactions for this variation include: natural circulation with relatively low secondary level BCM, and the S.G. tube leak transient.

Applications. Selected events observed with minimum AFW wetting would be contrasted to the behavior with maximum wetting. The code models of AFW effects could then be revised to reflect the wetting change, and exercised through the events of interest. The code sensitivity to AFW wetting would thus be established and may suggest model improvements. Neither (minimum nor maximum) wetting model replicates the plant AFW effects, however. The 3-40

suggested code comparison at wetting extremes simply highlights the code adequacy in this area, because of the atypically shortened horizontal dimensions of the model tube bundle. Additional code work would be required to apply this AFW wetting information to plant predictions.

MIST Atypicality #4: RCP Two-Phase Characteristics. The MIST model pump design sacrifices two-phase performance for mass and energy closure, single-phase perfor-mance, and mechanical reliability. Model pump two-phase degradation will be prototypical only by coincidence. But model pump degradation will be character-ized by testing. This degradation data may then be used in conjunction with a systems code to predict system behavior thorugh a degrading-pumps event. A code so benchmarked, with a plant pump degradation model substituted for that of the model pump, can then be applied to the plant with increased confidence. This utility of the model data notwithstanding, it should be reiterated that model degraded-pump behavior is unlikely to shed any light on plant pump behavior.

MIST Atypicality #5: Downcomer Flow and Density Fields.

Atypicality. The MIST Baffled Quadrant DC preserves plant elevations of each zone by flow-stream interaction. The model DC voiding and condensation typi-Cality is hindered by excess fluid volume over the upper DC elevation and by excess metal associated with the RVVV piping. Both influences should be amenable to prediction, but the interactions of the four RVVVs with the voiding and condensation events may cause difficulty. The model DC volume is excessive above the top of the core (to obtain the annular configuration without excessive DC wall retardation of flow). The Model DC volume below the top of the core is approximately power-to-volume scaled.

Local DC hydraulics are only coarsely modeled in the MIST Baffled Quadrant DC arrangement (although it is the best of the candidate arrangements in this regard). Axial and tangential ir ecoverable pressure drops are misscaled; their ratio is approximated. Axin' pressure drop in the inner flow loop con-taining the DC will be compensated elsewhere. Wall effects are atypically operative, but they are readily modeled.

Testing. The MIST downcomer and RVVV performance will be examined during characterization testing. Test conditions are selected to highlight the RVVV and downcomer performance, rather than to obtain transient typicality. These characterization tests will consider vapor flowing through the vent valves and

< 3-41

I also single-phase liquid circulation. Test conditions include both symmetric and asymmetric loop conditions; the RVVVs will be controlled to actuate independently, ganged, and manually; tests will be performed both with and without HPI and leak flow. These tests are necessarily limited by the time allotted to characterization testing. This allocation could usefully be increased to expand the range of test conditions and the types of RVVV-DC interactions to be examined.

Downcomer (and/or RVVV) nodifications could be made to accentuate a character-1stic of particular interest. For example, the existing DC baffles could be modified to perturb the tangential hydraulic resistance. Such modifications would be relatively costly, however.

Application. The MIST RVVV/DC Characterization Tests will provide detailed model component (steady-state) performance over a range of imposed boundary conditions. This information will facilitate the detailed code benchmarking of model RVVV/DC interactions. Although these MIST interactions are not transferable to the plants, the code modeling information and the prediction techniques required to adequately represent MIST will apply to the plant modeling effort.

MIST Atypicality #6: Downcomer Tangential Resistance. This atypicality has been included in the preceding discussion of downcomer flow and density fields (Atypicality #5).

MIST Atypicality #7: RVVV Simulation.

Atypicality. The resistance of the model RVVV arrangement is composed of frictional and form losses versus form losses in the plant. The model RVVV should impose most of the plant RVVV resistance at the slit-orifice located adjacent to the discharge into the downcomer. To accomplish this, the RVVV piping is oversized and the valve resistance is minimized. The orifice will be sized to obtain the remaining resistance.

The model RVVV will open at a differential pressure of 0.125 psi. The simu-lated valve will obtain full open head-flow response between 0.125 and 0.26 psi, but the plant RVVV head-flow response varies over this range. For differential pressures greater than 0.26 psi, the model will obtain power-scaled flow rates for plant-similar fluid conditions.

3-42

The two-phase behavior differs between the MIST RVW and that of the proto-type. The RVVV should experience two-phase conditions only when the RV liquid level approaches the elevation of the RVVV. The plant valve inlet height is roughly five times that of the model. Also, the plant RVVV velocities are greater than those in MIST, tending to increase the effective vertical dimen-sion of the plant RVVV via vapor pull-through and liquid entrainment; the model RVVV is less likely to entrain than that of the plant.

The four-RVVV design chosen for MIST obtains the quickest response for the valve stroke time of the various design options. There is a time delay associated with the flow of the fluid through the piping, however. For con-tinuous RVVV flow, the transit time should.not have any effect. Fluid transit I delay time does impact RVVV effects at valve actuation. In order to minimize

) this impact, the guard heaters surrounding the RVVV piping lengths will main-tain the RV.VV fluid at the RV upper head conditions.

Testing and Application. Two MIST tests vary RVVV control. The valve actuating setpoints are to be increased in one test, the valves are to be 1 maintained (manually) open in the second test. Both tests otherwise corre-l spond to the boundary conditions of the Nominal (Composite) MIST transient.

{ RVVV Characterization tests are also planned. These tests and their applica-l tion have been discussed in Atypicality #5 (Downcomer Flow and Density Fields).

MIST Atypicality #8: Piping Metal Mass.

{

Atypicality. Model wall effects are amplified by the atypically large ratio of wall heat transfer surface area to contained fluid volume. Therefore,

) model heat losses and metal mass are closely controlled. Heat losses are minimized by using full-system active guard heating. Metal mass is minimized by avoiding flanges and conducting supports and by using minimum pipe wall thickness consistent with code requirements. The resulting ratio of metal volume to fluid volume in the model is roughly 35% greater than that in the plant. This excess model metal dampens system pressure changes, but these metals' effects are readily modeled and predicted.

OTIS results have indicated that piping metal has little impact during the depressurization and draining phases of the post-SBLOCA transient. But 1

3-43 1

. - . - - - - ~ - _ .- _ - - . - _ - -

l (upper-elevation) metal stored heat (and the superheating of the adjacent l

g vapor) govern the repressurization upon refill and thus become significant relatively late in the transient. The wall conditions at this time depend on the detailed fluid-wall-ambient energy balance over the several hours pre-

]

ceding refill.

l -

Testing and Application. A MIST test with the guard heaters de-energized is i planned. Additional tests could be conducted with altered guard heating.

I Such tests would serve to better quantify the role of piping metal stored j energy in the refill and postrefill events. This test information would

indicate the necessity for the code modeling of detailed system piping heat transfer, and would thus be directly applicable to the plant modeling and j prediction effort.

MIST Atypicality #9: Low Pressure Injection. MIST has no low pressure injection (LPI) system simulation. LPI system actuation is most prevalent at the low pres-sures achieved with larger (small) oreaks. LPI system operation also affects operator actions during cooldown.

MIST Atypicality #10: Secondary Side. The MIST secondary side characteristics

, limit the rate of secondary side depressurization and thus the ability of MIST to simulate a steam-line-break transient. The MIST secondary system could be modi-fied to increase its maximum depressurization rate; for example, a large feed reservoir could be added such that the model secondary system could be depressur-l ized by discharging to atmosphere. However, MIST is currently limited to 10% of scaled full power; the earlier portion of the steam line break transient cannot be simulated. A combined tube rupture and (limited) steam line break transient is planned for MIST.

MIST Atypicality #11: Stratified Single-Phase Flow.

Atypicality. The MIST pipe diameters are reduced from those of the plant in order to maintain approximate power-to-volume scaling while preserving the plant elevations. In vertical piping runs, the reduced MIST pipe diameters i suppress liquid-liquid counterflow; the degree and duration of fluid stratifi-cation is thus atypically increased.

The shortened MIST horizontal piping runs combined with the reduced pipe diameters tend to suppress horizontal liquid-liquid counterflow as well. This I

3-44 l

l

- _ _ . - - . . ~ _ _ - . . . - - ___ --

. - ~ - . - _ - - - . _ - . - - - -- .. - - - _-. _ .- - - . .- - .

phenomenon is perceived to be most significant in the cold leg discharge piping. The OTIS results indicate that such counterflow (or an equally-  !

effective mixing process) did commonly occur in the single, small-diameter cold leg piping of OTIS.

l Testing and Application. MIST cannot be used to study counterflow variations l without piping modifications between tests. MIST does have an alternate leak l

l site at the top of the (B1) cold leg discharge pipe. (Thisalternateleak ,

site does not have the requisite leak flow control and metering systems.)

Tests using this pipe-top leak could be compared to tests with the customary

~

pipe-bottom leak site. Stratification and/or counterflow and mixing will '

I alter the leak fluid conditions, and system mass balance, between these r

j tests. Such information could be used to exercise the relevant code models r l (or to indicate the need for code capabilities in single-phase counterflow). f i This information could be applied to plant modeling via suitable separate-

\

1 effects test having larger pipe diameters.

1

' i

{

MIST Atypicality #12: Hot Leo Flow Regime. The discussion of Hot Leg Separation l

[ (Atypicality #1) also pertains to this Hot Leg flow Regime atypicality.

t i

i

! i

(

l  !

I i

j i

3 45 if

3.2 EPRIINTEGRALTESTFACILITYATSRIINTERNATIONAL(SRI-2) 3.2.1 Facility Description Objectives. The two major objectives of the EPRI-sponsored experimental program at SRI International are:

1. To provide experimental data for the qualification and assessement of safety analysis codes /models simulating B&W plants.

2. To perform separate effect tests in support of the IST program.

Main Features of the Facility. The facility is a 2x4 integral test facility modeled after TMI-2. All vertical dimensions are 1/4 of the prototypical heights. All flow areas in the primary loop, except that in the core, are 1/324 of the prototypical areas. Horizontal segments of the pipes are not scaled. The i overall view of the facility and of the reactor are shown in Fiqurgs 3-12 and 3-13. Thenominaloperatingpressureis0.7MPr(100psta).

, Reactor Vessel. The reactor vessel is basically a circular cylinder 2.58 m (102in)highand25cm(10in)indiameter. It has an internal downcomer with four vent valves. Similar to the prototypical valves, the hinges for the valves are placed on the inside of the downcomer rather than the outside (see Figure 3-14). Such an arrangement allows the valve to open under a 1/4 psia pressure difference (prototypical values) without having to design a valve of excessive weight. As it is, the valve is quite thick as can be seen in l Figure 3-14. Each of the four valves simulates the vent valve in the prototype.

Its opening area is therefore twice that of the scaled area of a prototypical valve.

An 18-element rod heater will be installed in the core region of the vessel. The electric heater is manufactured by RAMA Corporation. When supplied with 44C-volt, 100-amp, 3-phase, ac power, it will produce 88 kW of heat. The heater control allows continuous variation of the core power below its maximum output. Coupled  ;

with a software program, the heater will be able to provide the scaled decay heat history. Several thermocouples are incorporated in the heater to allow automatic tripping of the heater when the element temperature exceeds the permissible value.

Internal components in the upper plenum of the vessel are partially modeled. For example,anupperplenumthroudwithappropriateflowholesisincorporated(see 3 46

l l

- no =

(\h

.g 9

- -; g 3 ru_.=_. =_ ===- ,y ch @ t- - - -_ w@ . - @r - --l@ej k!l l L

(Dj gn- -- , "l 4L F e 66 0 = 6eO . =

1 I

I d

j ,o.

m,

.~ .m. u, cem

/

i

,A ....

/yg....

N I cm i t i , .4

eri .: 9 WC'X'*; t "r,%

l; ... .

N l U6dM.Ed7 ',7*f I *uma#=44we .w as %

. ,x .

i eM--

h1i 'Mc --

- ay.pr. *' r.a., , ,,,

' . ::,at. ..

I l T$$tQ ,x ,

~

= .'t':a:*:.* * * *

-- ca.u:

J l ,

. .m~ .

i v a., ....,., s u..-ta:1 ,_ _ ._...,,

n- Ather .. .. . ,,, u g[6-

,r ,

v ,

p0 p, f, i a ' l j taava oo -- % g ahu r,i).. e _ _ . . .

jgr-..t . . M, ((]ryY

< - *14,C y. h.'

.We't' i

1

) Wt t,:'t - "-

7 c,r.a 1 '

(, )

, (' m g,.y -**' ..

) ~ ".,%0..a*.***

,Lr d

,. .. . -t m u , ,

j 8' 4 N .y . .. .  ; ;.s,a-- P

• 4 5 x

' r.u' 'Tr.'.'W.:T I 'f)1 .me ..

J')'.

Figure 3 12. SRI-2 Reactor Coolant System Arrangement 1

4 3 47 I

_ , . ~ . _ _ ----._z, -, . . . , - - - , - - - , , m - - , - . . - . . , . , - - - - - - - - - - - - - - ~ , . - - - - , - - - - , - - - - . . - - - . . , - - , - , , - - -

T I,

I t

o E,

i 1

l *)6 62 .UTnYmYE YOIII'"

• i vs wts memo w= ***** f iq .

N

• t:.# /g" *

. d, .us 4

1 d * / . ,

.J.9?.it 1

rr m 63

[' <

w.. l

..u~..,..

ve rW.. /~l;o.+o. '

- W+a-c

x. -

tb.:<

1

.;. 7

y

.+w qJ e f 9' t,ftr e W.3* :e.

, . . , vv,-...yo N -.-M.N< 4,\. .;

P~ **. *

.s.~ . .y

s. \,,

RS s. ,

c-t w).$b i

w Ls. i. . ,.

.~. ,

5 , . -. ..

i e u 4 1*tu e>>es ,At d s i g' * **o *m i

c . .,.j ,t g v.

• i u

. . .. m

. .u .

m. m..,.,

.... s s,. &

teem s.e* 0 0 - .- _. -.,- - .

4

y m..

m, s- ~

v..,7 -

L

• u i'S te ? **'

)

L .m.

- t

~ .,,, ..... t

c. ..

,,,* e. hth.  %  %

i .

f N

{  %'" W:.wt:.r:.'

3 j

L....= = w F

1 I 4 i w~

Ws i l q m ,,, '

l '.'u-"n.

.J t N' 8 %

J. . . r 3'"L. ~V -s

&D e w

.a-3

e n. +v W. ,_

i

3. u w.~w ,

.,~hML/Ull/fN,/// / / ///H/L "

l Figure 3 13. SRI-2 Reactor Vessel Details i 1

1 i

3 48 i

i -

m . _ _ _ _ _ _ - - - - . _ - __ _ _ _ _ _ . _ _ _ _ _ . _ - - _ _ . . _ _ _ - _

Figure 3-15); the upper head region is segregated from the upper plenum by an upper head baffle plate with scaled openings to allow some fluid exchange between the two chambers (see Figure 3-16); and several vertical posts are installed in l the upper plenum to partially simulate the control rod guide tubes and to achieve the correct fluid / metal mass ratio in the upper plenum.  !

The flow distributor plate at the bottom of the core in the prototype is not modeled. Instead, a multiopening orifice plate (see Figure 3-16) is added to the top of the core to yield the correct hydraulic resistance value from the down-corner annular to the hot leg nozzle. The vessel is equipped with a high point ,

vent which also serves as a noncondensible gas injection port. The flow paths in the upper reactor vessel shell are preferred. The flow holes connecting the upper head to the outer annulus are shown in Figure 3-17. The lower vessel assembly is shown in Figure 3-18.

l l

Hot Leg. The hot legs are made of 2-inch Schedule 40 steel pipe. Their 10 is i almost5.1cm(2in). The nominal pair of hot legs have a radius of curvature at I the U-bend of R=4d, which is slightly larger than the prototypical radius of curvature. An alternate pair of hot legs have an R/d of 11.5 (see Figure 3-12).

This alternate configuration is designed to study the effect of U-bend radius on phase separation. The lip at the inlet of a hot leg is rounded to reduce the form l loss. Despite such efforts, the hydraulic resistance of the model hot leg is

, still more than twice that of the prototype.

I.

l Pressurtzer. The model pressurizer will be equipped with a 3-kW electric heater

! and a relief valve at its top. The loop seal between the hot leg and the pressur-12er is modeled with the elevations properly scaled. Fluid / metal mass ratio is l

preserved.

I l Steam Generator. Two counter-flow exchangers were purchased and modified to serve as once-through steam generators. Each steam generator has 48 tubes. The tube i dimensions are prototypical, i.e., 5/8-inch 00 and 0.034-inch wall. The tube support consists of three half baffles. The main feedwater, the downcomer, and j the annular flow path for the steam are not modeled. The steam outlet is located j near the top tube sheet. Opposite to the steam line is the auxiliary feed

! nozzle. It is designated to spray cool only 3 of the 48 tubes to stimulate the effect of partial wetting. Metal masses, in the form of thick-walled pipe, are welded to the plenum of the steam generators to achieve the correct fluid / metal mass ratio (see Figure 3-19). The hydraulic resistance of the tube section is l about 1/3 of that of the prototype.

I 3-49 3

l I

_ - - _ ,____m--_-_.-,-,~ ,..- ---_-,,,,,- --

s 9

I

-Downcomer gShroud Vessel \\ / '

wali]

'N . '/

\, / Upper plonum

. ,N - '/ _

.N /

\ '/

\

[/ gsw Vent h valvo xy r~-.b Pivot ,

s -

, g n I x

-N f

' Arm

.\ ~

.N -

S /

\ /

\'

\ /

\ //

.N

.\

\

./' <

/,'

Downcomerj .

annular Figure 3-14 SRI-2 Vent Valve Design f

3-50 D

te -

_ . . . - _ . . ~ . -

- ~ . -

?

..[

o -

.- = - ,

=

- s J -

/ >

s

_ ~

.- 1 s  ! ,-

~

,>~

, n_. - - - , . . . wen ww u cc:c j - c == a COM m SM TV.C i ' '

V j s e,- j d

' D

' %u r r -

a T T 3L b-tt I k, l' n .  ;

d t p I h' . ,

p i 1- f J l j g- =-

,UM ,1% em -= d '

l j - no - w,s,=. -

j 1

p i

4 i

/i ,

-i o- 6== n ,e, c- .aw,2v =-

m ) j " ' * " "

h. I 4rY (ts./'

a, H

W

.' mi r I aow a .,s;h' 6 - *M, - i-ser k-

1. 9N i

h f **

.A I

' .., A '

j!

= [ .L. .' 7- -[^e"* m

- .m  ! .

-- ~a .m .!;.Lom

, __  ; , -m w- pgy 2.

-- ~-'s _4t_ - . , m %, e T *'* **  !

/

' ' ~ ~ " ' '

g

[g

g: -- -=% -

'i

$N s s

'ED~%%4 ]l-,i~,i iT ik mm:r

,j .- m '4 3 o,Q t-l'e ==

g i vne .-

..=t=== g e

rx Nk, ,,\.,. 8 '  ;

i

t. i i 3 .

c__ . t ,

u 4

Figure 3-15. SRI-2 Upper Vessel Components ,

4 i l1

.I!!  !  !

S- i

!:l h

/s> S 1

, N)C/,g s =

i s

{

n[\tifh.R+!& ./ }

. u

/

't ;

~,

~

46 4 er u

i z

as C

I

$ 1'h"  !

Qs is i;1m!*th 3

!L ".L. >

sfi s,a 7

h g  !.

3 2

?

s .. m 5

f I3 f r- y f j88

' I  !

s*

si eg 3

/* g l r* *

/ 1 w I  ! a i 's -

6  ; i2 V Y l e xx _

s .N s 3. _ . _ .

g,

\

T N \\ ,

/

\ ' '}L .

eff$.

fi =>

j7 ilIb ~

! t pf ~

x w\34 p% ,/ - -

x

'. agg=r.9__ / a ==

c __ ,_

rf ,  ;

N ~f i , q

.. h .

\ - w, . ,

s 3-52 l

• ses .DC.

f55%3 W)

' COO

- = 0 00 l

tm i o 4_ >

1. nW 1

w n

-tihfMJ 4:

[ ~. , j [ ]

n ,l t

'l .s

~v e _, 1 ,

~, _-s

- / l N

/

,/ \

' ./ \

!( \,

o I i' \

Q 7

we .. .-

p-#, g=_- . .,4p

" " ~

..k .

\

y .-

, L- h,;

j o, .,

g \ a- /

/

g # %_ 821 m jY

. >x ti ' . _ , -~.

t's N.-/

i;

• m A -

4! NM rd1H ri oiu

..: T4 .~. T.L.

w i

,  ; t t***

v5 '-4y ,

b -

l '

L . ,: e r- .= 3. m,-3

- ..m

.~v-- -..o--

4,.t.; *2;%= =,P Ptsa ** -. *Cir. / ~~*

. tr s /

/+.s ./ ' ' %: .V'1 / W s sz?N y',..W ye1 4' L. N,h. s p\ .o c -- a /

.\ ,. .

~ . ,. .~c

) ,Q,,:',.,.;%,,

i '. >

• o.~ . u. - u a

/

yg . ,, u.. / -g w,

Figure 3-17. SRI-2 Reactor Vessel Upper Shell

if500 * - et * *O e'gy-

. - .,ag

,s ,

%h h

4 I k tlm , Y ****' '

j AQ" . .. ,2.",4.

.{ / ,

j ,o , m r' i N

!  !! k  ! _f- - ~~T:

\ Gi' I _il i l ll ll 5 /5 Y..

um

?p 8/ l, i  ; w -

s3 '

/, aW li 0.7 j/'{c$ls s

l f

fmh >  ;;- ' ~

y h'

'\Y

.d A Lj -l-j l +-- ll 3

ll p,

w - . - t ,

i '

e ., tem B i - ,

m '~ s A

l e g \ ~

%-o . -

N ,

a n

..l

i

= 7 \QQ s, "t, g i, .n. E ,..~

s

,, 3 1

.~

/ ~

4

. , o...

y  ;

s tra ,

1 . ll n _,__ s i

/ = x

" l7.=

l -

• "* h%W- - '.~,-,

/ \

- - _. e.~ . 1.:T:

.- ' EC-E

~~  !

m yg t-i 3.is .~ n l i

a- ;l l

! mr =~h m r-

. ..,o H %- ,

f g~. : , gnu. = me a w

- s.

u os 1 emeh tehd" MOO escasona h Figure 3-18. SRI-2 Reactor Vessel Lower Assembly u

!;i  ; . ,,

, i i

@m L

L .

ui-g W n- g -\

\

/ .i w A 1  ;' .

N

\

,O' L ' 0 ~ ;, I 9 N 7,

A =

F T

s

- 'T- k- .

-,1 A _

e g(r 4TD g

iU' J .

7 r >

f f~  :: s 3;C / A '

s i<Ri , /Aa m u

- - '~ .

i t n e

?

~ l w.W-

- n P

. S i

s.

s

% r 0

V g t o

. =C* " - 5 R

=o2K r a

e n

e R.

c P

e o

w%"

~

o G

m L%

"* t a

e S

3~.,

O e

D 2 L -

E e I

[:. - g.,. .

1 k w  %

R S

- _=

,/ {

zf g

_2 n .

y&.1 sn _w 9 .

X'

_ l,) -

/ __

9.

! 3 N ^.

"m e

. i

=- ~;

r W

l[i

+1 IT w

s T=. $-

g i

+, i F

u g

~

.f . h

_ 1s 4:

- X f

- s

- ,~

A /

..N _ N

%gi,b,

[4, N .n\ o .

wre T. P I.h = ,

laAl lp . Y ,

~ ._

l I

, a

.y * .

,M

,w ,f / u

-s.

- 1 .

s

/ . #'\ k o[

N - 1C2 - -l m

i l

= $ , /

/y, 1

g. :. 'l 4 r

u

!i

. Li

?. " .d j

'r m B

L

, y J  ?'

m*t =. f..

2."

,\ .

/

Il* '  ;.' '-r n- @b,.

I

=

yE lt4 i s: < '

3' <,;,I  ;).  ! .!

, ,* .iji. l,ll k1i ; i

Cold Leg. Also made of Schedule 40 steel pipe, the cold legs have a 1.5-inch ID. A centrifugal pump with vertical axis will be installed in each of the four cold legs. The lopp seal around the pump is properly configured. With a stalled pump and an in-line venturi meter, the hydraulic resistance of the model cold leg is slightly over the prototypical value.

Reactor Coolant Pumps. Similar to a prototypical RCP, the pump is a centrifugal type with a vertical axis. The flow enters the pump axially and exists radially. Each pump is powered by a 1/2-h.p. variable speed de motor. It is capable of delivering 20 % of the scaled full flow. The motor speed will be controlled to stimulate the scaled coastdown.

Auxiliary Systems. They include:

e The high-pressure injection system o Break flow measurement system e Tertiary cooling system for the secondary steam o Proportional control system for the AFW e Pressurizer heater control system o Provisions for tube-rupture simulation Planned Instrumentation. Figure 3-295-1 shows the types and the locations of most of the instrumentation. Temperature (marked by T in Figure 3-20) measurements are to be made with standard thermocouples. Eighty TCs will be installed on the facility. Pressure (marked by P in Figure 3-20) refers to absolute pressure measurement. Absolute pressure will be measured in the vessel upper head, the pressurizer, and the secondary sides of the steam generator. P in Figure 3-20 refers to differential pressure measurements. They provide measures of collapsed water levels. Additional differential pressure transducers are needed for various flow meters. Altogether, seventeen differential pressure transducers will be installed on the facility. Conductivity (marked by C in Figure 3-20) probes will be used to detect the presence of either steam or ' froth.

3-56 i

O g  ; .gD-f ..

g g .:. i e  ::. O

}p .@ t .?" i t

1 e

i ggg,.y. ,-

l -=-

l ll>. Ik

\'

.- m , .

( ,g i ,, ... _ ,

e rm:,

p 14 [N , g, y rs , ,

5 U ir" '"~'"

y

_/ x

- > t om ,3 i

e -.o 1, r e.,

<;rmp ,i =TN 't,1

- -re_ 3

i. v .,3 o.

• E' E 2. 3 ,

[b b s- .

1. • ~. ,n7 ,

ec

] J, Tese.a ms -

.: o r-Q - n

• N=d

• -- T(wamus.av)

(. ) ( )

v __,. ._ r v,c _ .

I. t I

, , t-3_ p , ') ,_ ,

et s r,c -

25.s

>-r es , _ , ,

'.r ' a- ,

1 g.g;y: ra.y =~- -. -

.u mm- ~ sm-u.cr y g - s cm ma 1 SetE twee b 5'D'J-q'1 Fet vEMEL t ps.S Figure 3-20. SRI-2 Loop Instrumentation l

3-57

In addition to the four categories of instrumentation discussed in the last para-graph, the following flow measurements will be made: primary flow rate in each of the four cold legs will be measured by a low-pressure drop venturl meter; the HPI, the AFW, the secondary condensate, and the condensed break flow rate will be measured by orifice meters. Cooling water flows to condense the discharged secondary steam and the two-phase break flow will be measured by rotameters. The

, power controller for the core heater also provides the electric power measurement.

Data Acquisition System. The data acquisition system software will be similar to the system used in a SRI UTSG 2-loop facility (!). The major differences in this UTSG facility will be the largely increased number of data channels. This section l will describe the system in general and focus on the changes from a 2-loop j i acquisition system. The system is based around a Digital Equipement Corporation 4

PDP 11/34A with 32K words of memory, 64 A/D channels, 2 D/A channels, 4 floppy disk drives, an IEEE-488 interface, and a 9 track digital tape drive. Since the facility will have more than 64 channels of data, multiplexing the channels will  !

be necessary. This will be accomplished by using a Hewlett Packard 3497A Data Acquisition Subsystem. All thermocouple channels will be monitored by this device since it provides superior performance on the low level signals. The 3497A will connunicate with the 11/34 over the IEEE-488 bus. The 11/34 will provide trigger and active channel information and the 3497 A will return the voltage readings of all thermocouples in ASCII format. All high signal level transducers will be con-nected to the 11/34 A/D channels.

, Additional hardware will also be required for process control since the 11/34 does not have enough D/A channels. Examples of process control are secondary water level, heater power ramp, etc. The exact details of the process controller are still to be decided. One possibility is an Apple IIe that the lab has access to. It would have more than enough computational pcwer to perform the tasks i

required. Also it has more than enough A/D and D/A channels for input and output.

3.2.2 Scaling Approach The design of the facility is based on the scaling rationale presented by

, Ishii (!, 7). The scaling approach, however, was modified in order to account for the nonprototypical pressure (0.7 MPa instead of 10 MPa). Details of this scaling j approach will be presented in a forthcoming report and will not be elaborated upon here. Several remarks, however, are in order:

3-58 '

Scaling Criteria. The scaling laws for both the single-phase and the two-phase flow are summarized in Table 3-6. Table 3-6 shows that, for a particular quantity such as the core power density, its functional dependence on the length scale remains the same for the single- and the two-phase. The numerical multipliers, however, are different. This difference is the result of nonprototypical fluid properties. Should the model operate at prototypical pressure and temperature, these differences would disappear. However, the budget available for this experimental program is insufficient to build a 14 MPa facility. Scaling compro-mises as a result of nonprototypical operating pressure are therefore inevitable.

The fact that the single-phase scaling laws are slightly different from the two-phase ones means that there exists a " discontinuity" in scaling. We would like to i point out that this discontinuity is in scaling only. More precisely, it affects how the model data should be interpreted when they are extrapolated to a full-scale system. The physical experiments in the model will not show a discontinuous behavior when the primary system goes from single-phase to two-phase.

Application of the Scaling Criteria. The scaling criteria shown in Table 3-6 leave a facility designer with two free choices: the model height scale and the flow area scale. After due considerations of the two-phase flow limitation--i.e.,

the minimum pipe diameter allowed, the headroom in an availbale laboratory, and the program budget--we have chosen the following scales:

LOR " I/4 ; AOR = 1/324 = 1/(18)2 i

Since the facility is modeled after TMI-2, these scales translate to a hot leg internal diameter (ID) of 5.1 cm (2 inches), a cold leg ID of 3.8 cm (1.5 inches),

and an overall height of the model to be about 6 m (20 feet). Because horizontal I pipe lengths have much less effect on the thermal hydraulics of an OTSG 2x4 i facility, they are shortened from the appropriate scaled values. The primary motive for that is to improve the hydraulic resistance matching which proves to be unattainable for many components around the loop. The flow area scale factor is ,

applied not only to simple pipes such as the hot legs, the cold legs, and the l aggregate area of the tubes in a steam generator, but also to the flow holes in the upper plenum shroud. On the other hand, the flow area in the core is not scaled. We simply take whatever is left after a multielement electric rod heater is inserted. An upper core office plate with multiple openings is designed to simulate the correct core resistance.

3-59

Table 3-6 SCALING CRITERIA FOR SRI-2 Single-Phase Two-Phase Component height (Lj/Lo)g = 1 (Lj/La)R

• 1 Flow area (Aj/Ao)R = 1 (Aj/Ao )R

• I Flow resistance (fh+K)R"1 (I + K)R

• I l Pipe wall (as/

i aj)R = 1 (a3 j/aj) = 1 ,

l Core power (q)R=0.42Lff/2 (q)g=0.12Lff/2 Velocity U R=0.56Lgf2 U R=0.96Lf(2 Time t R=1.8Lf(2 t R=1.04Lf(2 Note: Subscript R refers to ratio of model to prototype.

The numerical coefficients results from the fact that SRI-2 does not operate at prototypical pressure.

3-60 ,

A core heater is sized to give a maximum of 5% scaled decay power under single-

, phase conditions. The power turns out to be 88 kW. Because of the difference in scaling, this power represents a 17% scaled decay power under two-phase condi-tions. Additional details of the facility are given in the next few sections.

3.2.3 Scaling Compromises Investigated by SRI-2 As far as IST support is concerned, the major strength of the facility is its ability to perform separate effects tests at relatively low cost. Thus, it is expected that important additional information will be obtained in the following areas:

e Interruption /re-establishment of natural circulation

--HLUB separation

--Hot leg flow regimes

--RVVV simulation e Interloop interaction and oscilliations

--Effects of heat losses

--Downcomer configuration / resistance 1

--Scaling rationales The fact that the scaling philosophy and approach is different from MIST will prove extremely useful. For it will be possible to determine the effects of scaling distorsions on several key phenomena (e.g., effect of hot leg diameter on interruption of natural circulation, effect of " internal" vs. " external" downcomer and related RVVV operation, etc.).

On the other hand, the restriction to low pressure may preclude investigation of the downcomer tangential resistance issue. It should be pointed out, however, that the result of the facility, in combination with the results from UMCP facility can shed some light on this question since both facilities are designed with internal downcomers but with different tangential hydraulic resistances.

3-61

Thus, comparison of the appropriate results from these two facilities may help resolve the issue of the impact of this hydraulic resistance on the transient.

The facility is designed with 48 tubes in each steam generator. Thus, multi-dimensional effects are expected here more than in the MIST facility. However, the instrumentation in the steam generator may not be sufficient to study that effect.

In summary, it is believed that the facility will help address and resolve at least two key issues (interruption of natural circulation, flow oscilliation) and will provide a mean to evaluate scaling distortions and compranists in MIST (e.g.,

external downcomer, atypical candy-cane curvature). Limitations inherent to low j pressure and incomplete instrumentation, however, will prohibit a thorough analy-sis of all the issues (downcomer density fields, multi-D, and feed effects).

I We discuss in Section 3.2.5 in more detail, the relevance of the test facility results to plant and/or MIST transient analysis.

3.2.4 Test Procedure 4

The test procedure is described in Table 3-7.

3.2.5 Application of Data Vertical Flow Regime Transitions. We use the Dukler-Taitel flow regime transition model (3) at a nominal pressure of 0.7 Mpa (100 psia). The bulbly/ slug transition is defined by the correlation:

Jg = 1/3[jf + 1.15 v I (3-1) where 4

2 v,= [g (o2 ,oq ),1/4 l (3-2)

• f at P=0.7 MPa we get:

j g= 1/3[jr + 0.17 m/s] (3-3)

! 3-62 i

I--_ .= ---_-- - -, - - - - -

- - - _ _ _ _- . . . _ . m _ , _ _ _ _ _ _ _ _ _ . . _ _ __ _ __ . . _ ___ _ _ _ _ . _ . . _ , _

i' I

I i

, I Table 3-7 i SRI-2 TEST PROCEDURE i

Type of Test Initialization Operator's Actions Expected Results t 1  !

1. Steady-state Primary system full. Start core power at 88 kW Environmental heat loss can be sing'e-phase Pressurizer half full. (i.e. 5% full power) calculated, natural circulation Secondary level at 955. Aim for primary pr. of 100 psia VVs should not be open. .

Forced circ. to equalize temp. Adjust sec. pr. until hot No problem expected. f

No leak. leg temp. at 300*F.  ;

! No HPl. Af ter first steady-state 3 s-s operations achieved. #

reduce power to 26 kW. Leak flow measurements at the end [

to be checked against calculated l Readjust sec. pr. to get f 300*F in hot leg. leak flow rate and prz. emptying Af ter second s-s, impose rate.

asym. cooling by having different sec. pressures (keep hotter of the two

! hot legs at 300*F). ,

j At end of test, open a leak i for a short while.

w VVs expected to open when upper head I 8

2. Two-phase natural Start single-phase NC w. 26 kW !solate prz. prior to draining.

i $ circulation w.o.

leak & HP1 Drain a prescribed amt. of water from a leak port.

Adjust sec. pr. to achieve s-s.

Af ter s-s, increase or decrease is voided.

VVs may alter location of thermal

, l l No HPI. core power to observe change in center thus complicating the .

HL flow regime. establishment of s-s. ,

Further deplete water inventory  !

Initial single-phase NC should serve i to test system response. l 1 as a repeat of part of Test 1.

3. Two-phase NC with leak Start single-phase NC with 26 kW. If s-s is not achieved, alter s-s NC may not be achieved.

& HPI initiate a "10 sq cm" leak. sec. pr. or core heat to 1

' Initiate HPI. attemt a s-s. There is only one primary pr. at -

' HPI flow controlled by primary pressure. which leak flow is equal to HPI flow. If energy is not balanced at I that pr., s-s will not be achieved.

4. Two-phase NC with leak Start single-phase NC with 26 kW. After sufficient void is Test offers a chance to check HPI-

& HPI Initiate a "10 sq cm" leak, created in the primary, start leak heat removal.

hPI to be manually controlled. HPI and set its flow equal to .'

that of the leak flow.

If s-s is achieved. .try alter the primary pr. by changing secondary pr.

j If s-s is not achieved, adjust j - core power to attegt another n

5-s.

Stop AFW and drain both SGs to j see if HPI-leak is sufficient j to remove the core power.

i 1,

-- - - . . .= ..

_ _ . _ - - - . = - _ . _ _ -

Table 3-7 (continued)

Type of Test Initialization Operator's Actions E wpected Results

5. Boiler condenser mode -- --

Test hos. 7. 3. A 4 can all be adapted to BCM.

If s-s two-phase NC is not success-ful under a given set of conditions.

its chance of success is even less for BCM because of the concern for core coverage.

Test No. 5 could be a multiple test.

6. Effect of VVs on HL VVs locked close. Upon s-s two-phase NC, note VVs are expected to open af ter

] flow regime Same initialization procedures as the HL flow regime then unlocking.

Test No. 4 remotely unlock the VVs. Steam diversion could result in Attempt a second S-s. change of flow regime in the HL.

Second s-s should be a repeat of part of Test No. 4

7. Effect of VVs on HL Same as Test No. 6 but with 88 kW. Same as Test No. 6. Same as Test. No. 6.

flow regime (higher core power)

LJ

$s 8. Feed and bleed Establish a low primary pr. two- Stop AFW. s-s operation expected.

• • phase NC with 26 kW. When primary pr. reaches No leak. 80 psia, open "PORV* and No HPI. turn on HPI.

Set HPI flow equal to "PORV" flow and attempt a 5-s.

9. Loop asymmet ry Repeat the s-s result of Test No. 4 Reduce or terminate AFW in the This test may induce some interloop 1.e., sym. AFWs and leak flow intact loop; make it up in oscil., which may have occurred equaling that of HPI. the broken loop. in earlier tests.

If s-s is achieved, reverse the AFW flow rates in the two SGs.

10. SBLOCA in CL suction Start single-phase NC with 26 kW. Basically follow OTIS Test --

No. 220100

11. Repeat of Test No.10 -- -- --

12, 5BLOCA in CL discharge Same as Test No.10. Same as Test No.10. Ef fect of VVs expected to be more pronounced.

13. SBLOCA in CL suction Same as Test No.10. Same as Test No.10. Timing of loop flow interruption with larger radius HLUB may be altered.

14 Tube rupture Same as Test No.10. Need more information to plan --

this test, e.g., can the steamline in the broken SG be independently shut off ?

?

The slug / churn transition is derived from, L

j=/g6[4060-0.22] (3-4)

Where L E is " entry length." Slug flow will occur in the channel if LE is smaller than the length of the channel (L=3.56 m). Using a diameter D of 0.05 m, we observe this transition for j=j+ g jf = 1.062 m/s (3-5) l Finally, the slug / annular transition is given by:

Jg =3.1( )1/2v, (3-6)

, 9 hence, jg = 7.29 m/s (3-7)

The plot of the flow regime transitions is shown in Figure 3-21. The boundaries I appear remarkably insensitive to pressure.

Relation Betweengj and jg in the Hot Leg. Using the same analysis as for MIST design (2), it is possible to relate j to g jp for given conditions of power, Qc '

subcooling, and RVVV operation.

The model described in Ref. 2 is based on a simple energy balance in the core at steady-state. If Q is the core power, then:

Oc" RVVV h7g + mHLV(h g -h 3 ) + mHLF(h7 -h3 ) (3-8) l j where 3h is the core inlet enthalpy (subcooled), SRVVV is the flowrate through the reactor vessel vent valve, 5HLV and SHLF are the vapor and liquid flowrates in the leg, respectively.

Assume now thaty S is the steam flowrate at the core exit and that a fraction F is diverted into the vent valves, then:

SRVVV

• f b y (3-9a) 3-65

10' - l Z SRI-2 l m

- Plant $

- -- MIST f l->(Annular)

Kl gChurn/ Slug Bubbly l F l

\

& 10 -- g i Churn / Slug

{ 1

/ El \ E

/ El E f Churn / Slug gl I, g

/ \ 2 l. A=

=l jj

/ g '

Slug g(Churn) j f

10 I I I III/ I I\ l i i 11I lI I il i II 11 10-2 10 10 10' i Jg(m/s) l Figure 3-21. Flow Regime Tr:nsitions i

s

' 3-66 i

_ . . , - - w - , ,-

5HLV " (I-f) b y (3-9b) and finally Qc " by [Fhfg + (1-F)(h g-h5)l + SHLF(hf-h3 ) (3-10)

Oc " by Ihfg + (I-flohl + bHLFah (3-11) where ah = (h f -h s ) is the subcooling.

Now introducing the superficial velocities:

m (3-12a) j g =A HL#g (1-F)

(3-12b)

LF If"A HL*f we get:

Oc O 1 ah

• ITT + F} (3-13)

Aoghfg fg J + f fog ohhfg) d f g

Finally, dividing by a reference velocity U n, in order to obtain nondimensional numbers:

• 1 ah * *

+N sub df (3-14)

O c"(TT+iiglJ g f where:

• QC (3-15)

Oc " 2A HLoh g fg Uo jg = j g/Vo  ; j = jf/U a (3-16)

(3-17)

Nsub =

Now both Q* and N sub are preserved according to the scaling rationale adopted for the design of the facility.

3-67

The model subcooling enthalpy, (hs ),, is given by:

(hfq),(ao/o) p (3-18)

(hs )m = (h ) , - (ah)p (hfg)p (ao/o) i 9

where the subscripts m and p refer to model and prototype respectively.

~

Numerically, this yields (100 psia vs. 1000 psia):

, (hs)m = (hy)m - 0.11(ah)p (3-19)

Following Ref. 2, we assume an inlet subcooled temperature of T=200*F (i.e.,

(hs )p = 170.3 Btu /lbm. This results in:

1 (h),=257.5 3 Btu /lbm (3-20)

I which corresponds to a (secondary) saturation pressure of 56 psia.

l j The model power is obtained by the scaling rationale as:

l 1/2 0,= 0.12 Q p

[V ([L ] (3-21) p m i

1 l where (V p /V,) is the volumetric scale factor (1/1296) and L,/L p the length scale j (1/4). Thus:

! Q

! Qcm " 5 0 (3-22)

.l Returning now to equation (3-14) we see that, since Nsub is preserved in the facility design and operation and since F.1, the' relation between power and superficial velocities are essentially preserved in the model. That is, for the same fixed fractional (normalized) power, the ratio of superficial velocities in the facility and in the plant will be essentially identical.

Equation (3-14) is plotted in Figure 22 for F=0.9 and Q* corresponding to 2% of full power.

1 3-68 i

_ - - . . . . - - _ , - . _ . . . - ,,.-,-,_-_,_-_mc ,-- _ . _ - - ,. - . - . - _ _ , _ , - . . .--m - . _ _,, .- -.,_

10 _

- = 0.9 Slug / Churn

- NSUB = 11.5

~

Bubbly / Slug N

Plant 2% N i

, N g gSlug/ Churn

's N \

_ \

a s \

E 10 -

SRI 2% \

\

-r -

~

\ \

\

\

\ \

\ \

a

\ -

\ l g }

\

\

Bubbly Slug g i

- l 1 1 I I I I I 1 i f f I i l II i l i i l!I l III

10-2 10" 10 Jg(m/s)

Figure 3-22. Predicted Flow Regimes in Vertical Hot Leg for SRI-2 and Plant 3-69 i

Hot Leg U-Bend Behavior. The simplest approach is based on a HLUB Froude number (2):

2 Fr - (3-23)

Now, the facility scaling rationale imposes:

L 1/2

= 0.48 j (3-24) j,= 0.96 jp(p"p) p thus, (Fr),=0.23([R)(Fr)p (3-25) m Two different R,will be used in the facility. Thus, we have numerically:

(Fr),=( 7)(Fr), (3-26) l l

This result indicates that the model HLUB Froude numbers bound the prototype HLUB l

l Froude number.

t l

In Ref. 2, a prototype (Fr) = 1.7 was computed. This would result in (Fr),= 2.9 and 1.0, respectively.

1 A different approach to HLUB separation can be taken by considering the residence l time in the candy-cane vs. the bubble migration time (i.e., separation). If the j former is larger, then separation should be expected.

Migration time = (3-27) whereas Residencetime=f (3-28) 9 Therefore, Migration time

'"ResidenceTime"b0 v, Ili (3-29) c n 1 + no separation i c 1 + separation 3-70 i

___. _ , _- - - , . , ~ . . ~ ~___. _ __. . _ __ . . , . _ __

We have:

(J (v,)

'm "g)(j,) ,p ,

),,**( (y)m ' ( 3

• Cp p (3-30)

Thus, C

m " (0.165) 0.058 pC (3-31) by comparison, CMIST = 0.07 cp (3-32)

Thus, the model will afford a direct comparison with MIST and will provide an intermediate data point between MIST and the prototype.

Flow Oscillations. In this section we consider the effect of a small flow perturbation in one part of the system on the flow in another part of the system. This is accomplished by dividing the system into a number of sections, or components, and regarding each section as a transfer function in frequency domain (i.e., using Laplace transformer). The diagram on Fiaure 3-23(a) represents the system viewed as a system of transfer functions.

Each " box" represents a transfer function which should be determined.

Consider a single component [ Figure 3-23(b)]:

l l

l The meaning of the transfer function is that given a perturbation 6wy (in Laplace I

space) at the component inlet, we get a response 6w2 (also in Laplace space) determined by:

3-71

SG W RV VV SG

_ DC -

j i

Figure 3-23(a). Transfer Function for the Primary Loops dh, G(S)  : dh2 Figure 3-23(b). Transfer Function for a Single Component A A dW i dW2 m G(S) -

n

\(

H(S) <

Figure 3-23(c). Canonical Transfer Function 3-72

(3-33) aw2 (S) " O(S) 0*1(s)

Therefore, if we know G(s) we can readily determine the response to any given perturtgation. For complex closed loops, such as represented above, the transfer function algebra, well-known in control theory, allow to replace complex topolo-gies with " canonical" loop with feedback [ Figure 3-23(c)l:

Once the perturbation point and the response point are defined, the stability of the system can be analyzed according to one of several techniques from control theory (e.g., Nyquist analysis). Thus, the key to this approach is to determine the component transfer functions.

When two-phase conditions are present in the primary system, flow oscillations may occur due to density waves oscillations (DW0). The phenomenon has been exten-sively studied for BWRs (8) and more recently for CANDU reactors (9).

The study of this phenomenon is based on the void propagation equation first derived, in the context of two-phase flows, by Zuber and Staub (1_0, 11).

The void propagation eqttation can be established starting from very general assumptions and the phasic mass conservation equations together with the drift-flux (i.e., Zuber-Findlay) model. In nondimensional form it reads:

+cf =(1-a*)Q (3-34) where a* = Cg(1-y)a (3-34a)

Z* = Z/L (3-34b) t* = tUg/L (3-34c)

Ck= (Cg j + vg )) (3-34d) 4 3-73

, Co rL (3-35e) 1 0"h fguo i

~h Ua is a reference velocity, L is the channel length, r is the vapor generation 3

rate (inkW/m). It is assumed constant along the channel.

Equation (3-34) can be written along characteristic curves as:

h=(1-a*)Q (3-36) along the curves defined by:

h=C (3-37)

In addition, the suming of the two phasic mass equations yields:

h=Q (3-38) where j* = Coj/U a (3-38a) also, let V*j = Vgj/U a (3-39a)

U*

g = j* + V*j = C* (3-39b)

Now combining equation (3-37) and (3-38) we get (along dz* = C dt )

dt* = U gQ This equation can readily be integrated to yield:

U*2(t*+t*)

g = U*g(t)eQ*t g o (3-40) where the subscripts 1 and 2 refer to channel inlet and outlet, respectively.

t is the " time" taken by the fluid to travel from inlet to outlet. Integration of equation (3-38) yields:

3-74

U*2(1 t ) = U*1(o,t*) + Q* (3-41) hence, from equations (3-40) and (3-41):

t =kLog(1+ ) (3-42)

Q U gg l

Integration of equation (3-36) along the characteristic curve yields:

s(t*+t)=s{(t*)e-Q*t o (3-43) where:

s}=(1-a (3-44) Finally, starting from the well-known relation j W = Aog g + Aofjr (3-45) it is straightforward to establish: Wi=U*is$-V*j (3-46) where

• Wj W (3-47) j = Aofuo Substituting equations (3-40) and (3-43) into equation (3-46) yields:

W*(t +t[)=W{(t) (3-48) This last equation can be Laplace-transformed: W2 (S) " W (S),-st o (3-49) 1 and hence 6O(S)2

                       = G(s) = e-st o                                      (3-50) 6h(s)g 3-75

Equations (3-42) and (3-50) define the transfer function in components where two-phase flow occurs. Physically, equations (3-48) and (3-50) express that the perturbationpropagateswithanormalizedvelocity1/th(note: the normalized length of the channel is 1). The mathematical form is also reasonable (and in fact required) since one can always decide to divide a channel into two parts with transfer functions Gy and G. 2 The perturbation at the end of the second part should then read: 1 l 6W 6W M = M 6W- = G2 (s)Gy (s) (3-W 1 6W y 6W 6W g j I where aW is the perturbation at th exit of the first part. Clearly, G (s)2 must have the same mathematical form as Gt (s) and furthermore we must have: G(sth)G(S,th)=G(S,t{+th) 2 y (3-52) This requirement is fulfilled by the transfer functions computed in equation (3-50). In order to complete the discussion we assume that the phases are incompres-sible. Thus, whenever only one phase exists in a section: E(S)" 2 1(S) and , G(s) = 1 (3-53) From the point of view of the comparison between the facility and the prototype, it then follows that the model will have the same oscillation modes, patterns, and onset frequencies as well as the same stability maps provided:

a. The topological structures are identical;

b. Thedimensionlessparameterst[areidenticalforeachcomponent.

3-76

For the facility under consideration, the topolog?>tc.1-cenF.CCtions 'are preserved and, by virtue of equation (3-42), the preservation of t*oreduces to the preser-vation of Q* and U*, which are both imposed by the sealing rationale. In Practice, however, heat losses in the hot legs will be different than in the prototype. These losses will induce a slightly different transfer function ther6. The distortion is not expected to exceed 10% but this remains to be measured.*

                                                       ~ . _ .

I i

  *N te that when Q* is small (e.g., heat losses), t = 1/U*y   g (i.e., independent of 1

3-77

              --                    -.      .,.      ,                   , . , , . , , ...,7..._--

( s WY , y yu ,

                                                                                                                                                                                  +
                           -                                                                                                              O 3.3 UNIVERSITY OF MARYLAND AT COLLEGE PARK (UMCP)                                                                         hc4   c v,

e < ,, a 3.3.1 Facility Description N, s V n , This section sumarizes' the final design of the UNCP 2x4 Loop components. ( gx \- Additionally,theloopinstrumentationanddataacquisitionsystemisdescribed.D ' Theoverallloopassemblyisshown(elevationandplanviews)inFigure.3-74andf _ q-is drawn to scale. , b ,, ,

                                                                                                                                \o Primary System. The Loop is constructed entirely of stainless steel and perates;                                                                   ,

at a maximum pressure of 300 psi. It'As volume scaled by,a ratio of 1/560, short in height and proportionally compens5ded in cross-sectiorb1'areaO burinth design,'q emphasis was placed on: e T h f

a. Flexibility--through modularity ,

l , -

b. Expandability--with regard to instrumentation 7

                                                   +
                                                                                                                    \

c. Visual observation--using many view port's. y y ,

                                    , (                                                                                          ., /
                                                                                                }
                                                                                                                               % ',7                                                 -

The Loop consists of a reactor vessel with core barrel and annular downcomer, two ' i t, hot legs, four cold legs, two once'-through steam generators, and a pressurizer. (ca ( t ., Theprincipalfeaturesofthereactorvesselandcorebarrelareshown,,id,

                                                         ~

Figure 3-25. id height. The The vessel is 20 inches in dhmeter and 50.25 inc/es( downcomer gap is 1.25 inches. The core b$rrel can be skirted to de reasedown-comer size; an increase in size wookl require replacement of the core be fal. TNe ' ' -

                                                                                                                                                              , u ,, ,

vesselhas19topheadpenetra$?onsfortheheaterrodsandforinstrumeh.'ation,g

                                 ^

and four view pcrts which allow visual observation.of the reactor vessel vent

                                                             ..4      N valves.                                                   (, S
                                                           ,       e                                                                                       w The core barrel has eight reactor vessel vent (ffapper) valves. Penetr.,4.tiodp are provided through the vessel which allow selected valves to be h[ld at My open                                                                        >

p' position or fully closed, if desired, by externally contro11e'd bat scre,ws. The valves open due to a different_ial pressure, and the onset of opening sand the dynamiccharacteristicsofventval,eactioncanbealteredlyusing'dichined,h.' , q wy ', weights on the flapper. The vent valve Mesign is shown in Figure 3-26. a .,&  ; w z , 3, e r 3 L

                                                                                                                                              ,ih'jb The heat addition into the Loop is accomplished by means of 16, beater rods measuring i inch in diameter and 49 inch total length. Fifthd5 of the M heateh3; .

N . p t

                                                                                                                      ,    *g-3-78                                  '\                                                                                  B
                                                                                     .                          s b      s l'y'                                                                                   y I i

e h x

                                                                                                                                                            .r f                    _,                        ,

l s j 4 ,. c , .- --r s ca.o , l ,

                                                                                                =                                            t e

t PUESSAFER l i + s E I \ 8F I \

                                                       -      -             ,       .                                                  a es-s
                                                                                                                                           '{'
                                                                                 '                                          i                                                  qsm 3

9

                                                                                                                                                                      -  p. ... a ...

6 El'J

                         ' '                                                                                           U::O.;:."'
                                                                                                                         ' ' ~ ~                          '

l 1 emo

                                                                                                   *^

us j;r .A7..:7 - F _.._ . f C

                                             ,(, I- -=
                                                              ~

f,

                                                                                     -              e- - o n .-.                ,                     ;

e EEN r

                                                  .y                                                         ,                        }                  ,j

[a av " 4

                        '-                                                                                                          exim aa
                                                            -    ~
                                                                                =_F      ;j A                                                                                         \     ]
                                                                                       ~\
                                                                                    'l }
    ^n l          p                                                                          V r       "l*
                                                  .=: :--' :::::.;.%

l-  :* -

                                                  . .-                                                      <<                    . i.

er s

                                                          -              Figure 3-24.                UMCP Reactor Coolant System Arrangement
                                                     +

s. 3-79 4 9 4 4 I. _ _

I i l Ill a

                                                .~

J

                                                                                                             /s aG t

y. (!

                                                                       , 4
                                                                                                                               . ), .1                t x' , '

s '

                        ~c'.

Ns sk5 A a, J /

                              ~                                                                               .

s

          <     i~                                ,,      J;.
                                       ~

_e . s 4_ \( s5

                                                                                                   .            _ rL JbY~'             ~           -
                                                                                                                                                     \
                                                                         ,02                                %<                   j s              '

r d e#\ e

                                              /                                                                        .R                                                          m o
                                              /                                             .

c n

                                   ,N                                                                                                   1 D

w o _ w]9\ \' d n a l e s s e V r o t c a e R L

                                                                                                , y                                              '                    -           P C

Fjb,

                                                                         -                                                                                                        M 1

bTL jt !I U

                                                                                                                                                      /

F .

                                          ~-

H . W 5 2 M6Q - 2

                                    -    P-          -

1 3 e

                                               ,3                             _                       -                   _                    -     _         , ^ _ _              r u

M 14 {\ 4 ,l _ ' -~- g i . F I-r - -

                                                                                                                        .         '            ,V           I:\

I i' -l.  ! , !ii bI _

                      ,i'                  l.                                       ,   l                                                        5 i

m ~ - s m/ TI m V'-

                                                                   ,p, p llnB                            .                                       l gl a a!

m

                    -              j                                      -                                                                    s             .J               r k                                                -                                                        f
                                                                                                                                                        /J!
                                                                                                                                                                /
                                                                                                                                                                     /

I I L x il

                                               '                                                                               ]             -

(

                                                                                                                                                                        ~       .

y= lll

                    " 1 1/4 "          =

3/16 "

                  /
                  /    \ _1/8" i~
                                         \

(./_ } \ x n

: _r__ m_ \

s n _ _ , N

                    - m/
     *     ~
                           / =I
      'r L1                      \
                  /O                     N r}

P

                                          \
                                            \
                  /

4 L

                             ~

S* Figure 3-26. UNCP Reactor Vessel Vent Valvc 3-81

are used during operation leaving one heater as a spare. Only the bottom 24 inches of the heater rods are heated. The total maximum heat input is 202.5 kW. Five banks of 3 heaters each are presently controlled jointly, but have the capa-bility of being controlled individually for arbitrary power input. The heaters were manufactured by Watlow, and the control unit was made by Holmer. The heater control unit is compatible with the data acquisition and control system. One of the heater rods is shown in Figure 3-25. The hot legs are made of 3.5 inch (I.D.) Schedule 40 stainless' steel (304) piping and 300 lb weld neck flanges as shown in Figure 3-24. Three view ports are included in each hot leg: one view port in the horizontal section near the reactor vessel, one view port in the vertical section, and one view port in the candy cane. The view port in the vertical section can either be positioned near the candy cane or it can be positioned near the bottom part of the vertical section of the horizontal outlet. Details of view port sections are given in Figures 3-27 and 3-28. The design of the view ports is such as to minimize any I dead water region or flow disturbance. The unique design of the view port in the candy cane enables the investigator to see the very top part of the hot leg. This is important as it is there that the first vapor bubbles will appear in transition from one-phase to two-phase flow. The view ports will be used for direct photo-graphic observations of phase separation and also for velocity measurement. High point vents have been installed at each of the hot leg U-bends and will be used during filling of the Loop and during certain tests studying hot leg high point venting or spraying. Also, the Loop includes a safety valve on the pressurizer and a high point vent on the reactor vessel. The cold legs are made of 3.0 inch (I.D.) Schedule 40 stainless steel (304) piping and 300 lb stainless steel weld neck flanges. The view port in the vertical section can be positioned near the bottom of the steam generator, or at the top part of the cold leg, near the center of the steam generator. An illustration on the design of this view port are given in Figure 3-29. Provisions for HPI injection and spray lines and hook up to a water treatment system are on the cold legs. The K-factor (the flow losses due to pipe friction and entrance losses) of the model Loop can be changed (increased) by inserting screens or orifice plates at various locations within the Loop. For this reason a small recess has been machined into every flanged connection within the Loop. l 3-82 l l

              '                                                                         i                        7
                                                                      ~-

(o

                     ... . p pgp;s;
                               . .g                 j.7                       _.
                                                                                    .-Q       o7 7       --            -

p,(Qg g q. _ . _.4 s i O' &TQ l

                                                                                                                                    ..M
                                                                                                                                                    'txv_

i

                                                                                                                                                                    /0 o . o ,//
                                                                      . _ ~ . .u.,      -                        -
                                                                         '                 o-             -e                                         ,o , o e

i' ____..._...____7_,... g.; f e,,._._.. ., , , ... ... ..

                                                          /                                                                         N
                                                                                                                                                             -. . . g
            .                                      /                                                                                        s I            .T7.              ~ * ***
                                            .,.               .s,                                                       ,

t . . .. .. . l >- , . Dl. 4-l _. '- ,

,                          i                     .
                                                                                                                                                ./       ,

l . .-' , . ! /i , i\ . -/ .\ l i* I I ~ ~) p _ . . . . R'ts

                                   -~                                                                         "'
                                                    .,'?                                    E
                                                                                                                                          ,,(_
                                           .s        '..' .             s, i

hh' ! m t- . T .em R-

                                                                                                                                              " --i i                                    .\                Ly'                                 ,             g     7T7-                                   ,        !
                                               \a                                                               21A                     . _ ._

] M' ' %'

                                                                      /                                         fV                                     ii, 4
                                                 ,;--A                                                         u.*.31,6                                il' S
                                         ' o ' g,%                                                  s x      \hll3 din;- - - I
                                                .g/i ,                              s
                                                                                                                'sh-                  a =7, --- '
                                       /
                                                                     +A .r
                                                          /

l Figure 3-27. Design Details of UMCP Candy Cane View Port l 1 3-83 t

                 , . . . _ _ _           . _ _ . . . _ . . _ . . _ .                            . . _ . _ . -y. . . .
                                                           -      V'h                                                                         .

A fh mp

                                                                                          -h$11 tfll- .-----
                                                                                               ,     w                         . _s _ _ .
                                                                                                                                  ,t
                                *#      t

• f.i _ _._ _ _ . -

                                %p      %                        1

• 5  % l

                                                                                                                                    -           i

h. .C' '.' 4 4 _. I i Gv9l_2

                  ${' -         _._-

ilij

             --tf }t t t} }}Q.                                    ,>\. .'titti              n.          -                             1 w,                                              , .
               -                                                                              N
                                                  /                                i                                                  '

N_. \ 1 L ._ /

                                                                                   \,                        -
,                                                 \                                 /

I I I _ __ . .. I Figure 3-28. Design Details of UMCP Hot leg Vice Port l l 3-84

! l l l l i l I i 3 i i Y y --- i s 2 g v e t')

                                                                            ~f ,             '
                                                                                                    'i g t                      g

(' f t 7 J h #

                                   ~f      .

f b  % 4 ' Id ', O cn as s y 2 a

                 ?                          s     A       '                                             !                            3 k%-

f, $. 4wY-. j - p "[ 5 l l,  ? k

           *k M .w.

g-- 41f.- t

                                                                              //. f_

e p

           ;;                   $                                                        I                               A     8 3        /f                                          #a#?

s h i_ _ , f+# 4 I1 g 4 _ 3 3 f . p' i7 ' _. 'll k O

                                                                                                                                  '?   E l                                                i      $                                                e                "
      ,  y                             .-     d
                                                ',1_1                        5         4                                     O ht-
            !                                                I r-          -' ,4,i2 3           gm .r
            -                                 ~

ut -4 8s> 4 cn

                          /                                                                                              0 a

l. .

                                                   -1 3a,        :j h, ,k                       l x

J._ J r  :. v i q, I L.

                                                               .                                             ef3                        g t9*

242 c 0 o 7 ) 10 Y n a 7tp 4({pg

                                        '#)2 j                                                             J                                                       I c ,6en 4v                                                                     4 n  *' d,u o o .J sn.        Q4                                                     e2 1 h *s 'n                                                        e 4 w ]{    .
                                        .#'4 8t d'  - e.

f

                                                                                                          .ayg-(El 13,. 4 ($

l 1 I 3-85

\ The two steam generators each contain 28 stainless steel tubes, 12.8' in length, 1.18" 10, 1.25" 00. The tube sheet pattern is shown in Figure 3-30. The outer shell is 12" Schedule 20 (12.75" 00, 12.25" 10) stainless steel pipe. The middle tube sheet divides the secondary side into two equal half sections. This feature was included into the design of the steam generator to provide mere flexibility in the positioning of the thermal center. The top and bottom sections may either be connected together allowing inlet main feedwater to exit the top section or be controlled separately utilizing water, air, or steam as a secondary fluid. There are four nozzles near the top of the shell side for auxiliary feedwater spray. The pattern of 23 thermocouples located in the tubes combined with chang-ing spray nozzle characteristics is intended to allow AFW spray effects to be examined during selected transients. The pressurizer is 12" Schedule 20 stainless steel pipe and is 52" in length. Connection to the hot leg is with a 3/4" stainless steel pipe. The plant typical loop seal is included in the surge line. Two electric heaters (2.5 kW each) are inserted near the bottom of the pressurizer. Spray, instrumentation, and pressure relief valve penetrations are at the top of the pressurizer. The spray line is connected to the nearest cold leg. Secondary System. The Loop secondary system is a closed loop system that pumps water from the cooling tower, through the steam generators and back to the tower. System pressure is approximately 45 psi, and can be increased to approxi-mately 150 psi by replacement of the pump. As already discussed, the secondary side of each steam generator is split into two equal halves. Depending on the configuration, the secondary system pumps water into one or two inlets on each steam generator. The inlet flow rate is measured using a vortex flowmeter. Inlet and outlet temt.eratures are measured using thermocouples. Auxiliary System. Provisions have been made in the UMCP 2x4 Loop for the following auxiliary systems.

1. High Pressure Safety Injection--The HPSI pump takes suction from a temperature controlled water supply and injects into the cold leg.

3-86

g. e g..

_ l ** l**;  ; l"- l"

                                                                                                                                           ;;_l$             l'__ I"_ _1 "   _ I"    l"
                                                                                                +                                                                                                                                   a    o N[   /L   x h

g

                                                                                                                                                                                  '     '    i
                                                                                                  % '-         I        I      I         I l'4                                               -                                                                                                      g and p ugged                                                                    k                                                                j                                   =
                                    %=>                                                                    GMMM).e.e-                                                                                                 .$

A (T ( N

                                                                                                                                                                                                  -                          o   -  e i'/ 4oggg.L _                                                               ):                                      j'             D
                                                                                                                                                                                                                       =

in t' i - nom O (TYP.) \ I_ (Tj ~ (',] * (* +j , (+ 1. Plate is !" thick 3'4/ ,, 1 '/4 ,$

                                                                                                                                           ],[
                                                                                                                                                                                                                           ~

2. Plate #own is THREADED -

                                                                                                             ~
                                                                                                                                        -t-
                                                                                                                                                                      ] [+                 THREADCD                          n *o                *#          #

and plugged ( and plugged

3. 'Ihe center plates P."ORit 7 L (TYP.) k { " are carbon steel N) N+) s v

4. The end plates are stainless steel o

8" _ IG" Figure 3-30. Spacing of Tubes in UMCP Steam Generator

2. Auxiliary Feedwater--AFW is sprayed at one of four locations, at the top of each steam generator.

3. Blowdown Collection--For the SBLOCA tests, the leak flow is collected and measured.

Instrumentation and DAS. A design requirement for flexibility and rapid modifica-tion times was invoked on the loop instrumentation and DAS design. A relatively large number of penetrations have been installed in the loop to allow access for pressure and temperature measuring instruments as desired. Additionally, the DAS has the capability of being extended, if required, to handle approximately five times the present number of inputs. This flexibility in design is believed to be j very important to the successful performance of loop tests, j The main data acquisition system is a Hewlett-Packard (HP) 3054A Automatic Data Acquisition System. The 3054A is a computer based system consisting of the HP 3497A Data Acquisition / Control Unit interfaced to the HP 9816S desktop computer. The 3054A measures and analyzes data from the loop instruments and also provides control signals to the loop components, as required. Important features of the 3054A include full automation, on-line data analysis, precision measurements, and high speed data / program storage and access. The computer is also capable of interfacing with the University of Maryland main computer (UNIVAC or IBM) for additional off-line data processing, as required. The 305CA includes a number of software programs for data acquisition and manip-ulation. When preparing the loop main codes for instrument readings and data analysis, these programs will be easily used as subroutines. Having these soft-ware routines available has eliminated the writing of the basic data collection programs. A second DAS, consisting of an Apple IIe computer with D/A converters is used for heater control. This system will be interfaced with the main HP DAS. Local temperature measurements are made using sheathed thermocouple asssemblies. The thermocouple will extend out into the flow to give one temperature reading at that location. In addition, four element thermocouple rakes are used to measure temperature profiles along the centerline diameter. One rake is in the top of each U-bend. The other two are in two of the four cold legs, in the horizontal section near the vessel. 3-88 i l

The steam generator (primary) has, at present, one five element rake, extending from the top, in one tube. The rake consists of five thermocouples spaced one foot apart. Two additional vertical rakes and several single TCs have been selected for installation very shortly. Also, horizontal rakes, measuring selected tube inlet and outlet temperatures will be installed. Pressure measurements are made around the entire loop using electronic dp cells. A pressure measurement can be made at each thermocouple location. Also, four cells will be positioned radially around the vessel and will monitor downcomer pressure. These dp's are used to measure circumferential fluctuations as a result of asymmetric vent valve opening. Absolute pressure measurements are located in the pressurizer and reactor vessel. These readings will be used as heater power trip signals and as auxiliary I system (such as HPI) trips. A technique to measure the expected low single-phase flows in the cold leg is presently being investigated. Consideration is being given to hot wire / hot film, heated T/C, LDA, tracer techniques, and micromotion. Table 3-8 shows location of the temperature and pressure measurement points on the loop. Flow measurements can be made at any view port location. Instrumentation locations are shown schematically in Figure 3-31. 3.3.2 Scaling Approach This section presents an overview of the scaling rationale used in the design and construction of the facility. The overview includes the basic equations and parameters for both one-phase and two-phase scaling, and continues into proposed simulation of transients using the facility. The determination of the transient simulation initial conditions and boundary conditions are evaluated. It is important to note that the transient analyses to be presented are considered by the UMCP staff to be the most feasible approach to employ at present time. The UMCP staff has reviewed a recent TRAC calculation for a plant and assumed this information to be the best available data describing plant SBLOCA behavior. As such, conclusions to be derived in the following analysis is larglely based on the TRAC calculation. It is conceivable that, in the course of the experimental tests, the analyses will need to be somewhat modified. 3-89

S L E N ) N S S ) A U E V H L S K R C P U A - L R 4 0 ) P / 4 8

                                                                                 )
                                                                                           )   W                   P
                                                                                                                      ,  )

2 2 8 4 ( 2 (6 O (

                                                         -                       1             4)                  O S        (         L(2                     L     T Y               G                  E      ()         )       -   N T               E        S         S .       4       0  )  0     E I               L        G         SNR(         )    1   5 1     M L             s          E         EEE           0   (  2  (     U I            '  D L                VGZY0                (        R B            CL                           IR1        S     S     T A      /OSTSRMRA(                                    E   S T     S P            TCEOEOTUD                               K   L R     N A                    KHKTSSNL                        A   L O     I C      DRA AC                             SOA        R   E P IURORAOECT                                  CWW S            UO W                  EWREO             C     E     O n

A LF2T2RTPST / P I L D F- - - - - - - T D V F i o E t K a _ A T c R o L T, C n T T / y/ P o T, / i t a a / t 4 N" ,'

                                                           ,=

n e i

                                           - /             ,=

m u r

                         .                                                                                                       t T,                    -                                  ,-                          ,

o s n ~ P .- ,= I

            )                                                                                                                     p

( o

-                                        -          _               j                 )               )

L P o 4 4 - T, ( ( C M P T, P T, P T, P O T, T, U P P N J 1 3 3 e r

                                                 -                                                         -                       u g

i F lT, E)l P KT lT AM 7, ., RL ,p T, T, E T. C y P

                                                        /4 T(                                 J"                - )

E _ - K A ~ R _

      )

CT _

   /M                                                              ,

TL E 4 ( w1O il! .ll i l l l l 3 j ,  !! i

Table 3-8 UMCP INSTRUMENTATION LOCATION THERMOCOUPLES-Cold Leg (each) 9 - SG Outlet to vess1 inlet Hot Leg (each) 4 - Vessel outlet to SG Inlet Vessel 3 through top head to inside of core barrel 4 radially at two elevations Rakes 4 element at each U-bend 4 element in two cold legs Steam Generator Primary (each) 5 element rake (6' long, from top) in one tube Planned: 5-10 element rake, extending entire length Additional rakes (horizontal and/or vertical) Steam Generator Secondary (each) i Inlets and Outlets DIFFERENTIAL PRESSURE CELLS Capable of measuring pressure at each T/C location Present: AP around entire loop Hot Leg (each) 4 between vessel outlet and SG inlet SG (each) I between inlet and outlet

Cold Leg (1 of 4) 2 between SG outlet and vessel inlet Vessel i between vessel inlet and outlet 4 radial for RVVV opening 1

3-91 +

  , , - , - , - - ,,s-    --      -    -

re, -

                                                     ---vv       -r    .   --       v,    --

• w m- -.---+-m- w- e ---e >=- - - --

The overall scaling logic for both single phase and two phase begins with consid-eration of the fundamental laws of conservation of mass, energy, and momentum. j for single-phase natural circulation, these equations relate the balance of . buoyancy forces, due to density differences, to flow losses in the loop, due to friction and form losses. 4 The two-phase flow scaling logic was developed as follows. Using the conservation equations, the usual set of nondimensional parameters, such as Froude number, sub-cooling number, and so forth, are developed. Using these parameters, values for each are calculated and compared to those of the prototype. The comparison shows that most model values compare well (factors of =1-2) to the prototype. There-fore, this indicates that, during a transient, the flow regime for the crucial component (hot leg, downcomer, etc.) should be preserved. , 4 1 Transient analysis is an essential link in the effort to simulate high pressure

(=2000 psi) transients with low pressure (=300 psi) experiments. The problem does not exist for MIST, which is a full pressure simulation, but for any lower pres-sure systcm, the following challenges must be faced

1. How is the " pressure scaling" performed?

i I l

2. How are the initial conditions, including initial pressure and subcooling, set?

3. How are the trip points set so as to properly maintain the main features

) of the transient?

The one-phase and two-phase scaling is more fully brought forward in the next sections, and is followed by the analysis of transient simulation.

One-Phase Flow Scaling. The volume flow rate at any cross section in the loop can be deduced from appropriate statements of conservation of mass and energy and from basic thermodynamic identities. As shown in Ref. 12, the flow rate or volume-power relationship for steady state is 1 l Q=[ ]1/3 (3-54) 4 i 3-92

while for the non-steady case 1/2 #(fok*)(aogr.) t) [ Q(t)=( ) tanh g, (3-55) P where K K K K 1 2 3 4 5 K* =A 7 +A 3 +A y +2Aq + Ay g 2 3 4 5 v v V 2v 4 v i 2 3 5 M* = Ao[7A 2+ 7A 3+ 7 +4 (2AA 5)2 + 7! g and the indices 1, 2, 3, 4, 5 refer to the core barrel, the hot leg, the steam generator, the cold leg, and the downcomer, respectively. Equations 3-54 and 3-55 can be used as scaling equations. For the steady case the scaling becomes: 0 hprototype,[J)3(b)p[M)[1m) 20s (3-56)

                                                                                      *         "p q model                    Om (h),              K The ratio 1,/t which refers to the difference in elevation of the effcctive thermal centers of the core and the steam generator is fixed by the maximum build-ing height available in the laboratory in which the model loop will be housed, and a realistic estimate of the prototype. With this ratio essentially fixed and the apparent limited possibility to vary, the K* ratic (only increases in K* are possible), the only other adjustable parameter is the volume ratio Q /Qma                                                               p Selecting a volume which is too large leads to large power requirements (expen-sive!) and also to large diameter components (also expensive). On the other hand, selecting a volume which is small leads to small diameter pipes for the hot and cold legs and a very narrow downcomer. It is here that two-phase flow scaling arguments enter the selection in addition to economic considerations to stay within the given budget. A volume ratio of Q /Q,=                                                      p 500 was selected. The thermal l

power ratio gbl9mtherefore becomes 3455 if the pressure in the prototype is 600 psi or 1209 if the pressure in the prototype is 2100 psi during steady or , quasisteady natural circulation. 3-93 4

 - .- n,,-...---   , , , , . - - - --,            -                    , - - ,            ,-.n        . , - . , - - - . -  - - - - , - - , - , , . - . _          . , - . -,e,->- -

For one-phase non-steady state natural circulation, Equation 3-55 gives a time constant t to be modeled. It is seen that t model/t prorotype becomes: y g g t model g rprototype,[q)[_g)[_Q,) (3-57) y where

• V V V 2v v 1 2 3 4 5 V =

lgAg + g A + 7 A + (2A )2 + A7] 2 3 4 S Equation 3-57 shows that volume scaling every component individually in the same l

;    proportion as the overall volume scale leads directly to identical time constants for given K* and Q ratios.

i l A maximum power output of the model core of 200 kW and a model loop pressure of f 300 psi will permit simulations of the natural circulation in the prototype start-ing within 2 seconds after prototype shutdown from full power, assuming the q prototype pressure drops rapidly to 600 psi or a simulation starting at 4 seconds after shutdown, assuming the prototype pressure remains at the operating pressure level, i Two-Phase Flow Scalinq The Significance of Similarity Parameters 1 This section discusses the relation between similarity parameters and the conser- ! vation laws with which the flow phenomena can be described and predicted. This relation is best demonstrated on hand of a simple example, see Zierep (Ref. 13). Consider two-dimensional, steady, viscous, one-phase flow. Analysis based on the l fundamental principle of conservation of momentum yields the so-called Navier. Stokes Equations: uh + v =-fh+v( 3x

                                                         +

ay

                                                                )                                                     (3-58) 4

, 3-94 i l

and h+h=-fh+v( ax

                                                                                                          +

ay

                                                                                                                 )                                                                   (3-59)

These equations can be brought into dimensionless form by introducing the following dimensionless variables: i x' = { , y' = { , u' = h , v' = h. and p' = 3pDref - (3-60) j where i and a are suitable reference lengths, U is the free stream velocity and apref is the pressure differential between two suitable points in the flow field. Equations 3-58 and 3-59 then become: l

                                                                                  ~

i t au' 1 2 av' APref 12 3 , v t a 2y, g 3 2 7 , p + qa v, p = U ou,2 7 o

• U= (i ax,2 + 67 ay,] ay 2 (3-62) 1 j In Eqs. 3-61 and 3-62, three dimensionless parameters occur, Ut op v and (3-63)

                                      =1 = f                   2=                                           "3= ou, j                                The first parameter is the geometrical scaling factor; the second is the Reynolds number, and the third is Euler's number.

j Equations 3-61 and 3-62 lead to the following similarity and modeling state-ments. Two flow fields which have the same dimensionless parameters at, "2' j and v3 have the equal velocities (u' and v') and pressures (p') at the same reduced coordinates (x' and y').

]

2 Equations 3-61 anmd 3-62 also show the importance of the different dimensionless i parameters; for instance, a large Euler number points to the importance of the pressure forces in modeling the flow field. l 3-95 l

The above development showed that a group of similarity parameters can be derived from a basic conservation equation. Conversely, the use of this particular group of similarity parameters in modeling infer that the particular conservation equa-tion is applied to model the particular phenomenon. At first glance it may appear that developing all scaling parameters necessary for complete modeling in thermal-hydraulics is automatically accomplished by following a method such as the one shown above. In reality, however, much physical insight is required for meaningful and correct modeling. First, the most important or dominant physical phenomena have to be recognized since otherwise the results are either wrong, or trivial, or so complex that they are useless. Second, the phenomena have to be correctly described and related to the conservation equa-tions. Third, the so developed dimensionless scaling parameters have to be correctly interpreted as far as their meaning of the individual terms is concerned (e.g., the characteristic length in the Froude number may refer to the pipe diameter, or the pipe length, or the liquid depth in stratified flow, or the bubblediameter,etc.). The similarity criteria for two-phase flow during natural circulation and the modeling of small break LOCA has been treated in depth by Ishii and Kataoka (6) and by Zuber (14), respectively. The similarity groups are: 5 = Flux due to Phase Change Phase Change No. Npcbg = du aHgoog inlet Flux Subcooling No. N 0"sub ao " Subcoolinq sub " AH Latent Heat 79 { 2 0 Inertia froude No. N F r = gg t 4 >, oM = Gravity Force V Drift Flux No. N d= (or Void Quality Relation) o Density Ratio N, = og/o 3-96

1 1 1 + aox/o a Friction No. N g = h (1 + aux /u )0.25 () g Orifice No. N g

                                                = K(1 + aox1 j.5 , ) [ )2 Time Ratio               T{=(      )

50 Heat Source No. Qsj=(oC3 p3u naHsub ) I

It is readily seen that the above characteristic dimensionless groups cannot be l used simultaneously as scaling parameters for the entire system. For instance.

l for the heated section one cannot simultaneously satisfy Froude Number and Phase Change Number scaling. For the case of determining rate of loss of coolant, the ! scaling becomes somewhat more complex and more oriented towards the scaling of specific flow phenomena in the pipe before the break as shown by Zuber (14).

;            The dimensionless two-phase scaling parameters and the resulting system character-l             1stics implied to the UMCP loop is given below.

I

Liquid Inventory 1

J In order to preserve the liquid inventory ratio in each loop, then: Vold fraction, ajp = 1 f Volume, ViR = 1/500 Vol. Flow Rate QiR = 1/500 Subscript R refers to ratio of model quantity to equivalent prototype quantity. ] i ) ' ? 2 3-97 i

, Geometrical Dimensions: Since ViR = 1/500

1

! ViR"LAR iR alR " TD6 L 1R i Then LiR = 0.212 for H.L. i

!                                                  L1R = 0.175 for C.L.

LiR = 0.308 for S.G. LiR = 0.167 for 0.C. To descirbe system natural circulation, the thermal center distance (L) is { needed. A nominal value of LiR = 0.3 is selected. ] Velocity. In Ish11's (6) scaling principle, UR=/Qisbasedupontheassumption that the velocity is an independent variable. In a natural circulation loop, ] j velocity is determined by the buoyancy force, inertia, and resistance and thus

;                                        cannot be independently determined. The only velocity that can be independently controlled is the HPI velocity. Thus, we should consider both natural circulation velocity and forced velocity.

i (1) Forced velocity: l l To preserve inventory ratio, then i i 1 1 1 OR " T6g , a 1R " 165 g O 1R d -"l 1R " a1R iR i t (2) Natural Circulation Velocity (Bouyancy-Driven):

3-98 I

_ , - . . - . - . . . _ _ _ _ ,- - _ - _ . - - - - - - - - - _ - - - . - - - - - - . ~ . - - - . - - - -

Consider the steady-state condition, where the inertia term is neglected, assuming homogeneous flow, then aofg l gg = f [L j!o,(inhotleg) o ,= (1-x)ofg + og = (1-a)ofg 2 Dgofo aDq d"i fo, " (1-a)f l

                /a  R Op diR " (1-a)Rk l

Thus

jiR " k II IR=1 and R"1 If the whole system resistance is considered

(

• 2 2 i

aofgL th9"Kd8m"i i Then, Ja,R " #lth,R if (K a)R " I Froude No. (based on length): i Fr g= { } 4 ) For forced velocity, 2 L (FrL)R * ( )R"LR 4 1 3-99

                                          , .. _ ~ .          . _ . -,_ _ _ _ _ _ _ . _ . _ _ . ,.

i for buoyancy-driven velocity, j 2 (Frg)R = ( )R

                                                                     =1     if jR " #k
                                                                     =[D    =2 if jR " #k R

t i Froude No. (based on diameter): Fr f r flow pattern D= 4 For forced velocity. L 1 , (FrI = = 0.8 l DR=#

 !                                                                        R
                                                                                '(h) i

For buoyancy driven, (Fr)R"I 0

II I " 'k R

                                                                     = 1.4           if                  jR " 'k Drift-flux No.:

Ndrift " Yg j/d Vg ) = 1.4 ( 98")I/4 1

                                                                            *f a

R(Ap)R l[4 V gj,R*{(opp)2 l o R= = 2.09 (of)R=1.259 l aoR = 1.249 1 ' y gj,R , [2.09 x 1.249l1/4 = 1.13 ) (1.259)2 l 3-100 i { t

t 1.13 N Drift,R " J R JR= /Q is the most appropriate for phase separation and NDrift.R = 2.3. Phase-change No.: l 49oo l C aV gg [ yPCH , du aH M . D. AH fg l n fg og og i . 4 e N PCH is a volume ratio preservation parameter. Thus to force NPCH,R " I' I 9R(AYfq)R N j PCH,R"%(aHfg)R i

;                                    At Pp= 1000 psi, P , = 200 psi                                                                 (aVfg)R = 5.35 (aHfg)R = 1.299

) (aHfg)R 1 1 OR = g.299 g = 0.485 x 10-3 Q R*I*(avg l Decay Power 5% 2% 1% i 4

!                                    QP,kW                                                         10x10                              4x10 4                          2x10 4 i

gm,kW 48.5 19.5 9.7 i 3

Subcooling No.:

!                                                                    aH i                                                        N sub g sub . aH fg o g 1

(ao)g=1.249 , (o)g=0.195 g (aHfg)R

• l*299 i

j We force (Nsub)R

• I' i

J 3-101 l

     ----m. , - . _ .n-   m_,.--      ---,--e, , - - , - , , -  .     <------w-=..-,4+-=                       ,,-~ew.m..ee-yr-~c-,              ,-,*-,-,rw---,_-..---n                 or, - - ,m.,-- p - ..g,---- -~-.--m,_ry,

I I (aHf ,)RI#g)R (aHsub)R" (ao)R Time ratio (for bubble migrating in U-bend): i ' rise D/Vej j '

• travel curl j O dR R 0.1 0.5

("t)R"(Rcur}R gVjR Friction No. Np : j Mp = K*a 2 i (

N FR*(K*)R(ag/

l For H.L. , ag2 = 10-4 l 6.58 FR = K* x 10-4 =7.57 x 10-2 x 10-4 = 0.87 ! N l l 4 1 Note: K* used were based upon single-phase calculation. Secondary adjustments need to be made if Martinelli 2e-multiplier is used. i The approach taken in this systems scaling logic is to require the void fractions  : tobethesame,(a),=(a),. The relative vapor content within the loop is determined by 4 phenomena during the LOCA test, simultaneously. i I

1. The vaporization due to depressurization.

l

2. The vaporization and consequent energy loss due to inventory loss.

3. The balance between the Decay Heat input, producing steam. l

4. The condensation of steam during HP! activation, f Consequently the pressure histcry is characterized by the following transient  ;

sequences as described in the next section. l l 4 i i 3-102 - I I

3.3.3 Scaling Compromises Investigated The MIST scaling compromises to be addressed by the UNCP Loop are:

1. Interruption of natural circulation--Hot leg separation

2. Establishment of Boiler-Condenser Mode--AFW multidimensionality

3. Reestablishment of natural circulation--Downcomer flow and density fields; RVVV simulation; flow stratification

4. Long-term cooling--Downcomer flow and density fields and RVVV

5. Interloop interaction--Steam generator metal; downcomer flow and density l fields; downcomer tangential resistance; RVVV simulation; flow stratification

6. Combined primary and secondary side blowdown and SGTR--AFW multidimensionality The UMCP 2x4 Loop intends to address all of the above issues. However, because each issue involved specific physical phenomena, the controlling parameters are different for each case. In Table 3-9, the dominant dimensionless groups for each issue are listed. For planning tests to address each atypicality issue, partic-ular attention will be given to these dominant groups to ensure that the ratio of these groups, in model scale to their respective plant scale, is close to one.

I The atypicalities can be addressed in two ways, sensitivity studies within a given test facility, and interfacility comparison of results obtained from different configurations and conditions. The sensitivity study within a given facility can assess the effect of the variation of a given parameter (e.g., the flow distri-bution between loops). Since one parameter can be varied at a time, with other configurations and conditions being unchanged, a true parametric sensitivity can be achieved. However, since the ranges of vairations are limited, it is difficult to extrapolate to a wider range. For those areas where wider range variation needs to be covered, interfacility comparisons must be done. For example, if one wished to determine the pressure effect on natural circulation resumption, the three facilities can yield data at 2000 psi (MIST), 300 psi (UMCP), and 100 psi 3-103

i Table 3-9 DOMINANT DIMENSIONLESS GROUPS FOR EACH SCALING COMPROMISE (UMCP) , Dominant Dimensionless Issues Group Function Location j H. L. Separation: Fr Flow Pattern Hot leg ' s Flow Pattern Hot Leg N PCH Vapor Generation Core N 0 rift Phase Separation Hog Leg N, Time ratio for Hot Leg separation U-bend Section Boiling-Condensation Mode: NPCH Vapor Generation Core 1 M Sub Vapor Generation Core N PCH Condensation S.G. 4 Nu,Cond Condensation S.G. j Pjet/P o Jet-spray in AFW AFW S.G. j Re, Nu Flow in tube in each tube l Re-establishment of H.C.: N PCH Vapor Generation Core, S.G. Qg3 Heat Source Core, S.G. Fg Flow pattern H.L. j F L Natural circulation H.L.  ! i i l Ti Time ratio for H.L. I stored head for i Gr Two-dimensional flow H.L. l Long-term Cooling: NRichardson Natural Circulation System Loop N Fric ion Resistance System Loop j orK}yss KO,C Internal Circulation D.C. i .i , 3-104 i

Table 3-9 (continued) Dominant Dimensionless Issues Group Function Location Interloop Interaction: N PCH Vapor Generation Core Nsub Subcooling Core Inlet T*j Transient response S.G. and other parts Q39 Heat Source S.G. and other parts NResistance Resistance Loop NuCond Condensation on D.C. and C.L. We Drop Size D.C. and C.L. j Fr0 Flow Pattern C.L. ) I J i i j i i 5 l j 4 i 1 3-105 l l 1

    .__,..._,.__.-_--_-_---.__...._.---e.-                                 . . _ - . - - - , , . , _ . - - _ - - _ - - . . - - . . _ _ . - - . - . . . . . - - - - _ - . , _ _ _ _ _   .

v r, ,

                                                                       .(-

s. s , ,

u h T ii (SRI) for comparison. If one wishes to compare hot leg j'iameter ef *ect on[p%Ie separation,thethreefE111tiescanyielddataof3-1/2'in.'(UMC));f1[t.'(IRI), and 2.63 in. (MIST). t k s )u

                                                                                     ; s.             t      's 3.3.4    Test Procedure                                                                

Simulation of Transjents. The plant simulation by TRAC is u:ed as the baMs for

                                                                       \

the transient simulation of the UMCP facility. , i 1 \ i Thepressureversustimedecaycurveconstitutesthefrair[worktodescribethe l various phases of the SBLOCA event. Three regions are id7r>tified: i ,, T

1. Rapid pressurizer-dominated depressurization. This first transient initiates when the break is initiated (t=0) and, terminates when the HPI j flow is started (t=2035 sec). At this time the pnssurizer I

is assumed to 4 be about empty. Note that tire pressurizn i level'versus time curve is l \ linear in this first period. The time whhi the HPI flow initiates is 't ( given by the intersection of this straigh$ line, dita the zero level s ( line. In the UMCP Loop, the HPI flow will ' be L ltiated when the pressur. V N-  ; izerleveldropstoalevelthatcorrespondstibi . void scaled to 1/500' with respect to the plant. During this period ths primary loop is coni sidered liquid solid. , g ss

2. Natural circulation instability transient'. The next laadrark of the g plant simulation computation is the initiation of the Boiling 'i Condensation Mode in loop "A" (t=1960 sec). In the 1100 sec.onds -

( ' following the BCM in Loop "A", the HPI flow and the break mass flow rate s ., equilibrate. Consider that the HPI flow in the UMCP'Ltop is scaled , (1/500 ratio). Therefore, this second transient i.s ccnsidered to last s

                                                                                                                            ,        i.k until the break mass flow rate equals the HPI flow rate.          \

3. Establishment of Bolling Condensation Made s

This last period initiates when the break mass flow rate reaches steady state and continue until BCM is developed in both loops.,s In summary the trip points are as follows: 1 Break initiation 3-106

t (' i 2 End of the pressurizer level linear drop i 3 Break flow and HPI flow equilibrium 4 BCM in both loops Y In the following, the boundary conditions and the control strategy for each period s( are outlined. Ffqures 3-32. 3-33, and 3-34_ identify the trip points on the plant ,, simulation curves, g , f ,

                        ,    Rapid Pressurizer Dominated Depressurization. In this transient, the primary loop s)t, considered liquid solid and the vapor generation occurs only in the pressur-I'zer. The' level'1n the pressurizer is dictated by the break flow through a mass balance and the pressure is related to the void by means of an energy balance on

( the pressurizer (Appendix 8). I The UMCP loop will initiate this phase at 300 psi (maximum operating pressure) and at a temperature selected as the saturation temperature in the upper plenum / hot legs when the pressurizer is about empty (5-10% of initial level). ( I ' This temperature is estimated by means of a simplified analysis (heat and mass j balances on the pressurizer). Should the estimated temperature be high, flashing I in the uppor plenum / hot leg will be observed prior to the pressurizer liquid inventory. depletion. In the opposite case saturation conditions will not be reached when the pressurizer level has dropped to 5-10% of the initial value. Ther? fore, the water level will further drop rapidly. Note that the vapor genera-tion in the primary loop is responsible for the change in liquid level decay in the pressurizer as can be observed in Figure 3-32. This transient will be n ecuted " hands off." The controlling parameters are the initial temperature and cpressure and the break size. The duration of the transient is dependent on the l I brea(sizeonly,anditcanbeshownthat-thepressure/voidcurveisindependent i of time or breaksize. Real time will be maintained in this period and the total break flow will be scaled to the 1/500 ratio. The initial level of water in the pressurizer will be such to meet the conditions of real time and scaled total break flow. Note that the natural circulation in the primary loop is at steady state throughout the period, therefore, the heat input (core) and output (S.G.) is defined, and is scaled along with the plant heat / time decay. 3-107 t

s

                                                                          -s 6-5-
                                  @ t=0     see Break Initiation
                                  @ t=   190 sec End of Rapid Depressurizatio 4                                                         t
                                  @ t=Balance 3000ofsec Influx  and Efflux l

E  !

 ,                                                                              \

J d' I w w . J 2-E [ t

                                                                 /

I. / 0; L ^ *A Mh 1 2 g I 0

                                .                  ,                f 1000             2000             3 000 TIME (sec)                                   .

Figure 3-32. PredictedPlantPressurizerWaterLevd1 l l 3-108 i

I . l l BREAK 400000- t=0 sec i Break Initiation _ _ _ _ _ _. H P I 350000 - t= 190 sec

                           @EndofRapidDepressurization t=3000 sec
        ^ 300000 -         @BalanceorInriuxandErriux
        .T m

3 g 250000 -

u. ,-

        $ 20000 0-                                                                          ,

2 ' O150000-W ,

        @l00000-w                                         -

1- - z - 50000- ,- O-

                 ;              2                               3
           -50000                   '                    '

0 1000 2000 3000 TIME (sec) Figure 3-33: INTEGRATED MASS FLOW (BREAK FLOW & HPI FLOW) 3-109

16-

               @ t= 0 sec - BREAK INITIATION I4 -       @t = 190 sec - END OF RAPIDESSURIZATION DEPP
               @t = 3000 sec - BALANCE Or tux       INr s  Eretux I2~

ENSATION MODE

               @t =4500 sec - BOILING -- CONE g 10-n.

2 in 8-n: S V {0 6 - e Q. 4-2- I

                                                         @                  4 0               .             .
                                                    /               .

0 1000 2000 3000 4000 TIME (sec) Figure 3-34. Predicted Pressure of Vessel as Function of Time 3-110

Natural Circulation Instability Transients. The simulation of this second phase of the SBLOCA event is rather complex. The HPI flow is initiated and both flow rate and subcooling must be determined. Furthermore the heat input (core) and output (S.G.) must also be scaled adequately. The principles on which these parameters are controlled are the conservation laws and the two-phase scaling parameters defined by Ishii (6). A brief outline of the control criteria for each of these four parameters is included hereafter.

1. HPI flow rate. The system pressure and the break flow both concur in the determination of the HPI flow rate. Note that the void must be preserved and this fact uniquely determines the flow at any given instant of time.

2. HPI subcooling will be modeled in order to obtain the same subcooling number at the core inlet. This should preserve the same phenomenological sequence of events in the core.

3. Heat Input. The heat decay at the core should follow the decay curve of the plant simulation with the adequate power scaling. Furthermore, the heat input will also be modeled to provide closure to the heat balance for the overall primary loop.

4. Heat output. The heat output at the steam generator is the most complex parameter to control and to scale due to the nature of the UMCP steam generators, j The parameters to be munitored are the inlet and outlet conditions at the primary side of the SG. The flow rates in the four SG sections will be modulated to simu-late the different AFW and feedwater behavior in the secondary side. The UMCP facility has the capability to simulate any heat sink distribution. The modeling link between these input and output variables is the object of current research.

The overall mass and energy balances on the SG will dictate the controlling stra-tegy in order to achieve natural circulation interruption / resumption events similar to the plant simulation. 3-111

1 l l 1 A minimum for the pressure versus time curve is observed in the plant simulation at about 2500 seconds. A surge in the pressurizer water level is also observed at that time. This event is expected to occur in the UMCP model Establishment of Boiling Condensation Mode. This last transient is modeled in the same way described for the previous period. The structure of the parameter control strategy is identical to the one previously described. Furthermore, the first SG achieving BCM (LOOP "A") is far easier to be modeled in terms of the secondary side coolant flow rates (upper and lower section of the S.G.). l Experimental Approaches to Address Scaling Compromises. For each issue, identi-fication of the concerns and phenomena is attempted so that parameters of impor- l tance can be studied and measured. For each study, the experimental or analytical approach and the appropriate measurements are listed. A. HOT LEG PHASE SEPARATION

1. CONCERNS 20 effect on phase separation j 1

2. PHENOMENA TO BE STUDIED Secondary flow in larger pipes.

3. PARAMETERS OF CONCERN

4. APPROACH

a. Analytical Drift model will be used to determine phase separation

b. Experimental Measure a and 8 at each height to determine the drift flux equation coefficient C o 3-112

c. Instrumentation dp cells, flow measurement of liquid and gas in separate effect tests, and liquid velocity probe in loop with vapor velocity measured by video techniques.

B. RVVV EFFECTS (Part 1. Qualitative Study)

1. CONCERNS Effect of RVVV on flow circulation pattern

2. PHENOMENA TO BE STUDIED Effect of RVVV bypass on initiation of natural circulation Effect of RVVV bypass on SBLOCA flow paths Effect of RVVV bypass on steam-ECC interaction

3. PARAMETERS OF CONCERN Flow field and density field in downcomer Global flow distribution change following RVVV opening

d. APPROACH Visual observation Selective opening and closing of RVVV I

Instrumentation : Circumferential high frequency dp cells near RVVV C. RVVV EFFECTS (Part 2) (Downcomer Flow & Density Fields & Tangential Resistance)

1. CONCERNS 1

3-113

4 Flow and density fields may affect natural circulation and interloop instability especially with RVVV bypass

2. PHENOMENA TO BE STUDIED Flow distribution as affected by downcomer gap and distribution of cold leg nozzles and RVVV

3. PARAMETERS OF CONCERN Flow velocity and density at various locations

4. APPROACH .

1 Measure velocity by pilot tubes HTC, or visual means l Measure density by conductivity probes Measure tangential resistance by high frequency dp cells Measure loop flow rates

5. CONTROLLED VARIABLES RVVV opening and closing Cold leg flow rates and enthalpy Gap size (with inserts)

D. INTERLOOPINSTABILITIES(1) CONCERNS

1. Effect of steam generator, RVVV etc. on interloop instabilities 3-114 i

i

2. PHENOMENA TO BE STUDIED Scaling interloop flow distribution as affected by steam generator loading, steam generator metal, RVVV, etc.

3. PARAMETERS OF CONCERN Flow rates in each loop as a function of steam generator load distribution, RVVV performance, and steam generator stored heat

4. APPROACH

a. Analytical Two computer analyses (1) Finite difference approach 1

i (2) Circuit approach

b. Experimental (Scoping)

Variation of steam generator loads:

                             --Feedwater flow rate
                             --Feedwater subcooling
                             --Feedwater level
                             --AFW distribution Measurements:
                             --Flow rates of each loop
                             --Temperature and pressure distribution in each loop 3-115

c. Experimental (Steam generator metal effect)

S.G. stored heat can be changed by using insulation or heat tape wrapping Flow rates will be measured using flowmeters The stored heat is released by transient conduction; thus, a change in the time constant can also affect the result.

d. Experimental (RVVV)

Variation of RVVV opening provides different bypass to the loop flow Flow rate and dp's (especially circumferential) will be measured at each loop location as response to RVVV opening and closing E. STRATIFIED FLOW IN PIPES

1. CONCERNS Flow and temperature distribution in the hot legs and cold legs may be causing internal circulation in pipas, which may affect global circulation

2. PHENOMENA TO BE STUDIED Internal circulation and flow temperature stratification in pipes

3. PARAMETERS OF CONCERN Temperature and flow distributions 3-116 l

l l

4. APPROACH Controlled variables:

Pipe diameters (by changing inserts) Pipe temperature (insulation or wrapped _ heaters) Measurements: Temperature using TC rakes Velocity using pilot tubes or HTS Visual observation F. AFW MULTIDIMENSIONALITY

1. CONCERNS AFW affects boiling-condensation mode and affects interloop flow

, instabilities

2. PHENOMENA TO BE STUDIED Variation of condensation rates in various tubes as a function of location

3. PARAMETERS OF CONCERN Flow rate and heat transfer from each tube

4. APPROACH Measure temperature distribution at entrance and exit of each tube (TC distribution map)

Measure flow rates at tube exits (pilot tube or HTC) 3-117

  - - - - ,       -        r m r-      , --, ---          ,,c--,--  ~ , ,   -m ---, - - -
                                                                                          -m e

May use a plexiglass mock-up for visual study 3.3.5 Application-of Data To facilitate interfacility comparison, the predicted plant condition as predicted by TRAC has been chosen as the reference case. In order to relate to the plant conditions, it is necessary that the test facility data be normalized to the plant behavior through respective scaling principles. Then, the normalized results from various facilities can be compared according to their scales, conditions, or configurations. For example, the phase separation can be represented by the drift flux coefficient C, which is subject to the effects of velocity and void profiles and thus to secondary flow in a large pipe. If C is plotted against hot leg pipe diameter, the effect of diameter on the secondary flow can be assessed and results can be applied to the full size

plant.

For the UMCP 2x4 Loop, the test results can be scaled up to plant level for com-  ; parison with TRAC results in the following ways: (primed quantities represent those scaled up from model-scale results) P =P,p f( ) mo e o p=am Op = 500 Q, U p =U,- @A) p Tsat.p = Tsat(P') T'p = ,C,g p Cp ,p . 9m An example of scaling-up the depressurization curve in the early phase of SBLOCA from UMCP to plant scale is shown in Fig. 3-35. The scaling of coolant level in the pressurizer and the break flow rate are also shown in Fig. 3-35. 3-118

l .0 - PP, . Calculation TRAC Calculation 300 psi (FTototype) 2000 psi

                           % =0.5                                          O's,1= 0.5
           ~

Y' C e R,2

                                          '      '                                            '       's 0.I             '                         '     '                    '                '     -   '

n 0.01 0.1 1.0 Figure 3-35. Normalized Pressure as Function of Normalized Liquid Inventory tReferences are Initial Conditions)

Conclusions. The UMCP 2x4 Loop is different from MIST in the following respects:

1. Flow areas are larger, thus resistance can be better matched.

2. Larger diameters allow the possibility to study multidimensional effects. l

3. The downcomer is of annular design, allowing for study of density and flow fields in the downcomer.

4. The gravity operated RVVV allows realistic modeling of RVVV effects.

5. Adequate power allows a wide range of test conditions.

The shortcomings of the UMCP Loop are:

1. Relatively short in height

2. The designed maximum operating pressure is 300 psi.

For the first shortcoming, the ler.gth ratio being about 0.3, we have shown that most of the important two-phase dimensionless scaling parameters can be maintained to have a model/ prototype ratio close to (a,s,N pch, Frg , FrD ' K " I

• 4 For the second shortcoming, PoR=0.15, we have developed a pressure scaling rationale through conservation of internal energy and preservation of void (or inventory) fraction distribution so that the pressure transients can be simulated with proper selection and control of initial and boundary conditions.

Thus, we-believe the UMCP Loop design and scaling rationale (volume scaling, void fraction preservation, pressure ratio scaling) enables us to address MIST scaling compromises. 3-120 i -. -

Section 4 SEPARATE EFFECTS TEST FACILITIES

4.1 ARGONNE NATIONAL LABORATORY--FLOW REGIMES 4.1.1 Facility Description In view of the inherent difficulties associated with full-scale testing, scale models of prototype systems have been extensively used to predict the behvaior of nuclear reactor systems during normal and abnormal operations, as well as under accident conditions. The severity of the accident that occurred at the Three Mile Island Unit-2 plant has increased interest in this area. New scaling criteria for a two-phase system has been developed based on a rigorous perturbation method by Ishii et al. (!, Z).- This new approach has been used to evaluate the design parameters of the new 2 x 4 simulation loop under the MIST program (Z). In view of certain scaling difficulties and distortions of the large-scale simulation facility, a supporting experimental study to investigate the hot leg U-bend two-phase behavior and associated scaling problem has been initiated in 1984 at ANL. A simulation loop for studying the hot leg U-bend flow interruption and hot leg two-phase flow regime has been constructed based on the scaling criteria developed by Ishii et al. (!, Z). The overall loop schematic is shown in Fig. 4-1. This l two-phase flow loop is designed such that it can be operated either in a natural f circulation mode or in a forced circulation mode using nitrogen gas and water. l The 5 cm 10 riser simulates the vertical section of the hot leg. At the end of a j riser section there is an inverted U-bend and the flow is turned downward into a gas-liquid separator which simulates the once-through steam generator. At l present, the loop is 5 m in height and the test section is made of. Corning Pyrex glass tubes with a U-bend of 9 cm and 15 cm are planned. I Attached to the bottom of the riser is a stainless steel plenum which holds a multiple nozzle gas injector. The nozzles are made of hypodermic tubing molded I into an epoxy plate. These stainless steel tubes, having nominal 0.015 cm ID and I a nominal 0.03 cm UD, are used for nozzles. The gas is injected vertically into ! the riser through these nozzles. The gas flow rate and pressure are measured f between the plenum and 0.5 m3 accumulator. The gas (N 2

                                                                                        ) is supplied in high l       pressure cylinders, and reduced in pressure through the manifold regulator filling the accumulator. The accumulator dumps out variations in pressure and temperature

] as the gas is delivered from the regulator. The differential pressure is measured at five locations as indicated in Fig. 4-1. At low liquid flow (<2 m/sec) these pressure transducers give very accurate measurement of void fractions. The liquid flow is measured by a paddle wheel type flow meter which has very small ap. The two-phase flow section is all transparent, such that flow visualization, high j 4-1 1 i

        ---,_--_e.--,.   --, .-,_ ,- _     ,3-.-.         . .._,-,.,~-m-- - , - - . _            -_r, m-_..   -

                                                                                                                 -           =    .-    _.

Exnansion Tant 4

                        .a Gas k                        -

h -

                                                                                                            " l' Drain         p Q         5' Gas Senarator                       -L h         5' Test Section Cornina Pyrex Glass Tube (=2" ID)

(U-bend R = 3 1/2") a 4' 4 E, Paddle Pheel Flosensor - ((l' Press m r g ~~ 0.5, Regulator Water ( Friction Control Valve i h~ ;4, X ><1 O Tank ' T. ~~' > k Gas Gas Water - njector a a yA :cumulato - Suonly p

                                                                }                                             g
            #                                rc           x   /                                               TRotareter Water Filter                                                                  "i" g                                   Tip Tanks o Drain Figure 4-1.         ANL Basic HLUB Loop Design

speed still photography, and high speed cinematography are possible. The test section consists of various length Corning Pyrex glass tubes with pressure taps between them. It is designed such that various local instrumentations for two-phase flow measurements can be easily accommodated if they become necessary. The loop design is based on the scaling criteria developed in the present study (6, 7). However, enough flexibility is built into the design such that certain scaling distortions can be studied by changing some components. One of the important aspects is that the height of the separator, as well as the liquid level within the separator, can be changed in order to study systematically the effect of the thermal center of the steam generator. In view of the flexibility of the loop, the following basic loop dimensions are considered for the tests. l Hot Leg Diameter (cm) 5.1 5.1 10.2 Hot Leg Rise (cm) 550 350 350 U-bend Radius R/D 1.78 , 3 1.78 1.78 DrivingHead(cm) 100 - 300 100 - 200 100 - 200 The range of the gas flux is from 0 to 100 cm/sec with the expected liquid circu-lation rate of 0 to 100 cm/sec. The overall loop frictional resistance is changed parametrically by the friction control valve in the single phase liquid section corresponding to the cold leg. The liquid circulation rate can be adjusted to desirable values if necessary by the use of a small pump located in the bypass of i a cold leg. t Two different geometries for the hot leg inlet section are used. For the first series of tests, only the vertical section of the hot leg is simulated without a horizontal section. Thus the gas is injected directly into the hot leg vertical section (see Figs. 4-1 and 4-2). This will force the flow regime At the bottom of the hot leg to be a bubbly flow. Since in the prototype system it is considered to be impossible to develop a slug flow (1_5), the hot leg flow regime may be better simulated by eliminating a horizontal section which can induce a slug flow pattern. In the second series of tests, a part of a reactor vessel and the hori-zontal section of the hot leg are also simulated geometrically (see Fig. 4-3). This second series will give important information on the hot leg flow regimes in the other simulation facilities such as MIST. 4-3

X[ n i o t c - e S t

  ._                 e l

I n t h g

           ,        i a

r u t S y t i l i c a F L N A 2 _ 4 . e _

        =             r     _

u g i F _ a1

             >           i

                                             =

L g e l t o H d n a l I e s s e l V n ro oi tt ca al eu Rm i lS a i n t o ri at Pc e yS t il l a o it w cn ao Fz i L r N o AH 3 1:' ' 1

                    = ::   :

4 _. = e r . u g i F 4 _ b

                                         ,              t Il l,!     I i                      i

4.1.2 Scaling approach Base of Scaling Criteria

!    The available methods to develop similarity criteria for two-phase flow systems have been reviewed by Ishii and Jones (1@), and the similarity analysis for a two-phase flow system has been carried out by Ishii and Zuber (ll), and Zuber (18).

i The results based on the local conservation equations and ones based on the perturbation method were utilized by Ishii and Kataoka (!). The extension of the similarity analysis to a natural circulation system under both two-phase and

single-phase conditions was achieved by considering the scaling criteria from a small perturbation method and the steady state solution. For this purpose, the l relatively well-established drift-flux model and constitutive relations (19, @)

were used. l l l In the present analysis, the scaling criteria developed by Ishii and Kataoka (@) are used. The application of the above criteria to the conceptual scaled model design for the 2 x 4 forced and natural circulation loop system under single and/or two-phase flow conditions has also been carried out (Z). In what follows, the summary of the scaling criteria and its application to the planned ANL hot leg U-bend study is discussed in detail. Single-Phase Similarity Laws The similarity parameters for the forced and natural convection circulation loop systems can be obtained from the integral effects of the local balance equations along the entire loop. A typical system consists of a thermal energy source, energy sink, connecting piping system between components, and a circulation pump. The dimensionless variables and parameters used in the similarity study are obtained from the dimensionless balance equations (@). In these equations, the fluid properties are assumed to be constant except for the buoyancy term, where the Boussinesq approximation is used. The significant dimen-sionless similarity parameters of the system can be expressed in terms of the Richardson, Stanton, and Biot numbers, a dimensionless friction number, Fj , a dimensionless pump characteristic number, dF , a dimension less time ratio number, T},andadimensionlessheatsourcenumberQ39, as follows: 2 R s g a aTogt /u (4,1) Stj s 4 h jo t /o C p un di (4-2) 4-6 )

Bi - hj 639/kst (4-3) Fjsfj(i + i

                               )+K j                                          (4-4)

T$=(ais /8 i)(t o/uo) (4-5) and SI Q (4-6) si = os i psi "o aT o C < where u o

              , ATo
                    , and toare prescribed reference velocity, temperature difference, and equivalent length, respectively, for the system. t ei is the equivalent length for minor losses distributed over the ith section, i.e., bends, elbows, etc., and Kg is the singular loss coefficient defined at the inlet and outlet of the ith section, i.e., expansion and contraction coefficient.

In addition to the above defined physical similarity groups, several geometric similarity groups are obtained. These are: Axial Length Scale: Lj s tj/to (Lh 5 t h/Eo) (4-7) Flow Area Scale: Aj s aj/ao (4-8) It is noted here that the hydraulic diameter and the conduction depth 63 j are defined by dj = 4 aj/(j and 6j s a is /Ci where aj, a3 9, and (j are the flow cross sectional area, solid structure cross sectional area, and wetted perimeter of ith section, respectively. Hence, dj and 6j are related by I dj = 4 (aj/as i) oj (4-9) l The reference velocity uo, and temperature difference ATocan be obtained by using the steady-state solution. By taking the heated section as a representative section, one can solve the fluid energy equation for temperature rise. Thus, AT g = (go ig/p C pu g )(aso/ag) (4-10) where subscript o denotes the heated section. Substituting the above expression into the steady-state momentum integral equation, the solution for the velocity becomes 4-7

                 .                                        .. .                . . _ . ~                     . _ . . .                         _          _

2 a g (ho t o/o C,) th (ago/ag) un = 2 (4-11) (F4/Aj ) for the natural circulation loop. Equations 4-1 through 4-8 represent relationships among the daiensionless coefficients and the generalized variables of the differential equations. From the dimensionless form of differential equations, it is evident that if the similarity is to be achieved between processes observed in a model and in the prototype, it is necessary to satisfy the following identitles: AIR

• I j

(Lg/Aj )R " 1 LhR " I i 2 (F4 /A )p , g RR*1 (4-12) l Stir =1 T{R"I BiR " 1 Os1R " l 4-8

                          .~,w  - , - , - . . - - - < - - -         - - - - -           -
                                                                                               , , , , ,. ,           , ,,,         w,- --, ,   , . -- ,

It is apparent from the above set of equations that the complete transverse area similarity is required as expressed by Eq. 4-12. On the other hand, axial length similarity is required for the hot fluid section. The total loop axial length similarity is a somewhat weaker restriction than the complete axial geometrical similarity required by (t/t), 9 o L =1 (4-13) 1R=(t/t), j o However, for simplicity both the transverse area and axial length similarities are assumed where the energy transfer is important. l In view of the complete transverse area similarity, the dynamic similarity condition reduces to the overall friction similarity given by (F3)R = 1 (4-14) This expreses that pipe friction loss and the minor losses associated with the loss coefficient S L i i ei f ilaT + ar) e and Kj can be interchanged without changing the overall value of the pressure loss term. By adding or removing bends or by providing additional flow restriction in i the form of orifices, it should be possible to simulate a wide range of scaling conditions. j In view of the reference temperataure difference, Eq. 4-10, and the reference j velocity, Eq. 4-11, the Richardson number is automatically satisfied. It is noted I that the heat transfer coefficient is not an independent parameter, because it is 1 a function of the properties and flow field. Thus, in general, it is difficult to satisfy the Stanton and Biot number similarity. However, the thermal inertia ratio can be given by N thi*("oCasps's)1" 1/8 9 T j (4-15)

 ,                                                      p 4-9

The Biot number and Stanton number similarity conditions with the constitutive relation for h mainly simulate the boundary layer temperature. When the heat transfer mechanism is not completely simulated, the system would adjust to a different temperature drop in the boundary layer. However, the overall flow and energy distribution will not be strongly affected in slow transients typical of a natural circulation system. The violation of the Biot or Stanton number similar-ity within the liquid flow condition should not cause a major problem except at very rapid temperature transients. Thus, the thermal inertia ratio Nthi and time ratioT{areconsideredtobeimportantparametersforthesolidheattransfer simulation. Thus these two numbers are analyzed as primary scaling parameters in the following. Two-Phase Flow Similarity The similarity parameters for a closed loop system under a two-phase flow condi- l tion can be obtained from the integral effects of the local two-phase flow balance uations along the entire loop. Under a natural circulation condition, the majority of transients are expected to be relatively slow. Furthermore, for developing system similarity laws, the response of the whole mixture is important rather than the detailed responses of each phase and phase interactions. There-fore, the drift-flux model formulation is more appropriate for the derivation of system similarity parameters under a natural circulation condition. This is because the drift-flux model can properly describe the two-phase mixture-structure interactions over a wide range of flow conditions. 4 The important dinensionless groups which characterize the kinematic, dynamic, and energetic flow fields are given (6, 7) p.s follows: Kinematic Similarity Parameters: 46q t Phase Change Number, N pch I (du gaH gg o)( ) (4-10) V Drift-Flux Number, N 3 (4-17) d o Dynamic Similarity Parameters: 4-10

2 u Froude Number, N Fr 5 gt o og h (4-IO) f(t+1,) 1 + x (ao/pg) a 2 Friction Number, N7m d [1 + x (au/pg)]0.25 () (4-19) 2 3 Orifice Number, N g aK[1+x/2(3,f,g)j[ ) (4-20) Energetic Similarity Parameters: 1 I aH [ Subcooling Number, N sub i(aH S

                                                                                                                    )( )                                                          (4-21) f9                    9 TimeRatioGroup,T{s(                                                 )                                                                        (4-22)

Usi "o C p Heat Source Number, Q S ' s (4-23)

                                                                                             "si psi        C "o A"sub i

The Froude, friction, orifice, and heat source numbers together with the time ratio group have their standard significance. On the other hand, the subcooling, , phase-change and drift numbers are associated with the two-phase flow system. l It can be shown from the steady-state energy balance over the heated section that Npch and Nsub are related by i ] (f)x=Npg - N3g (4-24) 9 where x is the vapor quality at the exit of the heated section. 4-11 l l

1 l The drift number takes into account the drift effects due to the relative motion of the fluids. Since the vapor drift velocity V j gdepends on the flow regime (2_0), this group characterizes the flow pattern. It should be noted that  ! the constitutive relations for the relative motion between two phases, hence for  ! the drift velocity, Vgj should be specified in the above similarity groups. The j relative motion can be specified by a number of different forms. The representa-tive constitutive equation for the relative motion based on the drift velocity correlation has been reviewed and sur.narized previously (20). The important point is that it depends on the two-phase flow regimes and pipe diameter (2_0, 15). Fluid-to-Fluid Scaling and Simplified Criteria  ; In general, the solid materials as well as the working fluid may not be the same between a model and a prototype. In what follows, a scaled down model of a prototype with certain values of tR and aR is considered. Thus, "R

• Lom/ fop (4-25) aR = aom/aop Although it is difficult or impossible to satisfy all the scaling criteria presented above for a scaled down model, particularly with different fluid I properties, practical engineering criteria can be developed by considering only l the essential elements.  !

For a single-phase case from the heat source number, thermal inertia ratio and friction number similarity, it can be shown that l 1 )

                                      !                                     (4-26) uR*      R      o s ps R oC (h)R"(oC                                                           (4-27) s ps)R Furthermore, from the time ratio number similarity, I                                                           (4-28) 6R*1R       u'1/2 R

e

                                'Q i'

4-12

7 _.. . - - ty ,

                                                           \'                                              '
                                                                                                               , t .

s However, because of the restrictions imposed by Eqs. 4-9, 4-25, 4-27, this criterion cannot be easily satisfied for pipe sections. Note that-for a scaled down model, the flow area ratio aR is usual'ly very small, thus the tydraulic ' diameter ratio dR = / y . For major heat, transfer sections such as th'e simulated core or the steam generators, it is easier to satisfy both Eqs. 4-27 and 4-28 i because of the subchannel geometry. N v

                                                   \,

For a two-phase flow case, it is useful first to consider the effects of the proposed similarity criteria on the steam quality and void fraction. Under the l phase change number and subcooling number similarity, i.e., (Npch)R " IY"d (Nsub)R = 1 Eq. 4-24 implies that '1 ' l l L, (x M) =1 (4-29)

           'g R Furthermore, from the identity given by s

(a7)1 = 1 + (1 + Nd )/(* 9

                                             )                 t                      ( - 0) and the drift flux similarity, i.e., (N )Rd = 1,'one. Sets
                                            %                             s.

s k (af)R=1 ] (4-30) , t r These two relations given by Eqs. 4-29 and 4-31 are quite useful criteria in case of fluid-to-fluid scaling. t In view of the above, the Froude number, subcooling number, and pha change number similarity implies, , w. uR " ER

                 /                                                                    (4-32) y (aHsub)R"( t.H 5 ps )R     ,

N 8 ~.

      .s                         ,

I 4 ( I 4-13 1

                   ,  ,, ---                                 I , 4 --             r

and h r!? if i f hR * "R )R(ANfg )R (- } s By considering the thermal inertia ratio similarity, i.e., (Nth)R = 1, the heatilig j ratio becomes d E hp = t-I/2("I99h(5DS) (4-35) p R l l In addition to the above, the friction similarity l l l (Nf j)R = 1 and (Na g)R = 1 (4-36) should be satisfied. These can be approximately satisfied by (F9)R = 1 (4-37)

              , . which corresponds to the section-wise frictional similarity rather than the loop-a                     wisqfrictionsimilaritygivenbyEq.4-14.

g3 d ThetImescaleisgivenby

                    ~
                                        =t                                                                (4-38) i s

tR= R <l '

s 4'

Therefore,allthetimeeventswillbeacceleratedby/{inthemodelfor1R

                      <1. This is one of the most important results since it implies that the real time scale is not practical for a reduced length model.

If the same time scale to be maintained in the single phase and two-phase transients, then the heating ratio for a single phase case should be

\'

4 I l 4-14 o

                                                                                                 +   s

( . t s qt = t 2[8 s)g (single phase) (4-39) Thus, in view of Eqs. 4-35 and 4-39, the heating ratio for single-phase and two-phase cases may not be same if the fluid and solid properties tre not same, s Scale Model Limitation Since the accident occurred at the Three-Mile Island Unit 2i plant, numerohs issues

have been raised in regard to 2 x 4 loop NSSS designs. Some of the issues are generic in nature and others are design specific. In particular, concerns have beenraisedrelevanttotheBabcockandWilcox(N&N)NSSS1argelybecauseofsome N unique components and hardware features contained in the plant design in addition to the'2 x 4 loop geometry. It was suggested that unique design features of the B&W reactor may produce unique and complex thermal-hydraulic behavior during small  ;

break loss-of-coolant accident or during some other abnormal transient that could be misleading to an operator. Perhaps the most distinguishing feature of the B&W plant design relative to the , other PWR vendor designs:is the once-through steam generator (OTSG) as illustrated in Fig. 4-4. In the OTSG systen, primary fluid flows downward through the tubes in the steam generator. The secondary feed enters the steam generator secondary i riser section from the bottom. However, use of the OTSG in the B&W NSSS necessi-tates the use of a vertical hot leg, with a 180* U-bend at the top to royte the primary fluid from the reactor vessel to the top of the steam generator. ,X A detailed scale model study for the plant design has been. carried out under the present program and reported in Ref. 7. The most severe condition in terms of the thermal-hydraulic simulation is imposed by Eq. 4-37, because in a scaled model the hydraulic diameter can be much smaller in piping system. Therefore, for a given value of ag, calculations are centered to determine tR to meet fir

• 1 for each section as required by Eq. 4-37. A computer code which is capable of ca}culating - , i l

Fj for each section among other, system parameters for the prototype and ' scaled i model was' developed. The results gathered from a., series of computer calculations indicated that the hot leg simulation imposes the. strongest ctnstraint. Once the ^ necessary condition for the' hot leg is obtained, the other.' sections are easily adjusted by increasing the minor less coefficients. Therefore','in what f'ollows, , L the basic scaling criterion from tne hot leg is presented. 4 { i 4-15 s s

s 1 I I I

                                                                                                 )

2h _ _ _k _ _ jk -HOT LEG 0 - . 1  ; Q OTSG = D SN POSSIBLE TWO-PHASE 23 REGION

                                                         ~

SECONDARY SIDE

                                                                          ""o
                                                         @                }

l ' e REACTOR STEAM OUT .c j' CORE DOWN / * ' FEED WATER IN--Z:_____ COMER g _g l COLD LEGS (2 PER LOOP) i SINGLE PHASE REGION

                                 'cece, TWO-PHASE REGION figure 4-4. Typical Loop with Once-Through Steam Generator 4-16

i l The prototype hot leg flow resistance consists of the commercial steel pipe friction and the distributed loss due to elbows. The distributed loss factor, (fj tej/dj), was calculated to be about 1.2704 for the prototype. For a scale model, lower (fjtei /d j) value is desirable due to the requirement imposed by Eq. 4-37. Therefore, the friction factor for a drawn tubing is used. Further-more, the practically minimum values for the flow restrictions are used. Then the maximum possible ER for given aR is given by t R = 10.0 (1/aR)-0.672 (4 40) l t and the volume ratio by V R = 10.0 (1/aR)-1.672 (4 41) Hence, for a sample case of VR = 1/815, one obtains for the practically optimum case as aR = 1/218 (4-42) tR = 0.27 And in this case, the time scale is reduced by t t R= =/ = 0.52 (4-43) p Therefore, the key timings of various transients will be reduced by a factor of 0.52 in the scale model. The prototype hot leg diameter is 91.5 cm, vertical rise 14.25 m, and total length 21 m. By using these numbers, the corresponding model { dimensions are 6.2 cm, 3.84 m, and 5.6 m, respectively. In view of the commer-cially available glass tube sizes, d ofm 5.1 cm and 10.2 cm have been chosen. The i 4-17 l l 1

vertical elevation of 3.5 m and 5.5 m are chosen to simulate the prototype flow as well as to study the scale distortion caused by using a too long hot leg (MIST facility). The simulation of the MIST hot leg two-phase flow by the ANL facility is much easier due to MIST's smaller flow area in comparison with the prototype. Basically almost no restrictions such as Eq. 4-40 are necessary, since the hot ' leg diameter of the ANL facility is 5.1 or 10.2 cm which is very close to that of MIST facility. 4.1.3 List of Issues Addressed by the Project The major issues addressed under this project are the interruption and reestab-lishment of the loop natural circulation. These phenomena are studied in detail by focusing on the simulation of a prototype system and atypicalities of MIST and other facilities in terms of the hot leg U-bend flow separation, hot leg flow regimes, and void distribution. Because of a very small hot leg diameter and very large t/d ratio, the flow regime and void distribution in the MIST hot leg can be significantly different. This will also influence the phase separation at the U-bend. 4.1.4 Test Procedure The detailed description of the test facility for the adiabatic tests is given in Section 1. Two different series of tests are planned. These are described below. Series (I) Without Horizontal Hot leg Section or Vessel (1) Phase 1 Hot Leg Diameter 5.1 cm Vertical Height 550 cm , 350 cm U-bend Radius R/D 1.78 , 3 Gas Flux 0 to 100 cm/sec Separator Level 3 Levels Friction Valve 4 Positions (2) Phase 2 Hot Leg Diameter 10.2 cm Vertical Height 350 cm U-bend Radius 1.78 Gas Flux 0 to 50 cm/sec Separator Level 3 Levels Friction Valve 4 Positions 4-18

Series (II) With Horizontal Hot Leg Section and Vessel Hot Leg Diameter 5.1 cm Vertical Height 550 cm , 350 cm U-bend Radius 1.78 Horizontal Length 120 cm In all of these tests, first the position of the friction control valve and the separator water level are predetermined. A test is started from a relatively high gas flux and a steady-state natural circulation is established. The void fraction, flow rates, pressure drops, hot-leg flow regimes, and U-bend flow separation are recorded. Then the gas flow rate is stepwisely reduced and each time a new steady state condition is reached. Eventually the natural circulation termination point is reached. After the occurrence of the flow termination, the process is r,eversed by increasing the gas flux. The test is terminated after the flow circulation is firmly reestablished. Thus the relation between the gas flux and the natural circulation rate, flow interruption point, and flow resumption point are measured together with other standard two-phase flow parameters. The above procedure is repeated for each friction control valve position and water level in the separator. From this the parameteric effects of the loop frictional resistance and driving head can be experimentally established. 4.1.5 Application of Data Adiabatic Simulation of Hot leg and U-bend Flow The adiabatic gas-water loop described in Section 4.1.1 is used to study the hot leg and U-bend two-phase flow phenomena. In this adicbatic simulation and flow visualization study, the boiling in the prototype or MIST is simulated by injec-tion of gas into the simulated hot leg. This makes the measurement of the gas l flux very accurate, whereas in heated experiments such as performed in the other ! facilities, the flow is coupled with heat transfer and the direct measurement of the gas flux is not possible. The simulation experiments are carried out by parametrically changing the gas injection rate. The gas injection rate and the prototype hot leg inlet quality are related in the following form. l l The quality in the gas injected system is given by o ([q x)m =(f)m(,aa"jlu a j 3 fo m (4-44) l l 4-19

where ufo is the liquid volumetric flux. From the quality similarity given by Eq. 4-29

                                                 )                                                            (4-45)

(fg x)p=(f)m(od g gg + u fo m If the density ratio in the model is very small or-(o g/o),<< 1, then the above equation can be simplified to ( fo[j m] =(xf)p g (4-46) Equation 4-45 or 4-40 relates the simulation loop parameters to the prototype condition. The time scale and velocity scale are given by Eqs. 4-38 and 4-32. The same formulation can be used for simulating the MIST or other facility by the ANL facility. In place of the prototype value, the values for these facilities l should ba used. The interpretation of the data can be carried out by using the basic scaling criteria developed in Section 4.1.2. Thus Time Scale t R"Ek! Velocity Scale uR"ik! (4-47) i Void Scale (8)O f R

                                                   =1 The value of tRis determined by knowing the simulation facility length, t

Using these criteria, the data on void fractions, flow rates, and pressure drops 4 can be translated into prototype values. Furthermore, the information on hot-leg flow regime, U-bend flow separation, flow interruption and resumption can be i recasted in terms of the prototype values. The same can be done for MIST and 4-20

other simulation facilities. Furthermore, the experimental parametric study on the frictional resistance, thermal driving head, and various geometries such as the hot leg diameter, height, and U-bend radius will clearly show the effects of scale distortions on these variables in definite numbers. Another focus of the tests is the hot leg flow regimes and flow regime transi-tions. In the Series (I) tests, the gas is injected into the vertical section of the hot leg directly. The inlet section is always in the bubbly flow with a rela-tively uniform bubble size. From this controlled inlet condition, the entrance effect on the two-phase flow regimes and regime transitions can be easily studied visually. It is considered that the bubbly flow is the predominant flow regime in the hot leg of the prototype system. Hence Series (I) tests are very important for understanding the flow in the prototype system. On the other hand, in Series (II) test, the vessel and horizontal section of the hot leg are also simulated. This geometry may tend to produce a slug flow at the bottom of the vertical section of the hot leg. Although in the prototype system these slug bubbles should disintegrate immediately due to interfacial instabil-ities (1_5), this may not occur in the integral simulation facilities such as MIST, UM, and SRI loops. Therefore, it is quite important to identify the effect of the vessel and horizontal section on the two-phase flow regimes. The most significant advantage of this facility is the transparency of the test section which permits direct visual observation of all the flow phenomena under strictly controlled conditions. Various parametric effects and scale distortions can be studied experimentally in a short time. Furthermore, according to the above scaling criteria, it is quite simple to recreate the flow conditions which may have occurred in the hot leg of a high pressure simulation facility by impos-ing the necessary boundary condition. This will greatly enhance Ue physical understanding of MIST tests. Phenomenological Modeling The data from these parametric studies will be used to develop a two-phase flow model for hot leg and U-bend flow. A parallel analytical study (15) is currently being carried out to determine the pipe diameter and entrance length effects on j the flow regime. From the combination of the data and analysis, the final two-phase flow model will be recommended. l 4-21 l

4.2 SAIC--LARGE PIPE FLOW REGIMES < Flow regime characterization and flow " carryover" in the candy-cane section of

- PWRs with once-through steam generators during a small break (SB) loss-of-coolant

! accident (LOCA) are issues that are currently receiving considerable attention. In a SBLOCA the hot leg two-phase flow charateristics can significantly influence 2 the transient behavior. The correlations which are available to predict the two-phase flow regimes in the hot leg geometry were developed from data obtained in

relatively small diameter vertical pipes (s5 cm) with large length-to-diameter l ratios. Typical approaches for flow pattern prediction in the past have been to J

coordinate experimental observations by plotting transition boundary lines delin-eating flow regimes on two-dimensional plots (3, 22). However, extrapolations of these correlations to reactor conditions and geometry with large pipe diameters (91.5 cm) and small length-to-diameter ratios (4-6) is somewhat questionable. Furthermore, complexity of the flow pattterns coupled with limited theoretical basis and subjective judgment by the observer have greatly contributed to dis-agreement among various experimental data and have prevented the formulation of a generalized map (23, 24, 25). This has resulted in limited applicability of each

;                                     map to the narrow range of its corresponding experimental parameters (i.e., pipe 4

size, fluid properties, etc.). Thus, flow pattern prediction during a SBLOCA still remains highly uncertain, i , Some of the most important shortcomings of flow pattern prediction include the lack of sufficient data for large diameter pipes and the effects of entrance region geometry and pressure on flow regimes. In this context, we have begun an experimental program starting with 10.2-cm (4-inch) diameter glass pipe to characterize flow regimes in a PWR hot leg geometry. These results show that valuable information can be obtained from atmospheric pressure air-water flow tests in reactor geometries. Presently, we are continuing our experimental program using 30.5-cm piping to obtain additional data which is needed to char-acterize the physical differences between flows in small and large bore vertical 4 pipes. This document describes the test facility, scaling, issues addressed in this 4 experimental program, test procedures, and finally how the results can be used in support of analysis of MIST data and/or plant behavior. f s

4-22 i

  . - . _ _ . , _ . . - - - . - ~ . .            . - - - - ~ ~ . . - - _ - - . - . - . _ - - .        _ . _ - . - - - . _ - _ - . _ _ _ _ . , _ , _ _ _ - - _ - _ _ - _ . - ,

4.2.1 Facility Description The test facility used for the two sets of experiments is the SAIC/EPRI Air-Water Two-Phase Flow Loop. This test facility consists of a main support structure on which various sizes of piping may be attached. The loop is essentially an atmo-spheric pressure system with horizontal and vertical test sections of about 10 feet in length. Water capacity is 400 gpm at 90 feet head, provided by a 4x3 stainless-steel close-coupled centrifugal pump. The pump is connected to the flow loop with bellows at the inlet and outlet to minimize vibrations. Air capacity is 500 ACFM at 100 psig, provided by an Ingersol-Rand SSR2000 helical screw air compressor with an aftercooler. A DIAL-AIR model R41-0C-000 air regulator set to the desired pressure limits the air pressure in the loop. For the two experiments described in this work, the loop was configured as shown in Figure 4-5 for the 4-inch pipe tests, and as shown in Figure 4-6 for the 12-inch pipe tests. The water reservoir for the flow loop consists of a 1700-gallon tank, a 300-gallon tank, and a 150-gallon tank. The 1700-gallon and the 300-gallon tanks are connected with 3-inch piping and valves, and an 8-inch elbow connects the 300-gallon tank to the bottom of the 150-gallon tank. The pump draws water from the 150-gallon tank and water from the loop is returned to the 300-gallon tank. This arrangement was selected to minimize flow disturbances at the pump inlet. The 1700-gallon tank is the supply reservoir. l In the remaining portion of this section, the specific flow path configurations for the 4-inch setup and the 12-inch setup are described. The pipe network shown in Figure 4-5 is built with 3-inch and 4-inch diameter glass pipe. The vertical candy cane section is all 4-inch pipe. This size pipe was initially chosen to match experimental facility existing capabilities, and the 12-foot height was selected to allow maximum buoyancy effects within laboratory height limitations. A 120-gallon tank located at the outlet of the candy cane provides a more real-i istic exit flow condition of the test section in that the water level in the tank can be adjusted by using internal overflow pipes of various heights. For most of i the 4-inch pipe diameter tests, the level in this tank was adjusted such that the hydrostatic pressure at the top of the candy cane was a few inches of water above I atmospheric pressure. An air vent at the top of the candy cane allows the system pressure to be equal-ized to atmospheric conditions during system fill and drain. System pressure was continuously monitored at various locations with an ENERPAC 0-30 psig pressure gauge to ensure that the design pressure was not exceeded. 4-23

  - .-   _ _    _m    _ . _ _ _ .    .        _ _.        _ _ _ _ _ - _. _. . _ _ _ _ _ _ . _ _ _ . ,         . - -_            m                   _
                                                                                                                                                                          ._   ..______.....___.___m                  _              . ..

I f AIR VENT P , _: __- TOP TANK PRES $URE w TAP s TO TOP TANK C NTROL #3 PRESSURE TO DRAIN TO HOT LEQ f PIPE TAP #2 Vie vil i V121 M EASURE. MENT v14 NK FLOW TOTAtl2ER = I Hz 3.4 m (141') DRAIN Vtr DETAILS S & C (8: l' PIPING. C: 2* PlPING) a DETAIL A b V5 fr 2 - AIR VS gggULATOR N^ --e AIR INLET V10 SEE DETAll V8 - T IV21 AN A IO FICE 3 RESER R 1 VENTuntE _ hi TAP # THERMOCOUPLE COMPRESSOR

                                                                                                                           .M Vi g            _ _V i

SEE DETAIL 8 SEE DETAIL gV2e TO DRAIN Figure 4-5. EPRI/SAIC Test Facility (4-Inch Pipe) 1 4

_ _ . . _ . _ _ . _ _ _ .,___._mm...... .__m_-__._______.o_m_..._m__.__,___._._ _m_ . . . _ . _ . . - - . --- - - - - _ . - - - - - - - - - t i 1 j i .i l i i ] Ovfwtow Am atate j r a s y i Ret t 78 I f TO WeeCN j ..LE. VEL 7 e ,cs... _- g PA.E.S.Su.m.t

                                                                                                                                                                                                   . on , , . , . ..,

e. aana ,

!                                                                                                                                                                                         e,870P IANEL) 4I I
                                                                                                                                                                  . .. _f ,               ee es es j

4 ll ll catenaA DEwssionsETER

                                                                                                                                                                                  ;                                  O
                                                                                                                                                                                      ,                                 y  (voe FRAC 7s000 4,

FN) 9 S Fe I', Ut WS 4 ft. 88 1 r (r IXL s Am mEto "'"*" Xv. v3 ye T , ACE

                                           -e    ____ W W           =

g _

                                                                                                                                                                                                                            - ..**Liyt l

]  %

                                                                                                                      ^
                                                                                                                           ,                                            7"E,S,5g coMason WAfta AEstRwCW                                        .A,,,       ,A.,
                                                                                                                                                    ^,               = ' ~ -

i

?

s - IA88st 8 w - TAPet 5.5 rt - 2

                                                                                                                                                                       - ~ 7 -.

) . i - A

                                                                                                                                                                               \ m_asJ.EC.T.IOII                                                                             -

I t Figure 4-6. EPRI/SAIC Test Facility (12-Inch Pipe) i i s i I f k i i 1 i t _ -

A check valve in the air supply line downstream of the pressure regulator and in the water line just upstream of the air injection location prevents flow reversal in these lines. A 40-gallon measurement tank, connected to the low point in the flow loop allows the measurement of system water volume at the conclusion of each test. Also, butterfly valves (V2, V3, V8A) installed in the loop allow quick isolation of the test section and subsequent measurement of air and water volumes trapped in the test section. These measurements allow the void fraction to be determined. Water is pumped into the main loop by proper positioning of valves V1 and V21. A 2-inch pipe is used to carry the flow from the pump to the inlet of the test section. A replaceable segment in this line accommodates various size venturies for flow measurement. 4 The experimental setup for the 12-inch diameter candy cane tests is shown in Figure 4-6. The water and air capacity, and the water reservoir are the same as l described previously. In addition to a tank at the outlet of the candy cane (representing the steam generator), another tank at the inlet to the candy cane (representing the pressure vessel) was provided to more closely approximate the 4 B&W reactor geometry. The vertical section of the candy cane was made of clear PVC, and the three elbows were made of glass. With viewing parts in the two tanks, visual observations may be made at all locations in the simulated primary system flow path. l The 40-gallon measurement tank and the butterfly isolation valves were not used in . these tests. Instead a narrow beam gamma densitometer, mounted on an adjustable j platform, was used to make void fraction measurements at several elevations in the ! vertical section of the candy cane. The tank representing the reactor pressure vessel has two water inlets and one air inlet. The air inlet and one water inlet are at the bottom of the tank and the second water inlet is at the top. By proper positioning of valve V3, water can be directed to either inlet. Inside the tank, an air injector produces uniformly dispersed air bubbles. This injector / dispenser is simply a 1.5-inch dimeter PVC pipe with 0.5-inch diameter holes (10) drilled in it and connected at the bottom of the tank. An air vent at the top of the tank allows air to be released during system fill. 4-26 l

The water tank representing the steam generator was constructed such that the candy cane connection is at about mid-height of the tank. The water level in the tank can be adjusted by internal overflow pipes of various heights. The tank also has an external overflow located near the top level. With this design, the hydro-static pressure at the top of the candy cane could be held slightly above atmo-spheric pressure. Water and air flow rates are determined by measuring the pressure differential across a venturi and an orifice plate, respectively. Two venturis, Models 2-658 and 1-409 manufactured by Gerard Engineering Company, were used for water flow measurements. Three orifice plates with bore diameters of 0.159-inch. 0.707-inch, and 1.414-inch manufactured by G. M. Cook Associates were used to measure air flow rates. The pressure differentials were measured by two identical Gould differen-tial pressure transmitters, Model PD3000, having a range of 0-30 inches H20. ! In order to calculate the air flow rate in SCFM, corrections must be made for temperature and pressure. An iron-contant thermocouple, inserted upstream of the venturi, measures the air temperature. Because the air compressor has a built-in air cooler, the temperature increase above ambient is minimal and, therefore, the temperature correction is small. A pressure gauge on the regulator is used to monitor the air pressure. With the air pressure and temperature known, correc-tions are made according to tables supplied by the orifice manufacturer. Pressure in the hot leg was measured at three locations (shown in Figure 4-5 and 4-6) using three Celesco Model P70 pressure transducers. In the experiments with the 12-inch diameter pipe, a gamma densitometer was used to determine local void fraction at different elevations in the vertical section of the hot leg. The densitometer uses a collimated beam of gamma rays from a CS-137 source. 4.2.2 Scaling Approach for the 12-inch diameter test facility, the hot leg geometry was scaled to approximately one-third of prototypical values. For the 4-inch tests, the geometry was not scaled explicitly. However, physical phenomena influencing flow regime and flow separation such as a horizontal section, a 90* elbow, a vertical section and a U-bend were modeled. This procedure was adopted for the 4-inch tests partly because of the scoping nature of the tests and partly because we wanted to expand the data base for phenomenological understanding of flow regimes in large vertical pipes and reactor geometries. 4-27

                      ._       ._       __       . - .        .-     __           ~    .-

I , Flow regime maps for pipe flows have been expressed in terms of various nondimen-sional parameters (23, 2_4, 2_5). Based on engineering judgment and available literature, these nondimensional parameters were examined in defining the test conditions for this experimental program. The rationale for this approach as opposed to initiating a fundamental scaling study was the phenomenological nature of these tests and the usage of the results for code validation purposes. Following Taitel and Duckler (3), gas and liquid superficial velocities were selected as the pertinent parameter in defining the tests matrix in both series of tests. Test conditions were selected to cover a wide range of gas and liquid superficial velocities (.01 m/s < j g< 1.5 ms, O < jj < 0.5 m/s) expected in a reactor geometry during a small break (58) LOCA (<2% decay power). For the 12-inch tests jj was kept below 6 cm/s since it was recognized to be the limiting j liquid superficial velocity at which breakdown in natural circulation becomes an  ! issue. Tests at high gas velocities (jg > 1.5 m/s) were also performed for comparison with semianalytical predictions. 4.2.3 Scaling Compromises Investigated i The flow regime characterization project addresses two MIST atypicalities--the hot leg separation and the hot leg flow regime. Both these atypicalities affect the issue of interruption of natural circulation at MIST. The hot leg flow regime atypicality also effects the reestablishment of natural circulation issue. These atypicalities result from the small pipe diameters of the MIST facility. 4.2.4 Test Procedure With the loop supply tank (water reservoir) and both Tanks 1 and 2 filled with

,               water, the loop pump is started and the loop supply water is recirculated into the

{ supply tank. Then valve V1 is opened slowly and water is introduced into the loop. The flow is then recirculated via the hot leg and the top tank overflow pipe to Tank 2. Once the system is filled, the water flow is stopped and a 10-second reading of all instrument outputs are recorded. These data are used to check the pressure transducers and the gamma densitometer calibrations. The instrument calibrations are also checked with the loop empty at the beginning of each day of testing. With the pump running, value VI is adjusted until the desired water flow rate is obtained. At this time, the air pressure regulator is set at 10 psig for the 4-28 4

  . , . , _ . -          - - - - -,, .,-w ,   - ,--    , , - , . - - - , - - ---..--, , - , -      --

4-inch tests and at 20 psig for the 12-inch tests and valve V8 is gradually opened and adjusted until the desired air flow rate is established. The loop flow is then allowed to recirculate for about 15 mir.utes to ensure steady-state conditions. In earlier shakedown tests, it was determined that steady-state can be reached in a much shorter time period. The criteria for steady-state conditions were set based on comparison of test results with identical air and water flow rates but different test durations. When additional test time did not alter the results, it was determined that steady-state condi-tions were reached. Finally, each test data are recorded for 5 minutes. For the 12-inch test, addi-tional data are recorded while the densitometer elevation is varied. At the con-clusion of each test, the air and water supplies are simultaneously (time error of

 = 1 second) shut off and the volume of water trapped in the test section is measured. The collapsed water level is then determined.

For the 4-inch tests, the volume of water trapped in the test section is measured by draining the test section water content into the measurement tank via valve V14 (Figure 4-5). To reduce uncertainties in the measurement, the measurement tank was initially filled with water to a height of 2-inches. This excluded the necessity of calculating the volume of the drain flange region at the bottom of I the tank. { For the 12-inch tests, the system (pressure vessel and hot leg) volume as a func-tion of water level in the hot leg and the pressure vessel is predetermined by direct measurements (weighing the water). After the days tests are terminated, water in the loop is drained by proper positioning of the loop valves. 4.2.5 Application of Data The hot leg separation and hot leg flow regime atypicalities in MIST result mainly from geometrical considerations. Although MIST is a high pressure facility, the pipe diameters representing the hot leg in that experiment are an order of magni-tude smaller than prototypical values. This atypicality has raised concerns about extrapolation of MIST results for prediction of flow regime and flow separation in a prototypical hot leg. 4-29

             .-              - ~_      -       - _- -.                                       . -                        . . . =_.       .         .

T j-i This SAIC test data will be used to formulate computer code models for natural circulation in PWR hot leg geometries. These codes can then be used to determine I under what conditions natural circulation will break down and reestablish itself. Pipe diameter effects in these models will be verified with SAIC 4-inch and 12-inch test data. Pressure effects will be verified with MIST and other data

,               including SAI 4-inch data.

. Flow Regime Currently, the most'widely used flow maps are based either on empirical correla-tions (2,2) (Figure 4-7) or simplified semianalytical analysis (6) (Fiaure 4-8). Since a basis for scalability of empirical correlations can not be justified due to lack of phenomenological models, for this work we will focus on examining the i. semianalytical approach. More specifically, we plan to develop a flow regime map j which is compatible with the semianalytical modeling of adiabatic two-phase flow in vertical pipes presented by Taitel and Dukler (3). Their analytical basis was { selected partly due to the fact that their approach has been relatively successful I f in predicting the existing experimental data (generally small pipes), partly j because it is based on phenomenological models, and mainly, because if verified, j " Scale up" can easily be implemented. If the result of our. tests do not agree 4 with Taitel and Dukler (3) predictions, then, either the existing correlations ! have to be modified or new correlations have to be developed for extrapolating MIST results to prototypical plants. I i { in what follows the flow regime transition correlations based on Taitel and j Dukler (3), are summarized. The semianalytical correlations presented here are f based on Taitel and Dukler's most recent publication on the subject (3). I i Two-Phase Flow Patterns and Transition Mechanisms in Vertical Pipes (3) l I l. Figure 4-8 shows a generalized flow map for an air-water two-phase flow in a ] vertical tube. To quantify the map,_Fiqure 4-9 illustrates the results for a 5 cm i diameter vertical pipe. The underlying mechanism for each transition region, j according to Taitel et al. (3), is as follows: I: i 4-30 1 1

  -- . - - -              . , _ .  . -    - - - - - - - . . , . . , - . , . . - - - - .          . ~ - . . - , . , , , -          . - . - - - , -

6 - 10 ~,-106 - 5 10 - 10 5 - _ 4 10 -

                      -10 4                                                      ,-,' '

Annular ,/ spy 103 _ - 10 3 s Annular - t . 8 E 2 _10 ,,,,________{___,________ .

 "g Churn,,_j 4*
                                                              ,s                 '

10 - '

                    -10                                 ,                          i        Bubbly                   -

l \ 4  % 1 s-1 - i '

                   '                        #                                                                       ~

i Bubbly-slug . 1 10-I_ ..10-1 SIU9

                  -s 10-2_s%                                                                                                              -

a 2' 2 4 6 N, kg/s2 m , ,01 1,0 10,3 1,0 10,5 10 s , , , , , e o i - s s a

      "                                                      2            3                 4          5          6 1       10               10            10               10           10        10 2

lb/s ft Pg j, 2 Figure 4-7, Flow Pattern Map for Vertical Flow (Hewitt and Roberts (22)) 4-31

I I I I _. FIllELYDISPERSEDBUBBLE(II) / c B.

                                                                              \
                                                                                \

BUBBLE (I) A

                                                               \                  g i
                                             '                                        l d                                              s                  \
                                                    \                  g               i s                   SLUGORCHpRN                       i s                   \ (IV)         g               i
                                        \                   I            l              D
                                          \

l D l

                                            \                             l D                           l         ANI!ULAR
     -                                        g n                         i           IV) g                          ,

I D I 1 i I g

                                                ,                         i              i
                                                                )

I I I I A _. increasing i fE I I

• Io i I I E I l 1 g I I I g l I 3

{I i i i 11 g d g Figure 4-8. Flow Pattern Map for Vertical Tubes (Taitel and Dukler (3)) i 4-32

I I I i 10 - flNELY DISPERSED BUBBLE (II) c

                                                                                                         /

B.

                                                                                            \

s \ A N

                                                                                                \

BUBBLE s i 1.0 - II) s g , s I g s SLUG OR CH RN

                                               \

s N \ (IV) 1 I 5 \ \ l D

                                                     \
 .,7                                                                1                    D          I I                  ANNULAR D                                l 0.1  -

I I (V) g g I 1 SLUG D l (III) g I 1 i i I g i f I l t I i l 1 0.01 l E I l l 1E/D = 50 100 2g0 500 I I l l 3 I I i I e 11 I 0.1 1.0 10.0 100 3g (m/sec) Figure 4-9. Flow Pattern Map for Vertical Tubes (Taitel and Dukler (3)) 5.0 cm dia., Air-Water at 25 C, N/sq. cm. 4-33

i e Bubbly Flow at Low / Moderate Gas Flow Rates The necessary condition for the existance of this regime (Region I Figure 4-8) is that the terminal velocity of relatively large bubbles (correlated Harmathy (26) be less than the terminal velocity of Taylor bubbles (correlated by Nicklin (27). As a result, Region I of Figure 4-8 will not exist, if p 02 gj4 (, _,), s 4.36 (4-48) e In other words, Region I in the map will be covered by slug flow. The above inequality implies that, for air-water flow at low pressures, we will have bubbly flow in Region I if D ;t 5 cm (4-49) e Bubbly-Slug Flow Transition (Line A) 1 Relating liquid velocity, gas velocity, and bubble velocity and assuming that coalescence of deformed bubbles takes place at a void fraction of 0.25, the following equation for transition from bubbly flow to slug churn flow (curve A) is obtained jg = 3.0j - 1.15 g(pl~DG)" I/4 g 2 (4-50 "L It is noted that the flow regime transition is independent of pipe diameter. e Bubble to Dispersed Bubble Flow Transition (Line B) In bubbly flow regime, coalescence of bubbles and dispersion due to turbulence compete with each other. Where turbulent dispersion is dominant, the flow will be bubbly even for a a > 0.25. The following i equation specifies the threshold of the dominance of turbulent dispersion (LineB). l 4-34 i

0 0.446 J +j = 4.0 0 429(,7,()0.089 [g(,[_,)] (4-51) g g v 0.072 o There is a dependence on tube diameter for this transition. e Churn to Slug Flow Transition (Line D) l Churn flow regime represents an entry region phenomena leading to slug l flow further along the pipe. In churn low, the short unstable slug of liquid falls back and emerges with the liquid coming from below causing it to approximately double its length. In this process the Taylor bubble following the liquid slug overtakes the leading Taylor bubbles and coalesces with it as the slug between the bubbles collapses. This process repeats itself and the length of the liquid slugs as well as the lengths of the Taylor bubbles increases as they move upward, until the liquid slug is long enough to be stable and form a competent bridge between two consecutive Taylor bubbles. Between the inlet and the position at which a stable slug is formed, the liquid slug alternatively rises and falls, and this is precisely the condition of churn flow. Modeling the phenomena and making simplifying assumptions, the transition height for slug to churn is given by: 1 (4-52)

                                        =40.6(h)+0.22 There is a dependence of this height on tube diameter and entrance length, i                       e    Transition to Annular Flow Annular flow is characterized by the continuity of phase along the pipe in the core. The liquid phase moves upward partly as a wavy liquid film and partially in the form of drops in the gas core. Annular flow cannot be sustained unless the gas velocity in the gas core is sufficiently high to lift the entrained droplets. The minimum gas velocity for entrainment can be found from a gravity / drag force balance for a stable droplet. The size of the droplet can be specified from the balance between forces due l                             to the surface tension and gas impact. This results in the following equation for transition to annular flow (Line E).

i l 4-35

1/2 1/4 = 3.1 (4-53) log (oL - DG II i As observed, the result is independent of tube size. Using the semianalytical results of Taitel and Dukler (3), transition boundaries were calculated for a prototypical plant, MIST, and SAIC experimental condi-tions. Figure 4-10 shows a comparison of transition boundary lines from bubbly to slug flow. These results basically indicate the effect of pressure on bubbly / slug transition line, but do not include the effect of pipe diameter and/or entrance region on the flow regime. Preliminary results from SAIC tests indicate that differences also exist due to diameter effects. If further analysis of results show that the diameter effects are the same order of magnitude or larger than pressure effects, it is clear that new models need to be developed for MIST data j i extrapolation to prototypical plants. ' I Figure 4-11 shows transition boundary lines from churn / slug to annular flow. Again, the diameter effect is not predicted. Even though annular flow is not interested in MIST, the data can be used to develop a more generalized map. Since the MIST data is from a high pressure facility, comparison with SAIC data will also allow an evaluation of pressure effects. Hot leg U-Bend Separation Vapor segration at the U-bend depends on bubble rise versus throughput t hes. The time for fluid to travel through the U-bend can be expressed as: nR

                      *1

• T where R = U-bed radius of curvature 4

i J = mixture superficial velocity The time for bubbles to migrate to the top of the pipe is the vertical distance divided by bubble velocity. This may be written as: i 4-36

Bubbly / Slug Transition

                                                                /

1.0 _,,

                          ~

Prototype Experiment 4 3. (high pressure) steam-water y (air-water)

,                 w       -
                                            !!            .h 0.1    -
                                           $m              *

' ~ 0.01 i i i t ik iil i , i i i i n il i i i iiiii

                        .01                                  .1                          1.0                                  10.0 i

jg (m/sec) Figure 4-10. Predicted Bubbly to Slug Flow Transition Lines for Plant vs. Experiment i l 1 l 4-37

Prototype (high pressure) steam-water 1.0 - Experiment [ (air-water) O s - E Churn / Slug to u L " Annular Transition 0.1 _ a

                                                                                                         \.5          ,a
      .01                    i    t t 1    ii i II        i   i    iiitii l                     r   i e    i i i i il     '

i i iiiiin

          .01                                     .1                        .98                                          14.57 Jg         (m/s)

Figure 4-11. Predicted Churn / Slug to Annular Flow Regime Transition Lines for Plant vs. Experiment

d

                                          '2 " Tr where d = pipe diameter V = bubble rise velocity 7

l The relative tendency of voids to accumulate at the U-bend is then estimated by j the ratio of traverse to bubble rise time, or

                                                       *1 r=-
                                                       '2 Rewriting the above in terms of ratios of prototypical plant values to model values one may write:

4 S1 (t-traverse time) = ff=ff(h)2 l S2 (t-bubble mitigation) = h i f S(voidaccumulation) = f{f=fhh j Where capital letters refer to plant values and lower case letters refer to the model and j J & J = mixture volumetric flux d & D = pipe diameter r & R = U-bend radius of currentive j q & Q = msss flow rate 1 Substituting the appropriate physical parameters, value of S was calculated to be near unity for SAIC 12-inch tests. This implies that void accumulation is about i the same in the experiment as in the plant. I l j To investigate interruption in natural circulation, emphasis will be focused on i correlating void fraction and liquid collapsed level in the vertical section to l the amount of liquid carryover in the U-bend. More specifically, amount of liquid j carryover at a given gas superficial velocity as a function of collapsed liquid i

;                                                                                                                                               4-39 a

1

j level and liquid superficial velocity will be determined. The results can then be used to develop new models and validate codes for MIST data extrapolation to

prototypical plants.

i i 1

,                                                                                                                                                                       i 1

I i l 4 i t i l J

4-40 l

_3

                                                                                                    ,(.            s
                                                                                                                                               /          1 5

k

                                                                                                                                             \

Qq's lj 46 fs AJ 4.3 SAIC--MIST AUXILIARY FEED STUDY One of the requirements in the MIST program is the development and validation of models and correlations to be used for analyzing the MIST data. To meet this h l need, a separate effects test facility operating at atmospheric pressure is being ,' designed (L8). The objective of these tests is to provide data and understanding of subcooled auxiliary feedw'ater spreading in the superheated steam region at the j top of the steam generator secondary side. The information will be incorporated l in models and correlh,t' ions used to analyze t$e prototype MIST experiments. The proposed model facility will use the same test cross-section geometry as the MIST ( facility. However,gthe model tests will be conducted at near atmospheric pressure such that [words missing?] and transparent windows will be installed for [ visualization.

The test conditions will be selected to simulate major phenomena expected to occur I

in MIST. Some of these physical phenomena that need to be simulated are: c'oun-tercurrent flow at the tube support plate holes, water droplet entrainment in j upward steam flow, spreading of the auxiliary feedwater on the uppermost tube' ( -(

supportplate,'andtubewettingcharacteristjcs. The primary-to-secondary, heat transfer is influenced by these phenomena. The heat transfer rate, in turn, '

I strongly influences the primary and secondary side exit temperatures, the steam ' i ( ] boiloff rate, and the secondary side refill rate. , s ) " i I 4.3.1 Facility Description , I i A schematic of the conceptual test section design is shown in Figure 4-12. The experimental setup will utilize a 4-foot section of a 19-inch bundle. Hot pres-

surized water (250*F) and room temperature water will be the working fluids for i s- J '

i thesimulatedprimarysystemcoolantandtheauxiliaryfeedwater,respectively. Slightly superheated steam (215'F) at atmospheric pressure will be 03ed to simulate tolloff from the water pool in the bottom of the OTSG. A schematic of i j the conceptual experimental design is shown in Figure 4-13. } i Quantitative measurements of flow and temperature will be made at key locations in the various flow loops such that mass and energy balances can be established.

 ,                           Also, temperature measurements will be made at the outlet of selected tubes to f                             determinetheoverallheattraNfefcdefficientpertube. Windows in the upper 4

j portion of the tube bundle will allowg visual observations. 4 {

                                                               !                                                                                          e s

i 4-41  ? l ' i

> g, s. ,

__ _ _ . _ _ _ _.._ _-_ _ _. _ _ . _ _ _ _ _ _ ____ _. _ . . l._ _._ . . _ J

t n OOO 0000

                                                  =            00000 0000                                                A.,<,ARY i

3 000 ffE& DATER tuollLE p PRIMARY INLET l . - _ . _ . _ _ . . . . _ _ U (f 4 - AUIILIARY F tid'JAlt R g ,,__% N0llL[ 1 N. -

                                                                                                                    ~   ~ _

I s 'I//I//$/NY///5////////////h/E I/L 1U8C SUPVDR1 FLAll l f

      ,                                         $;gm     -  _ _          . .
 \',                                            INLII                                                                       % DRAIN
        }       t
  \,

U

        ~

PRIMARY Outlti Figure 4-12. SAIC Auxiliary Feedwater Test Section Schematic l J l

                                  ,.                                                  4-42 s.
         +

g .

          .I__,. _ - _ _ _ - .        _   _     __       _ _ , . . _,                  _ _ _ _ _ - -                                           ,- _ __ _ , - . , . .

                                                                                               ~.

9 ,4

                                                                                                  - g~                           \                                                                                 !:
                                                                                                                      \

t t

                                                                                                                ?
                                                                                                                                                                                                                      , ,x . '        )

.j i f /

                                                                                  '                                                                                "                                                                                             l

[ r l

                                                                                                                                                                                                                                         ", '                         y'

( .

                                                                                                                                                                                               ,1
                                                                                                                                                                                             ' f: (.         s PRIMAAf WATER p ,p                           THROTTL1h4                '

P 250*F

    '                                                          V                                                      30 psia CYCLCNE T/C     ggggggg q                  SEPARAT04                                                                                                       t l

V p STEAM OUT Aut!LIARY

                                                                                                        $lh?[$

t fil3 k 70'F  ! I'111 kATER L TRAP r BYPAS5 . . i .gVALVE T/C

• l  ;

HOTWATERmT! i s

                                                                                                                                                                                                                                                          \

l s 4 q, i (. - r J TH20TTLIhG SUPE 2 HEATER YAOI STEAMP215'Ff) LTEAN' h g j f GENEAATOR i -  : y CRITICE

                                                                                                                                       ~~

) 4 PAIMAAY SYSTIM  !' ! WATE4  % i / j i n . lT/C AuxtuARY FEE 0 ' T/C ' ' , AEATC4 DRAIN ,

                                                                                                                      .                                           /               ,                  .

Figure 4-13. ' SAIC Auxiliary Feedwater Conce,7tual- Loop Design _ o k ( 4 , e e 4

                                                                                                      ^
                                                                                                                                                                                                                '?

r  !

                  ^

4-43

                                                                                                                                                                                                                                        ~'

_ . _ _ . _ _ _ _ _ _ _ _ . . . , - , _ _ _ . . . , . . . _ _.,_m_, _ _ _ - . . , _ , _. . _ . _ _ __

4.3.2 Scaling Approach Steam condensation in the vicinity of the feedwater jet and upper tube support plate and its effect on other phenomena (spreading, wetting, and entrainment) was first examined. Condensation of steam was estimated from a simple energy balance. Such an analysis showed that the condensation rate would be typically a few percent of the feedwater flow rate. However, since more detailed modeling and/or experimental data are required to estimate the amount of steam condensation with my degree of certainty, this issue is being examined further. If calcula-tions and/or experimental data show that a substantial amount of steam is condensed in the vicinity of the feedwater jet, then auxiliary feedwater and steam flow rates will be adjusted accordingly to compensate for this phenomena. Under prototypical (MIST) operating conditions it is expected +, hat countercurrent flow and in the limit " flooding" to occur at the tube support plate holes. Clearly, flooding at the tube support plate holes is dependent on the numbei of l holes through which the auxiliary feedwater is expected to pass. In this context, flooding criterion (2,9) was examined and, assuming uniform spreading of water on the tube support plate, the maximum number of holes at which flooding is still possible was determined. The result of these calculations are shown in Figure 4-14. This figure suggests that if the auxiliary feedwater is not spread over more than about nine holes, flooding at the tube support plate holes is a strong possibility. In order to ensure observation of such pher.omenon in the nodel, " flooding" similarity (2_8) must be maintained. This is achieved by letting: ((j*)1/2 + (j*)1/2 ,[(j*)1/2+(j*)1/2g where the subscript p refers to the prototype (MIST), m refers to the model, and (j* and jj are dimensionless volumetric fluxes for vapor and liquid, respectively. The model will be operated at approximately the same feedwater flow rate as that of the prototype which in turn requires (j* to be approximately the same for both the prototype and the model. Since the model will operate at atmo-spheric ~ pressure and the MIST facility operates at approximately 1000 psia, calcu-lations accounting for the different densities show that the required steam mass flowrate for the model must be about one-eighth of the steam mass flowrate in MIST. The difference in density also results in a higher steam velocity in the cradel . It is noted that the above flooding similarity also satisfies the dynamic similarity of the drag force on the droplets. 4-44 l \ - - . ___ .-

l 1.4 1.3 - 1.3 - 1.2 - 7, 1.1 - m

          +                                                1          -

m .9 -

                                                  .8                  -

Flooding Criteria [29]

                                                  .7                  -
                                                   .6                     '         '      '        t        i          I      i        i       1 0     2       4        6        8        10        12     14       16      18       20 Number of TSP Holes for Liquid Flow Figure 4-14. Flooding Prediction (29) at the Tube Support Plate it

The possibility of Leidenfrost phenomenon was also examined in the prototype. Inferring from primary side temperature traces obtained in the GERDA facility, which is very similar to the MIST facility, no Leidenfrost phenomenon is expected to occur in MIST. In the model, the vapor temperature and hence the tubs wall temperature were also calculated to be well below the Leidenfrost temperature. Both primary and secondary side heat transfer phenomena were considered for the overall analysis. The primary side heat transfer coefficient is maintained in the same range between model and MIST by using the Dittus-Boelter correlation. Also, since the tubes in the model are prototypical, tube conductance remains approxi-mately the same. The secondary side heat transfer coefficient is not scaled explicitly. However, physical phenomena influencing this coefficient such as wetting of the tubes and spreading of the auxiliary feedwater on the tube support plate are being modeled. Results of these tests will thus provide a phenomenological under-standing of the secondary side heat transfer process. Finally, it is expected that the liquid will break up at the tube support plate and the tube walls by the upward steam flow and form droplets. Some of these droplets will be entrained in the steam flow and eventually be carried out of the test section. Thus droplet formation and entrainment of droplets in the main steam flow were considered in the vicinity of tube support plate holes. Bounding calculations were carried out to simulate these phenomena. Similarity of droplet formation was examined by considering the Weber number criteria. The critical 2 droplet diameter (largest droplet size) may be obtained from DC " N'c'/(D Vg g ), where We c is the critical Weber number, o is surface tension, og and Vg are vapor-density and velocity, and D cis the maximum droplet diameter achievable without breakup. Next, the critical diameter to lift the droplets was calculated by balancing the drag force and the gravity force. This yields dc = 3/8 CDogV2 g /(otg), where CD is drag coefficient, at is liquid density, and g is the gravitational constant. By equating these two diameters, a limiting velocity is determined. At this velocity, the steam flow is capable of entraining the largest droplets that are formed. These calculations also show that the limiting veloci-ties between MIST and the model are scaled by approximately the same factor as the bulk steam velocities. 4-46

Table 4-1 SUP94ARY OF SCALING ANALYSIS FOR SAIC AUX FEED TESTS Prototypical Value Parameter (MIST) Model Value Aux. Feedwater 100*F 65'F Temperature Aux. Feedwater Flow 360 lbm/hr 360 lbm/hr Maximum Secondary 0.1 lbm/sec 0.013 lbm/sec Steam Flow Pressure 1000 psia 14.7 psia Saturation Temp. 545'F 212'F Steam Density 2.3 lbm/ft 3 0.038 lbm/ft 3 Liquid Density 62 lbm/ft 3 62 lbm/ft 3 Total Cross Sectional 0.0506 ft 2 0.0506 ft 2 Area Tube Support Plate 0.0204 ft 2 0.0204 ft 2 Flow Area Steam Velocity 0.8 ft/sec 7 ft/sec Steam Velocity Through 2.5 ft/sec ~17 ft/sec Tube Support Plate SurfaceTension(o) 4.5 x 10-3 lbf/ft ~5 x 10-3 lbf/ft 4-47

  . .                        . ,.     . -               =         - _                               . .     -

1 Table 4 1 presents the results of the present scaling analysis. It is noted that I for similarity of physical phenomena, the steam mass flow in the model is much smaller than that in the MIST facility, although the steam velocity is signifi-cantly larger. This results from the large difference in steam density in the two test facilities. i In summary, a simple scaling analysis has been performed to ensure similarity of expected phenomena related to auxiliary feedwater injection in the MIST facility  ! and in an atmospheric pressure model currently under design. A summary of perti-nent scaled parameters and their comparison with MIST values is shown in Table 4-1. 4.3.3 Issues Addressed by the Project l l The objective of the auxiliary feedwater project at SAIC is to investigate AFW spreading and wetting in a tube bundle which is identical in cross section to that used in MIST. This project will address the AFW multidimensional atypicality of MIST which in turn affects the establishment of the boiler-condenser mode issue l l and also the interloop interactions and oscillations issue of the MIST program, j l 4.3.4 Test Procedure

Test conditions will be selected to simulate a wide range of steam velocity, auxiliary feedwater rate, and asymetric boiloff rate expected to occur in MIST during a SBLOCA. In all tests primary system tube flow will be held constant and steady state reached before auxiliary feedwater is injected. The test will then continue until steady state is reached while AFW is injected at a constant rate.

A tentative test matrix has been selected to vary the steam boiloff rate between 5-40 lbm/hr., and the auxiliary feed between 0.5 to 1.5 gpm through three I nozzles. Primary system flow rate will be held constant at about 3 gpm per tube. The test matrix will be finalized as soon as new information becomes available to us and after some scoping tests are carried out. 4.3.5 Application of Data l 1 l The AFW multidimensional atypicality in MIST stems from a geometry effect. In the . MIST experiment the number of tubes per AFW nozzle was approximately scaled the same as prototypical values. However, the ratio of internal to peripheral tubes in MIST is considerably lower than in a prototypical steam generator. Thus, the spreading and tube wetting of water from the AFW nozzles to the internal part of the tube bundle can be significantly different in these two geometries. 4-48 _ - . _ _ _ _ _ _ _ _ _ _ _ . _ . . . __ . ~ . _ , , _ _ - - _ _ _ _ _ . __

                                   ~_         _ . _ - .               -                 .         .     . _ .    -_-

l~ SAIC's auxiliary feedwater (AFW) project is attempting to address the extent of

! this AFW atypicality by investigating the spreading and wetting characteristics of AFW on the uppermost tubesheet support plate of a tube bundle having the same-cross-sectional geometry as that used in the MIST experiments. The SAIC experi-ment is low pressure to allow visual observations. Additionally, individual tube inlet and outlet flow temperatures will be measured which will allow tube heat transfer coefficients to be determined. These heat transfer coefficient deter-minations, coupled with the visual observations, will show the extent of AFW spreading in the tube bundle. If the AFW does not spread to the internal portion of the tube bundle and wet those tubes, then the AFW multidimensional atypicality in MIST may not be significant at all. However, if many tubes are wetted, new models must be developed and incorporated in the codes for analysis of MIST data                        .

S to account for this atypicality. The result of these tests will also yield valuable information for heat transfer modeling in prototypical steam generators. SAIC's test data will be used to develop computer code models. These codes will then be run to validate MIST data. Since the MIST data is from a high pressure facility, comparison with SAIC data will also allow an evaluation of pressure effects. l 4-49

      --      - - . . _.   . --  .     . . . . - .      . . . . ~ , - - - - . . . . . ,   . . - . _ - .       . . -

I Section 5 ANALYTECAL MODELS AND CODES l l

                             , ~ . . _ _ _

5.1 TRAC-PFI/ MODI COMPUTER CODE The thermal-hydraulic computer codes represent perhaps the best overall scaling tool for assessing the impact of the atypicalities associated with the various integral experient facilities associated with the IST program because the codes generally contain a reasonably complete set of field equations to describe the phenomena involved in the facilities. This statement is not intended to question the value, validity, or insight provided by more conventional scaling methodol-ogies; nor is it intended to suggest that any one code is necessarily superior to others. The computer codes have the capability to model the various integral facilities and the plant prototype at the correct scale and' fluid conditions. In this way the effect of various atypicalities in the experiments can be assessed against plant performance. However, the calculated results are dependent on the constitu-tive relations and assumptions built into the codes. In particular, the codes use simplified flow-regime maps for determining interfacial drag, heat transfer, and mass transfer, and generally rely on a one-dimensional representation. The transient reactor analysis code (TRAC) is one of several codes providing analyses in support of the IST program. This code can be used to address several of the MIST atypicalities, the resolution of which are not amenable through the other integral facilities. The code calculations are not subject to all of the facility-imposed limitations. 5.1.1 TRAC Description The TRAC code is an advanced, best-estimate systems code for analyzing light-water reactor accidents. The Office of Nuclear Regulatory Research of the U.S. Nuclear Regulatory Commission sponsors the development of the code at the Los Alamos National Laboratory. The TRAC series of codes is intended to provide an advanced best-estimate predictive capability for the analysis of postulated accidents in light-water reactors. Current TRAC versions from Los Alamos provide this analysis capability for pressurized water reactors (PWRs) and for many thermal-hydraulic test facilities. The TRAC code is based on a finite-difference representation of the field equa-tions necessary to describe the thermal-hydraulic behavior of fluid systems. The code features a three-dimensional treatment of the reactor vessel and its asso-ciated internals and a one-dimensional representation of the flow in piping networks. The basic features include two-phase nonequilibrium hydrodynamic 5-1

models, flow-regime-dependent constitutive relations, optional reflood-tracking capability for both bottom-reflood and falling-film quench fronts, and consistent treatment of entire accident scenarios, including the generation of consistent steady-state conditions. The early TRAC versions were designed primarily for the analysis of large-break loss-of-coolant accidents (LOCAs), with the latter versions adding capability to analyze small-break LOCAs and non-LOCA transients. Since the inception of the development work, we have released several versions of the code for general use, with <!ach new version containing corrections, improved constitutive relations, new models and features, etc., relative to its prede-cessor. Reference 30 describes the latest version, TRAC-PF1/M001. This code utilizes a two-fluid, six-equation hydrodynamic model in both the three-dimensional VESSEL component and the one-dimensional components. The VESSEL component uses a semi-implicit numerical scheme, but the one-dimensional com-ponents use the new stability-enhancing two-step (SETS) numerical scheme to eliminate the material Courant limit on the time-step size. Because of the SETS numerics, the code can provide greatly reduced computer central-processor-unit (CPU) times when analyzing slow transients with input models consisting only of one-dimensional components. The code is modular in both its internal architecture and in its interface with the user. The code architecture is not important to the current discussion and is described in Ref. 30. From a user perspective, the code uses one-dimensional components (PIPE, TEE, PLENUM, PUMP, VALVE, PRIZER, STGEN) to represent the reactor piping loops and the secondary and safety systems. The PIPE, TEE, and PLENUM components are the basic building blocks for constructing piping net-works. The PIPE and TEE components, which may consist of one or more fluid cells, include the momentum-flux terms in the momentum equation, and the TEE component provides for a single branch in the network. The PLENUM component provides for multiple network branches from a single cell and does not include the momentum-flux terms. The PUMP, VALVE, and PRIZER components are special-purpose components based on the PIPE component to represent pumps, valves, and pressurizers, respectively. The PUMP component uses homologous pump curves to model the behavior of centrifugal pumps. The VALVE component adjusts the flow area and hydraulic diameter at one interface between fluid cells to represent the opening and closing of valves. The most important feature of the PRIZER component permits the user to neglect make-up and let-down systems during steady-state analyses. The STGEN component allows the user to describe the flow paths in steam generators with PIPE and TEE components, with heat slabs to connect the primary and secondary ! 5-2

sides. Although there is an ACCUM component, the PIPE component provides a better representation of accumulators. The CORE component, based on the PIPE component, provides a one-dimensional treatment of a reactor care, with the fuel-rod model consistent with that in the three-dimensional VESSEL component. The FILL and BREAK components specify flow- and pressure-boundary conditions, respectively. The user can specify reactor systems and experiment facilities with any combina-tien of one- and three-dimensional components, subject to certain restrictions on the connections to VESSELS and PLENUMS. A transient, one-dimensional heat-conduction solution describes the structural mass in the system, including the steam-generator tubes; this conduction model allows for multiple materials and heat sources in the structural mass. (The ACCUM component does not include heat structures, and the VESSEL component does not permit modeling heat sources in the structural mass.) The code uses a two-dimensional transient heat-conduction solution to represent fuel rods (or simulators) in the core. The model assumes azimuthal symmetry, treats the axial conduction explicitly, and treats the radial conduction implicitly. The code uses a variable fine-meshing scheme in the axial direction to track steep axial temperature gradients in the rods; this model improves the calculation of dryout, rewetting, and reflood processes. The code couples the structural heat slabs and the rods to the hydraulic calcu-lation with a set of heat-transfer correlations and associated logic to partition the heat transfer between the liquid and vapor phases. The heat-transfer correla-tions include forced and natural convection to liquid, nucleate boiling, transi-tion boiling, film boiling, and forced convection to vapor. l The code uses a flow-regime map to calculate interfacial drag and interfacial heat and mass transfer. The map attempts to define various flow regimes based on void fraction and mass flux; the map includes stratified flow in both horizontal and inclined pipes. Under stratified-flow conditions, both the interfacial drag and interfacial area are minimums. Other flow regimes include bubbly, dispersed droplet, annular mist, slug, and churn turbulent. The TRAC-PF1/ MODI code includes a critical-flow model to obviate the detailed noding previously required in the vicinity of breaks. The code also contains additional conservation equations to permit tracking a noncondensable gas (nitrogen) in the gaseous phase of the system and a solute (boron) in the liquid 5-3

phase. Among other things these two additional fields can be used as markers to track the movement of liquid and vapor, e.g., the high-pressure-injection liquid. The code contains a general trip and control capability to facilitate modeling automatic and operator control actions. Two code-development efforts currently under way with TRAC will provide additional flexibility in the analyses. The heat structures will be generalized in a manner consistent with the treatment of the fuel rods, and thus the radial and axial conduction, together with the axial fine meshing in the structural heat slabs can be considered. The generalized heat structures will permit greater freedom in specifying node locations within the slabs, surface areas, and slab radius. The second effort involves extending the SETS numerics to the three-dimensional VESSEL component; this change in the numerics will permit the user to investigate the j macroscopic multidimensional behavior in the inlet annulus, the steam-generator secondaries, etc. 5.1.2 TRAC Scaling The traditional scaling methodologies rely on dimensionless parameters in an attempt to preserve phenomena at different scales. The dimensionless parameters are the coefficients that result from nondimensionalizing the field equations describing the phenomena of interest. However, if one considers more than one or two phenomena, one generally must make compromises that adversely impact the scaled phenomena. Basically, the experimentalist knows which things are scaled correctly and which are scaled incorrectly, but he has difficulty extrapolating integral experiment results to full scale. The TRAC code contains a reasonably complete set of two-fluid field equations to describe the phenomena, and thus provides a detailed, complex scaling tool to describe any combination of liquid water, steam, and air within the limits of the code's thermodynamir properties of water. Although the code is not constructed to work directly with the experimenter's dimensionless parameters, it is capable of reproducing the scaling considerations within the limitations of the field equations, the constitutive relations, and the assumed dimensionality; the major limitations are the constitutive relations, the assumed dimensionality, and the lack of a turbulence model. Because the code is not necessarily constrained by the same limitations forced by the facility design and because the code represents a general scaling tool, the 5-4

code provides a unique tool with which the experimentalist can explore the effects of the facility atypicalities. The approach involves correctly scaling down the prototype plant, or any part thereof, and comparing calculations at correct scale with facility calculations. Conversely, the facility can be scaled up and compared with prototype-scale calculations. In either case, the differences between sets of calculations can be attributed to the atypicalities of the flow in piping networks. 5.1.3 Scaling Compromises Investigated The TRAC code can provide useful insight into most of the atypicalities of concern to the MIST program; however, in those areas which question the details of flow regimes and other constitutive relations or which require the code to calculate more than one temperature and velocity for a given phase in a given cell (thermal stratification and countercurrent flow in a liquid-full cell), the flexibility of the code to address the scaling issues is limited by the existing field equations and constitutive relations built into the code. For purposes of addressing the MIST atypicalities, the main thrust of the code applications is in areas in which the other experiment facilities are substantially constrained by the same atypi-calities as MIST. Because of the small scale of all the integral facilities, they are in some way limited in their representation of structural masses and heat-transfer areas. The small scale and the desire not to exaggerate frictional, flow-regime, and separa-tion effects result in pipe sizes that are not geometrically scaled. Safety considerations then force heavier walls than desired, although some reduction in wall thickness can be achieved by reducing operating pressure and temperature. The bottom line is that the heat-transfer-area to fluid-volume and the structural-volume to heat-transfer-area ratios are distorted. The code is not constrained in the same way as the experiment facilities, and the user is somewhat free to specify the heat-structure area, thickness, and radius independent of the fluid-volume and safety considerations. (The generalized heat structures will be even more flexible in this regard.) This flexibility permits the code to address directly the effect of excess structural mass in the steam-generator secondary on the interloop interactions and oscillations. 1 l Similarly, the flexible specification of the heat structures and the capability to model heat sources inside the heat structures permits the code to model the excess mass and heat-transfer are associated with the primary piping and the effects of 5-5 i

the guard heaters, if necessary. The code can then calculate the effects on the reestablishment of natural circulation and long-term cooling. With regard to the RVVV simulation, the code can define the valve to operate in a large variety of ways. If the one-dimensional valve component is to represent the B&W characterization of the actual RVVV with a loss coefficients as a function of pressure drop across the valve, a straightforward code update will have to be made; the three-dimensional VESSEL component already includes the capability to use the B&W RVVV characterization. If the real valve is sufficiently character-ized, the code can simulate any of the current valves in the experiment facilities representing the RVVVs and also the correctly Maled B&W RVVV. The code automa-tically couples the RVVV simulation into the calculations of the interloop inter-actions and oscillations, the reestablishment of natural circulation, and the long-term cooling. The flexible nature of the STGEN component to model (within a one-dimensional constraint) the flow paths and heat structures associated with the steam-generator secondary permits the code to complete the secondary side volumes and flow paths that are ignored in the 19-tube OTSGs in the MIST facility. (In the near future, the combination of the three-dimensional SETS numerics in the VESSEL component and the generalized heat structures will permit a macroscopic multidimensional analy-sis of the steam-generator secondary.) This modeling capability together with the fact that the code can also represent the full steam-line rupture and the over-filling of the steam-generator secondary allows the code to address at either full scale or MIST scale the effects of a combined primary and secondary blowdown in combination with a steam-generator tube rupture. However, one may question the appropriateness of the interfacial-drag calculation in the steam-generator secondary. 5.1.4 Analysis Procedure In order for the code to show the effects of the atypicality, at least two calculations must be made, one correctly representing the facility and the other correctly representing the desired scaling. Tne differences between the calcula-tions can then be attributed to the atypicality. This process can be performed for each atypicality singly or in groups to explore the synergistic effects of the atypicalities. In the limit, one can see that the number of required calcula-tions, some of which can be very long, is large. Therefore, we should be very careful in defining the required calculations. Even without performing any such 5-6

calculations to explore the atypicalities, we know that computer codes can be used to bridge the atypicality gaps that are common to all of the integral facilities. 5.1.5 Application of Code Results In the limited areas discussed in Sec. VIII.C the TRAC code provides a unique capability to assess the impact of the atypicalities of MIST against the desired scaling. Within the accuracy of the calculations and the limitations of the field equations and constitutive relations, the code can provide a quantitative measure of the impact of the scaling flaws; the more conventional scaling methodologies can only indicate the scaling error but not the impact of the error. With the code we can decide if the scaling compromises encompassed in the MIST facility are acceptable in an integral sense. Ultimately, we will use the data from the MIST facility to assess the quality of the code calculations. If the data comparisons indicate that the code can ade-quately calculate the behavior of the MIST facility, we have increased our confidence in the applications of the code to plant-scale transients because the same set of field equations and constitutive relations will have been used for both scales. This approach ultimately relates the MIST data to the power plants. 5.2 TETRATECH/EPRI TWO-PHASE FLOW PUMP MODEL Reactor coolant pumps (RCP) play an important role in both steady state reactor operation and in transient or accident conditions. During a LOCA, the reactor core may undergo two-phase flow conditions, and pump performance characteristics change drastically from that for single-phase flow. Pump performance is known to degrade severely in two-phase flow. The degree of pump performance degradation depends on various physical, geometrical, and thermal conditions of the system. ! In reactor safety analysis, it is, therefore, important to predict pump perfor-mance under such varying conditions. Pump performance characteristics are germane to the initial transient, to the pump bump, and to pump-running transients. With regard to net positive suction head (NPSH) requirements in two-phase flow, the problem of possible cavitation formation in addition to the incoming two-phase flow may need attention at low pressures. However, it is not an easy problem to address and probably will require some testing. Currently, major system thermal-hydraulic analysis codes such as RETRAN, RELAP, and TRAC use the simple empirical correlation for two-phase pump performance developed by the Aerojet Nuclear Corporation. This correlation is based on 5-7

limited test data obtained from a small radial-flow type pump that is atypical of reactor coolant pumps. The deficiency of such an empirical correlation is serious as discussed in Ref. 31. Such an empirical correlation may be only valid for the particular pump type and conditions from which the correlation is derived. For the purpose of analyzing MIST tests, the pump performance model selected should be able to handle the particular geometry of the MIST pump and various thermal-hydraulic conditions. It should be noted that the primary consideration for selecting a pump type for MIST facility was leak-tightness and minimal heat loss and no effort was made to preserve specific speed (2). Indeed, the MIST pump has a specific speed of 477 in British units. Therefore, this type of pump has different characteristics even in single phase flow compared to either the ANC e pump (specific speed of 926) that produced the widely-used empirical correlation or a prototypic PWR RCP (specific speed of 4200) which is of mixed flow type. This fact points out two potential issues: how one can extrapolate the pump behavior and performance observed in MIST facility to that of a reactor pump, and which correlation or model should be used to predict the performance of the MIST pump. The first issue is not simple to answer. The second issue can be resolved by using a newly developed analytical model (32, 33). This new model was developed under ERPI sponsorship with aim to eliminate empiricism that is inherent in the existing models. The model is based on the first physical principles. It takes into account the effects of local void fraction, condensation, compressibility, and slip velocity and is able to handle any type of pump. The pump model is described below. 5.2.1 Tetra Tech /EPRI Analytical Two-Phase Pump Model Description The model (32, 33) utilizes a one-dimensional control volume approach. The liquid and vapor are assumed to follow a stream tube, which is assumed to be known from a rated condition. Figure 5-1 shows the control volume used in this model. Here, the streamline, s, and the direction normal to the streamline, n, on the stream surface are used as the curvilinear coordinates. The control volume is composed of ds, dn, and a unit thickness in the direction normal to the stream surface. The cross-sectional area normal to the streamline is therefore ledn. The model is based on two mass conservation equations, one for each phase, the mixture momentum equation, the gas momentum equation, and the mixture energy equation. 5-8

                                          /

j j / dmrwt s i

                 ;.       ai b            '

4 y 11 Yds

                                                                                                             % dr I     '                                                                                          I dm r=2 cosy i\ ,#         r
                                        .j                                        ,I,I                        1 ds si@'

i /_ rwa dm ,

                                                                                                   - ~I g                                                                                            g
           '\                  gs **"0'***'                                       \\                      l
                 \              \          p + (dp/ds) ds             *a           g\                    /
                  \                                              , i
                                                                                    \                   /
                    \                                         u,     #,2                  \

g 'g /

                                   \                                  v2
                                                                                ~

da sin $' cosy = dr

                                     \                                                                         dr/ds = sing' cosy
                                       \
                                         \

(b) Geometry of stream surface element. (a) Force diagrams on the fluid stream element. Figure 5-1. One-Dimensional Control Volume Method for Rotating Machines 1 5-9

The formulation taken from Furuya (33) is presented in the following. The mass conservation equations for the liquid and gas phases are given by, respectively; d g lo g- (1-a) wgAl = bg A (5-1) h [o g a wg A] = hg A (5-2) f where o's are densities, a the volumetric void fraction, w's the relative flow

; velocities, r the g    condensation rate, t and g denote the variables belonging to liquid and gas phase, respectively.                                                        -

The momentum equation for the two-phase flow mixture is written by balancing the momentum of the gas and liquid with the pressure force and centrifugal force 2 2 h[o(1-a)wA]+h[oawA] g (5-3)

         = - h A + [og (1-a) +go al rw hA where o denotes the static pressure, r the distance measured from the center of blade rotation, and w the angular velocity of pump rotor.

i On the other hand, the momentum equation for the gas phase is given t 2 h(oaw g A) g + Cy ,o g wgh[(w-w)aA] g g a A rw 2

         =-haA+o g                         +   oaA g                               (5-4)

(wg-wg ). wt -*g ' where C y , denotes the virtual mass coefficient, Cd the drag coefficient, and rg the radius of gas bubble. C vm is the virtual mass coefficient, which is set at 1/2 for spherical bubbles but chosen to be zero as the flow condition changes from uniform bubbly flow to churn turbulent flow, see papers by Hench and Johnston (34) i and Furuya (3_2). In the present analysis, this transition was set to start when the local void fraction a became 0.25. C vm was then linearly decreased to zero at a = 0.35. 5-10 4

The ratio of drag coefficient Cd to the bubble radius rg used in.the present numerical analysis is provided below:

               = 8.64 (in-I) or .34 (mm-1); 0 s a s 0.25 9                                       (bubbly flow region)
               = 2.79 (1-a)3 (in-1) or             ; 0.35 s a 3
                   .11 (1   a) {b )                   (churn turbulent region) i
               = Linear interpolation of the values at a = 0.25 and a = 0.35 (transition between the bubbly flow and churn turbulent region)                           ; 0.25 s a s 0.35 The Cd /"g value for the bubbly flow region is the one obtained by Haberman and Morton (35) whereas that for the churn turbulent region is the empirical value obtained by Zuber and Hench (36).

Assuming thermodynamic equilibrium between the phases, the mixture energy equation is written: l 5-11 l

hlo(1-a)(fw2 g . 3,),,,3j

          +hloa(jw    g        +9 '9} "9.Al 2                             (5-5)
          - [og (1-a) w g +o g a wg ] rw                A
          +[(1-a)h(pw)+ah(pw)]A=0  g                g where i denotes the internal energy.

By subtracting the mechanical energy part from the energy equation with the help of the momentum equations and using the mass conservation equations, one can express the condensation / evaporation rateg r as follows:

      .                  -1                 d(hqc g) 9                    ,2,,2 8g "*g        ds h gg[g       +

2 9 d(hgc) g

             + og(1-a)w s             ds 1

_A(,g d[paA] , ,& d[p(1-a)Al) (5-6) ds ds

             -f(ow(w-w)h[w-w)aA))

gg g g g g hoa(w-w)2 g g g ,t ,g where hgg = gh -h g , i.e., the difference in enthalpies between the gas and liquid phases at the saturated conditions. Equations 5-1, 5-2, 5-3, 5-4, and 5-6 completely determine the five unknowns wg , wg p, a, and r . g When gr s 0, the above formulation reduces to the case of noncondensable gas analyzed in Ref. 32. Fig. 5-2 shows a typical result for air / water system taken from Ref. 32. Head is shown to degrade gradually at low inlet void fraction and drops drastically at approximately a = 20%, corresponding to a flow regime change 5-12 1 1

o. 1.3 , . . . . . . . . 1.2 go L.1 -

                    ,*       ,             e ej",z    t0 N      ,            g       .8
            *3  -

0 1.o e . o

g i

2 .6 7

                %o 72
                               ,     o 1

j

            .5
            .4
                             ,                                            1.o                                        o 0                   D               #                                  1 g        .3  -

go

                                                             ,        f .2                                         a   .-

2 -

                                                       , , , , _ / /                                                 o
                                              ,5]ojip tgo
                                               ~

e 2 -

                                    *                                                       . ..                  o a                hoW                            ,,                    oy
            .1  -                                                  *                                                    -

o

           .,p       i         b        i        e             e   r       e                        i             e 0.0    .1        .2      .3        .4            .5  .6      .7                     .B              .9   1.0 Inlet Void Fraction, a G Creare air / water data o B&W air / water data
                              - Present theory Figure 5-2. Comparison of Present Theory for Homologous Heads with Creare and B&W Data for v/a Between 0.8 and 1.2. From Furuya (Reference 32) 5-13

from bubbly to churn turbulent. flow. Figure 5-3 and 5-4 show results for a steam / water system for a mixed flow pump and a radial flow pump, respectively. Compar-ing to Fig. 5-2, one noticeable feature is the increase in pump head at low void fraction. Such increase in pump head at low void fraction is due to vapor conden-sation effect within the pump impeller as explained in Ref. 33. 5.2.2 Application of Model A reasonable analytical model is at last available that can predict the pump performance under two-phase flow conditions. The model is based on the first physical principles and does not rely on empiricism as other models do. The model can be used in analyzing any type of pump, and is thus particularly useful in studying the performance characteristics of the MIST pump which is not of the same type as the prototypic PWR primary coolant pump. The model takes into account the effects of geometry, condensation, local void fraction, relative velocity between the phases, and compressibility. Comparison of the results obtained by the model with test data shows a reasonable agreement. 5-14

l 1.5 4 . 4 4 a 6 4 4 4

                                                                           ,CREARE -

n _ "N O C-E - [ o

                               /                                         - THEORY _

1.0

           $.          *o                           Steam / Water System 1.0 h3                 o i

i 1.2 o

0.8 ~

1.0 E 0.5 -

                               .               o*%,.                                 -

8 h o' % S O -

                   ~

k B

                                                      . 8         .

3 x W' ,

                                                                     ,    *. 1.2 0.0         '      '        '       '      '    '        '   '*     '

O.1 0.2 0.3 0.4 0.5 0.6 0.7 08 09 1.0 INLET VOto FRACTION, m y Figure 5-3. Comparison of the Head Between the Theory and Experimental Data for a Mixed-Flow Pump. From Furuya (Reference 4) 5-15 e

1.5 , , , , , a i i , e SEMISCALE ~ e - THEORY -

                                         .E._. 0J                                                                          -
        "a        _
                                  / ""                                                                                      _
        ; i.o                  _-
                                                  - '8                                                                      .

steam / water System g , 1.2 0.s I

        <                                                                                   t.0

a. _ -

O 30.5 -

        ,I         .                                                                                                       _

8 - e-8 m - o e _p s - -

                                                                                                         ,,,           e .

x 0.0 ------4--------------o-- _ e

• _
• e e * *

           -0.5           '     '       '                  '      '   '     '             '                         '

O.0 0.1 0.2 0.3 0.4 0.5 06 0. 7 08 0.9 1.0 INLET VOID FR ACTION. 4g Figure 5-4 Comparison of the Head Between the Theory and Experimental Data for a Radial-Flow Pump. From Furuya (Reference 4) 5-16

i i Section 6

SUMMARY

AND CONCLUSION r i f l t F

SUMMARY

AND CONCLUSION Two key coordination meetings, jointly sponsored by NRC and EPRI, were held in order to refocus the objectives of the IST support projects. The major product of these meetings is the effective integration of eight projects into the resolution of the IST licensing issues (1). The essence of this inte-gration is shown in Table 1-1. The support projects are all considered essential to the success of the MIST program. This report describes the individual IST support projects and their contribution to the resolution of specific IST issues raised by MIST scaling compromises. The IST support projects discussed in this report include integral facilities, separate effects facilities, and analytical projects: Integral Facilities Sponsors MIST NRC/EPRI/BWOG/B&W SRI-2 EPRI UMCP NRC Separate Effects Facilities ANL NRC SAIC-Flow regimes EPRI SAIC-Aux feed EPRI Analytical Projects TRAC-PF1 Code NRC TETRATECH-pump EPRI The scaling compromises introduced by MIST design may not allow a complete evalua-tion of all issues raised by the Test Advisory Group (1). The specific MIST scaling compromises and their relation to the major program issues are listed in the rirst column of Table 1-1. The table also shows the specific issues to be investigated by each support project. 6-1

                                                                            \

Thus, the effects of downcomer flow characteristics ar.d react:r vessel vent valve simulation on reestablishM nt of natural circulation, long-term cooling, and interloop interaction will be investigated by the three integral test facilities f # and by a methodical comparison of their respective results. In addition, three-dimensional TRAC calculations for the downcomer region will provide an indication s of flow fields in this region and its impact on transient behavior. i e Separateeffectfacilities(AHL,SAIC)willbeusedincoordination[withtheinte-gral test facilities to address the scaling questions related to hot leg flew regimes and U-bend phase separation. MIST auxiliary feedwater characteristics will be examined in another SAIC separate effect test facility in order to provide additional information regarding flow distribution in the secondary side. The TETRATECH/EPRI analytical pump modal,l based on conservation %uations, will

                                                                   ~

' help model the two-phase performance of the MIST pumps and wi11 provide the basis

                                                                                     . t to extrapolate to plant conditions.                                                      '<

The impact of excessive metal mass (relative to fluid volume) in MIST will be investigated by a parametric study with the TRAC code in combination with results from the MIST characterization tests. The MIST design of the secondary side limits the depressurization rate and, therefore, prohibits an optimum simulation of a double-ended steam line break. Again, the MIST test data (for a partial steam line break) will help qualify the TRAC code which will then be used to extrapolate to the more severe accident. 1 ItisalsoveryimportangtorecognizethattheH[STfacilityitselfcanbeused to analyze the impact of its own atypicalities. This can be done with parametric type tests already planned and with additional tests alreadyhdgeted but still requiring technical definition. ' g s There are two scaling compromises which could not be addressed by the support projects: 3

a. Pump cavitation: The MIST pumps NPSH is atypical. However, this is (

believed to be relatively insignificant to the major issues addressed by MIST. y 6  % 6-2 f t

                                                \          '

             .(

15 . Stratification of single-phase liquid in the cold leg: This atypicality cannot be fully assessed at this time due to code deficiencies in modeling. The key faatures of Table 1-1 are: o!' 1 It represents a consensus of the IST group, e It identifies specific areas where support projects are needed and how the data can be used. e It responds to almost every issue affected by MIST scaling compromises. The integration implied by Table 1-1 requires coordination of work among seven organizations working on eight different projects. The complete success of the program largely depends on the individual and timely success of each project. This coordination effort is an excellent illustration of how relatively small projects can directly imosct resolution of licensing issues (1). Recommendations e In view of the essential role assumed by the IST support projects in

         ,        helping to resolve licensing issues, management commitment and support is crucial.

i e A similar coordination effort should be undertaken to review the NRC or l EPRI projects aimed at the development of phenomenological models. The review should consider how these models might help in the resolution of IST issues, and how best to coordinate the individual efforts, e The support projects should be periodically reviewed by the IST group in order to assess progress, to transfer information, and to coordinate technical output as needed. e Confirmatory studies would be de.sirable, to ensure that the two MIST design compromises not covered by the support projects do not have a strong impact on the resolution of IST issues. 6-3

o i,' s

                          .J.

4 t s_

                                                                                                                                                 ,.             i
             ;         e        The purpose of the reporttis to assess the contribution:,of the support l'

project in resolving issces raised by MI5T design compromises. The support project, however, will produce a wealth of data and technical ,. insight which can also be used either to improve analytical models or to j assess computer codes.

                                                                                                                                                                  .i
                                                                                                                    ;                                              [ J.
                                              .I        .
                                                                                                                   - 3.
                                                                                                            .1'l
                  .)                                                                                                           ;\

r

• t i

T t

                                                   '6 '

{ O f., 5 I, i

                                                                                                                              ;)

s

                                                                                                                                                                             ?

! t 4

                                                                                                                                                                        ~=

8 g" o

                                                                                       %                                                                                           *s i

_l g '

, 5 4

I i a + 4 a #, t , , / k p 5 i k'

                                                                                                 .e l._

64 4 l , 3

4. ,

1 t

                                                                                      >         L                                      e
          .                  1 j                                    s                 >

'___t + l - r - s , _ , , - _ _ _ . . . ,_ , . - ., ,

i Section 7 REFERENCES f i 1 I l l l s

I t

1. " Test Advisory Group Final Report--Integral System Testing Program for the B&W-Designed NSS," BAW-1787, June 1983.

2. " Multi-Loop Integral System Test (MIST) Facility Specification,"

NRC-04-83-168, EPRI RP2399-1, Babcock & Wilcox.

3. Taitel, Y. and Dukler, A. E., "Modeling Flow Pattern Transitions for Steady Upward Gas-Liquid Flow in Vertical Tubes," AIChE, vol. 26, 3, pp. 345-354, (May 1980).

4. " MIST Design Review Board Final Report," B&W Report #84:4091-02:01:01, November 1983.

5. Kiang, R. L., " Decay Heat Removal Experiments in a UTSG One-Loop Test Facility," EPRI NP-3472, May 1984.

6. Ishii, M. and Kataoka, I., " Similarity Analysis and Scaling Criteria for LWRs Under Single-Phase and Two-Phase Natural Circulation," NUREG/CR-3267, May 1983.

7. Kocamustafaogullari & Ishii, " Scaling Criteria for Two-Phase Flow Natural and Forced Convection Loop and Their Application to Conceptual 2x4 Simulation Loop Design," NUREG/CR-3420, ANL-83-61, August 1983.

8. Lahey, R. T and Moody, F. J., "The Thermal-Hydraulics of a Boiling Water Nuclear Reactor," ANS/AEC Monograph Series (1977).

l

9. Ardron, K' H. and Krishnan V. S., " Stability of a Two-Phase Natural Circula-l tien Loop With Figure-of-Eight Symmetry," Proceedings Multiphase Flow and Heat Transfer Symposium III, Veziroglu, T. N., ed., Miami Beach, Florida.

April 18-20, 1983.

10. Zuber, N. and Staub, F. W., "The ?ropagation and the Wave Form of the Vapor Volumetric Concentration in Boiling Forced Convection System Under Oscillatory Conditions," Int. J. Heat Mass Transfer, vol. 9, pp. 871-895, 1966.

11. Zuber, N. and Staub, F. W., "An Analytical Investigation of the Transient Reponse of the Volumetric Concentration in a Boiling Forced-Flow System,"

Nucl. Sci. Eng., vol. 30, pp. 268-278, 1967.

12. Sallet, D. W., Hsu, Y. Y., Pertmer, G. A., et al., " Final Design Report for the UMCP 2x4 B&W Simulation Loop," The University of Maryland, August 1984.

13. Zierep, J., Ahnlichkeitsgesetze und Modellregeln der Stromungslehre, G. Braun Verlag, Karlsruhe, 1972.

14. Zuber, N., " Problems in Modeling of Small Break LOCA," NUREG-0724, October 1980,

15. Kocamustafaogullari, G., Chen, I. Y., and Ishii, M., " Unified Theory for Predicting Maximum Fluid Particle Size for Drops and Bubbles," NUREG/CR-4028, ANL-84-67, October 1984

16. Ishii, M. and Jones, 0., Jr., " Derivation.and Application of Scaling Criteria for Two-Phase Flows," Two-Phase Flows and Heat Transfer, Proc. Nato Advanced Study Institute, Istanbul, Turkey, vol. 1, p. 163 (1976).

17. Ishii, M. and Zuber, N., " Thermally Induced Flow Instabilities in Two-Phase Mixtures," 4th Intl. Heat Trans. Conf., Paris, paper 85.11 (1970).

7-1

18. Zuber, N., " Problems in Modeling of Small Break LOCA," in Heat Trans. in Nucl. )

Reactor Safety, S. G. Bankoff and N. H. Afgan, eds.,-p. 3. Hemisphere Pub. ' Corp. (1983).

19. Zuber, N. and Findlay, J. A., " Average Volumetric Concentration in Two-Phase Flow Systems," J. Heat Trans., vol. 87, p. 453 (1965).

20. Ishii, M., "One-Dimensional Drift-Flux Model and Constitutive Equations for Relative Motion Between Phases in Various Two-Phase Flow Regimes," Argonne National Laboratory Report ANL-77-47 (1977).

21. Zuber, N. and Findlay, J. A., " Average Volumetric Concentration in Two-Phase Systems," Trans. ASME J. Ht. Transfer, vol' 87, p. 453, (1965).

22. Hewitt, G. F. and Roberts, D. N., " Studies of Two-Phase Flow Pattern by Simultaneous X-ray and Flash Photography," AERE-M2159, (1969). I

23. Hsu, Y. Y. and Graham, R. W., " Transport Processes in Boiling and Two-Phase 1 Systems," pp. 160-186, Hemisphere Publication Corp., New York, (1976).

24. Delhaye, J. M., "Two-Phase Flow Patterns," in Thermohydrau11cs of Two-Phase Systems for Industrial Design and Nuclear Engineering, J. M. Delhaye, M. Giot and M. L. Reitmuller, eds., Hemisphere Publication Corp., pp. 37-70, New York, (1980).

25. Collier, J. G., Convective Boiling and Condensation, pp. 8-25, McGraw-Hill Book Company, New York, 1981.

1

26. Harmathy, T. Z., " Velocity of Large Drops and Bubbles in Media of Infinite or j Restricted Extent," AIChE, vol. 6, p. 281, (1960).

27. Nicklin, J. D. and Davidson, J. F., "The Onset of Instability on Two-Phase Slug

• Flow," Inst. Mech. Engr., (London), Proc. of Symp. on Two-Phase Flow, ,

Paper 4, (1962). l

28. Kim, J. H, Sursock, J. P., Hashemi, A., " Scaling Analysis for Auxiliary Feedwater Simulation in MIST Facility," ANS 1985 Annual Meeting, Boston, j June 9-14, 1985. Trans. ANS, Vol. 49, PP. 444-445, June 1985.

s

29. Wallis, G. B., One-Dimensional Two-Phase Flow, McGraw-Hill. 1969.

30. " TRAC-PF1/M001: An Advanced Best-Estimate Computer Program for Pressurized

} Water Reactor Thermal-Hydraulic Analysis," Los Alamos National Laboratory

Report LA-10157-MS, NUREG/CR-3858, September 1986.

i 31. Kim, J. H., " Perspectives on Two-Phase Flow Pump Modeling for Nuclear Reactor j Safety Analysis," Cavitation and Multiphase Flow Forum--1983, ed. by J. W. Hoyt, ASME, June 1983, pp. 29-33. 4 l 32. Furuya, 0., " Development of an Analytic Model to Determine Pump Performance Under Two-Phase Flow Conditions," EPRI NP-3519, May 1984.

33. Furuya, 0., "An Analytical Model for Prediction of Two-Phase Flow Pump j Performance--Condensable Flow Case," to appear in Cavitation and Multiphase j Flow Forum--1985, ed. by J. W. Hoyt and 0. Furuya, ASME, Albuquerque, June j 1985.

7-2 J

I

                                                                                                     'A .

34. Hench, J. E. and Johnston, J. P., "Two-Dimensional Diffuser Performance With Subsonic, Two-Phase, Air-Water Flow," Journal of Basic Engineering, ASME, March 1972.

35. Haberman, W. L. and Morton, R. K., "An Experimental Investigation of the Drag and Shape of Air Bubbles Rising in Various Liquids " David W. Taylor Naval Ship R&D Center Report 802, September 1953.

36. Zuber, N. and Hench, J. E., " Steady State and Transient Void Fraction of Bubbling Systems and Their Operating Limits. Part I, Steady State Operation "

General Electric Company Report No. 62GL100, 1962. 7-3

Appendix A FACILITY COORDINATION LETTERS FROM NRC/EPRI l 1 i l l l l i

EPRI Electnc Power Researen Institute February 25, 1985 Ab Hashemi (SAI) Bob Kiang (SRI) M. Ishii (ANL) Gary Pertmer (UM) Thad Knight (LASL) Bob Turner (B&W) Gentlemen: As you know, the meeting we held in Washington D.C. on February 20, 1985 led to a distribution of responsibilities and tasks aimed at resolving the IST issues associated with MIST atypicalities. The essence of the meeting is summarized in the attached table which relates the MIST , atypicalities to the IST issues and also shows how the l various support projects can potentially help resolve these issues. We are requesting that you study this table and provide us with a document describing how you are proposing to address the issues and how the project can contribute to their resolution. I. FORMAT The docuinent should contain the following sections:

1. Facility ,3r Code /Model) Description This section will describe very briefly the major characteristics of the facility (or model). The purpose here is to give the reader a reasonable " feeling" of the " tool" you are ,

proposing to use rather than providing excru-ciating detail. (Figures here are often worth a thousand words.)

2. Scaling This section should set the stage on how you plan to use the results for extrapolating to plant or to MIST. It should provide sufficient technical detail to enable the reader to follow the argument you will build in Section 5.

3412 HilMew Avenue, Post oftce Box 10412, Palo Alto CA 94303 Telephone (415) 855-2000 Washongton office: 1800 Massachusetts Ave. NW. Suite 700 Washir'gton, DC 20036 (202) 872-9222 A-1

February 25, 1985 Page two

3. List of Issues Addressed by the Project This section merely lists the issues and atypicalities, as defined in the attached table which you plan to address in your project. Also, please comment, as the case may be, why you are not addressing all the issues /atypicalities you were assigned.

4. Test Procedure This section describes the general approach and techniques used to acquire the data. Relevant information includes anticipated thermodynamic state, data collection, etc.

5. Connection with Plant and/or MIST This section is by far the most important. The objective here is to describe in sufficient technical detail how the information you generate will be used. How will the data be extrapolated or used to resolve the issues or address the atypicalities listed in Section 3. If a good case cannot be made, then the project will be considered irrelevant as far as this issue is concerned.

II. SCHEDULE A coordination meeting is scheduled for April 3-4 in Washington D.C. It is important that a draft document be available before or at the meeting (preferably before). We should discuss every item of Section 5 during the meeting and collectively comment on each other's documents. The meeting is intended as a working session, not a beauty contest! A revised document will be required by the end of April. All the documents will be assembled into a joint EPRI/NRC report, co-authored by all involved. This report will then form the basis and justification for further work in the Individual projects. A-2

February 25, 1985 Page three Your cooperation in this effort is important to the success of the program, and is greatly appreciated. Y[ Sursock/EPRI M. Ydung/RKC J. P. JPS/vb cc: O. E. Bassett (NRC) W. Beckner (NRC) P. Bochnert (NRC/ACRS) R. Carter (B&W) I. Catton (UCLA) R. B. Duffey (EPRI) l A. Hosler (WPPSS) l R. Jones (NRC/NRR) J. Kim (EPRI) W. B. Loewenstein (EPRI) - (E) L. Shotkin (NRC/RES) H. Sullivan (LANL) T. Theofanous (Purdue) N. Zuber (NRC/ RES ) i A-3

ATTACHMENT 1 TABLE'l TEST FACILITIES AND OTHER PROJECTS IN SUPPORT OF IST f

1. SRI-2 (Integral Tests) (EPRI)

2. University of Maryland (Integral Tests)

3. Argonne National Lab (Ishii - Flcw Regimes)

4. SAI Flow Regimes (EPRI)

5. SAI Auxiliary Feed (EPRI) ,

,                                   6.               MIST Undefined Tests (5 Tests) i

7. Tetratech Pump Model (EPRI)

8. TRAC Computer Code

9. Phenomenological Models 9

l 4 i l f i i , A-4 i

l l

   +*                                          UNITED STATES E'           'o,%                 NUCLEAR REGULATORY COMMISSION i             . ,I                        wAsHmc. Tow, p. c. zones
 \,...../

EAR 11 1985 Mr. Robert Turner Babcock and Wilcox Company Lynchburg Research Center Post Office Box 1260 Lynchburg, Virginia 24505 Dr. Robert Kiang Standford Research Institute 333 Ravenswood Avenue Menlo Park, Califcrnia 940?5 Dr. Gary Pertmer University of Maryland i Nuclear Engineering Department College Park, Maryland 20747 Dr. Thad Knight Los Alamos National Laboratory Post Office Box 1663 Los Alamos, New Vexico R7545 Gantlerer:

Subject:

Intecrel Facility Information Needs for Report on Scaling and Interfacility Test Comparison Durino nur IST Coordination Maeting on February 20, 1985, the need to establish an interfacility test compariser between MIST, SPI-2, and the University of Maryland at College Park (UMCP) was again identified. It was recognized by the rieetino participants that this comparison is needed to:

1. Investigate how the scaling philosophy and distortions in our integral facilities impact phenomena and various facility states; and

2. Pru ide a quantitative method for assessing the capability of our codes to address scalino effects. This code assessment on scale

                                             ~

would help to determine a confidence level for full-scale code predictions. Also, resolution of the IST program issues will require some integration of the three integral facilities to address MIST scaling and design atypicalities. In an effort to address these needs, it would be very beneficial to develop two documents that would describe a coordinated integration of efforts which includes specific objectives / methodologies. The first report would focus on A-5

the available experimental and analysis programs to examine MIST atypicalities. An outline for this report has been transmitted to each of the appropriate parties for their input. The second report would describe in greater detail each facilities scaling philosophy, techniques, and evaluation; and provide a common basis for 2 comparing test facility results. The primary purpose of this letter is to provide a list of specific information needs that pertain to the MIST facility, the SRI-? facility, and the UMCP facility that would feed directly into the i second report. INFL has been requested to collect this information and ) integrate it into a document which addresses specific scaling details and  ! implications with regard to testing, test comparisons, and data applications for each of the three integral facilities. The general outline of the report is perceived to be written as follows, l , 1. Description of background, issues, and purpose for each facility.

7. Statement of report purpose - provide a unified integral approach to resolving issues (scaling).

3. Detailed description of scaling philosophy and rationale supporting each facility design.

4 Evaluation of facility scaling limitations and distortions - design conpromises, atypicalities, and drawbacks as a function of phenomena and/or transients of interest.

5. Examine methodology or techniques for scaling up to or down from full-size plant (initial and boundary conditions).

l

6. Examine methodology or techniques to interrelate each of the three integral facilities. Recommend an interfacility test comparison I

based upon MIST defined composite test matrix. To reduce INEL's effort in collecting relevant facility documents, we would request that each facility provide by the next IST Coordination Meeting on April 3-4, 1985 the following to Mr. John Martinell, EG&G Idaho, Incorporated, Post Office Box 1625. Idah. Falls, Idaho 83416.

1. Facility design reports and relevant facility drawings.

2. PMG meeting minutes art technical meeting minutes.

i

3. Relevant reports on scaling philosophy and techniques.

In addition, the enclosure identifies specific parameters which need to be determined by each facility. We request that you devote serious attention on your entries onto this table to ensure that the follow-up scaling and i interfacility comparison evaluation by INEL is accurate and successful. i A-6

Your cooperation and interest in developing this report is essential to ensuring its value and timely delivery. We intend to issue the report as a joint NRC/EPRI document coauthored by INEL, B&W, SRI, the UMCP, EPRI, and NRC. If you should have any questions please do not hesitate to call. Sincerely, D 9-

                                       -Jean Pierre Sursock Electric Power Research Institute
                                        ///f           lVA Michael W. Young Reactor Systems Research Branch Division of Accident Evaluation U.S. Nuclear Regulatory Commission

Enclosure:

As stated cc w/encls:

0. E. Bassett, NPC L. M. Shotkin, NRC N. Zuber, NRC W. D. Beckner, NPC 4 R. Duffey, EPRI I W. Loewenstein, EPRI

1. Catton, ACRS T. Theofanous, Purdue Univ A. Hosler, Supply System R. Carter, B&W J. Gloudemans, B&W A-7

Appendix B INVENTORY / PRESSURE RELATIONSHIP DURING EARLY DEPRESSURIZATION l L

Relationship between coolant inventory ratio and pressure ratio during depressurization and draining of pressurizer.

1. Description of the Phenomenon:

- During the initial phase of depressurization, the pressurizer is drained with coolant inventory reduced from V                                   V
          =o              at t=tg to          =a     at t=t g,1                              t,2        2 where t denotes the time when the pressure is reduced to that corresponding to the f

saturation pressure of water at exit of core, and water starts to flash some hot spots in upper plenum, upper head, or in hot leg. Throughout this period, the coolant is subcooled in most of the system, except in the pressurizer which has been maintained at Tsat (Psystem). Thus, void only exists in the pressurizer, and the volume yielded by the break flow is replaced by the vapor volume c.reated by I flashing of saturated liquid inside the pressurizer. Thus, we can focus our attention to the pressurizer, treating it as a control volume: I i 2. Conservation of Energy The first law of thermodynamics states that l i du = ag - aW Since aq=0 (heater turned off), and AW is the work to push out the coolant d(Um) - Ut dmj = P/Pt dmi (2) I where m = Vp, mi = Vi ot mg = Vgog (3) 50 = U11 5 + Ug9 5 (4) Vi+Vg=VT (5) dmt = (dV1 )ot (6) B-1

From Eq. (2), (3), and (4) , d(ou) dV d(ou) dV 1 h=Vg + (8U)g +Y i (7) dt dt *IAU)t From. Eq. (6) and (7) d(ou) dV d(ou) dV V g dt +(oU)gg+V t dt + (pU)t dt I dV g odV g () U

        " *t t Y

• W l

Since d5/g = -dVg , Vg=V7 - Vj (9) d(ou) dV t d(ou)g odV g (v7 -V) 1 de - (ou) 97+vi dt dt " 0 1 1 All the above terms vary with dp during depressurization:

                            +V I V

T [(ou)1 - (ou)glh=[(ou)g+pl h (10) We can convert t-function into p-function by eliminating h. If we wish to retain t as variable, we only have to use a proper multiplier, which is determined by break size. hasthe l Eq. (10), as is, shows p as a function of Vs. Let call d(oU)q d(ou)g i dp *A dp =e [(cU)g - (ou)g] = e - a = b (ou)g = ap + C l l

Eq. (10) becomes dag a + bat = [(a+1)p+c] y dal dp I ) bag +a " (a+1)p+c fh[1n(bag +a)}=yf,h[in[(a+1)p+c)) (13) or e-a (a+1)pt+c {((e-a)agt+a)a12+"l (*+1)P+c! 2 ()

3. Numerical Values:

For the range p = 200 -300 psi, the coefficients are approximately: 3 a = 2.383 Btu / t 12.87 b = 67 (15) c = 54 psi Since (a+1)p nc. Eq. 14 can be approximated as l P bagt + a {*!a12,a)h 2 b (16) Thus, if the coolant inventory fractions are preserved, the pressure ratio is preserved. l If att = 0.5 a12 = 0.1 B-3

then Pt = 300 psi P2 = 250 psi For the 2000 psi range prototype, the coefficients a, b, and c vary drastically with pressure. Thus, the p-a relationship cannot be readily integrated in one step. However, the calculation can be performed in small increments. The result of the calculations for both the model and prototype are shown in Fig. 6. From Fig. 6 it is obvious the p-ratio scaling principle is applicable for both high pressure and [words missing?l, although an appropriate multiplier should be used to bring the P-ratio in line of each other.

4. Conclusion From the First Law consideration and from numerical calculation (Eq. 16), one can conclude that the pressure ratio (P/Pinitial) is a c.,od scaling approach to extrapolate the model results to prototype level.

B-4

NRCFORM336 U S. NUCLEL4 h EluL1 TORY COMMISSION l KEPORT NuMBE.8 fr ssfaed er TfDC esa For Me, et assy/ (2 84) Eff3S BIBLIOGRAPHIC DATA SHEET NUREG-1163 SEE INSTRUCTIONS ON 7 E REVERSE 2 TITLE AND SU8TsTLE 3 LE AVE S Coordination of Safety Research for the Babcock and Wilcox Integral System Test Program [ 4 DATE REPORT COMPLETED MONT.

                                  ,                                                                          f                              l
                                                                                                                                                        , EAR a AuT ORi5,                                                                                              [ January                                1987 8 DATE REPORT ISSUED-M. W. Young and J.                  . Sursock                                                                MarcT'"
                                                                                                                                                          ^~

T .E-,OR. No ORaAN,2 AT,0N M A.E A~o uNo A ooR us ,,~,,,,, ,, C .

                                                                                                      /      . PROxCT,T A5 mOR A u~,7 Nuona l       1957 Division of Reactor S em Safety Office of Nuclear Regul ory Research                                                                    *  *'N oa oa ANT hu-na U. S. Nuclear Regulator onnission Washington, D. C. 20555 to 5PON50RINu GRGANi2ATION NAvt ANO MAluNG A        E55 ftacivorI Co8e/                                    11a TYPE OP REPORT Same as 7 above.                                                                                           Topical Report D PERIOD COV ERED ssacess we eress February 1985 - April 1986 12 SUPPLEMENT AR Y, Ng T[5 1

13 Av57R ACT (200 werdg er assf This report. describes the MIST facili nd all the Integral System Test (IST) support projects sponsored by the USN and by EPRI. These support projects have been deemed to play an essential role helping resolve issues raised by MIST scaling compromises. Each support pr t is described in detail and application of the expected data to resolution o is es is discussed. l l The combined effort of MIST and se n oth support projects will resolve virtually all questions addressed te,e I program. 1 , 14 DOCUMENT ANALY5is - e RE **ORD5 DESCRiPTORS is ava:LAgiufy Steam Generator (Once Throug "'"* Two-Phase Flow Unlimited Transients-testing Reactor Safety cua'" c"'c^ noN

 .STI,(R$1bl980.IntegralSystem,TestFacility)                                                           (                                     U$~c~lIssified e r..s -,>

Unclassifled IF NYW9ER OS PAGES 10 PMcCE

n I UNITED STATES \ weca pouan ctus nare NUCLEAR REIULATOTJ COMMISSION " WASHINGTON, D.C. 20666 '**** us

                                                                              'sen'c * **
                                                                          ,,L^*",,,Dh, OFFICIAL BUSINESS PENALTY FOR PRIVATE USE. 4300 120555078877 US NRC          1 1AN1R2 ADM-UIV 0F PUB SVCS POLICY &

W-501 PUB MGT dR-POR NUREG WASHINGTON DC 20555 w}}