ML20135H699

Forwards Licensee Specific Notes & Procedure Changes to Encl Rev 1 to TR 017, TMI-1 Nuclear Generating Station Natural Circulation.... Procedure Changes Will Be Fully Implemented Prior to Startup from Next Refueling Outage

ML20135H699
Person / Time
Site:	Rancho Seco
Issue date:	09/19/1985
From:	Reinaldo Rodriguez SACRAMENTO MUNICIPAL UTILITY DISTRICT
To:	Thompson H Office of Nuclear Reactor Regulation
References
GL-81-21, RJR-85-343, TAC-47170, NUDOCS 8509240141
Download: ML20135H699 (4)

Text

$SMU-SACRAMENTO MUNICIPAL UTILITY DCTRICT O 6201 S Street, P.O. Box 15830, Sacramento CA 95852-1830,1916) 452-3211 RJR 85-343 AN ELECTRIC SYSTEM SERVING THE HEART OF CALIFORNIA September 19, 1985

' DIRECTOR OF NUCLEAR REACTOR HEGULATION ATTN HUGH L THOMPSON JR DIVISION OF LICENSING U S NUCLEAR REGULATORY COMMISSION WASHINGTON DC 20555 DOCKET 50-312 RANCHO SECO NUCLEAR GENERATING STATION UNIT NO. 1 GENERIC LETTER 81-21, NATURAL CIRCULATION C00LDOWN By letter dated December 14,1984 (RJR 84-563, Mr. R. J. Rodriguez to Mr. John F. Stolz), the District indicated that it was planning to adopt the GPU Nuclear approach for natural circulation cooldown. The topical report on the GPU Nuclear approach has now been received from the Owners Group. A copy of the topical report and the District's notes, which indicate the applica-bility of each section of the report, are attached.

The methodology identified in the report, along with the Rancho Seco specific notes and procedure changes,-form the basis of the District's approach to natural circulation cooldown without reactor vessel head void formation. The procedure changes identified in the attached notes will be fully implemented prior to startup from the next refueling outage.

If we can provide any further information, please contact Robert Little at 916) 732-6021.

\' 4 R. J. RODRIG'JEZ, ASSISTANT GENERA MANAGER, NUCLEAR Attachments d2 g 10

e---  ;

l i

RANCHO SECO NOTES  !

l ON TOPICAL REPORT 017  ;

I

SUMMARY

The analysis methods used and the results contained in Topical Report 017 Revision 1, "TMI-l Nuclear Generating Station Natural Circulation Cooldown-Analysis Without Reactor Vessel Upper Head Void formation," are appropriate for the Racnho Seco Nuclear Generating Station. The TMI-1 and Rancho Seco raeactor vessels are virtually identical. The analyses contained in Topical Report 017 were run to reactor coolant system (RCS) conditions of 204*F to envelope dif ferences in decay heat removal (DHR) system cut in temperatures and pressures.

The TMI-l cooldown times are based on DHR cut-in at 325 psig with a corresponding saturation temperature of 429'F. The. Rancho.Seco DHR cut-in occurs at 250 psig with a corresponding saturation temperature of 406*F. The results of Topical Report 017 indicate that for an RCS cooldown rate of 10*F/hr, it would require 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> to reduce the coolant temperature in the reactor vessel head to 406*F. For an RCS cooldown rate of 50*F/hr, the time required to cool the reactor vessel head to 406*F is on the order of 7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br />.

1.0 INTRODUCTION

This section of Topical Report 017 is valid for Rancho Seco.

1.2 NRC Concerns and Reauirements This section of Topical Report 017 is valid for Rancho Seco.

1.3 Rancho Seco RCS Natural Circulation Cooling Procedure B.4. Section 6.0 Operating procedure B.4, Section 6.0, addresses RCS natural circulation cooling for Rancho Seco. The procedure currently identifies a cooldown rate limited to 100*F per hour due to pressurized thermal shock considerations. The procedure will be revised to reflect a cooldown rate greater than 10*F per hour but less than or equal to 50*F per hour which will prevent upper head voiding.

Section 6.1.7.1.2.of procedure B.4 includes the following notes with respect to the possibility and mitigation of upper head voiding:

NOTE: During the natural circulation cooldown, it is advisable to cooldown/depressurize as slowly as plant conditions allow to minimize the possibility of void formation in the reactor vessel head region or hot legs. The depressurization should be in steps not to exceed 70-100 psi.

i l

CAUTION: During the natural circulation cooldown, monitor pressurizer level and makeup tank level. A sudden j increase in either level while pressure is constant or l

decreasing could indicate voiding in the RCS. In '

addition, unexpected pressurizer level changes during makeup or spray operations could also indicate voiding in the RCS. If voiding is suspected, stop any i

depressurization and repressurize to maximum pressure allowed by applicable cooldown curve and raises OTSG levels to 95 percent on the Operate Range. Do not continue cooldown on natural circulation unless required v by plant conditions. If continued.cooldown is required, cooldown and depressurize as slowly as possible for current plant conditions. Monitor natural circulation closely during the cooldown.

Section 6.1.7.2 of procedure 8.4 identifies the requirement to 1 neintain a 50*F subcooled margin during cooldown. The subcooled margin will ensure that steam voids ~will not form in the hot legs 4

and that interruption of natural circulation is avoided.

1.4 TMI-l Natural Circulation Cooldown Analysis Without Reactor Vessel Upper Head Void Formation This section of Topical Report 017.is valid for Rancho Seco with the exception that the Rancho Seco DHR system cut-in conditions are 250 psig and RCS temperature of 280*F. To cut-in the DHR system, the temperature of the fluid in the' upper vessel head must be less than the saturation temperature (406*F) which corresponds to 250 psig.

2.0 REACTOR VESSEL UPPER-HEAD C00LDOWN PROCESSES This section of Topical Report 017 is valid for Rancho Seco.

3.0 HEATING 6 MODEL OF TMI-l UPPER HEAD o

This section of Topical Report 017 is valid for Rancho Seco (including Sections 3.1 through 3.6).

4.0 ANALYSIS RESULTS 4

This section of Topical Report'017 is valid for Rancho Seco. '

4.1 Volume Averagina of the Too Foot of the Vessel Head-This section of-Topical Report 017 is valid for Rancho Seco.

l 4- d

I l

l

4.2 RCS Cooldown at 10*F/hr This section of Topical Report 017 is valid'for Rancho Seco with the

. exception that it would take approximately 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> to cool the head to the Rancho Seco DHR cut-in point without coolant flashing.

4.3 RCS Cooldown at 50*F/hr This section of. Topical Report 017 is valid for Rancho Seco.

t

c .

TR 017 Rev. 1 Page /

l TMI-l NUCLEAR GENERATING STATION NATURAL CIRCULATION C00LDOWN ANALYSIS WITHOUT REACTOR VESSEL UPPER HEAD VOID FORMATION TOPICAL RE'PORT 017 REVISION 1 PROJECT NO: 5400 41274-D. C. Bucheit (Energy,Inc.)

A. A. Irani (GPU Nuclear)

P. W. Lynches (GPU Nuclear)

July 2, 1985 APPROVALS:

4:7 ~' 7 >"

SECTION MANAGER ' 0 ATE h I66' Y-8M DEPARTMENT MANAGER DATE

. cl CA ? ,' ! h A ^ n

- vy 1 L _..

.c TR 017 Rev. 1 Page la DISCLAIMER OF RESPONSIBILITY This document was prepared by or for the GPUN Corporation. Neither GPUN Corporation nor any of the contributors to this document:

4

a. Makes any warranty or representation, express' or implied, with -

i respect to the accuracy, completeness, or usefulness of the information contained in this document, or that the use of any information disclosed in this document may not infringe privately owned rights; or,

b. Assumes any responsibility for liability or-damage of any kind which may result from the use of any information disclosed in this document.

4 T

4

}

i 4

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

TR 017 Rev. 1 Page ;L

, TABLE OF CONTENTS PAGE

1.0 INTRODUCTION

4 1.1 Background 4 1.2 NRC Concerns and Requirements 4 1.3 TM.I-l Natural Circulation Cooling Procedure OP 1102-16 5 1.4 THI-1 Natural Circulation Cooldewn Analysis Without Reactor Vessel Upper _ Head Vold Formation 7 2.0 REACTOR VESSEL UPPER HEAD COOLDOWN PROCESSES 8 3.0 HEATING 6 MODEL OF TMI-I UPPER HEAD 14 1

3.1 HEATING 6 Code 14 3.1.1 HEATING 6 Limitations 14 1

3.2 Upper Head Model 15 l 3.3 Assumptions 15 3.4 Steady State Heat Transfer 22 3.5 Initial Conditions 23 3.6 Boundary Conditions 24

. 4.0 ANALYSIS RESULTS 29

. 4.1 Volume Averaging of the Top Foot of.the Vessel Head 29 4.2 RCS Cooldown at 10F'/HR' 30 4.3 RCS Cooldown at 50F'/HR 33 5.0 OPERATIONAL GUIDELINES 36

6.0 CONCLUSION

S 41

7.0 REFERENCES

42 4 APPENDIX A 43 e -

yy'a - - - - -

TP. 017 Rev. 1 Page 3

SUMMARY

This report contains an analysis for TMI-I which simulates a natural circulation cooldown without reactor vessel upper head void formation. A ,

two-dimensional cylindrical model of the TMI-l RV upper head was developed and used as the basis for the thermal analysis. Computer aided analyses were performed by means of HEATING 6 (a multi-dimensional, generalized heat conduction code) to demonstrate the t'ermal h response of the RV head to RCS cooldown rates of 10F'/hr and 50F'/hr. The results-indicate that for an RCS cooldown of 10F*/hr to 204*F, it would require 22 hours2.546296e-4 days <br />0.00611 hours <br />3.637566e-5 weeks <br />8.371e-6 months <br /> to reduce the coolant temperature in the reactor vessel head to 429'F*. On the other hand, for an RCS cooldown of 50F*/hr to 204*F, the time required to cool the reactor vessel head to 429'F* is on the order of 7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br />.

The results of these analyses may be' correlated into pressure versus temperature curves which can be incorporated into existing cooldown procedures.

429'F is the saturation temperature which corresponds to the decay heat removal system cut-in-point - 325 psig RCS pressure.

• \

TR 017 Rev. 1 Page if

1.1 BACKGROUND

On June 11, 1980, the St. Lucie reactor was shutdown due to a loss of component cooling water to the reactor coolant pump seals which also required shutdown of the reactor coolant pumps. The cooldown was ,

accomplished by natural circulation. At approximately four hours into the event, charging flow, which was initially being divided between the cold legs and the aux 111ary pressurizer spray, was diverted entirely to the auxillary spray to enhance the depressurization and reduce the system pressure on the pump seals. At that time, abnormally rapid increases in pressurizer level were observed. Detailed evaluation and follow-up analyses have indicated that the increases in pressurizer level indicaticq were the results of steam void formation in the upper head region o' the reactor vessel. The steam void was produced at the instant the system pressure dropped below the saturation pressure corresponding sto the upper head coolant temperature. Under conditions of natural circulation, coolant in the upper head is expected to be in poor thermal communication with coolant in the plenum. Consequently, the coolant temperatures in the RV head will tend to remain elevated above temperatures indicated by hot and cold leg instrumentation.

1.2 'NRC Concerns and Requirements Because of the unexpected. occurrence of the vold during the St. Lucie event, and the failure of the operators to immediately recognize the void w

%- - . , e

1 1

~

1 TR 017 Rev. I Page 5 formation and take corrective action, the question of.whether such void formation is properly accounted for in safety analyses has been an area of concern to the NRC. These concerns relate to a) procedures and training to enable operators to avoid void formation (if possible), or recognize and properly react to reactor vessel head voiding during i natural circulation cooldown and b) the possibility that significant head voiding increases the susceptibility of the plant to more serious accidents. In particular, these issues are contained in Reference 1 l

, which requests the following information:

1. 'A detailed description of the natural circulation cooldown procedure and its basis (it should include guidance on possibility, prevention, and mitigation of upper head voiding and natural circulation interruption).

2. Demonstration by analysis or otherwise, that; a) Use of this procedure will not result in upper head voiding b) If voiding occurs, the procedure will prevent any voiding at the hot leg elevation 1.3 THI-1 Natural Circulation Cooling Procedure OP 1102-16 Operating procedure 1102-16 addresses RCS , natural circulation cooling.for TMI-1 in which an RCS cooldown rate greater than 10F'/hr but less than or equal to 50F'/hr is prescribed as a means of preventing upper head voiding. Section A.2.6 and C.4.3 include guidance on possibility and mitigation of upper head voiding as quoted below:

_ - _ 4 _

TR 017 Rev. 1 Page 6,

, Volds may occur in the Reactor Vessel head while depressurizing the RCS due to head water temperature being higher than RCS temperature. This condition may i

be evidenced by an increase'in pressurizer level while reducing RCS pressure even though an ad, equate saturation margin is indicated between T (hot leg temperature)

- and RCS pressure. Reference IE Circular 80-15.

f i

If void formation should occur in the RV head, the head bubble should be condensed before continul.ng the cooldown.

I If an RCP cannot be bumped, then RCS pressure should be i

held constant or slightly higher until the head bubble

! has condensed'as indicated by the return of pressure control to the pressurizer.

n

In addition,'the procedure will prevent any voiding at the hot leg elevation by Steps A.2.5 and 8.2.5

4 Steam volds at-the top of the hot legs can interrupt natural circulation. This is prevented by establishing and maintaining at least a 25'F subcooled margin after '

Reactor Coolant Pump trip.

}

f 5

1

TR 017 Rev. 1 Page 9 1.4 TMI-l Natural Circulation Cooldown Analysis without Reactor Vessel Upper Head Void Formation In order to formulate procedural requirements to prevent coolant flashing in the reactor vessel head, an analysis was performed to simulate a

natural circulation cooldown at THI-1. A two-dimensional cylindrical model of the RV upper head was used as the basis for the thermal analysis which uttilzed HEATING 6 (a multi-dimensional, generalized heat conduction code).

i Analyses were performed for RCS cooldown rates of 10F'/hr and 50F'/hr to 4

, determine the minimum time required to reach the decay heat removal system cut-in point (325 psig and RCS temperature of 300*F) based on the system design. In order to employ the decay heat removal system, the temperature of fluid in the vessel head must be less than the saturation temperature (429'F) which corresponds to 325 psig. In addition, the RCS 4

temperature was allowed to decline to 204*F in both cases. The results of these analyses are presented in Section 4.0.

/

i

8 a i

TR 017 Rev. 1

] .

Pagejf 2.0 REACTOR VESSEL UPPER READ C00LDOWN PROCESSES During natural circulation, the fluid in the RV upper head will remain i

relatively stagnant since the reactor coolant system loop flow rates will

be signift.cantly smaller than during forced circulation. The plenum

cover and structural compon,ents (shown on Figure 2-1) tend to isolate the RV head fluid from coolant in the plenum. Coolant will enter the head region at low velocity through the CR0 guide tubes which extend approximately 20 inches above the plenum cover. -Consequently, little

) mixing is expected between entering coolant and the majority of fluid in the RV head dome. The head metal and water will cool slowly by means of l- the heat transfer processes delineated below and illustrated on j Figure 2-2. Patterns of fluid circulation (which will likely result from f the effects of natural convection cooling) are depicted on Figure 2-1.

The head cool ng processes are:

Heat transfer from the exterior surface of tne mirror insulation to containment is considered to occur by means of free convection and radiation. -

1 Heat transfer across the three (3) inch thick mirror insulation is 4

considered to occur by means of conduction.

t Heat transfer from the exterior surfaces of the vessel head through the air space to the inside surfaces of the mirror insulation is d

expected to occur by means of natural convection and radiation.

RY C0MPONENTS TR Oli Rev 1 Page T i

(

W C."r I; "2 C""3 r -

L " ) CADAbusiw $ ,

.N 3 w=

3 wmL A l

y f tsmet j) l =

\ '1 i h ( * $

a

/ ( <

s -

j

, l l i )<

nam v a fl(~ f~~

m&

emea suveme=e 3 i (,

o, 4 '4 l s-q -

g JnWa:u?

{ AmqMW l

-l 2

%, - - y ,

,--gi i t

s

_  : ';;A

%W na= d NW

RGun 2-1

[.

RV HERD NERT T09MFER ""; 1,o i

I kjk j m4T4 ni d

j .

i g

g f ~

$4fd 7, 3 11FI

/- F N:s, s

</ l '

\

/

f 4 .

..... e._

i

' /

/ rw of ane

+

f>

~

/ _DRAWIh6 LEGEND

/ @,c,@,@#,camia unua

/ @.nsauria y &

- @. awaxrex -

Y= @. weme am, 7%MArtama neuae e-z \

TR 017 Rev. 1 Page il Heat will be conducted along the metal walls toward the CRD housings located in the vessel head.

Heat from the upper head wall will be conducted along the vessel steel toward the coolant inlet which is located in the annulus i

between the vessel walls and the core thermal shleid. Heat will also be transferred from the vessel walls to the cold and hot leg nozzles.

Heat will be transferred from the air space to the CR0 housings by means of natural convection.

During the cooldown of RV head metal, heat will be transferred from

,RV head coolant to the vessel walls by means of natural convection.

In addi_ tion, heat will be removed from the upper head water-by convective heat transfer between the coolant and the CR0 lead screws.

Heat transfer between layers of upper head water at different temperatures will include convective effects. The conditions under which convective heat transfer'is expected are elaborated upon in Section 3.3 i In the short period of time following the 4 RCP trip, heat will be transferred from the plenum cover and its structural members to the plenum coolant by means of forced convection. Throughout the l remainder of the cooldown period, coolant which enters the upper I

l

~

TR 017 Rev. 1 Page /J, head via CR0 guide tubes is considered to mix with and displace the coolant in the vicinity of the CR0 guide tubes. Mixing effects were restricted to the coolant volume defined by the 20.5 inch axial dimension (the guide tube height) above the plenum cover as shown on Figure 2-3.

Energy transport will occur from the upper head by means of effluent coolant which will pass through the outlet annulus to the hot leg.

Coolant which exits the upper head will be replaced by coolant flow

from the CR0 guide tubes.

1 During the cooldown, the specific volume of coolant in the upper head will decline with temperature. Consequently, the RCS will i provide over 5000 lbms of makeup coolant to the head until the decay heat system is employed.

i i

e

1. 4 TR 017

! Rev. 1 b .rson m c ' '

i 1W30/27kW

- g/ _

i g

mN  %

M 1 MM Q g"h!?[&

y

,F 4,; a q -c p {< & hy l

4 -( ,

i

/,p-e

/,

/ -

m .4 N

" /

/

\4 l

l l

l t

RGum 23

TR 017 Rev. 1 Page/Y 3.0 HEATING 6 MODEL OF TMI-1 UPPER HEAD 3.1 HEATING 6 Code HEATING 6 is a multi-dimensional, generalized heat conduction code which may be applied to problems which require steady state and/or transient solutions (*'. The thermophysical properties of various materials may.be considered to be anisotropic, temperature or time dependent. Boundary conditions may be applied by means of the following:

Coefficients for convective heat transfer natural convection forced convection Coefficients for radiative heat transfer Temperature Heat flux User supplied subroutines allow flexibility in the manner in which the methods of solution may be developed. Both explicit and implicit finite difference '..aerical techniques are available to the user.

3.1.1 HEATING 6 Limitations Since HEATING 6 is a nodal heat conduction computer code, it is limited by the following:

The mesh applied over the geometry of interest is fixed in size. The mass which occupies each unit mesh is assumed to be concentrated at the gecaetric center.

If each unit mesh is considered an elemental control volume, mass transport across control surfaces is non-existent.

~

. ~

m

~

i

-s TR 017 Rev. I Page AS~

In HEATING 6, the variation of material density as a function of temperature may inhibit or prolong the convergence of the iterative solution procedure. Consequently, the density of~  !

coolant in the upper head was assumed to be a constant (50.0 lbm/ft') throughout this analysis. This value tends to increase the thermal capacitance of nodes at high temperatures.

In order to include convective effects between layers of t

coolant at different temperatures, an effective thermal conductivity for the coolant may be defined by means of the following:

K ,, - h.c'x L where h c is the heat transfer coefficient for natural convection and L is a characteristic length of the nodal geometry.

3.2 Upper Head Model The geometry of the reactor vessel upper head was approximated in a two dimensional cylindrical coordinate system. The 90* vessel symmetry about the vertical centerline was taken into consideration. During the development of the thermal analysis, care was taken to account for the heat transfer paths and processes discussed in Section 2.0 Iriaddition, the following corrections were applied to various convective heat transfer coefficients used in this analysis in order to account for any differences in component surface area between the actual and model RV' head geometry:

1. The interior surface area of the RV head; I

i i

l .

• I

~

TR 017 Rev. 1 Page /6,

2. The exterior surface area of the RV head;

3. CRD lead screws and housings.

The primary components of the thermal model are: the plenum cover, the mass of upper head coolant, the vessel walls and mirror insulation.

Model component dimensions were estimated from construction drawings or established by calculations. Geometric regions and nodes were further used to' describe the head in detail. Figures 3-1 and 3-2 illustrate the regions and materials of the HEATING 6 RV upper head model. They are discussed in detall below.

The carbon steel vessel walls were modeled from the mating surface down to a point just above the hot leg nozzle. Water from the cold leg which enters the vessel by means of the annulus downcomer was also included.

The plenum cover and CRD guide tubes were also modeled. A variable boundary temperature (based u'pon an energy balance between coolant entering and leaving the upper head) was applied to the region defined from the top of the plenum cover to the top of the CRD guide tubes.

Application of the boundary condition restricted the effects of coolant mixing to the first 20.5 inches above the plenum cover. Heat transfer to this region from the coolant directly above (adjacent nodes) is considered to occur by means of natural convection. Above the region of.

adjacent nodes, convective heat transfer effects were also included.

The CRD lead screws and housing were also modeled. The 69 CRD lead screws and housings'were grouped into four separate concentric cylinders. The lead screw and housing cylinders were distributed radially and aligned on center with the CRD guide tubes.

• .y A p }

TR 017

%b ih EtB

$8h. =

O.

I u

-s a .

, , ~ -i h

E}  %, Tf54s .*

y 5=2 a N

' a m w s=

I. ( '% 21 k'*\ kh  % w

_; 'I b I [*%%R I ki  %

l g' a, WatW W*

) ,g j '@ s, L4 a W

• b"k W8 d

~5 a e  % M "-

o.kgokeakej.okfo41 okfoknaknoa

~

1 -

ESTIfffft F ml-I UPPUt VCSSEL M MIN 3tf Map or rye Mergniets g

I" "

J' /s s- ss rust rys, fug, ryg, LEGExp SEL-

• L -

sc OS- CAg6tw Srett. .

$~ ss G G G* g SS-Sw mss Mst.

NM - W4 TEA S.L G7 9 cs LsE cs cs G - Gyp (Asgarsgseo sg YG5 M5 /koi NW 6' cs G - 5 f,~_ fgfg#

9 9He 40 no no fu zusz-zusexwoa N I~

g k 4 j CS o

sx Q _ss cs 5- G G G sL tr 0 ,'

yg .

R 5-ssi d ss! W cs

• G I

~

cs G o l

a W W 5 5 3 g y 3 I

0.0 15.0 30.0 15.0 80.0 F5.0 90.0. 105.0 120.0 R

F/GURE 3-2 d f",3 o gG J _ _ _ _ _ _______ _ ___

1 km 1

TR 017 Rev. 1 Page /1 The reactor vessel head geometry was approximated by means of two cylindrical shells. The top of the head was modeled as a disk. Radial dimensions of various head components were fixed from construction drawings. The axial dimension of t,he head was established from a calculation which considered total coolant volume to be on the order of 522' cubic feet.

The vessel insulation was modeled as a cylindrical shell with an outer radius of 103 inches. The top of the insulation was modeled as a disk.

Since the distance between the head and insulation actually varies with the head radius, a mathematical average distance was used to fix the

~

position of the insulation.

3.3 Assumptions The following modeling assumptions were used in the analysis:

1. The spherical head was approximated as a disk on top of two cylindrical shells.

$. The' resistance to heat transfer offered by the 0.125" stainless steel cladding on the interior of the vessel was assumed to be negligible.

3. The effects of mixing between stagnant fluid in the upper head and coolant which exits the CR0 guide tubes is assumed to occur in the first 20.50 inches above the plenum cover. An energy balance over that region was used to establish a time dependent temperature which

~

=

Wi -

l TR 017 Rev. 1 Page .20 was applied as boundary condition at the top of the CR0 guide tubes. The effects of coolant flow from the CR0 guide tubes were considered limited to that region.

4. Heat transfer from the plenum cover to coolant in the plenum was based on a forced convection heat transfer coefficient for fluid flow over a flat plate.

5. Heat transfer from upper head coolant to the head metal was assumed to occur by means of natural convection.

6. Heat transfer from the CRD housings to the service structure region

. and containment were not included. A boundary temperature of 120*F was applied to the CRD housing 3 1/2 feet above the RV head. The temperature was based on a simplifted heat transfer analysis performed on a CRD lead screw,

7. Heat transfer to/or from the CR0 housings to the air located between the vessel head and insulation was neglected.

j 8. A short subprogram was used to include the effects of convective j heat transfer in the axial and radial directions between nodes at different temperatures. Heat transfer between nodes in the radial i

direction is discussed in Item 9. For the axial direction, the relative position between a specified node and its neighbor were l- determined first. Convective effects were included if the following criteria were satisfied:

l l

L .

m

ms

. TR 017

. Rev. 1 '

Page .2/

a) Neighboring node located above the specified node - temperature of the neighbor less than that of the specified node.

b)- Specified node located above the neighbor - temperature of the specified node less than that of its neighbor.

Otherwise, the convective effects between nodes were considered to be limited to laminar natural convection.

9. A somewhat less :enservative but realistic assumption is that some convective heat transfer is expected between adjacent nodes in the radial direction. Since cooler (denser) fluid will tend to fall in warmer surroundings, it is not unreasonable to expect fluid velocity components in the radial direction. Consequently, it appears-realistic to carry over the influence of convective effects to the radial direction.

10. The thermophysical properties of the stainless steel mirror insulation were estimated from its thermal conductivity and knowledge of its construction.

11. Prior to the onset of the transient, the temperatures of upper head coolant and metal were initialized to 604*F. This temperature includes measurement uncertainty, and is the approximate temperature of coolant in the hot leg under conditions of 100% power.

12. Subsequent to the 4 RCP trip and flow coastdown, natural circulation flow was assumed to become. stable and invariant at 3% of rated RCS flow at 100% power.

a usd 3-

. TR 017 Rev. 1 Page JJ.

13. . Coolant flow into the upper head from the CR0 guide tubes was assumed to be 8% of the RCS flow at any given time.(

14. A four pump coastdown was assumed to occur following the reactor

. trip. The effects of the coastdcwn were included only in the energy balance described by Item 3 Section 3.3. Operation of the RCPs after a reactor trip will tend.to substantially reduce the temperatures.of upper head coolant and metal. Since the temperatures of coolant and metal in the RV head tend to follow the hot leg temperature.during forced flow conditions, the amount of temperature reduction in the upper head will be dependent upon the duration of RCP operation assumed after the reactor trip and the post-trip cooldown rate'of the RCS.

15. The ambient (reactor building) temperature was assumed constant at 120'F.

16. The insulation was considered to be completely sealed against air leakage.

3.4 Steady State Heat Transfer The steady state heat transfer rate from the RV head was estimated from

.the manufacturer's specification for the heat flux associated with the insulation. Convective coefficients for the flims which act at the exterior surfaces of the insulation and between the vessel head and the interior surfaces of the insulation werc determined by means of an Iterative procedure which included the folicwing constraints:

1 4I$

D

~

• ~

w9 l

. TR 017 Rev. 1 Page.2$

1) Radiative heat transfer was assumed to take place between the exterior surfaces of mirror insulation and the containment surroundings. The insulation was assumed to be a gray body which

~

. transferred heat to black surroundings maintained at 120*F. An emissivity for #18-8 polished stainless steel (evaluated at 200*F) was assumed to be adequate.

2) The flim resistances between the RV head metal and coolant were considered to be zero at steady state.

3) The RV head metal was considered to be at a temperature of 604*F.

4) The temperature change across the insulation (approximately 380F*)

was determined from:

a) Heat flux specified by the insulation manufacturer. The heat flux was assumed appilcable for steady state conditions.

b) Thermal conductivity of mirror insulation - assumed to be independent of temperature.

5) Containment air temperature was assumed to be invariant at 120*F.

The heat transfer coefficient applicable between the vessel and insulation was initially assumed and iterated upon untti all constraints were satisfied.

3.5 Initial Conditions Several of the initial conditions imposed on the thermal model at the onset of the reactor trip have already been outilned in Sections 3.3 and 3.4. Consequently, they will be summarized by the following:

"]

-no -

. TR 017 Rev. 1 Page JM/

The plant was assumed to be operating at steady state conditions at 100% power.

Insulation temperature was taken to be 408*F.

Cold leg temperature was considered to be 555*F.

These are shown on Figure 3.3.

At time (t - 0+), the transient was initiated with the following events imposed:

Reacter trip with insertion of control rods achieved.

Following the reactor trip, 4 RCPs are tripped.

Subsequent to the trip of the 4 RCPs, RCS flow was taken to decay with pump coastdown.

3.6 Boundary Conditions Various boundary conditions were applied to components of the upper vessel head model. The types that were used are delineated below and illustrated on Figure 3-4.

3.6.1 Specified Temperature As' explained in Section 3.3, a constant temperature (120*F) boundary condition was maintained at the top of the CRD lead screws and CRD housings (BCS). A time dependent temperature (which is a function of the RCS cooldown rate) was appIted between coolant in plenum and hot leg metal (BC2).

A time dependent boundary temperature was defined at the top of the CR0 guide tubes. The method used to establish boundary condition (BC6) i s delineated in-Section 3.3. Item 3.

4

4 I

i

! csunut a w-i umn vessia. Henn eenem

, MRP of DK EWmA OWDir7 ens I i- N I l

l o N-

~

I

/

I - l4 II 9

teswo, o 1 ~ M *F l- 2 ~ 408 *p g i i f f 9 i iE 1 1 g- - -

1 y NOTE:

1 -

Mera, wam; ao.nsrxsym 9 i g tl i IEMMR47ukts ser 10 &t*4 l

N E' i bo4'FMD 406*F RESMCnyEty.

, %SWE 7EMPEngmg5

g. I 'gg g '

i i

i N3TR/80lbY/S use n THEImrm cbgomm

! o 1

- ) M TRA g n ygy 7,gg ,o,

1. E' NE M7 Fae Mysygg, I f.

i -

i 1

o. 2

$~ f 1 4 >

1 o.

i 8

i

~

9.0 15.0 30.0 15.0 08.0 75.0 90.0 105.0 120.0

! R a

2' <5' d i

l FtGuAnt 3-3 *L$

.! h. .

S

1 l

l

)

ESTIfefft (F int-t UPPUt VESML M WWEWtf 1

q Nap of TNs Bwaa9ey Cweirious N' f&.T y o

l- f. Bc9 oct Bcy Ac9 n

i .. . .. . ... .. ...... .......-

o l

i

[ Bett ,

. 1 Bett

. . .\ . . . . . . . .n .t Bet I.2 GENO a

......... act- raceownc7mv

1  : sci-nus mesnur l

l- y  ; gcjg 3 M

i BC/0 WMAE4*MiAC \

Rs- wnm4*- ONHrc7/M

-  : pce ses c  :

~

a u,oce,a io e ,u a n ~S- W cowygcyy( y Q I su Nu ren L..... .............;  ;

Bcs-figgp ygy,gggyygg i $~

Scr 6

f Sc Bc7 I i R$,gc7,ges ygyug,gm l & 11~ NaryRoi a wnScyg (

. ...  ; topsyyy y g g h.

Sh Agc, \B c/ 122 Bct3-paruggiawna.7,gg

\

• l ' NOTE BC6 bicllAGEs TIME DEMNDENT*

fcz &3- l 1 80tJNPhEJJ TEftllERMTUD=. SEE i

0' 1  ; t >

SECT /M 3 3.

Ber h  ;

E , , , , ,

'- b I a e.e is.e is.e e.o es.e n.e so.e ses.e nas.e R

i i

Fisuse 34 ;c y y

'id o (M

e -

-s-TR 017 Rev. 1 Page 27 3.6.2 Convective Heat Transfer A heat transfer coefficient for forced convection was applied as a boundary condition.to the lower surface of the plenum cover (BCl).

Similarly, a natural convection heat transfer coefficient was applied in the annulus downcomer between the vessel wall and the ,

core thermal shield.

In the upper head, natural convection heat transfer coefficients were applied at all metal-water surface interfaces.

Natural convection coefficients were also applied to the following model components:

To the outside surfaces of the_ insulation To the air space between the vessel head and the Inside surfaces of the insulation As mentioned earlier, the convective coefficients were calculated by means of an iterative procedure performed on the conduction and convection heat transfer relationships. A known steady state heat flux across the Insulation was used in the procedure to predict temperatures on the interior and exterior surfaces of the insulation. Subsequent 1 / the temperatures were used to predict suitable convective coefficients. It should be noted that the convective coefficient used between the vessel and the inside

~;

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

N9b l TR 017 Rev. 1 Page JF surface of the insulation is considered to include radiative heat transfer effects. Since it was difficult to determine an appropriate radiative heat transfer coefficient for the-complex geometry of the air space, the lumped convective coefficient was adopted and the air space was mod? led as an area devold of material.

k 3

~--

sM TR 017 Rev. 1 Page J2T 4.0 ANALYSIS RESULTS u Two analyses were carried out to demonstrate the thermal response of the RV head as a function of the cooldown rate in the RCS. The cooldown rates imposed on the RCS were 10F'/hr and 50F'/hr. The 10F'/ hr rate was

-imposed for a 40 hour4.62963e-4 days <br />0.0111 hours <br />6.613757e-5 weeks <br />1.522e-5 months <br /> period. The 50F'/hr rate was applied for an 8 hour9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> period. .Thereafter, the RCS temperature was held at 204*F for two hours. The results are shown on Figures 4-1 through 4-4 and are discussed in Sections 4.2 and 4.3.

The volume averaged coolant temperature in the top foot of the vessel head was used to represent the overall coolant temperature as explained in the following.

4.1 Volume Averaging of the Top Foot of the Vessel Head Heat transfer by means of natural ' convection is a complex process which involves m,iss and energy transport at relatively low fluid velocities.

Fluid circulation in natural convection is attributable to buoyant forces which artse from temperature variations in the fluid. Consequently, free convective flow is compressible flow. In this process, the convective heat transfer coefficient is characterized by the Rayleigh number which is a product of the Prandt1 and Grashof numbers, the Grashof number being proportional to the ratto of buoyant to viscous fcrces.

Ourtrg.a natural circulation cooldown, convective cooling of upper head metal and coolant will result in buoyancy driven fluid circulation. With

1 emo i

TR 017 Rev. 1 Page 3d continued cooling of fluid layers at different temperatures, coolant circulation is expected to develop as depicted on Figure 2.1.

During the cooldown, it is expected that buoyancy driven fluid circulation will prcpagate from the head walls and lead screws toward the centerline of the head. As mentioned previously, with the decline of coolant temperature in the RV head, it is expected that over 5000 lbms of coolant will be provided to the head by means of the RCS. Consequently, it appears to be unrealistic to treat the cooldown of the head strictly as a conduction problem.

Since convective effects are included in both coordinate directions, the use of a volume averaged coolant temperature in the top foot of the upper head appears to be a conservative representation of the coolant conditions in the RV head.

4.2 RCS Cooldown at 10'F/HR Figure 4-1 demonstrates the rate of change of coolant temperature in the RV head in response to the 10F'/hr RCS cooldown rate. The results indicate that the coolant temperature decreases at the rate of 8.86 F'/hr over the 40 hour4.62963e-4 days <br />0.0111 hours <br />6.613757e-5 weeks <br />1.522e-5 months <br /> period. Figure 4-2 demonstrates the rate of change in the saturation pressure of head coolant throughout the cooldown. At this cooldown rate, it appears that it would take approximately 22 hours2.546296e-4 days <br />0.00611 hours <br />3.637566e-5 weeks <br />8.371e-6 months <br /> to cool the head to the OHR cut-In point without coolant flashing. The

' average change in coolant temperature over that time is on the order of 8.23F'/hr.

"3

6 5 4 e~+,es o. e=*y

1. we TR 017 Rev. 1 Page 3/

TEMMA972n? W MMA - 4/H/M

790RERATUM(93 b -70pmr)

ACS QtE M WNs

) ABlF7E-Abf-/0Y4A' m-gJ i

P 1

\

gn Joo-

@0-

$$ Y&h N i e e i i i i i i , i i 3 i e i i e i e i i 0 l$ $O $ $Y n su M 4-1 4

, g

~

TR 6 wn SMTUOff)\W W W) Rev. 1 Page /1 ggymyyyw dbAnt a, w:

,gog -

lj} $g-At5-A7WM 1300-1206-110 0 -

10%>-

900~

sa>-

ico-600-i Soo-400-300-p 200-

s. RM(MCC 0,

y y'

, , n

! o 4 10 15 )$ jo h h FMURE +2 .

~ ,

uw TR 017 Rev. I Page 53 4.3 RCS Cooldown at 50F'/hr Figure 4-3. demonstrates the thermal response of RV head coolant to the 50F'/hr rate of change imposed upon the RCS. As you will note, the RCS coolant temperature was held constant at 204*F after 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />. The results indicate that the average rate of decline in coolant temperature over the first 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> is approximately 31F'/hr. Thereafter (in the interim between hours 8-10), the rate of decline in coolant temperature shows the effects of the hold at 204*F.

Figure 4-4 demonstrates the rate of change of the saturation pressure of head coolant throughout the cooldown. At this cooldown rate, the results indicate that it would take approximately 7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br /> to reach the DHR cut-in point.

At this point, it should be noted that the head coolant saturation pressure versus time data obtained for this cooldown rate were compared to RCS pressure data obtained during the natural circulation cooldown event at St. Lucle. The results of this analysts compare quite favorably to the St. Lucie data in predicting the time at which onset of coolant flashing was detected. The comparison is elaborated upon in Appendix A.

L_

}

TR 017 Rev. I g a '

Page3/

%4 j ~" ,

t!

• 5

-n

, -4 k-

' k g _t

• ow N M v% 4 N 00 o- S 1

I as 4RRR444555E -4 i!Q -u

-w o M

- =

^

p ,

-t

-m l ~

_~

j l --

k i i 4 4 a L

1._ .

' '; h

,p .

Satmim W WIM gym seraemtwhtessra w 71M-W/teS 8:10xunwN:

1600-($Chf**f'-A5~ YM fp N

• TR 017

/Sb o Rev. 1

/$3g j Page .tf

/3 78 2

/JM 3 8 56 4 y200 610 S

1. 6 11W- 256 7

,py g 88 9 jcoo u m

%0-Boo-W 600-500-400 -

300-200-100-natm

, 1 i i 4 } A + A 4 A M6uM M

~3

4 TR 017 Rev. 1 Page 34 5.0 OPERATIONAL GUIDELINES The methodology in previous sections of this report will provide the basts for operational guidelines to be developed for a natural

~

circulation cooldown.

While it is desirable to avoid reactor vessel upper head voiding in non-emergency cooldown situations, the following should be noted to keep this phenomenon in perspective:

1. A natural circulation cooldown is an unlikely occurrence The two most likely reasons in which the RCPs will not be available are loss of power to the RCP motors and loss of services (cooling water and seal injection) to the RCPs. Expertence has shown that both of these are unlikely for an extended time.

2. A steam bubble in the RV upper head is not a safety problem A steam bubble in the RV upper head is not in itself a safety problem, but a plant control problem for the operator. The expansion of the vold into the hot leg to Interrupt natural circulation is unlikely since the regions below the upper head are subcooled, and expansion of the vold into these regions would result in condensation, thus restricting the vold to the upper head region only,

3. Controlled transition to natural circulation or Intermittent operation of RCPs In the highly likely event of a controlled transition to natural circulation, operation of the RCPs after a reactor trip will reduce

~}

,4 TR 017 Rev. 1 Page i f the RV upper head temperature. RCP operation for seve,ral minutes beyond the reactor trip will be an effective method in reducing the saturation pressure at which coolant flashing will occur. In addition, either intermittent operation or ' bumping' would be an equally effective alternative in reducing the potential for coolant flashing.

Figures 4-1 and 4-3 of the previous section demonstrated the thermal responses of the RV head to 10F'/hr and 50F'/hr coolcown rates in the RCS. The RCS could conceivably be cooled down at any rate up to 50F'/hr. Figures 5-1 and 5-2 represent the minimum (11miting) RCS pressures required to prevent coolant flashing in the RV head.

These curves were developed by correlating the RCS temperature against the appropriate head coolant saturation pressure. For a given RCS temperature, as.long as the RCS pressure is maintained above the saturation pressure line, coolant flashing in the RV head will be prevented.

Plotted on the saturation pressure versus RCS temperature coordinate system, the data obtained for the thermal response of the RV head to various RCS cooldown rates no longer appear explicitly in time.

Consequently, Figures 5-1 and 5-2 are pr'ocess diagrams. If the (P vs T) data for the 10F'/hr cooldown (Figure 5-1) were to be super-imposed onto Figure 5-2, it is apparent (for a given RCS temperature) that the 50F'/hr rate is more pressure limited. As you will note, the important parameter on the process diagram 15 the Z

-4_,_. a m ,

l g gg,, Page ff A

-an M*A5~ WHA i

i I O

1100-

/W M-

. 800 -

i D-W i

M" 300-200-100 - l i

C a00 N .M 400 00 b00 lbo (

F/GURE S-1

^5

Page31 DN

• are-k d f$ h y isrs so, 1611 .RFf 1363 M 3 09 4S4 H 4%

1200.- 9?S 354 -

au 30+

m 254

//d@

JH at.

3 tof

/Q)t>- # M W

ax>-

700-BOO-600-f*>-

200-Nonc*

/CS }kRT *)r 100- MD 1545 0 AW 2AtWRs oa 0

D M M [N_

100 b ,h h M doo Ma n a m s-z t _ , _ _ _ , _ _ . _ . _ _ _ . . . - .

TR 017 Rev. 1 Page WO

-slope of the curve, 1..e. the change in RV head saturation pressure with respect to the change in RCS temperature. With 604*F as one reference temperature, the slope of the 50F'/hr curve is obviously smaller than the slope of the 10F'/hr evaluated over the same conditions. As a result, system pressures at the 50F*/hr rate must be maintained substantially above those for the 10F*/hr case.

The data obtained from the 50F'/hr cooldown analysis will be included in the pressure / temperature limits for a natural circulation plant cooldown. Should there be a conflict between the 50F'/hr data and the fuel-pin-in-compression limits whereby the fuel-pin cumpression limits prove to be more limiting over a particular region of the process diagram, a composite diagram of the two will be developed as the Ilmiting curve over that region, s

~}

s, _ . . . ._ _ _ _ . . _ .

' O- . a s

TR 017

Rev. I Page all

6.0 CONCLUSION

S l Several important conclusions of this report are:

In order to initiate the DHR system, the conditions in the RCS must f

be no greater than 325 psig and 300*F. To prevent coolant flashing in the RV head during RCS cooldowns at 10F*/hr and 50F'/hr,-the head

will be required to be cooled for at least 22 and 7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br /> 1

respectively.

1 Provided the system pressure 15 maintained above the indicated f minimums during the cooldown (saturation pressure shown on process

, diagrams 5-1 and 5-2), flashing of RV head coolant will be prevented.

J I

• The results of 50F'/hr cooldown analysis (Figure 5-2) will be

, incorporated into the existing cooldown procedures. Should there be

! a potential confilct between this data and the fuel-pin-in- '

f compression ilmits over any P-T tegion, the more limiting P-T data will be used over that region.

4 1

1 I

1 i

I i

. .7

_.,- ._ ,_ . __o. -

,,.r ._ . _ .- , . ..- . ,. . _ _ , . _ , _ . . _ , . _ _ _ , _ _ _ _ _ . _ . , _ _ . _ _ _ . _ . , _ , . _ , _ . . . , . . . , _ _

y _ ,

TR 017 Rev. I Page alL

7.0 REFERENCES

1. NRC Letter, J. F. Stolz to H. D. Hukill, Request for Additional Information, Natural Circulation Cooldown (GL 81-21), July 20, 1983.

2. HEATING 6: A Multi-Olmensional Heat Conduction Analysis with the Finite-Difference ~ Formulation, RSIC # RSR-199.

3. TMI-1 FSAR, Updated Version, Volume 2, Chapter 4.

4. J. Smotrel, RV Internal Flow Velocity, B&W Oocument # 51-1146582-01, August 31, 1983.

5. Heat Transfer, 1st Edition. Holman, J. P. McGraw Hill Book Company. 1981.

6. Principles of Heat Transfer, 2nd Edition. Kneith, F. International Text Book. Company. May 1981.

7. Momentum, Heat and Mass Transfer. Bennett, C. O. and Myers, J. E.

McGraw Hill Book Company, Inc. 1962.

8. Convection Heat Transfer. Arpact, U. S. and Larsen, P. S. Prentice Hall, Inc. 1984.

9. Handbook of Heat Transfer. Rohsenow, W. M. and Hartnett, J. P.

McGraw Hill Book Company. 1973.

10. Heat Transfer. Sucec, J. Simon and Schuster. 1975.

11. Convective Heat Transfer. Burneelster,- L. C. John Wiley and Sons, Inc. 1984.

12. Heat Transfer Pocket Handbook, Cheremistnoff, N. P. Gulf Publishing Company. 1984.

13. Process Heat Transfer. Kern, D. R. McGraw Hill Book Company, 1950.

14 Fundamental of Classical Thermodynamics. Van Wylen, G. J. and Sonntag, R. E. John Wiley and Sons, Inc. 1965.

15. Analysis and Evaluation of St. Lucie Unit 1 Natural Circulation Cooldown, NSAC-16/INPO-2. December 1980.

16. ASME Steam Tabies 5th Edition. American Society of Mechanical Engineers. 1983.

L_. .

. 1

,9 TR 017 Rev. 1 Page 43 j APPENDIX A As mentioned in Section 4.3, the thermal response of the analytical model to the 50F'/hr RCS cooldewn was compared to RCS pressure data obtained during the natural circulation r.coldown at St. Lucle. The saturation pressare of RV head coolant (refer to Figure 4-4) was overlayed on to the RCS coolant pressure versus time history for the first 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> as shown on Figure A-1. You will note there appears to be reasonably close agreement in the time during which the onset of flashing of upper head coolant is thought to have occurred.

Reference 15 indicates that the onset of RV head coolant flashing appears to have occurred at 6:15 AM. The authors have also noted that the RV head at St. Lucie appears to have cooled at a rate of 14F'/hr over the first 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> of the event. The RV head' temperature versus time data for the first 3 hc.urs (refer to Figure 4-3) predicts the RV head to be cooled at the rate of 15F'/hr. You will also note that the cross-hatched area (beyond the intersection of the two curves) on Figure A-1 demonstrates the presence of steam volds in the upper head. In addition, temperature data (cited in Reference 15) obtained from the core exit thermocouples throughout the event at St. Lucie, indicates the variation in RCS coolant temperature over time to I

be substantially reduced beyond the 3 hour3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> mark into the event.

C

( '

3

TR 017 Rev. I Page f/

i l

1 s

2M g  : I I I I I '

i d, a, , n 3 I sa. ,

e e

i n e i i

f '\ i E'**

?

l l SfluCMs M W. M n n 2nos y3 y- -

Cearea'srease i i TWr-1:NAvisent CFAt2MAFDEW ' l d *' , anzuw l  ;

' {p preneure feeereer. g, , gypper/ay ,

1m tenwonee 83 (RV MG99)W. 77Mt -

g ti i i i i i atuu.v =. ew I i

'N  !  % !  !  !  ! l !l  !

I  !

leae:

j- y j , j  ; ;Imuo: i  ;

5 'I I! o sr.turJ@m l i.l,,,e, n,,elu se areelue r=l.reer e 4esau-mu d. iI 8

Iisao imo , , A*4m '

K. ,(iL!X . , , ,

isso i

i 1 TI i i

i i i

i e

i i

i i

l i

i_

',\L 1100 I ' ' i i I a i t t

1M leg.

a,t-- r /M AOst- -- - - A. 1 g

soo .

I

' ' .M

' 'I I

, , e.r ,e ,ee , . m.

800 i senseen 1200 noen erie 1830, reester_

i m

i l i Il] W "g,,,,g greemme esenesee #wn efemeenn VA N :

l l i , '. :

! es= = = oys =* = ore .; = e ,eeeeeer  ;  ;

, ,",, 1: W 'i l X  ; i i i i i A I i IN i i

! I t

e i  !!  ! %d  ! 1 l l

,,, i I ,

I! , ,  ; % y i -

!!  !  ! l l N-@,/ T  !

' ' I ' '

0 '? ' ' ' ' I '

O 1 2 3 4 S e 7 8 9 10 11 12 Time Meer Trip pieural 2:33 3:33 4:33 5.33 s.33 7:33 333 tas to 33 11:33 l 12:33 1:33 2:33 w e, = em i l

l l'

1 l

FIGUM R4 l

l l

'.,___________________...__.__...__----+----'--- " - ~ - - " - ' ~-" ~ ~ ^ ~

.