HL-5708, Provides Updated Response to RAI for GL 96-06, Waterhammer in Containment Coolers

Provides Updated Response to RAI for GL 96-06, Waterhammer in Containment Coolers

ML20196A973
Person / Time
Site:	Hatch
Issue date:	11/20/1998
From:	Sumner H SOUTHERN NUCLEAR OPERATING CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
GL-96-06, GL-96-6, HL-5708, TAC-M96819, TAC-M96820, NUDOCS 9811300207
Download: ML20196A973 (8)

Text

- -

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

.' gg

!' , Vee President Operating Company,Inc.

Hatch Project Support 40 invemess Parkway Post Office Box 1295

[- Birmingham, Alabama 35201 Tel 205.992.7279 Fax 205.992.0341 so m e=N m .

COMPANY Energy ro Serve YourWorld" November 20, 1998

- Docket Nos. 321- Hi 5708 50-366 ,

i Tac Nos. 'M%819 ~

M%820 l 4

U.S. Nuclear Regulatory Commission '

ATTN: Document Control Desk

~ Washington, D.C. 20555 l ll Edwin I. Hatch Nuclear Plant Updated Response to Request for Additional Information .

Generic Letter 96 Waterhammer in Containment Coolers i

Ladies and Gentlemen:

On June 30,1998, Southem Nuclear Operating Company (SNC) responded to a Request for Additional Infonnation (RAI) concerning post-accident waterhammer and two-phased flow in .,

- containment coolers per Generic Letter (GL) 96-06. In the response for Plant Hatch, SNC

{

committed to perform detailed analyses, participate in an EPRI program for identifying more realistic waterhammer assumptions, revise procedures, and pro 5ide an update of the SNC response by November 30,1998. His letter provides that update.

Hatch Unit 1:

The Unit I containment area cooling system is an open loop system, in which the possibility of i waterhammer cannot be easily prevented by procedure changes or by simple design changes. I Detailed analysis supported by the EPRI demonstration program is necessary to close the issue.

Hus, completion of ac' ions for Unit I will be deferred pending completion and review of the EPRI  !

technical basis report, which is currently scheduled for July 31,1999. Additional information will i be submitted following receipt of the EPRI report.

Hatch Unit 2:

' The Unit 2 containment area cooling system is a closed loop system, which must be manually l l' restarted following a LOCA. Procedure changes have been made which will prevent operation of l the system when conditions exist which could cause a waterhammer. The changes consider the y appropriate failure modes and effects. Bus, all actions for Unit 2 are complete.

~

9811300207 981120,;

DR ADOCK O 1

-- . - . - . .. = - - . .

i r

I 4 U.S. Nuclear Regulatory Commission Page 2 November 20, 1998 The enclosure to this letter provides a more detailed discussion. Should you have any questions in  !

this regard, please contact this office.

l l

Respectfully submitted, y

M M H. L. Sumner, Jr.

JAW /eb

Enclosure:

Updated Response to Request for Additional Infornation.

Generic Letter 96 Waterhammer in Containment Coolers cc: S.quth em Nuclear Operatina Company Mr. P. H. Wells, Nuclear Plant General Manager SNC Document Management (R-Type A02.001)

U.S. Nuclear Reaulatory Commission. Washington. D.C.

Mr. L. N. Olshan, Project Manager - Hatch U.S. Nuclear Regulatory Commission. Region II Mr. L. A. Reyes, Regional Administrator Mr. J. T. Munday, Senior Resident Inspector - Hatch t

1 a

f

. 4 HL-5708

Enclosure Edwin I. Hatch Nuclear Plant Updated Response to Request for Additional Information:

Generic Letter % Waterhammer in Containment Coolers Summary Information Containment area cooling for Hatch Units 1 and 2 is not credited to operate after a design basis accident. However, the piping system pressure boundary must remain intact to ensure containment

l. integrity.

, Hatch Unit 1:

l' Unit I containment area cooling water is provided by an open loop system, with water provided to fan cooling units via the safety related plant service water system. Due to the system design, waterhammer is possible in the fan cooling units / piping system whether or not the fan cooling units are operated after a LOCA with a Loss of Offsite Power (LOSP). Thus, SNC elected to perform a .

l detailed analysis to determine the potential for waterhammer and its affects on the piping system l l and support structure. SNC also elected to participate in the EPRI demonstration program, which l l should provide more realistic assumptions to use in the analysis. l

ne detailed analysis consists of a dynamic hydraulic model and a structural model. He i preliminary detailed analysis has been performed using conservative assumptions that maximize l waterhammer loads (pressures and forces) on the piping system and piping support structure. He l assumptions include the use of maximum sonic velocities, no credit for noncondensible gases, and instantaneous void collapse. Using these conservative assumptions, the preliminary ant. lysis has shown that significant forces can be generated and transmitted to the piping support av

1 structural system. Bus, SNC has determined that input from the EPRI program will proside useful information necessary to complete the analysis, provide more realistic waterhammer loads, and to provide the necessary infonnation to resolve this issue. Further response to this issue is thus i deferred pending completion and review of the EPRI technical basis report.

Hatch Unit 2 Unit 2 containment area cooling water is provided by a closed loop system, with water provided to fan cooling units by the drywell chilled water system. His system (which includes the chillers, i' pumps, and fan cooling units) receives a LOCA signal, which will automatically shutdown the system. Prior to restarting the system, the operator must manually override the signal using a -

LOCA override procedure. Previously, this procedure was revised to prohibit operating specific fan cooling units in which the water may be susceptible to boiling.

As requested by the RAI, SNC considered the variaus failure modes of the system and that

_ waterhammer could occur even after the containment has begun to cool. Dus, the procedure has

, been revised again to prohibit operating the entire system when boiling may have occurred in any of the fan cooling units / piping system in cocjunction with a LOCA.

HL-5708 E-1 l

~

Enclosure

] Updated Response to Reques? for Additional Information:

Generic Letter 96 Waterham.ner in Containment Coolers t

Detailed Discussion Hatch Unit 1:

Water is provided to the Unit I containment area cooling system by the safety related plant service water system (PSW). The system is an open loop design. In the event of a LOCA and coincident LOSP, the normal system response would cause the PSW system pumps to trip, then restan after i

power is available. He fan cooling unit supply valves (air operated with electrical pilot solenoids) l would open, then automatically close after power is available. Containment isolation valves (motor i operated) would remain open in all cases. The cooling unit fans would trip due to the LOCA

- signal, and would not automatically restan. A failure mode exists that could cause the control l

4 valves to fail open upon loss of air or power. i

! )

To evaluate waterhammer for Unit 1, SNC decided to perform detailed analysis, including various

) failure modes and effects. Note that this analysis will not be finalized until the results of the EPRI

{ program are available.

l i The first step was to determine the extent of voiding within the system. To do this, two " drain l j down" calculations have been developed. The first assumes that the fan cooling unit inlet valves

will close as designed after restoration of power. The second assumes that these valves fail open,

{' thus maximizing the drain down. A short time was assumed (about 3 seconds) for the temperature

of the water in the fan cooling units to reach the boiling point. The system drain down was then I

determined in a quasi-steady state manner with the pressure in the fan cooling unit tubes assumed  ;

i to be the saturation pressure corresponding to the drywell temperature given a LOCA. The total l l drain down volume is then the integrated value over the drain down period. His method assumes l perfect heat transfer from the drywell through the fan cooling unit tubes, and therefore provides a 1 conservative estimate of the volume of system voiding, independent of whether or not the fan is l j mnning, is coasting down, or has stopped during the drain down.

i Next, a hydraulic model was developed scing the Bechtel computer program HSTA (Hydraulic Systems Transient Analysis). To properly onsider failure modes, fan cooling unit inlet " valves open" and " valves closed" cases have been generated. Also, a second " valves closed" case is being

generated to determine the effects of subsequent operator action to restart the fans as currently directed by the Emergency Operating Procedures (EOPs). The HSTA model generated preliminary

forcing functions over each pipe mn segment along the piping system, which is the hydraulic force j on the pipe segment vs. time for the duration of the waterhammer event.

a l Third, the preliminary forcing functions were input into a piping / structural model using the Bechtel computer code ME-101. Using ME-101, several supports have been shown to need further evaluation. From this observation, several approaches are available for resolution:

1

, 1) perform more detailed modeling using a finite element code such as ANSYS to show that the y piping /suppon structure will yield without rupture of the piping system,

2) modify the piping /suppon structure to withstand the higher loads (mitigation),

i i

HL-5708 E-2

Enclosure Updated Response to Pequest for Additional Information:

Generic Letter % Waterhammer in Containment Coolers

[ 3) modify the system logic and valves to prevent water hammer, or l

L 4) reduce the piping / structure forcing function loads by reduction of the predicted waterhammer l

loads in the hydraulic model.

Reducing the predicted waterhammer loads (which would reduce the forcing function loads) is one of the primary expectations of the EPRI program. Thus, SNC has elected ira participate in the L EPRI program. He calculations, hydraulic model, and piping / structural model will be finalized and will be available for review upon completion of the EPRI program.

Hatch Unit 2:

Water is provided to the Unit 2 containment area cooling system by the di vwell chilled water system. He system is a closed loop design, with a surge tank maintaining some positive system pressure when the system is shut down. If a LOCA were to occur, the chilkd water system would trip and would not automatically restart If a coincident LOSP were to also occur, the cooler supply valves (air operated with electrical pilot solenoids) would open, then t utomatically close aAer power is available. Contamment isolation valves (motor operated) wouli remain open in all cases. De cooling unit fans would trip following the LOCA signal, and wout i not automatically l restart. A failure mode exists which could cause the control valves to fail open upon loss of air or power.

It was determined that the water in the fan cooling units could boil following a 1,0CA. His was 1 checked by performing a review of the elevation of the surge tank relative to the fan cooling unit  !

elevations determine boiling temperatures compared to elevated containment temperatures and associated steam in the containment.' Even if the containment temperature were then reduced to below the water boiling points, voids could still exist within the piping system du: to a time delay l for the piping to cool aAer the containment cools.  !

. l l The operator can restart the system aRer overriding the LOCA signal using a LOCA override  ;

procedure. This procedure had previously been revised to prevent operation of the upper level fan l l ,

cooling units following a LOCA, and to prevent operation of the system if the contHnment  !

temperature rose above the boiling point of all of the fan cooling units. However, this revision did i

not consMer the effects of potential failure of the control valves, nor did it consider the possibility of continued voids aRet the containment has cooled.

l Thus, as stated in the Summary section, the LOCA override procedure has been revised again.

The procedure now prohibits the operation of the entire cooling system (including chillers, pumps, i

and fan cooling units) if a LOCA exists and the containment temperature hu at any time been above the boiling point of the water in any of the fan cooling units.

i t

HL-5708 E-3

r Enclosure l

Updated Response to Request for Additional Information- I Generic 12tter 96 Waterhammer in Containment Coolers j I

l Computer Model Descriptions '

l HSTA l De computer program HSTA (Hydraulic Systems Iransient Analysis) was used for the water hammer

! analysis ofthe Containment Area Cooling System at Hatch Unit 1. His program is a generalized fmite l difference code developed by Bechtel Corporation and used extensively over the last 20 years for water hammer design and diagnostic purposes in nuclear and non-nuclear piping systems. He metixxl of l

charactensucs is used to solve the hyperbolic parual differential equations (ofcontinuity and momentum) to obtain the liquid velocity and pressure head at a known grid location. Dese flow mriables are then utilized to generate the dynamic forcing functions on specified pipe run segments. HSTA can model a complex piping system contauung one or more of the several different types offlow devices (boundary conditions) present in the system. Examples of these boundary conditions are: time dependent pressure / flow reservoirs, valves, branches, pumps, surge and air tanks, vacuum breakers, etc.

For the validation of the HSTA code, emphasis was placed on comparison with experimental or test data. His was supplemented by comparisons against independent numerically predicted results available. When comparing time-history predictions of pressure, velocity, etc. in a piping system, both the magnitude and frequency of the variable compared could be important depending on the piping response. For this reason the comparisons are given by directly superimposing the HSTA predicted on the measured (or calculated) variable time-histories and not just by defining the percentage agreement or disagreement between them.

De formal HSTA program documentation includes the validation of the following HSTA capabilities:

1. Validauon of valve actuation transient and pressure wave reflection / transmission at branches / area changes in complex piping networks is against data given in the standard text books by Wylie and Streeter, and Parmakian

2. Validation of centrifugal and reciprocating pump actuation, surge vessels and air tank mitigation devices is agamst various test data (available in open literature and in other Bechtel proprietary

{

data). j

3. Validation ofwater column separation and rejoining calculation schemes includes the validation of  ;

both conventional and line filing schemes described above.

De conwntional scheme is validated agamst test data from several careful laboratory experiments in Europe for two different piping geometries. Further validation is against laboratory tests done for water hammer predicbons in a piping system at a nuclear power plant in U.S.

For the line filling calculation scheme, the HSTA validation was performed as follows:

a) Compansons agamst laboratory test data from Europe.

l b) Comparisons against in-plant test data from a nuclear power plant in U.S.

l HL-5708 E-4

Enclosure Updated Response to Request for Additional Information:

l Generic Letter 96 Waterhammer in Containment Coolers l

c) Co...y.risc=s against predictions from a totally differat computer program that used a simpler calculation scheme and not the Method ofCharacteristics that is used by HSTA.

d) Cu...porisees against predictions from the conventional calculation scheme in HSTA for vapor l pockets smaller than a nodal distance in lerngth Besides the vahdation in the formal code documentation discussed above, the HSTA code participatal in the EPRI water hammer computer code evaluation program during the 1987-1992 EPRI sponsored research effort into water hammer In this program, the results from HSTA were compared against  !

those from several other participating codes for a set ofsix varial water hammer simulation problems. l In the problems for wiuch data (test and analytical) was available, HSTA results compared very well agamst such data.

,. i r

I The HSTA program methodology and its application has been published widely (six papers) both in i national and intenational conferences

1 ME-101 The ME-101 program is used to determine stresses and loads in the piping systems due to i restrained thermal expansion, deadweight, seismic inertia and anchor movements, externally }

applied loads such asjet-loads, and transient forcing functions such as created by fast relief valve opening and closing, fast check valve closure after pipe breaks in the main feedwater line, fast valve closure in main steam line, etc. ME-101 analyzes piping systems in accordance with ANSI and ASME codes.

The ME-101 program is a finite element computer program which performs linear elastic analysis of piping systems using the stiffness method of finite element analysis; the displacements of the joints of a given structure are considered basic unknowns. The dynamic analysis by the modal synthesis method utilizes known maximum accelerations produced in a single degree of freedom model of a certain frequency. The principal program assumptions are as follows:

1. It is a linearly clastic structure.

2. Simultaneous displacement of all supports is described by a single time-dependent function.

3. Lumped mass model satisfactorily replaces the continuous structure.

. 4. Modal synthesis is applicable;

5. Rotational inertia of the masses has negligible effect.

'Ihe results obtained from pipe stress program ME-101 have been compared with the following:

1. ME-632, computer program, seismic analysis of piping systems, VERB MOD 8, Bechtel

International Corporation, San Francisco, California,1976.

l l-HL-5708 E-5 i

l- - , . . , , - . , -

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

Enclosure Updated Response to Request for Additional Information:

Generic Letter 96 Waterhammer in Containment Coolers j

2. ASME Benchmark problem results, Pressure Vessel and Piping 1972 computer programs verification, American Society of Mechanical Engineers.

1

3. Longhand calculations-ME-101 is compatible with NRC Regulatory Guide 1.92. A synthesis l of closely spaced modes is provided based on equation 4 of Regulatory Guide 1.92.

l l

The verification report is on file at Bechtel. i l

Other Procrams If other programs are used in the final analysis, they will be described in the fmal response after the EPRI program is complete.

SNC COMMITMENTS AND SCHEDULE:

1. SNC will continue to participate in the NEI/EPRI program to develop a technical basis l document to demonstrate proper input assumptions are used in the analysis. The program is currently scheduled for completion by July 1,1999. ,

l

2. SNC will finalize the detailed analysis of the Unit I plant service water system upon completion of the EPRI program, and will update the NRC staff by letter, within 90 days of l completion of the EPRI program. j l

i HI 5708 E-6