ML18054A296

Forwards Response to NRC Questions 6,7,8 & 9 Re Plant Performance Testing of Relief & Safety Valves & Informs That Response to Questions 1 & 2 Will Be Submitted by 880915

ML18054A296
Person / Time
Site:	Palisades
Issue date:	05/25/1988
From:	Smedley R CONSUMERS ENERGY CO. (FORMERLY CONSUMERS POWER CO.)
To:	NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM)
References
TAC-44604, NUDOCS 8805310134
Download: ML18054A296 (15)

Text

consumers Power POW ERIN&

lllllCHlliAN'S PRO&RESS General Offices:

~945 West Parnall Road, Jackson, Ml 49201 * (517) 788-0550 May 25, 1988 Nuclear Regulatory Commission Document Control Desk Washington, DC 20555 DOCKET 50-255 - LICENSE DPR PALISADES PLANT -

NUREG-0737, ITEM II.D.1, PERFORMANCE TESTING OF RELIEF AND SAFETY VALVES (TAC NO. 44604) - ADDITIONAL INFORMATION NRC letter dated October 20, 1987 requested 10 items of additional information regarding NUREG-0737, Item II.D.1 - Performance Testing of Relief and Safety Valves.

Consumers Power Company letter dated November 17, 1987 informed the staff that we would respond to Items 3-10 before April 1, 1988 and to Items 1 and 2 before July 1, 1988.

Our.letter dated March 29, 1988 responded to Items 3, 4, 5 and 10, and informed the staff that due to circumstances beyond our control, we would respond to Items 6-9 on or before June 1, 1988.

The Attachment to this letter provides responses to Items 6-9.

Our response to Items 1 and 2 will be delayed until we have selected a supplier for the new valves.

We expect to make that selection in late July of this year and will submit our response to Items 1 and 2 before September 15, 1988.

Richard W Smedley Staff Licensing Engineer CC Administrator, Region III, NRC NRC Resident Inspector - Palisades Attachment OC0388-0073-NL02 8805310134 880525 PDR ADOCK 05000255 P,

1

• oco

\\ '

OC0388-0073-NL02 ATTACHMENT

. Consumers. Power Company Palisades Plant Docket 50-255 RESPONSE.TO NRC QUESTIONS 6, 7, 8 AND 9 REGARDING PALISADES PORV/SRV.

DISCHARGE PIPING ANALYSIS May 25, 1988 13 Pages

NRC Q~estion 6:

Provide more information on the verification of REFORC by EDS Nuclear.

Provide comparisons of the results for REFORC calculations and EPRI/CE data to verify this code is an appropriate tool to evaluate piping discharge transients.

Response

REFORC was developed by EDS Nuclear uhder contract to EPRI. :rt*s function is to compute hydrodynamic forces from the thermal-hydraulic time history data provided by RELAP5/MOD1.

REFORC was verified for use in nuclear safety-related analysis in accordance with the EDS Nuclear Quality Assurance Program prior to its release.

The EDS Nuclear Quality Assurance Program required internally developed

. computer codes to be verified through a comparison of computer output with output obtained from hand calculations for problems covering the ranges of possible input conditions.

REFORC verification consisted of comparing force data output from REFORC with hand-calculated forces for various fluid discharge problems.

The fluid discharge problems included fluid discharging from a resevoir through straight pipe, elbows, reducers, and tees.

Fluid conditions included superheated steam, subcooled liquid, saturated liquid with flashing, and two-phase flow.

Outputs verified were total, wave, and blowdown forces.

In order to provide a comparison of REFORC results to results computed by the EPRI RELAP5/MOD1 post-processor, Impell performed a new RELAP/REFORC analysis. This analysis was performed using the EPRI/CE input file data for Test No. 1411 contained in EPRI report number NP-2479 (Reference 4).

Test No. 1411 was chosen because it more closely approximates the Palisades PORV system configuration and fluid conditions than the other

• EPRI/CE tests. Plots of the REFORC output are compared to the plots of EPRI RELAP5/MOD1 post-processor output and experimental test results in Figures 1 thru 4.

As shown by Figures l through 4, the REFORC-EPRI calculated results are comparable to the results obtained with the RELAP5/MOD1 post-processor.

It should be noted that both REFORC and EPRI post-processor outputs represent only the hydrodynamic component of force exerted against a perfectly rigid, massless system.

By contrast, the test results depict the system's.dynamic response to all load components including piping configuration and pipe suport reactions.* As a result, neither REFORC nor EPRI RELAP5/MOD1 post-processor results are directly comparable to the.

test data.

.3000 ----

1500 --

0

~

Q.)

0 J..

0

~ -1500 ---

-3000 0.00 I

I I

I 0.10 I

I I

I I

0.20 Time I

I I

0.30 (sec)

I I

I I

1*

I I

0.40 u..

ID

...J REFORC Calculated Segment 1 Force* History CE TEST ND* 1411 S!OftfMT l fD!tC!. R!l.llPla80LlD.DATAaOAIH!D 4111 2121 g

I~

I I

1

!!.~!\\----+-*---------------1-----;---~

I a

\\

I~

I'

~

~11.

i

\\

I I \\ i\\ /

1

• I I I

,\\ I 11 I

\\

I I I

/1 "' I I\\ I..

I I

I I

I I I I I I I I 11,,

\\ A I II,, Ii I

\\ N ~

',l'/ * " v

" I f

~

jW

~

I I

I i

I I

I I

! L---------.1..---------J---------:~------:-L-:---------:

I Q.0 0.1 0.1 0.9 0.4 0.1 TIME l SEC l I

0.50 Fi111%* 4-8.

Comparison Plot of RELAP5/MOD1-Calculated SeplOllt 1 Force with Esperimental Data. Teat 1411.

Figure l

• ~

,.0 Q)

\\.)

M 0

~

.e 10000 -

~

5000 0 --

-5000 -

-10000 0.00 I

I

~

I I

0.10 I

I I

I I

0.20 Time l

I I -

0.30 (sec)

I I

I l

I I

0.40 REFORC Calculated Segment 2 Force History CE TEST NO. 141i S[Oft!NT 2 roRct. R[tJIPl*SDLlD.QRTA:OASH[O 4111 2/21 I

0 0

0 IO

. -*-r *---*---,..-*.. **-** ---..

I !

,i IU c

,,.t..J I"\\ i-,, ~!"'/\\I

~

v.,, v

,1, i

.. ~ ~-----+-----~----...,._ _________ ~

i 1 I I

1 0.50

' o.o 0.1 o.r o.s o.*

0.1 TIME C SEC l Plaure 4-10.

Comparison Plot of RELAP5/HOD1-C.1culated Sepient 2 Force with Experimental *Data. Teat 1411.

Figure 2

~

,.c -

Q)

CJ J..I 0

~

20000 --

10000 ----

0 ----

-10000 ----

-20000 0.00 v

I I

I I

0.10 I

I I

I I

0.20 Time I

I I

I 0.30.

(sec)

I I

I I

I 0.40 REFORC Calculated Segment 3 Force History a a a a N

CE TEST NO. 1411 SEOftENT 3 fORC!.

REUIPlaSO~lD.DATA*DASH!D 4111 Z/Zl r

I I,,..

I I

---

* --*- -~----,_.. ____ _

I r..J. a u

ac:

a

~

§ -

I I

' a.a l \\

I r1*

\\

f I

I i

\\

I I

I I

I I

I I

I

,, I I

I I

I,

~

I

~I

"'" I IJ

'J

\\

I

\\ /

\\ I

\\ I I

\\/

i I

I I

I O*I a.a TIME C SEC l

/"\\

I \\

I

\\

I i

I I

.. I 0.4 I

I 0.50 0.1 Flpr_e 4-12.

Coapariaoa Plot of RELAP5/JIOD1~1C1Llated Sepioat 3 Force with Ezperimeatal Data, Teat 1411.

Figure 3

~

Q

~

20000 10000 0

\\..-_

0

~ -10000 -

-20000 0.00 I

I I

I 0.10 I

I I

I I

0.20 Time I

I I

0.30 (sec)

I I

I I

I I

0.40 REFORC Calculated Segment 4 Force History IA.

CD

...J 1M o u

a::

0 IA. g g -

I

• CE TEST NO. 1411 l!llft!NT 4 'OltCt* R!UU'lalOl..lD.~ATAaDASH!D 4111 2/fl I

I 0.50 IL--~-l-~~-L-~----L~~~~--=

' a.a 0.1 o.r 0.1 0.4 0.1 TIHE l SEC l Flpre "4-14.

Coaparholl Pl~t of l!LAPS/llODl-C.lcalaucl Sepiat o4 Force witll. Ezperiaeatal Data, Teat 1411.

Figure 4 NRC Question 7A:

Insufficient detail was received on the key parameters used in the RELAPS/MOD) thermal-hydraulic analySlS. Additional informatio~ is needed on the following items:

Node Size:

The EDS Nuclear report submitted December 30, 1982 stated the short pipe segments (less than 5 ft) immediately downstream of the valves was modeled with 8 to 10 nodes.

Farther down the line, the nodalization was coarser to minimize the model size. This approach was used because hydrodynamic loads were felt to be smaller farther down the pipe.

However, a review of the Intermountain Technologies, Inc.

(!TI) report verifying RELAP5/MOD1 as suitable tool for analyzing piping loads shows that for Test 1411, a steam test with the Crosby 6M6 valve, the piping loads did not decrease further down the discharge pipe. Justify that the node size used in the analysis produced bounding piping forces or redo the analysis.with smaller nodes.

Response

EDS Nuclear calculation 0540-006~673-001 Revision 0 (Reference 3) documents the results of a sensitivity study which was performed to determine the validity of utilizing a 11coarser 11 nodalization in downstream piping. This sensitivity study consisted of comparing the RELAP5/MOD1 output from the **coarse node 11 model actually used for the load analysis with a model using uniform piping volume lengths of approximately 0.5 feet.

This study indicated that the RELAP5/MOD1 output was not significantly affected by the use of the coarser nodes.

Therefore, the hydrodynamic loads calculated by REFORC using "coarse node" RELAP5/M001 output is considered to have produced bounding forces.

NRC Question 78:

Time Step Size:

The EDS report stated a time step of l.O x lo-3 set was used in the analysis. This time step appears to be too large to ensure that piping forces were accurately calculated.

EG&G Idaho, Inc.,

experience has shown that the time step size should be based on

~

nC where Ax = minimum node size, Ca sound speed, and n = a multiplier that should be at least 2, since the shock wave from the valve discharge can travel at twice the sonic velocity. Therefore, assuming Ax = 0.5 ft, At ~ 2.0 x lo-4 sec would appear to be more appropriate. Justify the time steps used in the analysis result in bounding piping forces being calculated or redo the analysis with smaller time steps.

Response

As stated in Impell -calculation 0540-006-673-001 Revision O (Reference 3) the RELAP5/MOD1 analysis was performed using a time step value of 0.2 millisecond. This is consistent with the value stated in NRC question 78 above.

The 1.0 millisecond time step stated in Report No. 02-0540-1124 (Reference

1) refers to the time step used for the REFORC calculations. This time step was selected based on an inspection of the RELAP5/MOD1 output data.

The output data was judged to be sufficiently smooth to justify a 1.o millisecond time step.

In addition, the REFORC results shown in Figures 1 thru 4 were calculated using a time step of 1.0 millisecond. Since these results compare favorably with the EPRI RELAP5/MOD1 post processor results, it can be concluded that the 1.0 millisecond time step used in the REFORC analysis produced bounding hydrodynamic forces.

NRC Question 7C:

The transient cases analyzed included only steam di~charge through the PORVs and safety valves. Since the PORVs are used for low temperature overpressure protection <LTOP), they will be required to pass subccioled

• and saturated water during these transients. Justify not analyzing this transient condition or perform the analysis and provide the results for our review.

If the transient was not analyzed because it was considered to be bounded by steam discharge, compare the calculated steam discharge loads with estimated water discharge loads to verify the steam load is

.bounding.

Provide the comparison for our review.

Response

In order to determine if the LTOP saturated and subcooled water discharge loads are bounded by steam discharge loads additional calculations were performed* to estimate the downstream hydrodynamic l-0ads.

The fortes were estimated by considering an elbow subjected to both pressure and momentum forces.

The expression used to calculate the forces is as fo 11 ows:

• Force = PA + ~

gc where:

P * = pressure A = pipe area m = mass flowrate V = fluid velocity gc = gravitational constant Several conservative assumptions were made in order to obtain an upper bound estimate of the forces.

These assumptions were:

0 Downstream Pressure Although the actual initial downstream pressure is small, the upstream LTOP actuation pressure of 375 psia (Reference 8) was used in the downstream force expression shown above.

0 Mass Flow Rate The discharge rate (m) through the valve was calculated using an upper bound flow coefficient for the valve.

The discharge rate calculated with this method for the saturated water conditions (187,000 lb/hr) was found to be conservative with respect to the Moody Critical flow rate (185,000 lb/hr).

For the subcooled discharge rate calculation, the downstream pressure was assumed to

. be atmospheric iri order to yield the maximum <consetvative) flow rate.

A comparison of the magnitude of the estimated LTOP initiated hydrodynamic loads to the saturated steam discharge load from Reference l is as follows:

Fluid Condition CLTOP) Saturated Water CLTOP) Subcooled Water Saturated Steam Force Magnitude Clbf) 10.9 x 103.

4.8 x 103 14.5 x 103 CRef. 1)

The results shown above demonstrate that the steam discharge loads are greater than the saturated and subcooled water LTOP discharge loads.

NRC Question 70 PORV Opening Time:

The PORV opening time used was 170 msec, based on an average of the opening times measured during the EPRI tests. Since opening times as fast as 62 msec were measured, and the piping forces can be significantly affected.by this parameter, the opening time used may not have resulted in the calculation of bounding forces. Justify that the piping forces calculated with the 170 msec PORV opening time bpund those expected at the plant or redo the analysis with the faster valve opening time.

Response

The 62 millisetond PORV opening time stated above refers to water discharge test results shown in EPRI Safety and Relief Valve Test Report, Table 4.l.1-3b (Reference 5).

The ~nalysis performed for Palisades considered only steam discharge from the PORV's.

As stated in the EDS report, high pressure water discharge was not postulated since high pressure safety injection pump shutoff head is below both SRV and PORV set pressures. Therefore, the 170 millisecond PORV opening time used is appropriate for the Palisades case.

NRC Question 8A:

More information is needed on the following structural analysis parameters.

A.

Lumped Mass Soacing:

Provide the maximum and minimum lumped mass spacing.

Demonstrate the spacing is adequate to analyze frequencies up to 150 Hz.

Response

For the determination of mode shapes and frequencies, a simple lumped mass idealization is used in the SUPERPIPE computer program.

A mass point is placed at each node defined in the model and if necessary additional mass points are placed automatically between these nodes, thus subdividing lengths of pipe into. subelements.

The criterion for introducing additional mass points is as follows.

For a simple span beam of length L, stiffness EI, gravity acceleration~.

and weight w per unit length, the first mode frequency, f, ignoring shearing deformations, is given by f =

2L 2 Elg

(--)

w 0.5 If a long length of straight pipe is vibrating, each mode of vibration will contain a number of equally spaced nodes, and each length of pipe between nodes vibrates as a simple span beam.

Hence, for a frequency of vibration, fm, the node spacing (i.e., equivalent simple span beam length) will be Lm, where 0.5 11' Lm = (--)

2fm Elg 0.25

(--)

w If a simple lumped mass idealization is used, and if an accurate determination of the mode shape and frequency is required for a simple span beam of this length, then at least one additional lump~d mass is required within the beam length. That.is, to obtain accurate results for a mode with a frequency fm, the mass spacing should be no larger than Sm. where Sm = 0.5Lm By specifying in SUPERPIPE a mass point spacing compatible with a frequency of 150 Hz, the maximum mas~ spacing Sm was found as defined above.

As a minimum, mass points have been placed at all elbows, valves, tees and discontinuities in the pipe routing.

In addition, at least one mass point has been located between any 2 supports in the same global direction.

NRC Question BB:

The time step used in the analysis was said to be 0.001 sec. It was also stated that frequencies up to 150 Hz were considered in the analysis.

The time step used appears to be too large to accurately calculate frequencies up to 150 Hz.

Justify that, with the time step used, frequencies up to 150 Hz are accurately calculated or redo the analysis with smaller time steps.

Response

The cutoff frequency of 150 Hz was chosen in the analysis as a conservative upper bound for mass point spacing and for damping. This choice ensured that all possible bending modes had been considered, and that the enveloped frequencies were conservatively underdamped.

The integration time step of.001 seconds for the time history analysis was selected to provide accurate response of the higher frequencies of the system.* As documented in Impell Report No. 02-0540-1124 <Reference 6),

sensitivity studies were conducted to.determine an acceptable time step.

The piping model was subjected to time histories using time steps of 2, l and 0.5 milliseconds; The changes seen in results between 1 and 0.5 were*

insignificant; therefore Q.001 second time step was used in the anlaysis EDS calculation 0540-006-641-C-001 Revision 1 (Reference 7) documents this*

study.

The direct integration solution method was used for the force time history analysis.

SUPERPIPE allows the.system dynamic characteristics to be written as ~ set of differential equations of the form:

Mu + cu*+ Ku = P where M, C, and K represent the mass, damping, and stiffness of the system, u is the time-dependent displacement, and P is the applied load.

For. the direct integration option, the analysis is carried out directly on coupled equations of dynamic equilibrium without unc6upling irito normal modes, and therefore without truncating the high frequency effects.

This set of equations is solved in coupled format by generating the response of the system as a function of the response at the previous time step.

By asstiming that tha damping matrix is a linear combination of the mass and stiffness matrices, two unique frequency damping ratio pairs can be selected.

These values were taken as one percent of critical damping at both the fundamental structural frequency and at.the highest significant mode considered in the analysis (150 cycles/sec.).

The frequencies of interest - that is, those between these limits - are conservatively underdamped.

NRC Question 9:

In its December 30, 1982 submittal, Consumers Power noted that at some locations downstream of the safety valves the calculated stress exceeded the allowable stress when valve actuation is considered, but that at no locations do the combined sustained and valve actuation stresses exceed the faulted allowable.

However, in Table 4-2 of EDS report 02-0540-1124 the sustained plus valve actuation load was calculated to be 29811 psi compared to an allowable of 13894 psi (equal to 1.2 Sh>*

Even if the allowable is increased to 2.4 Sh, the faulted allowable, the calculated load will exceed the allowable. Therefore, provide a more detailed comparison of the maximum calculated and allowable stresses for the PORV and safety valve inlet and outlet piping and supports. Provide a system diagram that is clear enough and detailed enough to locate all the locations compared above.

For all overstressed piping and supports, such as that identified in Table 4-2 of the EDS report, discuss what will be done to bring the stresses to within code allowables.

Provide a schedule showing when any needed modifications will be made.

Response

The PORV piping system was reanalyzed in June, 1985.

Changes to the model and analysis procedures which have an effect on the pipe stresses and support loads include analyzing the entire piping system as one model as*

opposed to the two models previously analyzed.

The supports which were previously modeled as in-.line anchors are more accurately modeled in this analysis.

The seismic response spectra was corrected in the new analysis. These changes, along with incorporating more recent walk.down data, affected the piping analysis results.

Impell Report 02-0540-1124, Revision 2 <Reference 6) shows that, with these model changes, and with modifications to spring hanger settings, the piping is all below EPRI recommended allowables.

The piping stresses are all below the FSAR allowables except for Equation 6 which uses a 1.. 2 Sh allowable stress. This equation considers pressure, dead weight, and valve actuation loadings.

The valve actuation loading applied was an envelope of the two valve actuation load cases.

The EPRI equations consider this load case in two parts. First, the EPRI criteria uses a l.2Sh allowable for a relief valve transient. However, for a safety valve transient, EPRI uses a l.8Sh allowable. Using this criteria, all piping stresses are qualified.

The isometric diagram of pipi.ng system is shown in Figure 5~

I I

N I

H858.1 R2 CONSUMERS POWER COMPANY PALISADES Pl.ANT

• FIGURES PRESSURIZER REUEf VALVE DISOWIGE PIPING

• SYSTEM LAYOUT 442 441 10438 R2 y

x

REFERENCES:

l.

EDS Nuclear Report Number 02-0540-1124, "Evaluation of Palisades Safety and Relief Valve Discharge Piping," Revision 0, dated 11-30-82..

2.

NRC Letter from Mr. Thomas V. Wambach to Mr. Kenneth w*.- Berry, "Pali sades

- NUREG-0737, Item II.D.1, Performance Testing of Relief and Safety Valves

<TAC No. 44604}," dated October 20, 1987.

3.

EDS Nuclear Calculation Number 0540-006-673-001, "SRV Piping Thermal Hydraulic RELAP5 Analysis," Revision 0, 11-5-82.

4.

EPRI Report NP-2479, "Application of RELAP5/MOD1 for Calculation of Safety and Relief Valve Discharge Piping Hydrodynamic Loads," December, 1982.

s.

EPRI "Safety and Relief Valve Test Report" Research Project Vl02, Interim Report, April, 1982.

6.

Impell Report No. 02-0540-1124 Revision 2, March 1984.

7.. Impell CEDS Nuclear) Calculation Number 0540-006-641-C-001, "Model Generation Loading, Criteria and Pipe Stress Results", Revision 1, 12-29-82.

a.

Palisades Nuclear Plant System Lesson Plan, SH-PS2-6, Revision 1, 5/13/83.