ML20012C593

Forwards Response to Request for Addl Info Re Degraded Tube R1C55 in Steam Generator A.Athos 3-D T&H Analysis Used to Obtain Calculated Velocities for Same Test Conditions

ML20012C593
Person / Time
Site:	Zion File:ZionSolutions icon.png
Issue date:	03/13/1990
From:	Chrzanowski R COMMONWEALTH EDISON CO.
To:	Murley T Office of Nuclear Reactor Regulation
References
NUDOCS 9003220250
Download: ML20012C593 (9)

Text

y ~~'N Commonwealth Edison

/ 1400 Opus Place

,y

- 7 Downers Crove, Illin:Is 60515'

..\\

/

March 13, 1990~

Dr. Thomas E. Murley, Director Office of Reactor Regulation U.S. Nuclear Regulatory Commission Washington, DC 20555

Subject:

Zion Station Unit 2 Response to Request for Additional Information Degraded Tube R1055 in Steam Generator A BRQ_ pocket No. 50-304

Reference:

(a) March 2,1989, letter from G.E. Trzyna to T.E. Murley t

(b) February 12, 1990, letter from C. P. Patel to T. J. Kovach

Dear Dr. Murley:

Commorceealth _ Edison submitted a detailed assessment of degraded tube

-RlC55 in the Zion Unit 2A steam generator in reference (a). This assessment was requested by the staff to ensure that this tube, which is plugged due to a circumferential crack, will not cause damage to adjacent tubes. Reference (b) requested additional information regarding the assessment of the degraded steam generator tube. Attachment A to this letter providea the answers to the nine requested questions.

If any further questions arise, please address them to this office.

Very truly yours, WC.

R. A.

,hrzanowski Nucle.ar L ensing Administrator cc:

Zion Resident Inspector C. P. Patel - NRR s

Region III Office Office of Reactor Safety-IDNS c

9003220250 900313 DR ADOCK 05000304

[q) p PDC r: J e

<h

bit REQUEST FOR ADDITIONAL INFORMATION DA COMMONWEALTH EDIS0N COMPANY

. ZION UNIT 2 DEGRADED TUBE R1055 IN STEAM GENERATOR A-

Reference:

Commonwealth Edison letter dated March =2, 1989, to Director of Nuclear Reactor Regulation, NRC, with enclosed Westinghouse Report WCAP 12175.

Question 1 Section 5.'f4 of WACP-12175 briefly describes the qualification of the analytical turbulence excitation model using a prototypical two-phase test.

Please described how FASTVIB (see Section 6.1) was qualified to calculate

_y fluid-elastic stability ratios for prototypical two-phase conditions. What is

~

the expected accuracy- (in terms of +/= "x" percent) of the turbulent response model and the FASTVIB model for prototypical conditions?

Answer FASTVIB instability condition evaluation were qualiflod for prototypical.

two-phase flow conditions with full scale, U-bend tubes in a simplified geometry (rectangular array of U-bend tubes) in a model boiler test by way of the following procedure:

instability condition established (by increasing flow velocity) for a specific tube in a two-phase test U-bend tube array (with prototypical pressure and temperature conditions).

- total flow conditions at instability established boundary conditions and tube properties establish frequency and diameter with the 3

frequency at instability measured in the test.

- damping measured in the test for the specific tube tested and varying void fraction ATH0S 3-D T&H analysis was used to obtain calculated velocities (i.e.,

l-effective velocity) for same test conditions 1

'a value for the Beta constant was calculated using the ATHOS based effective velocity with all other paramaters (in the defining equation for stability ratio, see bottom of page 35 of the referenced letter and l

. equation-[1] and [2] on page 36) based on the test (see above).

t 4

I i

2'-

.The above-integral verification procedure lines the value of the derived Beta with the measured damping and applied analytical methods for calculating-velocity, density and void fractions for prototypical two-phase flow conditions. Thus potential errors in the analytical methods would be C

compensated by the derived Beta.

Overall, the values for Beta obtained by these procedures from two-phase tests are in good agreement with, although somewhat higher than, those obtained from air model tests of U-bend tubes.

For conservatism in analyses, the lower valves of Beta f rom the two phase tests are combined with the lower range of measured damping from the two phase 1

tests.

In terms of accuracy, it continues to be Westinghouse's intent to be realistic with potential (overall) errors on the conservative side when doing flow induced vibration evaluations for field conditions.

For two-phase U-bend evaluations,.we expect both our turbulence amplitude calculations and our stability ratio evaluations to be accurate within engineering limits with any potential error being on the high (conservative) side. For turbulence, this is accomplished by providing that the spectra constants (C1 and S) are obtained from the envelope of the many field and test measured spectra in our database. This database includes (but is not limited to) Westinghouse laboratory and field experience as well as experiments reported the literature. This database has continued to gros for over 20 years and reflects the accumulated experiences of the nuclear industry.

The critical flow velocity (equation [1] page 36 of the WCAP) is proportional

'to Beta and the square root of the damping parameter (Zeta). Thus, Beta affects accuracy much more than does Zeta.

For stability evaluations, potential errors should be on the conservative side because of the use of the lowest value for Beta reported in our database and ~the use of damping data obtained from prototypical two-phase flow, pressure and temperature

-conditions. Sources for U-bend data include tests performed in our air water tunnels, tests reported in the literature, and proprietary tests performed by Westinghouse licensees. The database for instability related information also continues to be a living document at Westinghouse.

In addition, the FASTVIB computer code has been verified against an earlier configured version of our tube flow induced vibration evaluation codes called FLOVIB.

It was demonstrated that they produce identical results for the same (input)' conditions. This is important here because of the long history associated with the qualification and verification of FLOVIB relative to make accurate flow induced vibration predictions. FLOVIB has been used as the basis for the evaluation of many designs which have proven themselves to be successful based on field experiences. Many of these evaluations have been presented in the open literature.

Question 2 Provide an assessment of the uncertainty associated with the stability ratio results in Table 6-1 which is introduced by uncertainties in the assumed damping coefficient and stability constant and by the uncertainties in the ATH0S ficw velocity, density, and void fraction distribution results.

0774T/RAC:wj

p 1

Answer Please refer to the answer to Question 1 for an assessment of'the uncertainty associated with the stability ratios in Table 6-1.

Note that since the overall procedure was qualified against prototypic two-phase flow conditions,

' the efforts of all parameters (including damping, Beta, flow velocity and density, etc.) are included and, therefore, it is not intended that they be addressed separately.- Based on the integral test verification procedure, the Beta and damping values are selected to bound uncertainties on the b

conservative side, that is, to overestimate fluidelastie instability. More information relative to qualification of ATHOS can be found in Appendix A (attached). There, many of these parameters ar e addressed separately.

Question 3 Considering the scenario of a severed tube discussed in Section 6.4, confirm that the Westinghouse model considered the U-bend segment extending from the severed location to the top support on the hot leg side (rather than simply the shorter U-bend segment extending from the severed location to the top support on the cold leg side).

Answer This response confirms that the Westinghouse model considered the U-bend segment extending from the severed ~1ocation to the top support on the hot leg side.

Question 4

. The staff notes that, dependent on the actual crossflow velocity distribution, a tube may initially undergo instability in a mode other than the lowest frequency mode. Has Westinghouse calculated.the modal effective velocity (MEVEL) and associated stability ratio for several of the lowest modes? Do the results given in Table 6-1 correspond to the lowest mode?

Answer For cases with pinned boundary conditions, Westinghouse has calculated the modal effective velocity and associated stability ratio for all modes with frequencies less than or equal to 200 Hertz.

Results for the worst of these (in terms of stability ratio) are presented in Table 6-1.

For each of the two cases (in Table 6-1) with fixed boundary conditions, worst results from the lowest two made are reported.

It is recognized that other modes associated with such a short stub of tube in the U-bend will have very high frequencies indeed, and it is judged that they will be stable on the basis of this fact and the low stability ratios reported (in Table 6-1) for the modes considered.

0774T/RAC:wj l

b D

L s

'N'^

-4 Question 5 Clarification of note (3) of Table 6-1 is requested.

For examples, when the l

authors stated " Actual.U-bend values would be lower than the valves listed for these case", are they referring to all the values or only the values for etability ratio and turbulent displacement?

L Answer Note (3) of Table 6-1 refers only to the case with (assumed) pinned boundary conditions and the " current" crack condition.

For this case (only):

there are no U-bend modes with frequencies less than 200 Hertz because of the very short U-bend length and radius of the R1C55 tube.

values of stability ratio and turbulence displacement given for this case are actually associated with the worst case (largest stability ratio) straight-leg mode (rather than any U-bend mode).

stability ratios and turbulence displacements associated with the U-bend modes will be smaller than their respective straight-leg counterparts (given in Table 6-1).

Question 6 "The linearly supported tube" (p. 35) was used for the fluidelastic inetability analysis, while " nonlinear, finite element, dynamic methods" (p. 37) were used for the turbulence responses of U-bends. -Why were two different models used?

Answer For the Zion R1055 tube all supporting analyses, that is, both turbulence and fluid-elastic analyses, were done with linear models (the same (linear) model was used for both analyses).

The first two paragraphs of Section 5.2 (p. 37) were intended to address various aspect of the qualification of our overall turbulence calculational methodologies. These qualifications were done with both linear and nonlinear analytical techniques.

Focusing on the fundamental elements of linear model turbulence. methodology qualification, it is noted that the constants C1 and S (see equation on P. 38 (with the typo by the NRC Staff corrected as in 7, below) were obtained from Westinghouse air and water tunnel test. These were

.used to compute the response of U-bend tubes with known boundary (support) conditions in a prototypical (pressure and temperature) two-phase flow test.

These computed response paramete.s were demonstrated to be in very good agreement with the test mea =" rod response parameters.

As indicated in the third paragraph of Section 5.2 (as well as by the equation of modal turbulence amplitude nn p. 38) a linear model is considered.

0774T/RAC:wj

' Question 7 l

On p. 38,.the equation for the response to turbulence excitation is independent of damping.

Is the effect of damping included in the parameter C?

l Answer The damping parameter was inadvertently omitted due to a typing error but was included in the analyses.

It is not included in the Ct parameter (which is invariant for these analyses). Turbulence response amplitude is inversely proportional to the square root of damping and the equation on p. 38 of the WCAP should be modified accordingly.

Question 8 There are errors in Eq. (1) [ Beta is missing in the numerator of the right-hand-side), Eq. (2) [the slash (/) symbol in the denominator of the right-hand-side should be deleted),.and the equation on p. 38 [should be subscript o on rho rather than superscript 0; should be delta-zj in denominator of right-hand-side).

In the nomenclature given on p. 36, n is omitted.

Answer All Staff comments are correct.

"n" is the mode number.

Question 9 The' staff believes it would be prudent to inspect the tubes adjacent to R1C55 as a part of each inservice steam generator tube inspection in order to confirm the analysis prediction that damage to these tubes is not occurring.

Please discuss your plans in this regard.

Answer The inspection of tubes surrounding R1055 is part of the normal Zion Steam Generator Inspection Program.

It is important to note that Row 1 Column 54 and Row 1 Column 56 Tubes, immediately on either side of the tube in question, are currently plugged and therefore cannot be inspected. We will inspect the adjacent tubes in Row 2 and R1 Column 53 and R1 Column 57.

Additionally, a random sampling of twenty-five Row-2 U-Bends in each steam generator will be tested using the motorized rotating pancake coil (MRPC). Row 2 tubes R2 Column 54, 55, and 56 will be specified in the MRPC testing sample for the 2a steam generator.

0774T/RAC:wj

r I

4 i

' APPENDIX A

SUMMARY

OF ATHOS3 CODE VERIFICATION WITH REGARD TO UNCERTAINTIES IN CODE PREDICTED FLOW VELOCITTY, DENSIIT AND VOID FRACTION DISTRIBUTIONS The ATHOS3 computer code was used to generate secondary side velocities, densities, and void fraction distributions in support of the analysis of degraded tube R1055 in generator A of Zion Unit 2.

ATHOS3 (Reference 1) is a three-dimensional two-phase code for steady-state and time dependent analysis of the thermal-hydrauli: performance of PWR steam generators.

It, and earlier versions (URSULA2 and ATHOSE), were developed and are maintained by the Electric Power Research Institute, with CHAM of North America being the primary code developer.

The assignment of specific uncertainties to each of the ATHOS3 predicted paramaters, in particular, to local flow parameters in the U-bend region is difficult as only very limited data exits. However, a number of measurements are available for many other parameters related to both the global and localized thermal-hydraulic performance. These data have formed the core of a number of extensive code verification studies completed over the past several years. These studies provide an overall indication that the code can be used with confidence in the simulation of recirculating steam generator performance. Verification has involved comparisons with measurements from both scale model tests and from full-size operating steam generators.

In the early 1980's, the ATHOS2 code was validated against an extension database collected from a 10 MWt model boiler termed the Westinghouse Model Boiler No. 2, MS-2 (Reference 2).

This was a scaled simulation of the full-size Model F steam generator, which, like the 51 Series generators installed at Zion, is a recirculating feedring unit. The thermal-hydraulic data-included the circulation ratio, primary-side temperature and pressure distribution and temperature and pressure distributions on the secondary side. With prototypic steam-water conditions, tests were performed for overall steady-state operating characteristics, flow conditions at the tubesheet, and a number of operational transients, to name a few.

The ATHOS2 code was also compared to the thermal-hydraulic data obtained from Westinghouse 51 Series generators in the French Bugey 4 and Tricastin 2 plans (Reference 3).

Code predictions were compared with steady-state measurements for circulation ratio, fluid temperature and pressure, etc., and transient performance parameters as a fluid temperature, water level, and steam flow.

0774T/RAC:wj

16,.

' Subsequently, the current ATHOS3 version was released and has undergone a number of detailed' comparisons with both model and full-size plant data.

Reference 4 Volumes 1 and 2, document code predictions and comparisons with local secondary-side velocity distributions obtained in both 2/3-Scale (cold water test) and 0.95-Scale (air test) test models. These models simulated the inlet. region of the preheater from full-size, non-feedring, Westinghouse steam generators.

Local velocity measurements were obtained from both the feedwater entrance region of the tube bundle and at sites adjoining non-tubed regions, in addition to locations at the second and third flow passes between the preheater flow baffles.

In 1988, an additional verification study was completed based on plant data obtained from operating Westingbouse Model F and Model D4 generators (Reference 5).

The Model F data comparisons included the following thermal-hydraulic parameters: cfreulation ratio, primary temperatures, circumferential distributions of downcomer fluid velocity and temperature, lower bundle secondary-side temperatures, and pressure-losses in the r

circulation loop. Steady-state data was obtained over a wide range of power (36 to 99%). Data from operational transients including a large step load-reduction and a turbine trip were also obtained and compared with ATHOS3 predictions. The Model D4 program included data comparisons for steady-state temperatures and pressure drops in the preheater, and downcomer and primary fluid temperatures.

Yet another comparison between ATHOS3 predictions and velocity measurements was documented in the ANS Transactions in 1988 (Reference 6).

This study involved comparison of ATHOS predicted U-bend tube gap velocities with test data obtained from the French " Leonard" scaled model cold flow test.. As such, it is of particular interest to the Zion degraded tube study.

In the U-bend region, it was shown that the agreement between the ATH0S tube gap velocity predictions and measures was quite good.

1 The preceding studies represent an extensive code verification program completed over the past several years. Without citing the results of any individual comparisons between the code predictions and measurements, it can be stated that, overall, the ATHOS3 code is capable of simulating the thermal-hydraulic performance of both feedring and preheat recirculating steam generators. The code can predict both the global performance, as indicated by such parameters as the circulation ratio (i.e., the bundle flow rae) and primary fluid temperatures, and more localized temperature and velocity distributions over a range of power levels and boundary conditions. These comparisons have exercised all aspects of the code modeling including primary-to-seconardy side heat transfer, flow and pressure modeling, and the specifics of the supporting empirical correlations. This good agreement

-provides confidence that the code performs an adequate job of predicting other auxiliary parameters, such as the fluid density and void fraction, which have

.not yet been extensively validated directly against measured data.

0774T/RAC:wj

ep _

m NN f;

o

., s f 52 REFERENCEJ r

(1)

L. W. Keeton, A.. K. Singha1, "ATHOS3: A computer Program-for-Thermal-Hydraulic-Analysis of Steam Generators", Volumes 1. 2, and 3, L'

EPRI'NP-4604-CCM; July 1986.

.(2)

G. W. Hopkins, O. J. Mendler, and A. Y. Lee, " Thermal and Hydraulic Code r

Verifications - ATHOS2 and Model Boiler No. 2 Data", Volumes 1, 2 and ' 3,-

-EPRI NP-2887,. February 1983.

l (3) -P. J. Masiello. " Thermal-Hydraulic Code Qualification: ATHOS2 and Data.

E from Bugey 4-and Tricastin 2", EPRI NP-2872, February 1983.

(4)

A. Y. Lee, P. J. Masiello, "ATHOS3 Computer Code Verification:,

' Volume 1:

2/3-Scale Test and Volume 2: 0.95-Scale Test, EPRI NP-5557, June 1988.

L(5)

C. W. Hopk3ns, " Verification of the ATHOS3 Code Against Feedring and Preheat Steam Generator Test Data", EPRI NP-5728, May'1988, i

(6)

A. Y. Lee, P. J. Prabhu, "Effect of Steam Generator Anti-Vibration Bars on U-Bend Velocity Distribution", ANS Transactions, Volume 56, 1988.

I i

I l

i t'

0774T/RAC:wj l

O

- a