ML20207S523

Technical Evaluation Rept for Gessar II Responses for Resolution of Humphrey Safety Concerns (Hsc) Re Mark III Containment Sys

ML20207S523
Person / Time
Site:	Grand Gulf, 05000000, 05000447
Issue date:	11/30/1986
From:	Economos C BROOKHAVEN NATIONAL LABORATORY
To:	NRC
Shared Package
ML20207S501	List: ML20207S500 ML20207S507 ML20207S510 ML20207S515 ML20207S520 ML20207S523
References
CON-FIN-A-3346, CON-FIN-A-3396 NUDOCS 8703190605
Download: ML20207S523 (51)

Text

v j

J',

e Technical Evaluation Report (TER) for the GESSAR II Responses for Resolution of the Humphrey Safety Concerns (HSC)

Relating to the Mark III Containment System by C. Economos November 1986 Department of Nuclear Energy Brookhaven National Laboratory

~Upton, New York 11973 Introduction The basis of this TER is BNL's evaluation of the information supplied via attachments to the letter from Glenn G. Sherwood of the General Electric Co.

to D.G. Eisenhut of the NRC, "In the Matter of 238 Nuclear Island GESSAR II Docket No. STN 50-447 - Containment Design Margins," dated July 26, 1983.

With three exceptions, this TER addresses only those HSCs that, by previous agreement with the NRC staff, are considered BNL's responsibility.

The three exceptions are the addition of_ the issues related to HSC 4.4, 4.5 and 7.1.

In earlier TERs (River Bend Station and Perry Nuclear Power Plant), these had been considered the responsibility of the respective Technical Monitors. Not-withstanding these earlier decisions, we feel that BNL input in these areas is appropriate and have therefore included our evaluation here.

Note that our conclusion in each case is that the issues raised have been satisfactorily re-solved.

8703190605 870306 PDR ADOCK 05000416 E

PDR

• - Humphrey Safety Concerns 1.1, 1.2, 1.4 and 1.5 1.1 Presence of local encroachments, such as the TIP platform, the drywell personnel airlock and the equipment and floor drain sumps may increase the pool swell velocity by as much as 20 percent.

1.2 Local encroachments in the pool may cause the bubble breakthrough height to be higher than expected.

1.4 Piping impact loads may be revised as a result of the higher pool swell velocity.

1.5 Impact loads on HCU floor may be imparted and the HCU modules may fail, which could prevent successful scram if the bubble breakthrough height is raised appreciably by local encroachments.

Evaluation The concern is that the presence of encroachments will tend to increase pool velocities and breakthrough height relative to the unencroached pool thereby causing increases in a variety of LOCA loads including bulk and froth impact and drag loads in regions above the initial pool surface.

As a result of the 1/10-scale simulations of pool swell conducted by the Con-tainment Owners Group (CIOG) as reported in References 1.1.1 and 1.1.4, it has now been established that, for the most part, the effect of encroachments is to decrease pool swell velocities relative to an unencroached or clean sup-pression pool.

BNL has reviewed all of this information and has prepared a separate Technical Evaluation Report on its findings (Reference 1.1.5).

The range of conditions and geometries examined in these tests should cover any variations inherent in GESSAR 11 plants.

Conclusion BNL considers these issues closed for the GESSAR 11 because the 1/10-scale tests have confirmed that encroachment effects on pool swell behaviour will not give rise to loading conditions that cannot be accommodated by the exist-ing design basis.

References 1.1.1 Sets of slides provided by the CIOG to the NRC from the following 1/10-scale tests:

(a) Clinton Tests F2R, R05, and F5; (b) Grand Gulf Tests E3; (c) Perry Test D2; (d) an unidentified clean pool test (see letter of 19 March 1985 from J.E. Torbeck to R. Pender, cc:

J. Kudrick of NRC).

1.1.2

" Comparison of Velocities and Thicknesses of Water Column at Contain-ment Wall for 1/10-Scale Pool Swell Tests," submitted by G.W. Smith of CIOG to H.R. Denton of NRC via letter of 15 May 1985.

l

_.______..._____________..______________________________________J

' 1.1.3 Mintz, S. et al., "Clinton Plant Unique Encroachments Final Test - Re-port," General Electric Report MDE-36-0285, Attachment to IPC Letter No. U-600006 dated April 29, 1985 from F.A. Spangenberg (IPC) to A.

Schwencer (NRC).-

1.1.4 Mintz, S. et al., " Perry Plant Unique Encroachments Final Test Repot GE Report MDE-10-0185.

1.1.5 Sonin, A. and Economos, C., " Resolution of the Humphrey Issues Relating to Pool Swell in Mark III Plants," BNL Technical Evaluation Report, February 1985.

6 I

l r

t

_4 Humphrey Safety Concern 1.3 Additional submerged structure loads may be applied to submerged structures near local encroachments.

Evaluation The loads addressed under this item fall into two categories:

(1) loads on submerged boundaries (drywell wall, basemat, containment wall) and (2) loads on submerged structures proper, such as pipes and beams.

The GESSAR II response indicates that a two-dimensional SOLAV simulation was employed to determine the effect of encroachments on these loads. An increase in the containment wall load was predicted which is within the design load specification.

Submerged structure loads were also found to be less than those currently used for design.

Conclusion BNL considers this concern to be resolved based on the stated margin between design and the loads estimated with a conservative (two dimensional) simula-tion of encroachment effects.

. Humphrey Safety Concern 1.6 Local encroachments on the steam tunnel may cause the pool swell froth to move horizontally and apply lateral loads to the gratings around the HCU floor.

. Evaluation The GESSAR II response states that a conservative potential flow analysis of the flow field through the HCU floor was performed.

This analysis assumed steady flow, i.e., the liquid droplets had velocities equal to air and all of the froth was allowed to pass through the openings in the HCU floor. The re-sulting lateral pressures were found to be 0.45 psid on the gratings with beams experiencing a maximum horizontal force of 1.7 psid.

A structural assessment of all affected structures was performed using these loads.

It is stated that resulting stresses are " minor."

BNL concurs with the claim that the potential flow analysis is conservative.

In fact, independent calculations conducted by the NRC indicate that over 90%

of the froth will continue in the vertical direction, impact on the HCU floor and lose all its velocity.

As the froth begins to fall back toward the pool, the horizontal component of the flowing air. will accelerate the froth to some extent but steady-state conditions are not expected to be attained. Conside r-ing that the additional stresses are modest even with a conservative flow analysis, BNL does not feel that additional effort needs to be expended on this issue.

Conclusion BNL considers this issue to be closed.

l

Humphrey Safety Concern 1.7 GE suggests that at least 1500 square feet of open area should be maintained in the HCU floor.

In order to avoid excessive pressure differentials, at least 1500 square feet of open area should be maintained at each containment elevation.

Evaluation By definition, this concern is irrelevant for GESSAR II.

Nevertheless, a GESSAR 11 response is provided.

It indicates that the free-flow open area at the HCU floor elevation is 1500 square feet and that the open area at other locations in substantially greater (2000 square feet).

Conclusion BNL considers this issue to be closed.

7 Humphrey Safety Concern 2.1 The annular regions between the safety relief valve lines and the drywell wall penetration sleeves may produce condensation oscillation (CO) frequencies near the drywell and containment wall structural resonance frequencies.

Evaluation As stated above, the concern is that additional and unaccounted for suppres-sion pool boundary loads may be produced by steam condensation at the exit of the sleeve annulus. However, the scope of this concern is expanded considera-bly by Humphrey in Reference 2.1.1.

There it is speculated that, due to reso-nant coupling between the sleeve annulus C0 and the sleeve annulus acoustics, the pressure loads at higher frequencies could be amplified.

The GESSAR 11 response is similar to that first supplied by all of the other applicants. Specifically, the same Action Plan items are listed and then ref-erence is made to the GGNS submittal (Reference 2.1.2) as applicable.

The methodology proposed there derives sleeve annulus C0 loads by conservatively scaling down the main vent C0 load data base.

The potential for resonant am-plification of these loads was not addressed.

The use of the main vent C0 data base for development of sleeve annulus loads cannot be rigorously defended because of substantial differences in geometry.

However, based on the relative size of the steam / water interface that would exist at the sleeve annulus, BNL would judge that any additional non-resonant loads that may occur will be second order relative to main vent loads.

This judgement applies to both C0 and chugging loads.

The added sleeve C0 loads that were first proposed are substantial. For exam-ple, the peak-to-peak pressure amplitude was about 20% of that used in the main vent load definition.

They are also applied unifonnly in the circumfer-ential direction which represents a sizable conservatism.

This is because there are roughly twice as many main vents as drywell penetration sleeves.

These modifications are clearly more than second order.

Thus, provided it could have been demonstrated that resonant amplification does not occur, the C0 loads which were specified would have been considered adequate.

In an attempt to demonstrate the absence of such a coupling, results from Gen-eral Electric's 4TC0 tests (Reference 2.1.3) were cited under the Clinton Power Station docket (Reference 2.1.4).

A review of this material indicated that the contrary was the case; that is, the data implied that the type of resonant coupling suggested by Mr. Hunphrey was not only possible but appar-ently had actually occurred.

In fact, it can be inferred from this data that resonance causes about a two-fold increase in the basic C0 loads.

As a result of this finding, a completely new methodology for the C0 boundary loads was proposed, again under the CPS docket (Reference 2.1.4).

This meth-odology has been developed by General Electric utilizing the B.rk I FSTF data base (2.1.5).

This new design loading results in a substantial increase in tbe pressure loading at the higher end of the frequency spectrum (20-50 Hz).

For example, in terms of an amplified response spectrum, the load intensity is about double the one first proposed.

These new loads are shown to be bounded by other design basis loads.

For example, on the drywell wail the sleeve C0 1

load when added to the main vent C0 is bounded by the chugging load specifica-tion. On the containment wall, the bound is provided, with considerable mar-gin, by the pool swell boundary load.

The applicability of the FSTF data to the sleeve annulus loads involves con-siderable uncertainty because of the great disparity in geometries between the two situations.

This applies not only for the steam-water interface at the respective pipe exits but, more importantly, for the acoustic path through which the mechanism that drives the C0 phenomenon is transmitted.

In fact, it is not completely apparent that for the FSTF case the system has actually achieved a condition of resonant coupling.

This is because of the complexity of the FSTF vent system; i.e. the eight downcomers are connected to a vent header, which in turn is connected to a main vent which then connects to a simulated drywell.

Because of this complexity, it is difficult to ascertain the effective relevant vent sntem natural frequency with sufficient preci-sion.

Despite the uncertainties cited above, there are several factors that may be cited that compensate for any possible inadequacy.

First, there is the quali-tative evidence from the 4TCO tests (Reference 2.1.3) that, even when resonant coupling clearly occurs, the load amplification is limited to less than a two-fold factor.

Also, the generally conservative application of the available results provides increased confidence in the adequacy of the load method.

For example, the pressure results observed on 24" diameter downcomers are taken over directly and applied to the much smaller,14" diameter, SRVDL sleeves.

Also, in developing the loads on the suppression pool boundaries, the effec-tive source (steam bubble) radius was taken equdl to that of the sleeve with-out taking into account the actual presence of the SRVDL itself.

As shown in a recent submittal under the CPS docket (Reference 2.1.6), this results in a margin of over 30% in the loads that were developed.

BNL believes the margin would be even greater if the steam bubble were modeled more realistically (Reference 2.1.2).

Additional conservatism stems from the use of conventional acoustics for determination of pressure attenuation from the source to the pool boundaries.

Dissipative mechanisms that are present in the suppression pool and neglected in the analysis, would further reduce the loads.

Insofar as the chugging loads are concerned, BNL has not received a descrip-tion of these, even though this information exists (Reference 2.1.7) and has been assessed by the Mark III Containment Issues Review Panel (Reference 2.1.8).

The findings of this panel were that the proposed loads were only about 6% of main vent chugging and are easily bounded by design. BNL is sat-isfied that this is the case.

Note that resonance effects are not expected to play any role in or influence the chugging phenomenon associated with the sleeve annulus.

Conclusions BNL considers this safety concern to have been satisfactorily resolved because the conservatively estimated new loads have been demonstrated to be bounded by other design loads.

References 2.1.1 Humphrey Engineering, Inc., Letter dated June 17, 1982 from J. M. Humphrey (HEI) to A. Schwencer (NRC).

2.1.2 IPC Letter No. U-0714 dated May 25, 1984 from D. I. Herborn (IPC) to A. Schwencer (NRC).

2.1.3 Bird, P.

F., et al., "4T Condensation Oscillation Test Program Final Test Report", General Electric Report NEDE-24811-P, May 1980.

2.1.4 IPC Letter No. U-600319 dated from F. A. Spangenberg (IPC) to W. R. Butler (NRC).

2.1.5 Fitzsimmons, G. W, et al., " Mark I Containment Program - Full-Scale Test Program Final Report", General Electric Report, NEDE-24539-P, April 1979.

2.1.6 IPC Letter No. U-dated from F. A. Spangenburg (IPC) to W. R. Butler (NRC).

2.1.7 Enercon Letter No. RWE-0G-060 dated May 25, 1983 from R.

W.

Evans (Enercon) to B. R. Patel (Creare R&D).

2.1.8 Mark III Containment Issues Review Panel, " Assessment of Humphrey Con-cerns", CREARE R&D, Inc., Technical Memorandum TM-928, July 1984.

c

--,,n,..,,,.,

Humphrey Safety Concerns 2.2 and 2.3 2.2 The potential condensation oscillation and chugging loads produced through the annular area between the SRVDL and sleeve may apply unac-counted for loads to the SRVDL.

Since the SRVDL is unsupported from the quencher to the inside of the drywell wall, this may result in failure of the line.

2.3 The potential condensation oscillation and chugging loads prode:ad through the annular area between the SPVDL and sleeve may app ly unac-counted for loads to the penetration sleeve. The loads may also be at or near the natural frequency of the sleeve.

Evaluation The concern here is that the steam condensation process (C0 + chugging) at the sleeve annulus exit will give rise to loads on the SRVDL and SRVDL sleeve analogous to the lateral loads experienced by Mark I and Mark II downcomers during postulated LOCA blowdowns and that these structures have not been de-signed to accommodate them.

The response provided in the GESSAR II submitted available to us states that these loads "will be defined" in a later submittal (September 30,1983). This latter submittal has not been made available to BNL as of this writing.

Among the three domestic Mark III plants

• that have responded to this concern (References 2.2.1, 2.2.2 and 2.2.3) all have utilized a common methodology to define the requisite chugging loads.

It derives from the Mark II method of Reference 2.2.4 as modified by the NRC Staff's Acceptance Criteria (Reference 2.2.5).

In each case BNL has found the specification acceptable.

A C0 lateral load is only specified for the GGNS (Reference 2.2.1).

This is because, among the three affected plants, only the GGNS sleeve geometry creates an asymmetric C0 load. As discussed in Reference 2.2.6, the load spe-cification, which we can suppose was derived with the General Electric Co.'s assistance, is found acceptable.

There is no reason to expect that the load specifications cited above would not be suitable for use in GESSAR II plants.

Nothing that we know of the standard plant geometries would invalidate the applicability of these methods.

Conclusion BNL considers these concerns to be resolved contingent on incorporation of the methods used by the GGNS, CPS and RBS within the main body of the GESSAR II documents.

References 2.2.1 to MPSL Letter No. AECM-86/0175 dated August 14, 1986 from 0.D. Kingsley, Jr. (MP&L) to H.R. Denton (NRC).

• The fourth plant (Perry) does not employ a sleeve that extends beyond the dry well wall and, furthermore, seals off the annular area.

r-2.2.2 IPC Letter No. U-0714 dated May 25, 1984 from D.I. Herborn (IPC) to A. Schwencer (NRC).

2.2.3 Enclosure to GSU Letter No. RBS-24,131 dated July 31, 1986 from J.E. Booker (GSU) to H.R. Denton (NRC).

2.2.4 Davis, W.M., " Mark II Main Vent lateral Loads," GE Report NEDE-23806-P, October 1978.

2.2.5 Anderson, C., " Mark II Containment Program Load Evaluation and Accep-tance Criteria," NRC NUREG-0808.

2.2.6 Economos, C., " Technical Evaluation Report for the Mississippi Power

_..__ and Light Company's Grand Gulf Nuclear Station, Docket Nos. 50-416 and 50-417 on SER Outstanding Issue (8) - Mark III Containment System Issues - The Humphrey Safety Concerns," BNL TER, September 1986.

Humphrey Safety Concerns 3.1, 3.3, 3.7 3.1 The design of the STRIDE did not consider vent clearing, condensation os-cillation, and chugging loads which might be produced by the actuation of the RHR heat exchanger relief valves.

3.3 Discharge from the RHR relief valves may produce bubble discharge or other submerged structure loads on equipment in the suppression pool.

3.7 The concerns related to the RHR heat exchanger relief valve discharge lines should also be addressed for all other relief lines that exhaust into the pool.

Evaluation The concern is that, besides the main safety / relief valves (MSRVs), there are a number of other valves that discharge fluids into the suppression pool.

As a result they could produce loads analogous to those associated with MSRV dis-charges and/or LOCA blowdowns through downcomers.

These loads have not been accounted for in plant design.

Even though we have included HSC 3.3 above, the GESSAR II response does not.

Why this is so is not clear.

It represents a radical departure from all other utilities' versions of this Action Plan (No. 6).

Nevertheless, the response that is provided is general enough to cover the general thrust of the concern.

The response consists of an assurance that suitable modifications will be made to the GESSAR II so that all applicants will be required to provide a descrip-tion of all ECCS relief lines that discharge fluids into the suppression pool and to demonstrate that they do not produce loads that cannot be accommodated by the relevant structures. Examples of these modifications are given.

This approach is certainly acceptable to us but, oi v;urse, it implies that each applicants' design will have to be reviewed individually.

One cannot help but wonder why the first S in the acronym GESSAR (which stands for stan-dard) is not rendered inoperative thereby.

However, this is not BNL's prob-lem.

The only exception we would take with the total response is that it appaars to exclude the RCIC turbine exhaust as a potential load producer.

However, here again this is not supposed to be a BNL problem by request of the NRC Technical Monitor.

Conclusion We consider this issue to be closed for those areas of this concern that are l

BNL's responsibility.

_13_

Humphrey Safety Concern 3.2 The STRIDE design provided only 9 inches of submergence above the RHR heat ex-changer relief valve discharge lines at low suppression pool levels.

Evaluation The concern is that because of the relatively small submergence involved, steam condensation may not be complete leading to steam bypass and failure of the pressure suppression system.

The GESSAR 11 submittal is a complete non sequitur in that it simply refers the reader to the submittal provided for HSC 3.1, etc.

We do not find this responsive to the thrust of the concern.

However, the response provided by all of the domestic Mark III plant utilities (see e.g.,

Reference 3.2.1) clearly is applicable for GESSAR II plants and provides a convincing basis for resolution of the issue.

Co; clusion BNL considers this issue closed.

References 3.2.1 HP&L Letter No. AECM-82/353 dated August 19, 1983 from L.F. Dale (MP&L) to H.R. Denton (NRC).

- Humphrey Safety Concern 3.4 The RHR heat exchanger relief valve discharge lines are provided with vacuum breakers to prevent negative pressure in the lines when discharging steam is condensed in the pool.

If the valves experience repeated actuation, the.vac-uum breaker sizing may not be adequate to prevent drawing slugs of water back through the discharge piping.

These slugs of water may apply impact loads to the relief valve or be discharged back into the pool at the next relief valve actuation and apply impact loads to submerged structures.

Evaluation The GESSAR II response here is identical to that provided for HSC 3.1 etc.

In fact thc response simply refers the reader to the response provided for the latter.

Conclusion BNL considers this issue closed.

~

. Humphrey Safety Concern 3.6 If the RHR heat exchanger relief valves discharge steam to the upper levels of the suppression pool following a design basis accident, they will significant-ly aggravate suppression pool temperature stratification.

Evaluation The GESSAR II submittal is, once again, a simple referral to that supplied for HSC 3.1 etc. As in the case of HSC 3.2, we fail to see how this is responsive to the concern.

However, the response pro'rided by the other Mark III utili-ties (see e.g. Reference 3.6.1) clearly applicable to GESSAR 11 plants.

They provide a convincing basis for resolution of this issue.

Conclusion BNL,gonsiders this issue closed.

References 3.6.1 MP&L Letter No. AECM-82/574 dated December 3, 1983 from L.F.

Dale (MP&L) to H.R. Denton (NRC).

[

. Humphrey Safety Concern 4.3 All Mark III analyses presently assume a perfectly mixed uniform suppression pool.

These analyses assume that the temperature of the suction to the RHR heat exchangers is' the same as the bulk pool temperature.

In actuality, the temperature in the lower part of the pool where the suction is located will be as much as 7-1/2 cooler than the bulk pool temperature.

Thus, the heat transfer through the RHR heat exchanger will be less than expected.

Evaluation To complete the statement of this concern, the following should be added;...

"and containment pressure and temperature greater than expected."

Hunphrey's basis for expecting a temperature difference of up to 7-1/2 F is unclear (we assume here that Mr. Humphrey does intend Fahrenheit degrees).

BNL agrees that in the event of a postulated LOCA, the reality will be a ther-mally stratified pool.

However, to decide what the difference between bulk and RHR suction temperature is requires an estimate of the degree of vertical stratification that will occur, together with knowledge of RHR suction eleva-tion.

The first of these requirements was established to the NRC staff's satisfac-tion during its evaluation of the GESSAR II containment loads (Reference 4.3.1).

After a lengthy, detailed, and sometimes heated review process by the various ini.erested parties (Reference 4.3.2), the worst case vertical tempera-ture profile proposed by the General Electric Company for design (Fig. 381-3 of Reference 4.3.3) was judged acceptable.

The basis for this judgement is given in Reference 4.3.1.

It implies that the profile is applicable only for a standard top vent submergence (-7.5 feet).

In responding to Humphrey concerns 4.7 and 4.10 (see later), it is indicated that the GESSAR II RHR suction is located at an elevation 5'-6" above the basemat.

Comparison with the temperature profile referred to above implies that the RHR suction temperature is essentially equal to the bulk tempera-ture.

Accordingly, this Humphrey concern is not relevant for GESSAR II plants.

Conclusion BNL considers this issue satisfactorily resolved for the GESSAR II.

References 4.3.1

" Mark III LOCA-Related Hydrodynamic load De fi ni ti on," NUREG-0978, August 1984 4.3.2 Transcript of the ACRS Subcommittee on Fluid Hydraulic Dynamic Effects Meeting of September 24, 25, 1981.

4.3.3 General Electric Co., 22A707, " General Electric Standard Safety Analy-sis Report," (GESSAR-II), Appendix 3B through Amendment 1, February 25, 1982.

. Humphrey Safety Concerns 4.4 and 7.1 4.4 The long term analysis of containment pressure / temperature response assumes that the wetwell airspace is in thermal equilibrium with the sup-pression pool water at all times.

The calculated bulk pool temperature is used to determine the airspace temperature.

If pool thermal stratifi-cation were considered, the surface temperature, which is in direct con-tact with the airspace, would be higher. Therefore, the airspace temper-ature (and pressure) would be higher.

7.1 The containment is assumed to be in thermal equilibrium with a perfectly mixed, uniform temperature suppression pool. As noted under Topic 4, the surface temperature of the pool will be higher than the bulk pool temper-ature.

This may produce higher than expected containment temperatures and pressures.

Evaluation The concern is similar to that associated with HSC 4.3 above except that here the issue is the difference between pool surface temperature and pool bulk temperature. Based on tne GESSAR-II temperature profile referred to previous-ly, this difference is 8'F, in rough agreement with the 7-1/2*F difference cited by Mr. Humphrey in HSC 4.2.

Apparently, this was the AT he was refer-ring to and it was mistakenly cited in connection with the Bulk-to-RHR suction temperature difference.

The GESSAR 11 response states that the GGNS submittal relative to this concern is applicable.

In this submittal (Reference 4.4.1), the issue is quantified by means of existing information and analyses.

The results show that the effects of a 7 to 8*F difference between pool surface and bulk temperature would imply an increase in peak containment pressure and temperature of only 0.1 psi and 3"F, respectively.

These modest differences are overwhelmed by the existing margins of 5 psi and 19*F that can be demonstrated to exist due to various conservatisms used in conventional containment response analysis.

Conclusion Essentially, what this entire exercise has demonstrated once more is that the use of a mean or bulk pool temperature is an acceptable simplification which facilitates calculation of containment response.

The result is not surpris-ing. BNL considers this issue to be closed.

References 4.4.1 Attachment to MP&L Letter No. AECM-82/353 dated August 19, 1982, from L.F. Dale (MP&L) to H.R. Denton (NRC).

Humphrey Safety Concern 4.5 A number of factors may aggravate suppression pool thermal stratification.

The chugging produced through the first row of horizontal vents will not produce any mixing from the suppression pool layers below the vent row.

An upper pool dump may contribute to additional suppression pool temperature stratification.

The large volume of water from the upper pool further sub-merges RHR heat exchanger effluent discharge which will decrease mixing of the hotter, upper regions of the pool.

Finally, operation of the containment spray eliminates the heat exchanger effluent discharge jet which contributes to mixing.

Evaluation The GESSAR 11 response states that the GGNS submittal relating to this concern is applicable.

This submittal is referred to in Reference 4.5.1 as Action Plan.14.

This Action Plan was to utilize the following " Program for Resolu-tion."

1.

Testing information will be submitted to demonstrate the effectiveness of chugging as a mixing mechanism in the suppression pool.

2.

Analyses will be submitted to demonstrate that the suppression pool does not experience significant stratification during containment spray or fol-lowing upper pool dump.

The first of these items addresses the concern that chugging does not provide any mixing below the top vent.

The test information that was supplied indi-cates that this is not correct.

Measurements obtained from so-called drag disks that had been installed in the PSTF facility indicated that flow rever-sals occur periodically in both the middle and bottom vents during the chug-ging phase of the steam blowdown.

Although this is a qualitatively useful finding, the attempt to quantify pool turnover time from this information can-not be taken seriously.

This is because the drag disk device requires careful calibration under even the best of circumstances (steady, uniform fl ow).

Under the unsteady, highly non-uniform flow conditions that prevail within the vents during chugging, the notion that quantitatively correct values of flow velocities can be obtained using this procedure is not credible.

Furthermore, even if one were to accept these quantitative estimates, their applicability for the case involving upper pool dump, which can increase top vent submer-gence from 7-1/2 to as much as 12 feet, would be highly suspect.

As for the second item, no analytical information has been supplied by MP&L, nor, for that matter, any of the other affected Mark III plants.

Accordingly, we took it upon ourselves to develop a bounding scenario using information developed from the applicants' FSARs and other sources (Reference 4.5.2).

De-tails of these analyses may be found in Reference 4.5.3 for the GGNS and Ref-erence 4.5.4 for the CPS.*

The conclusion in both cases was that upper pool dump will not seriously increase pool thermal stratification in these two plants. The primary reason for this finding was the high elevation of the RHR

• A bounding analysis for the Perry Nuclear Power Station was not carried out at the request of the NRC Technical Monitor.

The RBS does not have an upper pool so that this concern does not apply.

- suction relative to the worst case temperature profile (see discussion in HSC 4.2).

These elevations were 8 and 10-1/2 feet above the basemat for the CPS and the GGNS, respectively. The relatively high temperature of the upper pool when the dump occurs (about 125 F) also contributed to resolution of this issue for these two plants.

We attempted to resolve this concern for GESSAR 11 plants using the same bounding scenario.

Unfortunately, the relevant parameters for GESSAR II, as obtained from Reference 4.5.5, were such that resolution could not be demon-strated.

Specifically, the RHR suction elevation (5.5 feet), increase in sub-mergence due to pool dump (6.4 feet *), and upper and suppression pool tempera-tures at dump (100*T and 150*F, respectively) gave the result that the RHR suction temperature was almost 40*F less than the post dump suppression pool bulk temperature (138 F).

The overall consequence of a deficiency of this magnitude in terms of containment response is unknown to us but its existence precludes our writing off the concern.

Also, the increase in thermal strati-fication in terms of pool surface-to-bulk temperature difference ( ATSB) is substantial, from 8*F to about 20*F.

In fact, this 20*F difference exceeds the margin of 19 F that was demonstrated to exist due to various conservatisms used in conventional containment response analysis (see GGNS response to HSC 4.4 in Reference 4.5.1).

Application of a bounding or " worst case" scenario to resolve safety con-

~

cerns is a conventional procedure that has been used successfully many times.

However, when it fails to fulfill this objective a complete reexamination of the elements that make up the scenario is needed, particularly those that introduce significant conservatisms.

The most obvious of these conservatisms is the assumption that the upper pool water sinks to a hydrostatically stable level consistent with its initial temperature and the existing suppression pool thermal profile without any mix-ing whatsoever with the suppression pool water.

We are the first to acknowl-edge that this is brutally conservative and would surely like to relax this assumption.

Unfortunately, insofar as we know, there is no technical basis for quantifying the degree of miWg that may occur.

Another adjustment that can be considered is the temperature of the upper pool water at dump.

Con-trary to what would be expected intuitively, the safety concern involved here would be ameliorated by an increase in this parameter.

This is because the RHR suction temperature would also increase, thereby reducing the differential relative to bulk temperature ( ATRB).

Dumping at a higher temperature would also improve the situation insofar as ATSB is concerned.

We will quantify these stated effects in a moment. Another candidate for adjustment is the RHR suction elevation.

As we noted before, for the GESSAR 11 design this param-eter is considerably smaller than in the GGNS and CPS plants.

Here again, an increase would improve the situation for ATRB since higher temperatures pre-vail at the higher elevations; there is no effect on ATSB however.

Finally, we can contemplate reducing the total makeup volume that is dumped. Reducing this amount affects both ATRB and ATSB favorably.

To understand why this is so one needs only to recognize that reducing the makeup volume to zero (as in

• We highlight here the discrepancy between this value and the 4.5 feet implied h Mr. John Humphrey via HSC 19.1.

We have no intention of following up on this unless specifically directed to do so by the NRC staff.

the RBS case) precludes the potential for " aggravating" pool thermal stratifi-cation.

In Table 4.5.1 we provide some quantitative indication of these effects.

Of particular interest is the efficacy of increasing the RHR suction eleva-tions in tenns of reducing themal stratification insofar as ATRB is con-cerned.

Note that in Case C ATRB can be zeroed out by increasing thisparam-eter from the base case value of 5.5 feet (Case A) to 9.4 feet.

Whether or not this is a credible approach is not for us to say.

We note however, once again, that a value of 10.5 feet is used in the GGNS.

Our selection of the other parameter changes were based on the following:

for upper pool tempera-tures the value of 125'F (Cases D, E and F) corresponds to the " maximum makeup pool temperature" specified in the GGNS FSAR (Table 6.2-50); the reduction in upper pool makeup volume from 32% to 25% of the suppression pool volume (Cases G and H) also makes this parameter roughly consistent with that used by the GGNS plant (also Table 6.2-50 of the FSAR).

Conclusion If the relevant parameters in the GESSAR II design are as given, some potential for serious increase in suppression pool thermal stratification exists.

This potential can be reduced significantly by moderate and, in our judgement, reasonable modification of these parameters.

References 4.5.1 MP&L Letter No. AECM-82/353 dated August 19, 1982, from L.F.

Dale (MP&L) to H.R. Denton (NRC).

4.5.2 Gunter, A.D. and Fuls, G.M., "Clasix-3 Containment Response Sensitivity Analysis for the Mississippi Power & Light Grand Gulf Nuclear Station,"

Offshore Power Systems Report No. OPS-37A15, December 1981.

4.5.3 Economos, C., " Technical Evaluation Report for the Illinois Power Com-pany's Clinton Power Station, Docket No. 50-461 on SER Outstanding Issue (8) - Mark III Containment System Issues - The Humphrey Safety Concerns," BNL TER, January 1986.

4.5.4 Economos, C., " Technical Evaluation Report for the Mississippi Power and Light Company's Grand Gulf Nuclear Station, Docket Nos. 50-416 and 50-417 on SER Outstanding Issue (8)

Mark III Containment System Issues - The Humphrey Safety Concerns," BNL TER, September 1986 4.5.5 GESSAR II Nuclear Island Design, General Electric Co. Document 22A7007.

~.

Table 4.5.1 INPUT RESULTS Upper Pool Makeup Volume Suppression Pool Upper Pool

(% Suppression RHR Suction Post Dump Pool Thermal Stratification Case Temperature (*F) Temperature (*F)

Pool Volume)

Elevation (feet)

Temperature (*F)

ATRB (*F)

ATSB (*F) i l

A 150 100 32 5.5 138

-38

+21 B

8.0

-16 l

C 9.4 0

4 D

125 5.5 144

-19

+15 E

8.0 F

10.6 0

h l

G 25 5.5 140

-15

+19 H

8.0

-3 ATRB = RHR suction-to-bulk temperature differential.

ATSB = Pool surface-to-bulk temperature differential.

I, i

i 4

t 7

e

p

- HumphreySafetyConcerns4.7and4.lb' 4.7 All analyses completed for the Mark III are generic in nature and do not consider plant specific interactions of the RHR suppression pool suction and discharge.

4.10 Justify that the current arrangement of the discharge and suction points of the pool cooling system maximizes pool mixing.

Evaluation The concern here is that if the RHR system's geometric arrangement for the suction and return lines is not properly designed, the capability of the sys-tem to induce bulk mixing and remove thermal energy will be degraded.

The GESSAR II response is identical to the generic approach developed for the Mark III Containment Issues Group (Reference 4.7.1).

The key element of this study was the Perry one-tenth-scale tests (Reference 4.7.2).

In these tests, a number of related concerns were addressed systematically.

These included short circuitinq, development of bulk pool motion, ability to eliminate ther-mal stratificaaon and the presence of isolated recirculation zones.

BNL has reviewed this information in detail and has concluded that the Perry one-tenth-scale tests correctly simulate design conditions in GESSAR 11 plants, and, alcng with the other test information provided, include a suffi-cient range of parameters to insure their applicability.

Since the findings from these tests show that good mixing can be achieved, as well as the absence of dort circuiting BNL concludes that the GESSAR II RHR systems can be ex-pected to perform in a manner consistent with design assumptions.

Conclusion BNL considers the issues raised by these concerns to be satisfactorily re-solved for GESSAR II Mark III plants.

References 4.7.1 Quadrex Corp., "A Survey of Tests and Analyses on the Effectiveness of the RHR System in the Pool Cooling ttode," Report No. QUAD-1-82-245.

4.7.2 Gilbert Associates, Inc., "Model Study of Perry Nuclear Power Plant Suppression Pool - Final Report," November 1977.

. Humphrey Safety Concern 19.1 The chugging loads were originally defined on the basis of 7.5 feet of submer-gence over the drywell to suppression pool vents.

Following an upper pool dump, the submergence will actually be 12 feet which may affect chugging loads.

Evaluation The GESSAR II response states that the GGNS submittal relative to this concern is applicable.

In this submittal (Reference 19.1.1), physical arguments and analytical procedures are used to estimate the pressure field that would be generated on the suppression pool boundaries if the worst case chug from the Mark III data base were to occur with the top vent at a 12-foot submergence.

The results are compared with design on an ARS basis and shown to be bounded except for local loads in the frequency range 15 to 32 Hz.

For these condi-tions an exceadance of design amounting to 35% occurs on the basemat and about 15% on the containment wall.

The GGNS utility argued that this exceedance is not important because this is a local load affecting only the basemat and containment wall liners and that, because of the hydrostatic head to which the liner is :ubjected, "a negative pressure will never be imposed on the liner."

Also, "since the liner is backed by concrete, no natural modes of vibration are excitable."

Without passing judgement on the merits of these arguments, BNL notes the fol-lowing:

the use of an acoustic model in the analysis represents a significant conservatism; dissipative mechanisms not accounted for in such an analysis re-sult in pressure attenuation which is much greater than predicted; this has been borne out convincingly by experimental results: application of the worst case chug to all vents, which is done for local loads, also represents a very significant conservatism; in a more recent submittal by the Cleveland Electric Illuminating Co. to address the concern relating to the combined effect of upper pool dump and encroachment on local chugging loads (Reference 19.1.2),

it was shown that by postulating a maximum strength chug at the central vent 4

and average strength chug at adjacent vents, the design loads were capable of bounding the combined effect.

In summary, the margins inherent in the design load for chegging are very large. They can more than accommodate any increment in loading caused by cff-design effects such as increased submergence due to upper pool dump as well as geometric differences between GESSAR II plants and the GGNS.

Conclusion BNL considers this issue satisfactorily resolved for the GESSAR Il plants.

References 19.1.1 MP&L Letter No. AECM-82/353 dated August 19, 1982 f rom L.F.

Dal e,

MP&L, to H.R. Denton, NRC.

19.1.2 CEI Letter dated July 11, 1984 from M.R.

Edelman, CEI, to B.J.

Youngblood, NRC.

1

_24 HumdhreySafetyConcern19.2 E

_The effect of local encroachments on chugging loads needs to be addressed.

- Evaluation The GESSAR 11 response states that the GGNS submittal relative to this concern is applicable.

In this submittal (Reference 19.2.1), physical arguments and

- analytical procedures are used to estimate the pressure field that would be 9enerated on the suppression pool boundaries if the worst case chug from the

- Mark III data base were to occur at vents located below the GGNS TIP plat-fo rm.

The results are compared with design on an ARS basis and shown to be bounded except for local loads in the frequency range 12 to 30 Hz. For these conditions an exceedance-of design amounting to 60% occurs on the basemat and about 15% on;the containment wall.

l The GGNS utility argues that this exceedance is not important because this is a local load affecting only the basemat and containment wall liners.and that,

.because of the hydrostatic head to which the liner is subjected, it will not

~

experience a " negative pressure in the frequency range of exceedance."

Also, "since the liner is backed byLconcrete everywhere, no natural modes in this range are excitable."

1

~

Without passing judgement on the merits of these arguments, BNL notes the fol-lowing:.- the use of an acoustic model in the analysis represents a significant conservatism; dissipative mechanisms not accounted for in such an analysis re-sult in pressure attenuation which is'much greater than predicted; this has 4

been' borne out convincingly by experimental results: application of the worst case chug to all vents' below the encroachment also represents a very signifi-l cant conservatism;- in a recent submittal :by CEI to address the staff concern relating to the combined effect of upper pool dump and encroachment on local i

chugging loads (Reference 19.2.2), it was shown'that by postulating a maximum strength chug at the central vent and average strength chug at adjacent vents, the design loads were capable of bounding the combined effect.

1 In summary, the margins inherent in the design load for chugging are very large. They can more than accommodate any increment in loading caused by off-e design effects such as increased submergence due to upper pool dump as well as geometric differences between GESSAR 11 plants and the GGNS.

Conclusion BNL ~is satisfied that the issues related to this concern have been satisfacto-rily addressed'and are therefore considered closed.

References 19.2.1.MPAL Letter No. AECM-82/574 dated December 3,1983, from L.F. Dale (MP&L) to 14.R. Denton (NRC).

19.2.2 CEI Letter dated July 11, 1984 from M.R.

Edelman (CEI) to B.J.

l Youngblood (NRC).

l 4

a 1

-. ~.

'- Additional Safety Concern BNLl*

The effects of increased submergence and encroachment on local loads are addi-tive. ARS comparisons showing the combined effect should be provided.

Evaluation The GESSAR II response does not address this concern.

However, a generic re-sponse was developed for this concern by the Containment Issues Owners Group and included in a Cleveland Electric Illuminating Co. (CEI) submittal (Refer-ence B.1.1).

As indicated in the evaluation provided for HSC-19.1 and 19.2 above, by postulating the very reasonable assumption that a maximum strength l

chug occurred only at a central vent with average strength chugs at surround-ing vents, local loads were reduced to levels which were adequately bounded by design.

Conclusion The utilities have provided an adequate demonstration that the design load for chugging has sufficient margin to accommodate the increment in loading caused by the combined effects of encroachment and increased submergence.

BNL con-siders this issue to be closed. for all Mark III plants including GESSAR II pl ants.

References B.1.1 CEI Letter No. PY-CEI/NRR-0123L dated July 11, 1984 from M.R. Edelman (CEI) to B.J. Youngblood (NRC).

l

• This additional concern, which evolved.from HSC19 and was developed by BNL, is not formally recognized as a Humphrey concern.

It is therefore denoted by this special notation.

r

' Additional Safety Concern BNL3*

Upper pool dump increases the length of the water column within the main steam SRVOL.

This will tend to increase pipe thrust loads during SRV actuation.

Provide an analysis of this effect.

Evaluation The GESSAR 11 response does not address this concern.

However, it has been addressed by each of the utilities that are affected (CPS, GGNS, PNPS) and, in each case, the piping systems have been found to be capable of accommodating the increased loads. There is no reason to suppose that the GESSAR II design would be any less adequate.

Conclusion BNL considers this issue to be closed.

• This additional concern, which evolved from HSC19 and was developed by BNL is not formally recognized as a Humphrey concern.

It is therefore denoted by this special notation.

i'

.\\

o DRAFT SUPPLEMENT 2 TO THE SAFETY EVALUATION REPORT PERRY NUCLEAR POWER PLANT, UNIT 1, D0CKET NOS. 50-440/441 6.2.1.8 LOCA-Re1ated Pool Dynamics

Background

The Mark III pool dynamic loads were reviewed at the CP stage for Grand Gulf Nu-clear Station, Units 1 & 2 (Grand Gulf 1/2) and at the PDA stage for GESSAR-238NI.

The staff concluded at the time that the information available was suf-ficient to adequately define the pool dynamic loads for nuclear plants applying for Construction Permits.

Since the-issuance of the GESSAR-238NI Safety Evalua-tion Report (SER) (December 1975), GE has conducted further tests and analyses to confirm and refine the original load definitions.

To keep the NRC and Mark III applicants apprised of the current status of these tests, GE issued an In-terim Containment Loads Report (22A4365) in April of 1978 and revised this re-port several times before GESSAR-II was provided to the staff on March of 1980, which is GE's FDA submittal for their standard B0P design and is to be refer-enced by Mark III OL applicants.

Appendix 3B of GESSAR-II provides the final-ized pool dynamic load definition for Mark III containments, and is the basic document used for review by the staff and its consultants.

The Perry applicant has incorporated the essential features of Appendix 3B of GESSAR-II in various sections of their FSAR submittal.

Except as noted below, the applicant has adhered to all analytical techniques, assumptions, methodolo-gies and concepts contained in Appendix 3B of GESSAR-II.

Where Perry unique parameters differ. from those of-the GE -Standard plant, Perry parameters are used.

The staff has reviewed GE's pool dynamic load definitions to arrive at a finalized hydrodynamic load definition that can be utilized by Mark III contain-ment applicants for operating licenses.

The pool dynamic loads were reviewed under Task Action Plan (TAP) B-10 " Behavior of BWR Mark III Containment".

The results of this review are documented in NUREG-0978 " Mark III LOCA-Related Hydrodynamic Load Definition".

They are applicable to the Perry Nuclear Power Plant.

Description of Phenomena Figure 6-3 shows the sequence of events occurring during a DBA and the potential loading conditions associated with these events.

Following a postulated loss-of-coolant accident (LOCA), the drywell pressure increases due to blowdown of the reactor system.

Pressurization of the drywell causes the water initially standing in the vent system to be accelerated into the pool and the vents are cleared of water.

During this vent clearing process, the water leaving the horizontal vents forms jets in the suppression pool and causes water jet im-pingement loads on the structures within the suppression pool and on the con-tainment wall opposite the vents.

During the vent clearing transient, the drywell is subjected to a pressure differential and the weir wall experiences a vent clearing reaction force.

Immediately following vent clearing, an air and steam bubble forms at the exit of the vents.

The bubble pressure initially is assumed equal to the current.

drywell pressure.

This bubble theoretically transmits a pressure wave throus,.

the suppression pool water and results in loading on the suppression pool bound-aries and on equipment located in the suppression pool.

As air and steam flow from the drywell becomes established in the vent system, the initial vent exit bubble expands to suppression pool hydrostatic pressure.

GE Large-Scale Pres-sure Suppression Test Facility (PSTF) tests show that the steam fraction of the flow is condensed but continued injection of drywell air and expansion of the air bubble results in a rise in the suppression pool surface.

During the early stages of this process, the pool swells in a bulk mode (i.e., a slug of solid water is accelerated upward by the air).

During this phase of pool swell, structures close to the pool surface will experience loads as the rising pool surface impacts the lower surface of th,. structure.

In addition to these ini-tial impact loads, these same structures will experience drag loads as water flows past them.

Equipment in the suppression pool will also experience drag loads.

After the pool surface has risen approximately 18 feet above the initial pool surface, the water ligament thickness has decreased to two feet or less and the impact loads are significantly reduced.

This phase is referred to as incipient breakthrough (i.e., the ligament begins to break up).

To account for possible nonconservatisms in the test facility arrangement, the staff has determined that the breakthrough height should be set at 18 feet above the initial pool surface.

Ligament thickness continues to decrease until complete breakthrough is reached and the air bubble can vent to the containment free space.

The breakthrough process results in formation of an air / water froth and, for load definition pur-poses, is defined to happen at 19 feet over the initial surface. The incipient breakthrough height and the height at which froth loads begin have been set higher than predictions from test results for conservatism. Continued injection of drywell air into the suppression pool results in a period of froth pool swell.

This froth swell impinges on structures it encounters but the two-phase nature of the fluid results in loads that are much less than the impact loads associated with bulk pool swell.

When the froth reaches the elevation of the floors on which the hydraulic con-trol units for the control rod drives are located (approximately 22 feet above pool level), the froth encounters a flow restriction; at this elevation, there is approximately 25% open area.

The froth pool swell experiences a two-phase pressure drop as it is forced to flow through the available open areas.

This pressure differential represents a load on both the floor structures themselves and on the adjacent containment and drywell.

The result is a discontinuous pressure loading at this elevation.

As drywell air flow through the horizontal vent system decreases and the air /

water suppression pool mixture experiences gravity-induced phase separation, pool upward movement stops and the fall-back process starts.

During this pro-cess, floors anr1 other flat structures experience downward loading and the con-tainment wall theoretically can be subjected to a small pressure increase. How-ever, this pressure increase has not been observed experimentally.

The pool-swell transient associated with drywell air venting to the pool typi-cally lasts three to five seconds.

Following this, there is a long period of high steam flow rate through the vent system; data indicates that this steam will be entirely condensed in a region right at the vent exits.

For the DBA reactor blowdown, steam condensation lasts for a period of approximately one minute.

Potential structural loadings during the steam condensation phase of the accident have been observed, and are included in the containment loading specification.

As the reactor blowdown proceeds, the primary system is depleted of high-energy fluid inventory and the steam flow rate to the vent system decreases. This re-duced steam flow rate leads to a reduction in the drywell/ containment pressure differential which in turn results in ; sequential recovering of the horizontal vents. Suppression pool recovery of a particular vent row occurs when the vent stagnation differential pressure corresponds to the suppression pool hydrostatic pressure at the row of vents.

Toward the end of the reactor blowdown, the top row of vents is capable of con-densing the reduced blowdown flow and the two lower rows will be totally recov-ered. As the blowdown steam flow decreases to very low values, the water in the top row of vents starts to oscillate back and forth causing what has become known as vent chugging.

This action results in dynamic loads on the top vents and on the weir wall opposite the upper row of vents. In addition, an oscilla-tory pressure loading ccndition can occur on the drywell and containment.

Since this. phenomenon is steam mass-flux dependent (the chugging threshold appears to 2

be in the range of 10 lb/sec/ft ), it is present for all break sizes.

For smaller breaks, it is the only mcde of condensation that the vent system will experience.

Shortly after a DBA, the Emergency Core Cooling System (ECCS) pumps automatical-ly start up and pump condensate water and/or suppression pool water into the re-actor pressure vessel.

This water floods the reactor core and then starts to cascade into the drywell from the break (the time at >Aich this occurs depends upon break size and location).

Because the drywell is full of steam at the time of vessel flooding, the sudden introduction of cool water causes rapid steam condensation and drywell depressurization.

When the drywell pressure falls be-low the containment pressure, the drywell vacuum relief system is activated and air from the containment enters the drywell.

Eventually sufficient air returns to equalize the drywell and containment pressures; however, during this drywell depressurization transient, there is a period of negative pressure on the dry-well structure.

A conservative negative-load condition is therefore specified for drywell design. Theoretically, this condition can lead to the flow of water over the weir wall.

Structures located in the vicinity could then experience drag and impact loads.

Load definitions to account for these effects are included.

Small breaks, defined as breaks not large enough to automatically depressurize the reactor, do not result in bounding pool dynamic loads except for the thermal loading conditions on the drywell and weir walls. Thermal gradient load defini-tions are provided for the design of the walls containing the suppression pool.

Pool Dynamic Load Assessment (a)

Generic Load Definition - The staf f's review of the generic LOCA-related pool dynamic load definition was completed early in 1984.

The results of 4

this review and the staff's evaluation of the pool dynamic load definition are documented in NUREG-0978 " Mark III LOCA-Related Hydrodynamic Load Definition", which was published in August of 1984.

With only a few -..

exceptions, the staff found the load definition proposed by the General Electric Company in Appendix 3B of GESSAR-II to be acceptable.

A set of Acceptance Criteria was developed by the staff to cover those areas where the proposed loads were not satisfactory.

These were included as Appen-dix C of NUREG-0987. A brief description of these Acceptance Criteria is provided below.

(i)

Pool Swell Velocity - Pool swell velocity controls impact and drag loads on the structures between the pool rest elevation and the breakthrough elevation. GESSAR-II proposes a value of 40 fps at all elevations.

The staff requires use of an elevation dependent value which varies continu-ously from zero up to a maximum of 50 fps at elevations greater than or equal to 10 feet above rest elevation.

(ii) Pcol Swell Loads on Structures Attached to the Containment Walls - The GESSAR-II specification corresponds to steady state drag at a fixed ve-locity of 40 fps.

The staff's Acceptance Criteria require this to be modified to reflect the change in pool swell velocity given in (1) above and the inclusion of impact type forces when the structure is not im-mersed prior to pool swell.

A detailed procedure for evaluating the im-pact load is provided in the Acceptance Criteria.

(iii) Bulk Impact on Small Structures - The GESSAR-II methodology was found ac-ceptable provided the structures involved satisfied certain limitations related to structural natural frequency, size and location above the pool. The Acceptance Criteria require that when any of these limitations are not satisfied, the load specification be reviewed by the staff on a plant unique basis.

(iv) Froth Impact Loads - The GESSAR-II methodology was found to be unaccept-able.

An acceptable alternative was developed by the staff and its con-sultants and is described in detail in the Acceptance Criteria.

The new method differs from the GESSAR-II approach with respect to maximum froth impact pressure, temporal characteristics of the forcing functions and region of application.

(v) Drag Loads - The GESSAR-II methods are found acceptable provided that they are modified to account for the change in pool velocity given in (1) above and provided that they correctly account for the structure-wall in-teraction effect on drag loads.

(vi) Loads on Submerged Structures - The GESSAR-II methods are acceptable ex-cept for computation of acceleration loads on non-cylindrical structures and the evaluation of standard drag during the C0 phase of the LOCA. The staff requires that the Mark I Acceptance Criteria as set down in NUREG-0661 be used to develop these loads.

(vii)

Impact Loads on Structures Above the Weir Annulus - The GESSAR-II methods were found to be acceptable except for radial structures located within one foot of the top of the weir wall and all structures located between zero and 0.25 feet above the weir wall.

Detailed procedures for evalua-tion of the impact loads in these cases are provided in the Acceptance Criteria. _ _ _ - _ - _ - _ _ _ _ - _ _ _ - _ _ _ _ - _ _

u 1

(b)

Perry Nuclear Power Plant Plant Unique load Evaluation -

-(i)

Applicability of Generic Load Definition - The staff has examined the information supplied in the FSAR and has concluded that the Generic Load definition described in GESSAR-II as modified by the staff's Acceptance Criteria described _in NUREG-0978 is applicable to the Perry plant except as _noted below.

All major structures and components that would experience LOCA-related pool dynamic loads are within the range of appli-cability of the staff approved methodology in terms of geometry and rela-tive location =in the containment and the suppression pool.

The major features of the suppression pool geometry (main -vent submergence 'and ver-tical spacing, pool radial width and pool depth) differ slightly from the standard. plant dimensions but these differences are not considered sig-nificant in terms of their effect on pool dynamic loads.

The use of the generic methodology _ by the Perry applicant to develop the LOCA-related 4

pool dynamic load definition is therefore acceptable to the staff.

l (ii) Plant Unique Load Definition (11-1) ' Containment Load Due to Pool-Swell at the HCU Floor (Wetwell Pressuriza-tion)

This item corresponds to Section 38.6.1.6 of Appendix 38 of the Perry FSAR.

The GESSAR 11 methodology requires a ramp-type pressure loading on the containment when froth flows past the HCU floor.

The overall j

pressure differential that is specified is 11 psi as shown in Figure 38-57 of GESSAR II.

The applicant states that a somewhat smaller value of pressure differential ( Ap = 10 psi) was employed for design.

The applicant states further that this reduction is justified by plant i

unique differences in the containment geometry. Specifically, the Perry F

HCU floor is seven feet higher than the GESSAR II standard and the open area at the HCU floor is also greater than standard.

We agree with the applicant that these plant unique differences will tend to decrease the pressure loading, 'particularly the increased ele-i i

vation of the HCU floor.

Based on this consideration, we find the re-duced design load of 10 psi to be acceptable.

(11-2) Drywell Bubble Pressure and Drag Loads Due to Pool Swell (Section 1

38.8.1.2 of Appendix 38) l The applicant's method differs from that recommended by GESSAR II in i

that the method of images procedure for calculation of pressure differ-3 ential is replaced by direct application of the LOCA bubble pressure.

Since this results in a larger pressure differential, the proposed method is conservative and therefore acceptable.

(11-3) Condensation Loads (Section 38.8.1.4)

The applicant does not use the methods of GESSAR II for p.ipes and RHR strainers.

Instead, the applicant claims that the loads on these struc-tures are bounded by the LOCA air bubble loads and the SRV second pop actuation loads, respectively, for all relevant load combinations.

Using the information available in GESSAR II, we have independently

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

~

confirmed that this type of bounding does occur. Accordingly, we find this proposed approach to be acceptable.

(11-4) Chugging (Section 3B.8.1.5)

The comments provided for Item (11-3) above are applicable here.

(11-5)

Impact 1. cads (Section 3B.10.1)

The applicant indicates 1 that the GESSAR-II methods as modifled by the 2

Draft Acceptance Criteria were the design basis for a certain class of structures which are af fected by the specification. This modifica-tion required that, for structures less than four feet long, the pulse duration be reduced by a factor x/4 where x is the actual length of the structures; also, for structures closer than six feet above the pool surface, the pulse duration was to be reduced by a factor y/6 where y is the actual elevation of the structure above the pool.

Since the Draft Criteria have been superseded by the requirements of Appendix C of NUREG-0978, the applicant is developing a new evaluation using a methodology that will be acceptable to the staff. The intent is to demonstrate that, for relevant structures, the new methods imply loads that are less than those originally used for design, thereby verifying its adequacy.

l A description of these new methods was presented in Reference 3.

The im-pact loading for each component is to be determined as follows:

a.

Pulse Duration The im pul se duration, T will be determined using the Maise" pactor the Mark II criteria,(sec),

whichever is greater, b.

Impact Pressure The impact pressure will be calculawd using the hydrodynamic mass of the target which is a function of the target shape (flat or cylindri-cal) and orientation above the pool (radial or circumferential). The hydrodynamic mass is determined from NEDE-13426P,6 Figures 6-8 and 6-9 for radial and circumferential orientation, respectively.

The impact velocity as a function of height above the pool is deter-mined using Equation C-1 of the Mark III Acceptance Criteria (NUREG-0978).

V = SH (2.6 - 0.506 /H)

H < 10 ft.

V = 50 ft/sec.

H ~> 10 ft.

The total impulse will be calculated using equation C-4 of NUREG-0978..

I

= (Mg/A)

• V/(32.2 x 144) = impulse per unit area, psi-sec.

p 2

Mg/A = hydrodynamic mass per unit area, Ibm /ft V

= impact velocity, ft/sec Using the pulse duration (t) the impact pressure is detarmined assuming an isosceles triangle loading; P = 2

• I /T(psi) p The above impact load and pulse duration will be used to evaluate each compo-nent. This evaluation involves the following additional considerations.

1.

A dynamic load factor will be evaluated based on the natural frequency of the component.

In most cases the DLF used in the original design was the theoretical maximum.

The original design load (P*DLF) orig. will be compared to the revised impact load, (P*0LF) rev., and found acceptable if, for Radial Orienta-tion, Porig/ Prev 11.25 and if for Circumferential Orientation, Porig/ Prev 1 2.00.

2.

For components other than structural beams originally qualified using a dynamic load evaluation, either method I will be used for re-evaluation or the dynamic evaluation will be redone using the revised load and a radial Mg/A.

3.

Components that were originally designed to a lower load will be re-evaluated and qualified based on sufficient margin in the original design.

4.

For structural beams originally qualified using a dynamic load evaluation, a reevaluation will be performed using the Maise criteria.

We have reviewed this methodology and find that it conforms in almost all.

respects to procedura'. steps previously approved by the NRC staff in NUREG-0487 and NUREG-0661 (essentially the Maise criteria of Reference 4).

Two areas of _ nonconformance have been identified, however. These are the use of a triangular impulse to represent the impact load and determination of hydro-dynamic mass Mg, for circumferential1y oriented structures from Figure (6-9) of GE Report No. NEDE-13426P.

The methods approved by the staff require use of a versed sine representation for impulse shape and determination of Mg from Figure (6-8) of NEDE-13426P for all structures.

Both of the modifications used by the applicant imply a reduction in the derived load.

This reduction is estimated to be about 10-25% because of the use of a triangular pulse (depending on structural natural frequency) and about 60% due to the use of Figure 6-9 to determine Mg for circunferentially oriented structures.

Application of the correct procedures would therefore imply an increase of up to 25% in load for radial targets and a doubling of the load for circumferential targets.

However, since these deficiencies will be accommodated either by reevaluation using the actual Maise criteria (Item 4 above) or by demonstrating sufficient margin as in Item 1, we conclude that the procedure employed by the applicant is acceptable.

7

The applicant also states that impact loads on pool swell deflectors have been developed by utilizing methods derived from the Mark I (NUREG-0661) and Mark 111 (NUREG-0978) Acceptance Criteria.

We find the use of these methods to be acceptable.

(11-6) Drag Loads (Section 38.10.2)

At the request of the NRC staff, the applicant agrees not to utilize Figure 3B-75 of GESSAR 11 for abscissa values (Ratio a/b) less than unity.

This deviation from GESSAR II is not plant unique since it will be required of all Mark III plants.

It is included here because it has not been previously imposed by the staff's acceptance criteria.

(11-7) Loads on Expansive Structures at the HCU Floor Elevatior The applicant indicates that the HCU floor has a design pressure dif-ferential capability of 10 psi rather than the 11 psi differential imposed by the GESSAR II methodology.

The justification for use of the smaller value is identical to that cited in Item (11-1). We find the use of a reduced design load acceptable for the reasons also stated therein.

(11-8) References (Section 38.13)

The applicant adds five plant unique references to this section of the FSAR.

These additions are relevant and appropriate and acceptable to the staff.

The six references cited in this proposed insert are as follows:

1.

Attachment to CEI letter PY-CEI/NRR-0010L from fi. R. Edelman, CEI, to B. J. Youngblood, NRC, dated January 31, 1983.

2.

Draft NRC Acceptance Criteria for LOCA Related Mark III Containment Pool Dynamic Loads Appendix C of Attachment to NRC letter from T. P. Speis, NRC, to H. Pfef ferlen, GE, dated October 8,1982.

3.

Attachment to CEI Letter PY-CEI/NRR-0235L from M. R. Edelman, CEI, to B. J. Youngblood, NRC, dated May 16, 1985.

4.

G. Maise, " Suggested Acceptance Criteria for Impact Loads on Short Mark 111 Structures Close to the Pool",

Department of Nuclear Energy, Brookhaven National Laboratory, Upton, N. Y.

11973, February 15, 1984.

5.

U.

S.

Nuclear Regulatory Commission, Mark II Containment Lead Plant Program Load Evaluation and Acceptance Criteria", NUREG-0487, October 1978.

6.

General Electric Co., ' Ma rk III Confirmatory Test Program.

One-Third Scale Pool Swell Impact

Tests, Test Series 5805",

G.E.

Report NEDE-13426-P, August 1975. ____

r

. e,e e

EVENT POTENT!AL LOAOING CONDITION

• LOCA OCCURS a COMPR ESSIVE WAVE LOAOING ON CONTAINMENT ORYWELL PRES $URE RISES

~

e SONIC WAVE LOA 0 LNG OF ORYWELL 1F e JET IMPINGEMENT ANO 8088LE VENTS CLEAR AND VENT PRESSURE LOAOS ON THE AIR / STEAM FLOW STARTS CONTAINMENT e VENT CLEARING ANO VENT FLOW AP ON ORYWELL e OUTWARD FLOW AP ON WEIR WALL

~

1 P POOL SWELLS IN e IMPACT LOADS ON LOW STRUCTURES A BULK MODE

^

o ORAG LOAOS ON STRUCTURES m

IN AND ABOVE THE POOL 1r BREAKTHROUGH

~

if POOL SWELL CONTINUES IN A FROTH MCOE ANO ENCOUNTERS e FROTH IMPINGEMENT ON FLOW RESTRICTION AT HCU

~

HIGH STRUCTURES u

FLOOR e FLOW AP ON HCU FLOOR ANO AOJACENT CONTAINMENT

.1F

=

ORYWELL VENTING

/

e FALLS ACK LOADS ON COMPLETE STRUCTUR ES 1I

~.

STEAM CONDENSATION IN POOL AT VENT ExlTS b

e CONDENSATION LOADS t

e I

t l

BLOWOOWN ENOS e WEIR WALL AND ORYWELL LOAOS DUE TO CHUGGING t

ECCS FLOODING OF REACTOR e NEGATIVE PRESSURE ON WE WALL. ORYWELL. AND ITS

. VESSEL AND CRYWELL DEPR E S$UR f zATION PENETR ATIONS e NEGATIVE FLOW aP ON WEIR WALL i

t W-l* CONTA!NMENT PRES $URE LCNG TERM HEAT UP OF THE PCOL Figure 6-3.

Loss-Of-Coolant Accident Chronology (OBA) i

_9 t

s DRAFT SUPPLEMENT 2 TO THE SAFETY EVALUATION REPORT RIVER BEND STATION, UNITS 1 & 2, DOCKET NO. 50-458/459 6.2.1.8 LOCA-Related Pool Dynamics

Background

The Mark III pool dynamic loads were reviewed at the CP stage for Grand Gulf Nu-clear Station Units 1 & 2 (Grand Gulf 1/2) and at the PDA stage for GESSAR-238NI. The staff concluded at the time that the information available was suf-ficient to adequately define the pool dynamic loads for nuclear plants applying for Construction Permits. Since the issuance of the GESSAR-238NI Safety Evalua-tion Report (SER) (December 1975), GE has conducted further tests and analyses to confirm and refine the original load definitions. To keep the NRC and Mark III applicants apprised of the current status of these tests, GE issued an In-terim Contein.::nt Loads Report (22A4365) in April of 1978 and revised this re-port several times before GESSAR-II was provided to the staff on March of 1980, which is GE's FDA submittal for their standard B0P design and is to be refer-enced by Mark III OL applicants. Appendix 3B of GESSAR-II provides the final-ized pool dynamic load definition for Mark III containments, and is the basic document used for review by the staff and its consultants.

The River Bend applicant has included Appendix 3B of GESSAR-II in their FSAR submittal as Appendix 6A.

Except as noted below, the applicant has adhered to all analytical techniques, assumptions, methodologies and concepts contained in Appendix 3B of GESSAR-II. Where River Bend unique parameters differ from those of the GE Standard plant, River Bend parameters are used. The staff has re-viewed GE's pool dynamic load definitions to arrive at a finalized hydrodynamic load definition that can be utilized by Mark III containment applicants for op-erating licenses. The pool dynamic loads were reviewed under Task Action Plan (TAP) B-10, " Behavior of BWR Mark III Containment". The results of this review are documented in NUREG-0978 ' Hark III LOCA-Related Hydrodynamic Load Defini-tion". They are applicable to the River Bend Station.

Description of Phenomena

[

Figure 6-3 shows the sequence of events occurring during a DBA and the potential l

loading conditions associated with these events. Following a postulated loss-(

of-coolant accident (LOCA), the drywell pressure increases due to blowdown of l

the reactor system.

Pressurization of the drywell causes the water initially standing in the vent system to be accelerated into the pool and the vents are cleared of water. During this vent clearing process, the water leaving the L

horizontal vents forms jets in the suppression pool and causes water jet im-pingement loads on the structures within the suppression pool and on the con-tainment wall opposite the vents. During the vent clearing transient, the drywell is subjected to a pressure differential and the weir wall experiences a vent clearing reaction force.

Immediately following vent clearing, an air and steam bubble forms at the exit of the vents. The bubble pressure initially is assumed equal to the current drywell pressure. This bubble theoretically transmits a pressure wave through <

,m

t the suppression pool water and results in loading on the suppression pool bound-aries and on equipment located in the suppression pool. As air and steam flow from the drywell becomes established in the vent system, the initial vent exit bubble expands to suppression pool hydrostatic pressure. GE Large-Scale Pres-sure Suppression Test Facility (PSTF) tests show that the steam fraction of the flow is condensed but continued injection of drywell air and expansion of the air bubble results in a rise in the suppression pool surface. During the early stages of this process, the pool swells in a bulk mode (i.e., a slug of solid water is accelerated upward by the air). During this phase of pool swell, structures close to the pool surface will experience loads as the rising pool surface impacts the lower surface of the structure. In addition to these ini-tial impact loads, these same structures will experience drag loads as water flows past them. Equipment in the suppression pool will also experience drag loads.

After the pool surface has risen approximately 18 feet above the initial pool surface, the water ligament thickness has decreased to two feet or less and the impact loads are significantly reduced. This phase is referred to as incipient breakthrough (i.e., the ligament begins to break up). To account for possible nonconservatisms in the test facility arrangement, the staff has determined that the breakthrough height should be set at 18 feet above the initial pool surface.

Ligament thickness continues to decrease until complete breakthrough is reached and the air bubble can vent to the containment free space. The breakthrough process results in formation of an air / water froth and, for load definition pur-poses, is defined to happen at 19 feet over the initial surface. The incipient breakthrough height and the height at which froth loads begin have been set higher than predictions from test results for conservatism. Continued injection of drywell air into the suppression pool results in a period of froth pool swell. This froth swell impinges on structures it encounters but the two-phase nature of the fluid results in loads that are much less than the impact loads associated with bulk pool swell.

When the froth reaches the elevation of the floors on which the hydraulic con-trol units for the control rod drives are located (approximately 22 feet above pool level), the froth encounters a flow restriction; at this elevation, there is approximately 25% open area. The froth pool swell experiences a two-phase pressure drop as it is forced to flow through the available open areas. This pressure differential represents a load on both the floor structures themselves and on the adjacent containment and drywell. The result is a discontinuous pressure loading at this elevation.

As drywell air flow through the horizontal vent system decreases and the air /

water suppression pool mixture experiences gravity-induced phase separation, pool upward movement stops and the fall-back process starts. During this pro-cess, floors and other flat structures experience downward loading and the con-tainment wall theoretically can be subjected to a small pressure increase. How-ever, this pressure increase has not been observed experimentally.

The pool-swell transient associated with drywell air venting to the pool typi-cally lasts three to five seconds. Following this, there is a long period of high steam flow rate through the vent system; data indicates that this steam will be entirely condensed in a region right at the vent exits. For the DBA re-actor blowdown, steam condensation lasts for a period of approximately one.. -

o D

minute. Potential structural loadings during the steam condensation phase of the accident have been observed, and are included in the containment loading specification.

As the reactor blowdown proceeds, the primary system is depleted of high-energy fluid inventory and the steam flow rate to the vent systen decreases. This re-duced steam flow rate leads to a reduction in the drywell/ containment pressure differential which in turn results in a sequential recovering of the horizontal vents. Suppression pool recovery of a particular vent row occurs when the vent stagnation differential pressure corresponds to the suppression pool hydrostatic pressure at the row of vents.

a.

Toward the end of the reactor blowdown, the top row of vents is capable of con-densing the reduced blowdown flow and the two lower rows will be totally recov-ered. As the blowdown steam flow decreases to very low values, the water in the top row of vents starts to oscillate back and forth causing what has become known as vent chugging. This action results in dynamic loads on the top vents and on the weir wall opposite the upper row of vents.

In addition, an oscilla-tory pressure loading condition can occur on the drywell and containment. Since this phenomenon is steam mass-flux dependent (the chugging threshold appears to 2

be in the range of 10 lb/sec/ft ), it is present for all break sizes. For smaller breaks, it is the only mode -of condensation that the vent system will experience.

Shortly after a DBA, the Emergency Core Cooling System (ECCS) pumps automatical-ly start up and pump condensate water and/or suppression pool water into the re-actor pressure vessel. This water floods the reactor core and then starts to cascade into the drywell from the break (the time at which this occurs depends upon break size and location).

Because the drywell is full of steam at the time of vessel flooding, the sudden introduction of cool water causes rapid steam condensation and drywell depressurization. When the drywell pressure falls be-low the containment pressure, the drywell vacuum relief system is activated and air from the containment enters the drywell. Eventually sufficient air returns to equalize the drywell and containment pressures; however, during this drywell depressurization transient, there is a period of negative pressure on the dry-well structure. A conservative negative-load condition is therefore specified for drywell design. Theoretically, this condition can lead to the flow of water over the weir wall. Structures located in the vicinity could then experience drag and impact loads. Load definitions to account for these effects are included.

Small breaks, defined as breaks not large enough to automatically depressurize the reactor, do not result in bounding pool dynamic loads except for the thermal i

loading conditions on the drywell and weir walls. Thermal gradient load defini-tions are provided for the design of the walls containing the suppression pool.

Pool Dynamic Load Assessment (a)

Generic Load Definition - The staff's review of the generic LOCA-related pool dynamic load definition was completed early in 1984 The results of this review and the staff's evaluation of the pool dynamic load defini-l tion are documented in NUREG-0978 " Mark III LOCA-Related Hydrodynamic 4

Load Definition", which was publisned in August of 1984. With only a few l

exceptions, the staff found the load definition proposed by the General l '

. Electric Company in Appendix 3B of GESSAR-Il to be acceptable.- A set of Acceptance Criteria was. developed by the staff to cover those areas where the proposed loads were not satisfactory. These were included as Appen--

dix C of NUREG-0987. A brief description of these Acceptance Criteria is provided below.

(1). Pool Swell Velocity - Pool swell velocity controls impact and drag loads on the structures between the pool rest elevation and the breakthrough elevation. GESSAR-II proposes a value of 40 fps at all elevations. The staff requires use of an elevation dependent value which varies continu-ously from zero up to a maximum of 50 fps at elevations greater than or equal to 10 feet above rest elevation.

(ii) Pool Swell Loads on Structures Attached to the Containment Walls - The GESSAR-II specification corresponds to steady state drag at a fixed ve-locity of 40 fps. The staff's Acceptance Criteria require this to be modified to reflect the change in pool swell velocity given in (i) above and the inclusion of impact type forces when the structure is not im-3 mersed prior to pool swell. A detailed procedure for evaluating the im-pact load is provided in the Acceptance Criteria.

(iii) Bulk Impact on Small Structures - The GESSAR-II methodology was found ac-ceptable provided' the structures involved satisfied certain limitations related.to structural natural frequency, size and location above the pool. The Acceptance' Criteria require that when any of these limitations are not satisfied, the load specification be reviewed by the staff on a plant unique basis.

(iv) Froth Impact Loads - The GESSAR-II methodology was found to be unaccept-

- able. An acceptable alternative was developed by the staff and its con-sultants and is described in detail in the Acceptance Criteria. The new

.4 method-differs from the GESSAR-II approach with respect to maximum froth impact pressure, temporal characteristics of the forcing functions and j

region of application.

1 (v) Drag Loads - The GESSAR-II methods are found acceptable provided that they are modified to account for the change in pool velocity given in (i) above and provided that they correctly account for the structure-wall in-teraction effect on drag loads.

(vi) Loads on Submerged Structures - The GESSAR-Il methods are acceptable ex-cept for computation of acceleration loads on non-cylindrical structures and the evaluation of standard drag during the C0 phase of the LOCA. The i

staff requires that the Mark I Acceptance Criteria as set down in NUREG-0661 be used to develop these loads.

(vii)

Impact Loads on Structures Above the Weir Annulus - The GESSAR-II methods were found to be acceptable except for radial structures located within one foot of the top of the weir wall and all structures located between zero and 0.25 feet above the weir wall. Detailed procedures for evalua-j' tion of the impact loads in these cases are provided in the Acceptance Criteria.

, 4

=

=,-.e--

g r7y

,,,--,--e

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

---r

--w..

- - - ~, -

(b)

River Bend Station Plant Unique Load Evaluation -

(i)

Applicability of Generic Load Definition - The staff has examined the information supplied in the FSAR and has concluded that the Generic load definition described in GESSAR-II as modified by the staff's Acceptance Criteria described in NUREG-0978 is applicable to the River Bend plant except as noted below. All major structures and components that would experience LOCA-related pool dynamic loads are within the range ofappli-cability of the staff approved methodology in terms of geometry and rela-tive location in the containment and the suppression pool. The major features of the suppression pool geometry (main vent submergence and ver-tical spacing, pool radial width and pool depth) differ slightly from the standard plant dimensions but these differences are not considered sig-nificant in terms of their effect on pool dynamic loads. The use of the generic methodology by the River Bend applicant to develop.the LOCA-re-lated pool dynamic load definition is therefore acceptable to the staff.

(ii)

Plant Unique Load Definition - Impact Loads on Certain Structures between the Pool surface and the HCU floors.

The bulk impact load specification in the NRC Acceptance Criteria (NUREG-0978) states that the GESSAR-II methodology is acceptable, subject to the following limitations:

(1) Targets must have combinations of widths and natural frequencies such that Figures 38.33-1, 2, 3 and 4 of GESSAR-II indicate them to be in the "GESSAR conservative" region with respect to the V = 50 ft/sec pool velocity curve.

(2) There are no structures smaller than 4 feet long.

(3) There are no structures closer than 6 feet above the pool.

In plant designs where some specific structures may not meet limitations (2) and/or (3), the pulse duration must be shortened with an appropriate adjustment to the pressure amplitude. The load specifications for these structures will be reviewed by the staff on a plant unique basis.

The River Bend Station structures above the pool satisfy limitation (1).

How-4 ever, limitations (2) and (3) are not satisfied for all structures. The RBS FSAR has utilized a modified version of the "Maise" criteria l for the design of structures closer than six feet from the pool surface and/or shorter than four i

feet in length. These modifications involve first decreasing the impact pres-sure amplitude by a factor (V/50)2, where V is the slug velocity at the struc-l ture elevation, and then increasing it by the factor (.007/T) where t is the im-pact pulse duration determined according to the requirements of Reference 1.

The Maise criteria do not allow a (V/50)2 reduction in peak pressure ampli-tude. They do, however, permit of a somewhat smaller increase in pressure for impulse durations less than 7 msec; i.e.:

the requisite increase is not

(.007/t) but (.007V/50r).

It is not clear how the RBS version of the Maise cri-teria evolved, but the overall result is a load specification that utilizes a peak impact pressure amplitude which is a factor (V/50) less than that imposed in Reference 1.

, 1

To address the staff's concern relative ta this apparent non-conservatism, the applicant supplied additional information to demonstrate the adequacy of their approach.2 This information consisted of a comparison between the RBS design impact pressures on selected structures with those predicted by an alter-nate method claimed to have been previously aproved by the staff. These compar-isons demonstrated that the RBS method yields loads which exceed those derived from the alternate method by margins of 1.25 and greater for radially oriented structures and 2.3 and up for circumferential targets.

. A detailed description of the alterr. ate method is provided by the applicant in Reference 3.

The staff and its consultants have reviewed this methodology and find that it conforms in almost all respects to procedural steps previously approved by the NRC staff in NUREG-0487 and NUREG-0661. Two areas of nonconfor-mance have been identified, however. These are the use of a triangular impulse to represent the impact load and determination of hydrodynamic mass, M, for H

circumferentially oriented structures from Figure (6-9) of GE Report No.

NEDE-13426P. The methods approved by the staff require use of a versed sine representation for impulse shape and determination of MH from Figure (6-8) of NEDE-13426P for all structures.

Both of the modifications used by the applicant imply a reduction in the derived load. This reduction is estimated to be about 10-25% because of the use of a triangular pulse (depending on structural natural frequency) and about 60%

due to the use of Figure 6-9 to determine Mg for circumferential1y oriented structures. Application of the correct procedures would therefore imply an in-crease of up to 25% in load for radial targets and a doubling of the load for circunferential targets. Although these increases are substantial, they are still bounded by the margins demonstrated by the comparisons that were provided in Reference 2.

Accordingly, the staff finds this plant-unique load specifica-tion to be acceptable.

Summary The staff has completed its review of the LOCA-related pool dynamic loads for the River Bend Station and finds the load definition used by the applicant to be conservative and acceptable.

References i

1.

Maise, G., " Suggested Acceptance Criteria for Impact Loads on Short Mark III Structures Close to the Pool", Department of Nuclear Energy, Brookhaven j

National Laboratory, Upton, New York 11973, February 15, 1984 l

2. to letter dated June 3,1985 from J. E. Booker, GSI to H. R.

Denton, NRC.

t

3., ibid.

4 i

..__.,.____.,__.~,m.____._._,._____.__.___,__,m._.

• e

• se.,

EVENT POTENTIAL LOACING CONDITION' LOCA OCCURS e COMPRESSIVE WAVE LOADING ON CONTAINMENT DAYWELL PRESSURE RISES e SONIC WAVE LOAOING OF ORYWELL I f a JET IMPlNGEMENT ANO B U88LE VENTS CLEAR AND VENT PRESSURE LOADS ON THE AIR / STEAM FLOW STARTS CCNTAINMENT e VENT CLEARING AND VENT FLOW AP ON ORYWELL e OUTWARO FLOW AP ON WEIR WALL 1r POOL SWELLS IN e IMPACT LOAOS ON LOW STRUCTUR ES m

A B ULK MOD E

^

e ORAG LOAOS ON STRUCTURES IN ANO ABOVE THE POOL 1r BREAKTHROUGH

~

if POOL SWELL CONTINUES IP. A FROTH MCOE AND ENCCUNTERS e FROTH IMPlNGEMENT CN HIGH STRUCTUR ES FLOW RESTRICTION AT HCU "u-FLOOR e FLCW AP ON HCU FLCOR ANO ADJACENT CONTAINM ENT

...1r s

t ORYWELL VENTING COMPLETE e FALLS ACX LOAOS ON

^

STRUCTURES if I

STEAM CONOENSATION e CONDENSATION LOADS

+-

IN POOL AT VENT EXITS r

t 8LOWOOWN ENOS s

e WEIR WALL AND ORVWELL

. LOAOS OUE TO CHUGGING t

ECCS FLOOOfNG OF AEACTCR e NEGATIVE PRES $URE ON WEIR l

. VES$il ANO CRYWELL

,7-W ALL ORYWELL, ANO ITS l

OEPRESSURIZA TION P E NETR A TIONS e NEG ATIVE FLCW AP ON WElR WALL

~

LCNG TERM HE AT UP CF THE PCOL

e CCN TA!NMENT PRES $URE L OAO t

I

=

l Figure 6-3.

Loss-of-Coolant Accident Chronology (DBA) 1

-E-+

-g+

,p.-pgr-p y9- - - - -.,y,-=.

.4,.,

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

p -. -, -

.p

,.,--g.4-g,,..y--e.w m.

,,-,y.mem

.y-yp.

4,,

. - - -, pr

.~

= -.

7-DRAFT SUPPLEMENT 2 TO THE SAFETY EVALUATION REPORT CLINTON POWER STATION, UNIT 1, DOCKET NO. 50-461 6.2.1.8 LOCA-Related Pool Dynamics

Background

The Mark III pool dynamic loads were reviewed at the CP stage for Grand Gulf Nu-clear Station, Units 1 & 2 (Grand Gulf 1/2) and at the PDA stage for GESSAR-238NI.

The staff concluded at the time that' the information available was suf-ficient to adequately define the pool dynamic loads for nuclear plants applyino for Construction Permits. Since the issuance of the GESSAR-238NI Safety Evalua-i tion Report (SER) (December 1975), GE has conducted further tests and analyses to confirm and refine the original load definitions.

To keep the NRC and Mark III applicants apprised of the current status of these tests, GE issued an In--

i terim Containment Loads Report (22A4365) in April of 1978 and revised this re-port several times before GESSAR-II was provided to the staff on March of 1980, i

'which is GE's FDA submittal for their standard B0P design and is to be refer-enced by Mark III OL applicants.

Appendix 38 of GESSAR-II provides the final-ized pool dynamic load definition for Mark III containments, and is the basic document used for review by the staff and its consultants.

The Clinton applicant has incorporated the essential features of Appendix 3B of GESSAR-II in various sections of their FSAR submittal.

Except as noted below, the applicant has adhered to all analytical techniques, assumptions, methodolo-gles and concepts contained in Appendix 38 of GESSAR-II.- Where Clinton unique 4

parameters differ from those of the GE Standard plant, Clinton parameters are used.

The staff has reviewed GE's pool dynamic load definitions to arrive at a finalized hydrodynamic load definition that can be utilized by Mark III contain-i ment applicants for operating licenses.

The pool dynamic loads were reviewed under Task Action Plan (TAP) B-10. " Behavior of BWR Mark III Containment". The results of this review are documented in NUREG-0978 " Mark III LOCA-Related l

Hydrodynamic Load Definition".

They are applicable to the Clinton Power i

Station.

Description of Phenomena Figure 6-3 shows the sequence of events occurring during a DBA and the potential i

loading conditions associated with these events.

Following a postulated loss-of-coolant accident (LOCA), the drywell pressure increases due to blowdown of the reactor system.

Pressurization of the drywell causes the water initially standing in the vent system to be accelerated into the pool and the vents are cleared of water.

During this vent clearing process, the water leaving the horizontal vents forms jets in the suppression pool and causes water jet im-pingement loads on the structures within the suppression pool and on the con-tainment wall opposite the vents.

During the vent clearing transient, the j

drywell is subjected to a pressure differential and the weir wall experiences a j

vent clearing reaction force.

I Immediately following vent clearing, an air and steam bubble forms at the exit I

of the vents.

The bubble pressure initially is assumed equal to the current

! i

6 i

e drywell pressure.

This bubble theoretically transmits a pressure wave through the suppression pool water and results in loading on the suppression pool bound-aries and on equipment located in the suppression pool. As air and steam flow from the drywell becomes established in the vent system, the initial vent exit bubble expands to suppression pool hydrostatic pressure.

GE Large-Scale Pres-i sure Suppression Test Facility (PSTF) tests show that the steam fraction of the flow is condensed but continued injection of drywell air and expansion of the air bubble results in a rise in the suppression pool surface.

During the early stages of this. process, the pool swells in a bulk mode (i.e., a slug of solid water is accelerated upward by the air).

During this phase of pool swell, structures close to the pool surface will experience loads as the rising pool surface impacts the lower surface of the structure.

In addition to these ini-tial impact loads, these same structures will experience drag loads as water flows past them.

Equipment in the suppression pool will also experience drag loads.

Af ter the pool surface has risen approximately 18 feet above the initial pool surface, the water ligament thickness has decreased to two feet or less and the l

impact loads are significantly reduced.

This phase is referred to as incipient breakthrough (i.e., the ligament begins to break up).

To account for possible nonconservatisms in the test facility arrangement, the staff has determined that the breakthrough height should be set at 18 feet above the initial pool surface.

Ligament thickness continues to decrease until complete breakthrough is reached and the air bubble can vent to the containment free space.

The breakthrough process results in formation of an air /,<ater froth and, for load definition pur-poses, is defined to happen at 19 feet over the initial surface. The incipient breakthrough height and the height at which froth loads begin have been set higher than predictions from test results for conservatism. Continued injection of drywell air into the suppression pool results in a period of froth pool swell.

This froth swell impinges on structures it encounters but the two-phase nature of the fluid results in loads that are much less than the impact loads associated with bulk pool swell.

When the froth reaches the elevation of the floors on which the hydraulic con-trol units for the control rod drives are located (approximately 22 feet above pool level), the froth encounters a flow restriction; at this elevation, there is approximately 25% open area.

The froth pool swell experiences a two-phase pressure drop as it is forced to flow through the available open areas.

This pressure differential represents a load on both the floor structures themselves and on the adjacent containment and drywell.

The result is a discontinuous pressure loading at this elevation.

As drywell air flow through the horizontal vent system decreases and the air /

water suppression pool mi xture experiences gravity-induced phase separation, pool upward movement stops and the fall-back process starts.

During this pro-cess, floors and other flat structures experience downward loading and the con-tainment wall theoretically can be subjected to a small pressure increase. How-ever, this pressure increase has not been observed experimentally.

l l

The pool-swell transient associated with drywell air venting to the pool typi-cally lasts three to five seconds.

Following this, there is a long period of high steam flow rate through the vent system; data indicates that this steam will be entirely condensed in a region right at the vent exits.

For the DBA i

1 t i

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

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

reactor blowdown, steam condensation lasts for a period of approximately one minute.

Potential structural loadings during the steam condensation phase of the accident have been observed, and are included in the containment loading specification.

As the reactor blowdown proceeds, the primary system is depleted of high-energy fluid inventory and the steam flow rate to the vent system decreases. This re-duced steam flow rate leads to a reduction in the drywell/ containment pressure differential which in turn results in a sequential recovering of the horizontal vents. Suppression pool recovery of a particular vent row occurs when the vent stagnation differential pressure corresponds to the suppression pool hydrostatic pressure at the row of vents.

Toward the end of the reactor blowdown, the top row of vents is capable of con-densing the reduced blowdown flow and the two lower rows will be totally recov-ered. As the blowdown steam flow decreases to very low values, the water in the top row of vents starts to oscillate back and forth causing what has become known as vent chugging.

This action results in dynamic loads on the top vents and on the weir wall opposite the upper row of vents.

In addition, an oscilla-tory pressure loading condition can occur on the drywell and containment. Since this phenomenon is steam mass-flux deper. dent (the chugging threshold appears to 2

be in the range of 10 lb/sec/ft ),

it is present for all break sizes.

For smaller breaks, it is the only mode of condensation that the vent system will experience.

Shortly after a DBA, the Emergency Core Cooling System (ECCS) pumps automatical-ly start up and pump condensate water and/or suppression pool water into the re-actor pressure vessel.

This water floods the reactor core and then starts to cascade into the drywell from the break (the time at which this occurs depends upon break size and location). Because the drywell is full of steam at the time of vessel flooding, the sudden introduction of cool water causes rapid steam condensation and drywell depressurization.

When the drywell pressure falls be-low the containment pressure, the drywell vacuum relief system is activated and air from the containment enters the drywell, Eventually sufficient air returns to equalize the drywell and containment pressures; however, during this drywell depressurization transient, there is a period of negative pressure on the dry-well structure.

A conservative negative-load condition is therefore specified for drywell design.

Theoretically, this condition can lead to the flow of water over the weir wall.

Structures located in the vicinity could then experience drag and impact loads.

Load definitions to account for these effects are included.

l l

Small breaks, defined as breaks not large enough to automatically depressurize l

the reactor, do not result in bounding pool dynamic loads except for the thermal loading conditions on the drywell and weir walls. Thermal gradient load defini-tions are provided for the design of the walls containing the suppression pool.

Pool Dynamic Load Assessment (a) Generic Load Definition - The staff's review of the generic LOCA-related pool dynamic load definition was completed early in 1984 The results of this review and the staff's evaluation of the pool dynamic load definition are documented in NUREG-0978 " Mark III LOCA-Related Hydrodynamic Load i

Definition", which was published in August of 1984.

With only a few [

I l

exceptions, the staff found the load definition proposed by the General Electric Company in Appendix 3B of GESSAR-II to be acceptable.

A set of Acceptance Criteria was developed by the staff to cover those areas where the proposed loads were not satisfactory.

These were included as Appendix C of NUREG-0987.

A brief description of these Acceptance Criteria is provided below.

(i) Pool Swell Velocity - Pool swell velocity controls impact and drag loads on the structures between the pool rest elevation and the breakthrough elevation. GESSAR-II proposes a value of 40 fps at all elevations.

The staff requires use of an elevation dependent value which varies continu-ously from zero up to a maximum of 50 fps at elevations greater than or equal to 10 feet above rest elevation.

(ii) Pool Swell Loads on Structures Attached to the Containment Walls - The GESSAR-II specification corresponds to steady state drag at a fixed ve-locity of 40 fps.

The staff's Acceptance Criteria require this to be modified to reflect the change in pool swell velocity given in (i) above and the inclusion of impact type forces when the structure is not im-mersed prior to pool swell.

A detailed procedure for evaluating the im-pact load is provided in the Acceptance Criteria.

(iii) Bulk Impact on Small Structures - The GESSAR-II methodology was found ac-ceptable provided the structures involved satisfied certain limitations related to structural natural frequency, size and location above the pool. The Acceptance Criteria require that when any of these limitations are not satisfied, the load specification be reviewed by the staff on a plant unique basis.

(iil fruth Impa,t Loads - The GESSAR-II methodology was found to be unaccept-able.

An acceptable alternative was developed by the staff and its con-sultants and is described in detail in the Acceptance Criteria.

The new method differs from the GESSAR-II approach with respect to maximum froth impact pressure, temporal characteristics of the forcing functions and region of application.

(v) Drag loads - The GESSAR-II methods are found acceptable provided that they are modified to account for the change in pool velocity given in (i) above and provided that they correctly account for the structure-wall in-teraction effect on drag loads.

(vi) Loads on Submerged Structures - The GESSAR-II methods are acceptable ex-cept for computation of acceleration loads on non-cylindrical structures and the evaluation of standard drag during the C0 phase of the LOCA. The staff requires that the Mark I Acceptance Criteria as set down in NUREG-0661 be used to develop these loads.

(vii)

Impact Loads on Structures Above the Weir Annulus - The GESSAR-II methods were found to be acceptable except for radial structures located within one foot of the top of the weir wall and all structures located between zero and 0,25 feet above the weir wall.

Detailed procedures for evalua-tion of the impact loads in these cases are provided in the Acceptance Criteria.

(b)

Clinton Power Station Plant Unique Load Evaluation -

(1)

Applicability of Ceneric Load Definition - The staff has examined the information supplied in the FSAR and has concluded that the Generic Load definition described in GESSAR-II as modified by the staff's Acceptance Criteria described in NUREG-0978 is applicable to the Clinton plant except as noted below. All major structures and components that would experience LOCA-related pool dynamic loads are within the range of appli-cability of the staff approved methodology in terms of geometry and rela-tive location in the containment and the suppression pool.

The major features of the suppression pool geometry (main vent submergence and ver-tical spacing, pool radial width and pool depth) differ slightly from the standard plant dimensions but these differences are not considered sig-nificant in terms of their effect on pool dynamic loads.

The use of the generic methodology by the Clinton applicant to develop the LOCA-related pool dynamic load definition is therefore acceptable to the staff.

(ii) Plant Unique Load Definition (ii-1) Weir Swell Motion During Negative Drywell Pressure.

The applicant's design method,2,3,te differs in significant ways i

from that recommended in GESSAR-II (Section 38.5.1.5).

The differences are primarily in the assumptions that are made to develop a set of describing equations to define weir swell behavior. A detailed descrip-tion of these differences has been provided by the applicant in Reference 3.

Generally speaking, the differences are such that the applicant's assump-tions are less conservative than those made to develop the GESSAR-II load definitions. The result is weir swell behavior that is significantly less severe. Peak steady state velocities, for example, are reduced from 30 fps to about 18 fps, and maximum swell elevation from 13 feet to about

,5 feet.

The jet front impact velocity is also reduced somewhat (from 10 to about 9 fps).

The staff has reviewed the applicant's analysis in detail and finds that, although less so than the GESSAR-II methods, it still preserves signifi-cant conservatisms.

In general, the applicant's method may be character-ized as a conservative application of conventional first principles using reasonable approximations to render the analysis tractable. We find the method acceptable for characterizing the dynamics of the weir swell motion during conditions of negative drywell pressure.

(11-2) Weir Swell Dras Loads During Negative Drywell Pressure The applicant's method differs from that recommended by GESSAR-II (Sec-tion 38.5.1.5) for rectangular structures. GESSAR-II requires the use of drag coefficient Co = 1.2 together with the use of an equivalent diane-ter equal to the diameter of the circle that circumscribes the rectan-gle.

The applicant's method employs the experimental correlations of References 5 and 6.

The staff and its consultants have examined the applicability of this correlation for the present purpose and conclude that it is acceptable in general, and, in fact, implies higher loads than -

those dictated by the staff approved GESSAR-II method for most structure geometries and orientation of practical interest.

For these reasons we find the proposed method acceptable.

(11-3) Impact Loads on Structures Between the Pool Surface and the HCU Floors This item corresponds to Section 38.10.1 of the GESSAR-II load methodol-ogy.

The applicant indicates 2 that the GESSAR-!! methods of this sec-tion, as modified by the Draft Acceptance Criteria 7 were the design basis for a certain class of structures which are affected by the modifica-tion.

The modification required that, for structures less than four feet long, the pulse duration be reduced by a factor x/4 where x is the actual length of the structures; also, for structures closer than six feet above the pool surface, the pulse duration was to be reduced by a factor y/6 where y is the actual elevation of the structure above the pool.

Since the Draft Criteria have been superseded by the requirements of Appendix C of NUREG-0978, the applicant developed a new evaluation using a methodology which would meet the staff's approval. This new evaluation would then be used to demonstrate, for relevant structures, that the new methods imply loads that are less than those originally used for design, thereby verifying its adequacy.

A description of the new method was presented in Reference 8.

The impact loading for each component is determined as follows:

1.

Determine properties of small structure under consideration:

Orientation (radial or circumferential);

Width: W in inches; Height (of structure's bottom) above suppression pool: H in feet; Length:

L in feet; Dominant vertical natural period (usually the vertical fundamental natural period): To in seconds.

2.

Determine hydrodynamic mass of impact per unit area from G.E. Report NEDE-13426P (Reference 9 Figures 6-8 or 6-9) corresponding to struc-ture's orientation and width:

2 (M /A) in Ibm /ft,

g 3.

Determine the impact velocity of water slug V, corresp. to height, H:

V = (H/10) * (2.6-1.6/IT/f6)

• Vmax in fps, for 0 S H 110 feet; V = Vmax = 50 fps, for 101 H 118 feet.

4 Determine design impulse:

Idesign " IN /AI*IY)*

m ' I" ESI

• SUC' H

y

-6

5.

Determine impulse duration according to G. Maise (Reference 10. Equa-tions (5) or (9)) corresponding to structure's orientation, height above pool and length:

t in seconds.

l d

6.

Determine the peak dynamic pressure design value:

• I design P,,,=

, in psi.

d 7.

Determine the dynamic load factor, DLF, corresponding to (td/Tc) assuming a versed sine shape for the design impulse.

8.

Determine equivalent static design pressure:

PREV = (P

• DU)Maise design max 9.

Evaluate adequacy of original design, that is:

PREV

> 1.0 ?

is P ORIG where P0 RIG is the original equivalent static pressure used for design.

This method was applied to a total of 66 structures for which the values of all of the key parameters listed above were reported in tabular form.2 We have reviewed this methodology and find that it conforms in almost all respects to procedural steps previously approved by the NRC staff in

. NUREG-0487 and NUREG-0661 (essentially the Maise criteria of Reference 10).

One area of nonconfonnance has been identified, however. This is determination of hydrodynamic mass N, for circumferentially oriented H

structures, from Figure (6-9) of GE Report No. NEDE-13426P.

The method approved by the staff requires determination of Mg from Figure (6-8) cf NEDE-13426P for all structures.

The modification used by the applicant implies a reduction in the derived load.

This reduction is estimated to be about 60% due to the use of Figure 6-9 to determine MH for circunferentially oriented structures.

Application of the correct procedure would therefore imply a correspond-l ing increase of the load for circumferential targets.

Although this is a substantial increase, we have found that it is still bounded by the nargins demonstrated by the comparisons that were provided by the applicant in all but one case (Target No. 3 in Table 2R of Refer-ence 8). For this particular case, the applicant has reevaluated the design using the increased load estimates and states that the calcula r

maximum stresses were within the (static) allowable for the material.j d Accordingly, the staff finds this plant unique load specification to be acceptable. t u

. ~

e References 1.

Enclosure to IPC letter U-0698 from J. D. Geier, IPC to A. Schwencer, ML, dated February 17, 1984 2.

Attachment ~ to IPC Letter U-600020 f rom F. A. Spangenberg, IPC to W. R.

Butler, NRC, dated May 16, 1985.

3.

Attachment to IPC Letter U-600265 from F. A. Spangenberg, IPC to W. R.

Butler, NRC dated 4.

Attachment to IPC Letter U-from F. A. Spangenberg, IPC to W. R.

Butler from NRC dated December,__,1985.

5.

Bostock, B. R. and Matr. W. A., " Pressure Distribution Forces on Rectangular and D-Shaped Cylinders," Aeronautical Quarterly, February 1972.

6.

Bearman, P.

W.

and Trueman, D.

M.,

"An Investigation of Flow Around Rectangular Cylinders," Aeronautical Quarterly, August 1972.

7.

NRC Letter from J. R. Miller (NRC) to G. E. Waller, IPC, " Draft Acceptance Criteria for Mark III LOCA Related Pool Dynamic Loads," dated March 16, 1982.

8.

Burns and Roe, Inc., Technical Report, " Bulk Pool Swell Impact Loads on Small Structures Above Pool Surface at Clinton Power Station," Attachment to IPC Letter U-600020 from F. A. Spangenberg, IPC to W. R. Butler, NRC, dated May.16, 1985.

9.

General Electric Co., " Mark !!! Confirmatory Test Program.

One-Third Scale Pool Swell Impact Tests, Test Series 5805 " G.E. Report NEDE-13426P, August 1975.

10. Maise, G., " Suggested Acceptance Criteria for Impact loads on Short Mark !!!

Structures Close to the Pool " Department of Nuclear Energy, Brookhaven National Laboratory, Upton, New York,11973, February 15, 1984.

i j

11. Revised Version of Reference 8 above.

i l

3 l

t i '

E e

e,6 e

EVENT POTENTIAL LOADING CONDITION

• e COMPRESSIVE WAVE LOCAOCCURS m

LOACING ON CONTAINMENT ORYWELL PRESSURE RISE 3 o SONIC WAVE LOACING OP ORYWELL t

VENTS CLEAR ANO VENT e JET IMPINGEMENT AND 8080LE PRESSURE LOAOS ON THE AIR / STEAM PLOW STARTS CONTAINMENT e VENT CLEARING ANO VENT PLOW AP CN ORYWELL o OUTWARO PLOW AP ON WEIR WALL S

q g e (MPACT LOAOS ON LOW POOL SWELLS IN STRUCTUR ES m

A SULK MODE

^

e ORAG LOAOS ON STRUCTURES IN AND A80VE THE POOL if SREAKTHROUGH 1 P POOL SWELL CONTINUES IN A e PROTH IMptNGEMENT CN PROTH MCOE AND ENCOUNTERS

~

e PLOW AP ON HCU PLOOR AND HIGH STRUCTURES PLOW RESTRICTION AT HCU PLOOR AOJACINT CONTAINMENT 3,

4....

j ORYWELL VENTING p

e P ALLS ACK LOAOS ON

.41 t

's

• COMPLETE r'ana- *a STRUCTUR ES i '*

• a a = *; -, ',1r d,,,

3, s'

f a

STEAM CONDENS ATION e CONOEN$ATION LOA 03 IN POOL AT VENT EXITS

^

g f

eW11R WALL ANO ORYWELL OLOWOOWN ENOS

'u

, LOAOS Out TO CituCQiNG y

ECCS FLOCOING OF REACTOR e NEG ATIVE ents*,Ung on wg,g VESSEL AND CRYWELL M

WALL, OnyWELL,gno gyg PENE TR A f TONS OCPRIMUHl2ATION e NECATIVE FLOW a# ON WilR WALL LCNG TE AM HE AT LP or yng PCOL e CCNTAINVENT PHES$unt LOAO Figure 6-3.

Loss-Of-Cool.)nt Accident Chronology (DBA) 9

.