ML20056C508

Submits Comments on BWR Stability Issues Discussed in 920917 Meeting

ML20056C508
Person / Time
Issue date:	09/28/1992
From:	Jun Lee NRC
To:	Boehnert P Advisory Committee on Reactor Safeguards
References
ACRS-CT-2048, NUDOCS 9306240318
Download: ML20056C508 (3)

Text

f j

09/29/92 11:32 01 313 763 4540 DEPT StCL ENG

$002'

~

Y-Mf fpt klal'13 September 28,1992 Mr. Paul Boehnert Advisory Committee on Reactor Safeguards SER 2 9 W U. S. Nuclear Regulatory Commission Washington, DC 20555

Subject:

Thermal-Hydraulic Phenomena Subcommittee Meeting on September 17,1992

Dear Paul:

In regard to the BWR stability issues discussed at the meeting,I believe we are mak progress in the right direction to resolve the concerns related to them. We still have considerable difficulties in accurately representing and understanding the sensitivity o BWR oscillations to system parameters and the impact oflarge-amplitude oscillations on containment integrity. Nonetheless, I suggest that we focus our effort on timely detection ofincipient unstable oscillations and pmdent mitigation or control actions to suppress the oscillations.

If detection and suppression of power and flow oscillations can be satisfactorily accomplished, the oscillatory incidents observed at the LaSalle Unit 2 and Washington Nuclear Power Unit 2 (WNP-2) should not significantly affect the basic premises of the current ATWS rules. I offer the following specific comments:

1. The time-domain oscillation detection algorithm the BWR Owners' Group (BWROG) suggests involves determination of the oscillation period, amplitude growth rate, and amplitude. The algorithm involves several empirical parameters and appears rather complex.

It is also unclear if determination of anyone of the three oscillatory characteristics monitored will be sufficient for timely and unambiguous detection of unstable oscillations. In addition to further strenmiming the time-domam detection algorithm, I suggest that the BWROO consider frequency-domain algorithms, which could offer distmet advantages, as discussed extensively at the meeting. Limitations of the frequency-domain algorithms cunently in use are their slow response time for highly stable oscillations and inability to provide indications for regional, out-of-phase oscillations. With some development, I would imagine that both of these limitations could be eliminated to m*n the frequency-domain algorithm a desirable addition to, if not a complete re 31acement for, the time-domain algorithm under study. In all of these detection algorituns, proper considerations should be given to account for failed sensors, especially for detection of re instability is driven by flow oscillations,gional oscillations. Since the whole BWR I also would like to suggest that we include y

ma momtormg of flow oscillations, especially the distinct characteristics of density wave oscillations where the inlet flow owi11** out of phase with the outlet flow.

O

2. Since BWR oscillations are of concern ultimately in the ATWS context, the detection t

algorithm should also include a dmely indication of scram failures. The ATWS eg indication is supposed to be included in Emergency Planning Guidelines (EPGs) but I q

com am not famitir myself how it is implemented and how reliable it is. If an ATWS is vte ej g mm a mc s

30073 um3 -g

\\D c

x a

a e'

09/29/92 11:33 31 313 763 4540 DEPT Nt'CL ENG

@ 003 s

I 2

properly and timely detected, then the Alternate Rod Insertion (ARI) system and Standby Liquid Control System (SLCS) are to be actuated to minimize the impact of the ATWS. An immediate actuation of these alternate and diverse m' i negative reactivity insertion should be able to suppmss, if not completely eliminn incipient unstable oscillations. With this feature of the ATWS rules in mind, I.,

with the BWROG suggestion that no further action is required under the ATWS rula enei 3

to make sure, however, that the actual implementation of the ARI system satisfies the diversity requirement.

3. To control power oscillations in BWRs when the contml rods are operational, the first mitigating action should be insertion of as many control rods as necessary. I believe this is called the Select Rod Insert (SRI) system, although I have not been able to understand the distinction between this system and the ARI system mandated under the ATWS rules. Rather than relying on other control mechanisms, I suggest that

'i increased effort be spent on determining the number and location of the control rods for the SRI system to insure that a sufficient amount of negative reactivity is inserted.

Unless we have a complete scram failure, them should be sufficient negative reactivity available through the ARI or SRI system to suppress the oscillations without reactor shutdown. This would have been the case for the mcent WNP-2 incident. On the other hand, if we do indeed have an ATWS, the ARI and SLCS should have been actuated long before the confirmation of instability and further control actions to suppress the oscillations should not be perhaps contemplated.

4. I feel uneasy about the feedwater flow reduction as a control mechanism for power oscillations. In ideahzed computer simulations,it can be, and apparently it has been, shown to suppress oscillations. Computer simulations, as well as actual European experience, also show that an increase in the recirculation flow could accomplish the same objective. Both of these control maneuvers primarily alter the void content in the core and thereby affect the core reactivity and void coefficient of reactivity, but in the opposite directions. This indicates, more than anything else, the sensitivity of the BWR system dynamics near the core instability threshold, and extreme cam should be taken in controlling power and flow oscillations through flow control maneuvers. This is especially true in actual operating conditions wh:n other transients are likely superimposed on the ongoing nuclear-coupled density-wave oscillations. In addition to the water hammer effects that W. Lipinski brought up as a possible concern in feed flow control maneuvers, I doubt if the TRACO code could adequately model complex, multi-dimensional mixing streams in the upper plenum and feedwater sparger region.

For these reasons, I prefer, as discussed above, insertion of a sufficient number of control rods under the SRI system to feed flow control.

5. The TRACG numencal algorithm needs further improvement and verification to handle large-amplitude power and flow oscillations. The use of an explicit thermal-hydraulic algorithm is only a short-term solution and more effort should be made to fully utilize more robust algorithms and establish the adequacy of the current explicit algorithm.

The TRACG simulation of the ATLAS tests,inchiding channel dryout, locks good but the tests do not cover flow reversals of the type simulated under the BWR instability conditions.

Flow reversals, as well as large power spikes and channel dryout, introduce considerable uncertainties into the TRACG results for large-amplitude oscillations, including the extent of fuel dama dynamic average power as osediations grow.ge estimated and the lack ofinemase in th

(-

09/29/g2 11:34 C1 313 763 4540 DEPT NUCL ENG

@ 004 3

6. As amply discussed at the meeting, BWR stability is quite sensitive to a number of system parameters. Furthermore, LAPUR frequency-domain predictions of decay ratios are expected to have a 20% uncertainty even with the best characterizadon of system parameters. Hence, the exclusion region, suggested urder Option I for long-term stability solution, will involve large fuzzy areas and will not be of much practical use. Thus, Option I with a number of sub-strategies seems to be an exercise in futili.ty and should be discarded.

I hope the above comments are useful to the Committee.

Yours sincerely, John C. Lee xc: L Catton