ML20207T407
| ML20207T407 | |
| Person / Time | |
|---|---|
| Site: | Seabrook  |
| Issue date: | 01/12/1987 |
| From: | Theofanous T CALIFORNIA, UNIV. OF, SANTA BARBARA, CA |
| To: | |
| Shared Package | |
| ML19306D588 | List:
|
| References | |
| NUDOCS 8703230540 | |
| Download: ML20207T407 (38) | |
Text
-
s /-,
.b REVIEW COMMENTS on SEABROOK STATION STEAM GENERATOR TUBE RESPONSE DURING SEVERE ACCIDENTS (a draft NUREG report dated 12/15/86_)
and related sections of TECIINICAL EVALUATION OF THE EPZ SENSITIVITY STUDY FOR SEABROOK (a draft BNL report dated 12/5/86) by T.G. Theofanous Department of Chemical and Nuclear Engineering University of California Santa Barbara, CA 93106 January 12,1987 8703230540 870313 PDR COMMS NRCC r,,.;,,, _,3 CORRESPONDENCE pon
. ~.......
e f
+.
s.
-1.
Introduction It has been known for several years now that core melt accidents at high primary system pressure give rise to certain unique phenomenological suspects with potentially significant consequences on containment behavior. The basic phenomena involve strong natural circulation currents, of steam, within the primary system and if the system remains pressurized, eventually forceful melt ejection into the reactor cavity. A summary of the early history and my views on a number of related issues may be found in Attachments 1 to 4. A copy of some of the viewgraphs I utilized in the first discussion to claim that hich cressure scenarios cannot eersist to core melt is provided in Attachment 5. This presentation and a 4
subsequent one given in front of the ACRS (see references in Attachment ) contain the first suggestion of the need to depressurize the primary system once the severe accident progressed
^
closely to the point of no return. A recent letter by Dana Powers (of Sandia) providing some important source termWyectives on this issue of depressurization is given in Attachment 6.
Finally, a copy of an early paper of ours which is referenced in Attachment I and which provides a scoping out of the effects of natural circulation is included as Attachment 7.
^
In order to complete the perspective created by the above, I would add the following:
(a) The potentially important effects of natural circulation were, probably, first realized by 1
Vern Denny, who embarked upon a code development program while everybody else was l
developing 1-D forced convection models. The first published reference of this code development effort and some preliminmy calculations of a general nature were presented in the September 1983 meeting at Cambridge, MA. This paper concentrated on possible effects on hydrogen production only--i.e., no mention of steam generator tube ruptures, possible impact on avoiding high pressure melt ejection (and the Direct Containment Heating problem), etc.
(b) The first reference to the potential for steam generator tube rupture was given by L.
Winters in an ENC memorandum in July,1982. This was based on RELAPS calculations l
which anificially cleared up the loop seals allowing for circulation around the loops.
l 1-
=-
t s.
3 (c) When we tied-in the natural circulation phenomena to the Direct Heating and the Tube Rupture problems in the March 1984 Containment Loads Working Group meeting there was considerable skepticism (perhaps even some hostile reaction) by most participants. A~whole host of subsequent studies, as well as the May 1984 meeting reported by W. Lyon and given 4
as Reference 8 in the NUREG report under review here, were nromoted by this nresentation.
At this point there is mounting concensus that the system will fail, somewhere. prior to core slump in the lower plenum. But this is not a firrn conclusion nor is it known whgzm the system will fail.
(d) Early on, we pushed for a definitive position to be taken (and documented) on the conditions under which the loop seals will cease to exist sometime into such an accident. This has not been done yet. On the other hand, the model experiments in Westinghouse (it must also be mentioned here that some scaling consideration for such experiments were provided by D. Squarer in the 1983 San Francisco Winter ANS meeting) demonstrated that natural circulation can extend wellinto the steam generator tubes even when loop seals are intact. The quantitative aspects of this problem are not fully well understood yet; that is, important modelling and scaling question remain.
For over two years now my efforts to " move" the issue of depressurization in a generic fashion have been frustrated. As you will note from my recent, generic, letters (Attachments 1 r
L to 4) I believe this is a matter of the utmost urgency. It may well be that a specific licensing l
L action (such as the Seabrook) provides a better forum for coming to full grips with the issue.
I have serious reservations about the NUREG report both on matters of perspective and philosophy as well as on matters of implementation. I provided my own perspective first, because I think I can be more effective in my discussion of the report ifit is clear where I am coming from. But first let us address the BNL report.
-w p,.
4
- 2. Comments on the BNL report 1
The relevant portions of the BNL report include section 2.3, part of section 2.5 (pp'.27-28), and section 6.4.
Section 2.3 deals with the quantification aspects. The frequency of high pressure sequences in which SGTR "might have an effect" is taken as 4x10-5 per reactor year. 'Ihis number was estimated by the applicant, was considered reasonable by the NRC, and was used in the BNL study. I have no documentation on the basis for this number, but I note the absence of an estimate of its ancertainty range. The NUREG report mentioned that the applicant meant it to be bounding. Without it this number is not very useful. In any case, it appears that such sequences represent a significant portion (- 20%) of all core melt events.
Conditional probabilities for Steam Generator Tube Rupture were based on SAND 86-0119 addressing the Surry Plant, which in tum, were based on " expert judgement," and were found
" consistent" with an earlier NRC memo (p. 2-29 was missing from my copy so I do not know what this memo was). Thus the numbers were taken in the range 0.01 to 0.3, conditional on no depressurization, which, itselfis conditional on core melt, was taken as 0.2. First, a formal matter. I agree with BNL that if the experts were asked to quantify these matters for Seabrook they would, in all likelihood, come up with the same numbers. Still, at this point the BNL report should explicitly state whether the BNL team examined the Surry-Seabrook relationship with regard to the phenomena of interest here and what were their conclusions.
Now, on the substance of those numbers. It is not stated whether the 80% chance to depressurize includes operator actions for doing so or whether the 0.8/0.2 split simply reflects the experts judgement on the likelihood to fail the pnmary system prior to core melt, and p
elsewhere than the steam generator tubes. If the former was a significant consideration,I would call it into question on the basis that current systems and procedures cannot provide any assurance one way or another (it they did we would have no problem). If the latter was the significant consideration the numbers represent a pure guess which I would also call into
- -"r 7
w e-t--t---,'#asye--,,r-e---,y
--~, w vv -
r--
---ev-*-esw----
w-1 e-*--=-s+---r--+-w-v ae-rm..--wwe-
-s-re---w
-+www----=--=*-w--w--wwwww---e=+er..-
e-
+-.~, -- -. - -,,. --
e 5
question. In my view the way things presently stand system failure is a virtual certainty. The real question then is what fraction of those failures could be in the steam generator tubes. I believe one would be hard pressed at this time to show a chance less than o.f order 1. In fact, considering that certainly not all tubes would be in " mint" condition, and lack of detailed data and structural analysis under the relevant conditions, I would not be very surprised if this number tumed out to be even larger. 'Ihus I do not feel comfortable with the comparable BNL range of 2x10 to 6x102. In my view, for this kind of study, use of numbers less than 10-1 4
requires substantial understanding of all technical elements. I think part of the problem is that the process has been broken, artifically, into two parts,' i.e., the numbers would have been even smaller if a three-part process was considered.
In summary, then, a prudent estimate of SGTR probability would be of order ~ 10-1 Combining this with (the best estimate I suppose) frequency value quoted as 4x10 5 yields an estimate of 4x104 for bypassing containment (assuming that secondary side behavior was i
l included in the 4x10 5 value). Assuming one order of magnitude uncertainty (I would not be surprised if it was more) in the above frequency estimate and a factor of x2 in the conditional 4
rupture estimate we obtain an "utmer" estimate of - 10 ner reactor year. I think this is an unaccentably larne number.
. Section 2.5.(pp. 2-27 and 2-28) provides, in brief, the source term methodology l
employed and resulting consequences. Release category SIW (large early bypass) was chosen I
as a conservative assumption. I agree that this is conservative, perhaps too conservative.
Some retention during blowdown would undoubtedly occur; however, it would strongly depend on the rates of blowdown, past history in the sequence, and details of the kinetics of l
deposion and resuspension processes. None of these is well known, and I can sympathize l
with the BNL team for not making arbitrary curtailings in this source term definition.
The discussion of consequences provides a general perspective that SGTR does not dramatically alter the risk even if what the report calls pessimistic assumptions were to be used.
l
a c.
6 u
'Ihis is a result that needs a lot of explaining. It implies, basically, that early containment failure at the assumed scenario frequency, is not such a big deal and I certainly disagree with this perspective. Furthermore, here is the place to discuss uncertainties in frequency estimates and their impact. Further, the statement on what the NRC staff " believes" is nebulous and only serves to incorrectly diffuse the issue.
To summarize. my own Audne (as exnlained abovelindientes that allowine hich messure scenarios to nroceed unchecked is unnecentable. This judgement is based on NRC safety goals, on the pur,pective provided by the risk estimates in the BNL report together with my
" upper" estimate of containment bypass (~ 10 per reactor year), and consideration of 4
intangible aspects regarding public response to such potential accidents.
Section 6.4 summarized the overall sensitivity study results. There are several difficulties here.
(a) 'Ihe " combination" of sensitivity study results is strongly qualified that is "not rigorous and 1
could lead to inconsistencies." his is like disavowig the results altogether. If something is I
salvageable the authors should explain what it is and why it is useful. Explaining clearly how results were " combined" also would be necessary-I suspect they were simply added up.
(b) The results are given in graphical and tabular form, cut-and-dry, with no explanations.
Furthermore,I see no anempt at interpretations. One may argue that this is the job of the NRC, but, I believe, the analysts are best qualified in judging, at least, whether differences are significant. Furthermore, major trend differences should be discussed and explained.
(c) He final paragraph asserts that"the conservative assumptions regarding accidents during shutdown and induced SGTR have the most impact on the dose vs distance and risk estimates in PLG-0465." This does not say whether in the BNL team's opinion this impact is significant. I do not know whether this noncomminal position was intentional or simply a poor choice of words. There is a good chance it may be the latter, especially since it goes on to say that the optimistic assumptions have a minac impact and that there is considerable uncertainty in
-. - - ~
- ~.. -.. -
l s.
the areas explored in the report (implying that the "pessindstic" estimates entail a significant impact).
(d) If my deduction in (c) above is correct then there is a need to explain how the combined effect of a few rather menial (individually) effects can be so significant. That is, when uncertainties cannot be discussed properly even on an order of magnitude basis are factors of x2 in the individual results significant? and is the non-rigorously formulated " combined effect" significant? Of course, I do not believe the individual area of SGTR is insignificant.
Furthermore, I do not think the subiect of DCH was proneriv treated nor do I think it is not sienificant.
(e) If my deductions in (c) above are incorrect, the BNL team should indicate, more clearly, that the final result is insensitive to the assumptions in the sensitivity studies. Again, I would disagree with the conclusion, however.
- 3. Comments on the NUREG report My key comments on this report will be categorized under the headings: Perspective, Philosophy, Mechanistic Assessment,andImplementation. Additionalcomments willbe given under the heading " Details."
j 3.1 Question of Perspective There is a repeated reference attributing the genesis of the SGTR topic to a " strong" Seabrook containment and hence the unfolding of"previously neglected bypass paths that were masked and found now to contribute to risk" (p. 5, p.11, etc.) I think this is an unnecessary fm' esse of the issue at best, but badly inaccurate at worst. The simple fact is that the potential l
for this particular mechanism to fail containment was discovered only within the past two years and it is not fully appreciated to this day. Furthermore, the historical perspective provided in l
the report is rather incomplete. I suggest that instead of dancing around the issue (a lot of
8 unnecessary pages in the report on that ) it should be clearly and blantantiv introduced. tocether with the key milestones and references in its develonment. alonclhe lines shown in mv introductory section.
3.2 Question of Philosophy The report gives an unmistakable impression (i.e., p.11) that the Regulatory process operates in a never-ending fashion. That is, as plants get stronger we get more picky. 'Ihis is a wrong and highly undesirable impression. Rather, it must convey the impression that assessments are made considering all (that we know) competing risk factors in as absolute sense as itis possible
- The report acknowledges that it did not make an effort to assess the pros and cons regarding the need to depressurize the primary system in case of such accidents. This is a short-sighted view and I strongly recommend that it.be revised. In my view the benefits of i
depressurization are straightforward, the costs are minimal, and there are no downsides to it.
Along the same lines, the whole question of operator action is left hanging. In the absence of explicit procedures we should not count on an operator thinkine that it would be good to depressurize the system. And how will he obtain valve control in a station blackout?
?
i 3.3 Questions of Mechanistic Assessment The key point made in the report is that experimental evidence and its applicability are prelimi=y, that the applicant's Computer Program has not been verified nor documented, and that certain modelling assumptions are optimisti:. I agree with all three. In addition to these purely thermal-hydraulic aspects there are other more serious questions that are not even touched. These include: (a) a whole range of issues regarding structural response under I
highly complicated, very far from design conditions, (b) uncertainties regarding actual steam generator tube conditions (certainly they will not be " mint") at the time of the postulated
,_.__-,-._._..._._.--___..._.m.
r-i.'
9 accident, and (c) major uncertainties in fission product release and deposition (or a whole sequence of such phenomena) processes.
Concerning the assumed plant state the report correctly identifies the applicant's assumption, that the secondary side will be at 1,100 psi, as inadequate. After all is it not the case of bypass (open secondary side to the atmosphere) the case of real concern? Along these lines the report should also discuss what can happen if the tubes fail with a pressurized secondary side. For how long will the relief valves (on the secondary side) lift? How much steam will be expelled? What is the likelihood that they will fail? What is the likelihood, in case of a massive SGTR, that the secondary system will fail somewhere?
All calculations have, presumably, assumed intact loop seals. In my opinion SGTR becomes a certainty in the absence ofloop seals. Hence, a comprehensive, fully-documented, assessment ofloop seal behavior under the whole range of possible scenarios.is essential here.
3.4 Questions ofImplementation The report gives the optimistic impression that this problem area can be " studied away." I l
@ve this is an overiv ootimistic and inacer00date iudcement for the Reculatory to make at this time.
Furthermore, the procedure outlined to " substantiate this judgement" is' unrealistic and incomplete. How are we going to guess the structural condition of all steam generator tubes al the time of a hypothetical accident? How are we to verify structural evaluations under such highly overheated conditions? What experiments will give us adequate data to ensure proper i
i verification of the fission product migration process? To be sure, this last comment refers to additional heat loads on the tubes rather than to source term effcets. That is not to mention verification of scaling for any natural circulation experiments and associated numerical tools.
How long and how much money will it take to obtain a sufficiently good " substantiation" of thisjudgement?
l l
l
I t
o 10 3.5 Some AdditionalDetails
- 1. p.15. The whole discussion at the top of this page is misplacing the issue. Only RCS pressure matters and item 3 is particularly irrelevant.
- 2. p.16. The major contributor here is station blackout and there is no power to drive the pumps. The comment at the bottom, p.16 and top, p.17, is thus irrelevant.
- 3. p. 20. Reference is made to section 2.2.4 but there is no such section. Here it should state on what basis can PSNH count on operator access to PORVS and for what cases they cannot.
- 4. p. 21. The discussion here on loop seals and pump use is appropriate.
- 5. p. 22. Here the authors of the report visualize removal of a loop seal"as water was fo'ced out of the RCS via the break." This is not possible for such low rates of steam flow.
However, as superheated steam kept bubbling through the loop seal water, it would slowly boil it off, having1the same effect. Here one must be concerned with a race between system depressurization, loop seal evaporation, and steam generator tube heat up. After the loop seals break the steam generator tubes would nearly equilize with core temperatures very rapidly.
- 6. p. 34. How about failure modes other than creep ruptures?
- 7. p. 36. There is need to go much more into the PSNH estimate of 10-2 to 10-3 per demand to fail to depressurize the system. Here is where the major pay offis and here it is that some considerable progress needs to be made.
~
- 4. Conclusions The NUREG report is significantly limited in its perspective and philosophy in approaching this problem. The BNL study on the impact of SGTR is highly qualified and in my opinion does not bring out the problem as forcibly as it should. It is promising that the utility believes that they can reliably depressurize the system. I believe this is the most
G 11 promising avenue and it should be punued aggressively. The benefits will be far reaching not -
only for Seabrook but also for all other plants. Instead of this limited report I would favor a comprehensive one dealing with all aspects of high pressure scenarios and,an assessment of needs and procedures for depressurization.
O 9
G
[
..w,.--.. - - -- -
- m-i-+4 h'me Meeei.m a mesi. 4 me., egei.e a
_w r
Machmen4 i UNIVERSITY OF CALIFORNIA. SANTA BARBARA CEREELEY
- fhvis
- IRVINE
- Los ANCELES
- RIVERSIDE
- SAN DIECO
- SAN FRANCISCO SANTA B ARDA!M
- SANTA CRUZ DEPARTatENT OF CHES!! CAL AND
$ANTA BARBARA. CALITORNIA 93!06 Ntl CLEAR ENCINEEDINC November 21,1986 Dr. T. Speis, Director Safety Technology Office of Nuclear Reactor Regulation US Nuclear Regulatory Commission Mail Stop P-1122 Washington,DC 20555
Dear Dr. Speis:
s Re: Direct Containment Heating, etc.
Nearly three years ago,in the NRC/IDCOR meeting dealing with the Direct Containment Heating (DCH) problem I presented some simple calculations that led me to make the following deductions and recommendations.
1.
High pressure steam is very effective in redistributing the core heat throughout the primary system. That is, different parts of primary systems follow the core temperature within a few hundred degrees centigrade, which implies failure of the pnmary boundary prior to gross core meltdown. As a consequence, all high pressure scenarios should be expected to revert to low pressure ones.
- 2. Prediction oflocation of failure is rather difficult and highly uncertain. This is particularly so for the steam generator tubes, which may be in unknown states of degradation. Such failures could cause containment bypass (i.e., in-station blackout situation) at a rather inopportune time regarding outside consequences.
- 3. Given the above it would appear prudent to ensure timely, reliable, and predictable primary system depressurization. This can be accomplished through a variety of means, including
" fuse" systems or automatic (or manual) depressurization systems with dedicated power supply (or steam driven). The selection should be left to the utilities (vendors).
4.
Given item 2 above, the burden of proof for requiring such a system is not with the NRC.
Rather, the burden of proof for not having it is with the licensee.
Subsequent calculations of our own as well as such done by EPRI,INEL, LANL and SNL seem to support the view that early failure (prior to gross core melt) is very likely. Also, experiments run for EPRI at Westinghouse further demonstrate that steam generator tubes would be involved in this heat up process even in cases where the loop seal had not cleared.
Since my original suggestion I spoke often about my views on this subject. I also expounded on them during a seminar to the ACRS of" Severe Accident Phenemenology for PRA use."
Now I have the following additional points to make.
a Dr. T. Speis November 21,1986 Page 2
- 1. Meanwhile, the DCH issue has persisted-in fact it has escalated. High pressure dispersal has been pursued at SNL for over 4 years now and new work has begun at BNL, with a significant commitment of res'ources. I have not heard yet from either of these two research teams an engineering judgement that dispersal yielding a DCH contammentTailure is a figment of the imagination. On the contrary, all signals I can read in meetings, etc., indicate a real concem m tins area.
- 2. De installation of a system assuring reliable and predictable depressurization would instantaneously remove all DCH concems and associated expenditures. By nature the DCH problem is like that of energetic steam explosions, i.e., hard to obtain the kind of evidence needed to satisfy those more skeptically inclined. His provides an additionalincentive in favor of an engineered depressurization behavior. However, even if the DCH problem, somehow, went away, the potential for steam generator tube failure would still require that the action recommended here be taken anyway.
- 3. It is my understanding that others (notably Jesse Ebersole of the ACRS) have argued irl favor of a depressurization system in PWRs for reasons other than those stated here. Such potential additional benefits should be factored into the decision making process.
He purpose of this letter is to express and document my opinion that this problem area needs to be addressed by the regulatory authorities with the utmost urgency. Please feel free to call on me ifIcan be of any help.
~
nank you for your consideration.
Sincerely, (7D_
T.G. neofanous, Professor Department of Chemical and Nuclear Engineering Director, Center for Advanced Multiphase Processes and Safety I
TGT/h cc: W. Morris C. Allen F. ElTawils k
hChmed 2 UNIVERSITY OF CALIFORNIA. SANTA BARBARA 1
it gi it 3EREELEY
- DAY 33
- IRVINE
- LOS ANCELES
- ptVER$1r>E
- SAN Ditt;o. $ 4N FRANCTSCO g}
$ANTA SAR8A3M
- SANTA CRt2
\\.
- /
s DEPARTafENT OF CNE3t!CA!. AND SANTA BARBARA, CALITORNIA 93106 NUCLEAR ENCINEERINC December 8,1986 Dr. T. Speis, Director Safety Technology Office of Nuclear Reactor Regulation US Nuclear Regulatory Commission Mail Stop P-1122 Washington,DC 20555
Dear Dr. Speis:
Re: High Pressure Scenario, etc.
'Ihis is a followup to my Noveniber 21,1986 letter conceming my recommendation to develop a regulatory requirement for the timely and reliable depressurization of all Light Water Reactors under all possible severe accident conditions including station blackout. In my previous letter I expressed this concern from the Direct Containment Heating standpoint. Here I would like to expand on this aspect and to introduce some additional considerations that further support my recommendation.
1.
Current approaches to the DCH-dispersal problem are based implicitly or explicitly on the i
assumption that vessel blowdown will occur from a single instrumenuube penetration failure.
I believe coherent failure of many such penetrations is not only possible but rather likely under l
the assumptions of high pressure meltdown phenomena. Curantly there are no reliable l
methods to estimate the coherence of such failures, and considering the superposition of uncertainties in the phenomena leading up to this stage it is highly unhkely that any but gross l
bounding estimates would be possible in the foreseeable future. On the other hand the effects on dispersal can be quite dramatic. We have carried rough calculations of dispersal potential in i
the Zion geometry and concluded that this item alone can dictate behavior. In addition to impacting strongly the velocities in the Steam Generator compartments (where de-entrainment should occur if dispersal is to be avoided) there is potentially an important effect in scaling l
phenomena within the reactor cavity (i.e., dimensions of entrained particulate). We are currently addressing quantitatively both issues; however,I bring them up here because they are L
associated with important and inherent uncertainties and thus have to be weighted in the overall approach of addressing the DCHissue.
- 2. Even if dispersal was eventually shown to be a non-problem the high pressure blowdown would be undesirable from a hydrogen behavior standpomt. To fully achieve the benefits of igniters, both in terms of their performance, as well as in terms of our expectation of such benefits, we must avoid, I believe, rapid releases oflarge quantities of hydrogen-as would l
potentially occur in a high pressum scenario blowdown. This is particularly so for ice condensors, but large dry containments could be also affected.
---m
,,w-r a-
.,rv
,-,-.v.-,,,w,y
-,-,w--.---
e-,-- -
1e-r
---v w-v.v----
v,-
Dr. T. Speis l
December 8,1986 Page 2
- 3. Even the steam explosions area would be favorably impacted by avoiding the high pressure scenario. There are two aspects here. The one arises from the loss of strength of the upper internal structures and vessel head bolts under the steam natural circulation that accompanies the high pressure scenario. Over 500 MJ of mechanical energy credit would thus be lost in the energy panition evaluations of potential missile energies. That is, much smaller steam explosions (even those that do not fail the lower head) could potentially create missiles capable of a-failure. De other aspect arises because of the increased quantities of premixing expected in high pressures and confirmed by our recent calculations. Although there are some doubts concerning the escalation of triggers to explosions at high pressure it would not be prudent, in my opinion, at this time to discount the likelihood of such explosions. The stmetural aspects of the above have been documented in our four recent steam explosion papers. De premixing aspects at high pressure are presently written up in a separate paper.
4.
Finally, let me not forget that timely depressurization might even enhance the recovery potential. This is because it would bring mto action the accummulators with good potential to quench the core and thus. buy additional time for recovery of power.
In Conclusion Here we have a perfect example of how a major portion of the perceived risk as well as of the resource commitment necessary for its full apprecir. ion (not necessarily alleviation) can be eliminated through a rather straightforward system change. I believe the matter needs the most urgent attention by the Research and Regulatory authorities. From my discussions with
" Systems people" I am convinced that implementation of a highly reliable, manually activated, system would be rather trivial. A variety of fuse-type devices (passive) are also possible. I think the NRC should request ideas from the utilities and vendors and select the best one among them. One could even make a design competition-with awards-to generate additional incentives. Please let me know if you have any questions or ifI can be of any help.
Sincerely, T.G. neofanous, Professor Department of Chemical and Nuclear Engineering Director, Center for Advanced Multiphase Processes and Safety TGT/h cc: E.Beckjord D. Ross W. Morris C. Allen F. El Tawila D. Powers
~ _ _ _ _ _ _ _ _ _ _ _ _
... ___.._.~._._
e ggd 3
=
UNIVERSITY OF CALIFORNIA. SANTA BARBARA c
(18s;13.1v. Davis invtNz. s.os ANCEt.ES
- PRYER $1DE
- SAN DIEGO
- SAN FRANCISCO SANTA BARSARA SANTA CRtl%
DEFART3 TENT OF CHE3!! CAL AND SANTA SARSARA, C LIFORNIA 93106 NVCLEAR ENCINEEDINC December 11,1986 Dr. T. Speis, Director Safety Technology Office of Nuclear Reactor Regulation US Nuclear Regulatory Commission Mail Stop P-1122-Washington,DC 20555
DearDr. Speis:
Re: Depressurization System for PWhs Following up my December 8,1986 letter I would like to clarify that there is considerable flexibility in specifying the timing and duration of the depressurization operation. This flexibility, of course, is extremely important in meeting the overall objeenves with a set of design criteria for the depressurizanon system that are both easy and inexcensive to meet.
Furthermore, because of this flexibility it should be easy to achieve the hig,h reliability necessary while minimizing the consequences ofinadvertent actuation.
Sincerely, T.G. Theofanous, Professor Department of Chemical and Nuclear Engineering Director, Center for Advanced Multiphase Processes and Safety l
l TGT/h l
cc: E.Beckjord D. Ross B. Morris C. Allen F. El Tawila D. Powers l
l l
T y
r y
--y,.
-~,e
+
~-
-,,,.-i-.vre-,.,
,,.-w
-,-e--,
s, w,,----,
-m
+ --
)
McChmenk N
{.
UNIVERSITY OF CALIFORNIA. SANTA BARBARA
!; w $
8 SANTA BAR3 ARA
- SANTA CRil2 3EREELEY
- DAVI$
- IRVINE
- LOS ANCELES
- MIVER$1DE * $AN DIECO. SAN FRANCISCO j~ #)1 s
DEPARTafENT OF CHEMICAL AND SANTA BARBARA, CALIFORNIA 93106 Nt* CLEAR ENCINEERINC December 19,1986 Dr. W. Morris Office of Research U.S. Nuclear Regulatory Commission Washington,DC 20555
Dear Dr. Morris:
Re: Further Thoughts on Depressurization.
We have carried out some simple calculations to illustrate the flexibility in timing and duration of depressurization needs mentioned in my December 11 letter. De primary system volume was taken as 270 m. Steam pressure and temperature were taken at 15.5 MPa and 445* C 3
(100* superheated), respectively. De system pressure transients for a spectrum of vent areas are shown in the attached figure. Note that for typical designs the accumulators will come on at 2
- 4 MPa condensing a great deal of the remaining steam. Bus, a vent area of-6 in should prove more than enough. Clearly, there should be no problem providing controlled venting of this size with the required high reliability.
Even if it took 15 minutes to depressurize, the core would sustain no significant clad oxidation 2
damage (adequacy of vent area, as small as 2 in, in this respect would also appear conceivable). Thus, full core quench, with attendant repressunzation would be expected. A new boil off, heatup cycle would then follow, for which the depressurization characteristics mentioned above should also be quite adequate. This time extension would provide additional opportunities for recovery of power and termination of the accident.
I recommend that more detailed depressurization transient calculations and associated scenario determinations be made as,part of the design effort (by vendors and utilities). I also recommend that you appomt a committee consisting of D. Powers, J. Kelly and myself to carry out a similar assessment for the NRC.
j Sincerely, h
T.G. Leofanous, Professor Department of Chemical and Nuclear Engineering Director, Center for Advanced Multiphase Processes and Safety TGT/h Attachment cc: E.Beckford C. Allen D. Ross F.El Tawila l
T. Speis D. Powers
-w-
-,,-.,i--,--y-
,y-y.,,,
v.
.y_.
y
.-.y,
y
,-,9--,
m
--.y_,.--,,_
. _..,,...,.-y.
j+-acllatenf 5
... : -Y':...Q i.
I.
- 1..
n f *,, *
~.
.3
~
[E' VERE RCCthEUT.
=-
y CO UTRIN M EUT P li E UO M E NO LO G Y '
~
~
FOR P R O 11 R E l L I i T i c Rlt k l9 fC Eff M E NT T.
G. T A'c o fa n o u i Purdue I
Pr esen foHon h a n A C '? 5 Se 'n n o
3/21] R4
t N ll 4.f!.
kl.'.L. Q y P}.E.HQ4t JCE N A RIO1
.r
~
'L o 6~A.i S% ].oD P
Jccnario ie.
Vertcf faR) an d ' m ell e xilt of Pwo Pre u u re. ~
(b,cPrtyvrl2ed Prin urf JydeN).
N.dm.E.'YR...$.N.5.iE.M.YJ..
1.e. 7MLq (Yakon Rfatkou/)
E mp f t ra@onJ o f Mij A P
fcen on*a,.
A asin u(6lor) o'o'not cenc 6n,
.o birref Camia/n ment //cohnf ) fuel blJpr/al L
o a n d doriud OxNaficy,,
P/T Lo adt.
G e
6 i
I e
6
. A '.;. m *
.....~ _ _..
t at 3
p a
N,s 8
^.
o t43 58 hc s
v O
no 3.h
=.a n t
w 8
Ih k
M4
.b &
C
(
0 2
N
- o. 3 Q
o E F 4
=
w
~
N k
) g t
_t o
s a
y 2
do 5N u
ev a
S b
t
'N=
A 7
8
=
4, o
e
,u u
,r
,A g
O
~'
y c
I 0
-0
. a. <
f
,p,,,,
- N) i f
~
g a
s,
[
s
. $ d' 4 k]R E. I pg c. < 1s F
s
- 3. r.
o n.
.U n. a b.
u
.e.h s
i
- . 2 k
a e
s.- E%
o o
e o
t s
y I W.d. g
,0 ' g b
@ w U s.
-)
w-g'
. ~ g. E e
,b 8
w a
4L e M +(
9
's%E 6 m'
% g,. V. NF* 1 3 4 4 41 R
8-o o
s.24 E
q' # *,. s
< c,
h,% M
'J
.s w Q~ 9 s,,
a b
g w.g s
c q, s
'o a
u n - e.
c w
~3 e
u 2
e o,
m%
Q O N.
9
&h h
H-
. {
g9
. * ~s
- (h I, ~g h, -Q@
- $$$t$
E=$u
=
e
.s s%
2 w
t e
u s.
s3
=
2.
'-4 h 4
(m.,
%c v c
C-1 3
S e $
e M.c
~..q - s,%
p 4 e.~v 3
l b
s
'a
=g a
e 4
2
[
p
--?%
.c D d*
4.O
~
b r e
~
.o.,
,u
.a
,.ar C.
~
s a
,h
{
.h
$M ep q.4 y
e' e
2 Q
~
r..
m.
.c e
3
~
THE M ObELL Ts 3 ;
Rs m,cp, g
m,Cyg y.r%cp,
.n.......,
2 2
d), C c Ne $ e
[9e (f 4 p
f,,,
Q', f E'
9 m,c,-
e T,
R, M = d 4 7,-7, y
I
..J+-
( 47 '. 7, -7 4
4 d4
@2 d Tr - Tx-T, A
M't e
1
. ~ g h ( f 2 y,,i k
A z,6.
2,6 T.= c, e co s 2 t. + c c cin 2z b 4 cg i + Q i
z x
=. ;
1,6 z,s Tp=.[c, + k ( c;2,% 2x)} c coi z 6
+fy 2, (ea,-c,q)] c nn;
+ c t +.e, c, + cy a
-[j n's r T) ~ 'h ]*-
c.
e
.+
^.
R, R
R,
.k.'s g
C, To (R, + R + R )
c, = Tu~ C4 C; a c, h _ c,}z, t
2 c, = W/c a, c =.E.
A @'
4 Coks C Q
- L u Rg Ry b a Q +
AL E2
+R C=
tl e,
+
R,
y e,
"Zeh44-Al[la T,=-
i m
_-.__._________,_.m._
oe4-s 4m.4m g.
Tp-9 2
8
- u u
C; a
e v
w g
n y
'l q*
e.
m IS, k
h 2
o a
?M T
h k
4 IM Q
f Y
?
d
'h Q,
bu
^
,a 2
h *Y E
N c2 u
gN u?M A
o
=
M4;14g9 G
F w
s c
4 e
w E
Q W
1$
~S O
c
-?
sb O
O C
e a
~O no a
2 en
=
3 X
4 4g y=
+
M b
q g
w o.
A k
8 e%
b w#
Uo g
w
~
i*
4 c.[
t g
4, i o in l
2.\\,1 k
T Ls O
o&M N
M w
W 2 W
.E a
?
E 4
- b. F4 F +
w 1
E e N-c -2 7 e
e g
o g
4 A
o o
.a 8
~
\\,da).L
~)
g lU-vF1.fEL diRlULATICA)
CASE
- ""l
?=
[* 5^)
3,cco h *"'**L fn,, 90 + 40 Eln/s
& tnj,3 so-s17
( 3?-su e
l
' bT~ Go*F '
67(*FJ 830 l lbl
~
i-(
AT
[ f"i Eraltta$ltig
~
l Te f
i.
1,***
4 l
Tg 3.G(Ton
},}
( C t 'l'>
g
- - m )
l l
T i
.].
e ut I
i,,
h $ f (cF) 41,
~
T S
Tc n 2 r*
' d y
L J
~
fel: Noninnt 5 ^' A X k 5 c
~
,/
l y (OD4*CIL?6hlTC.
n c.
!. 2 4
s e
to 4
b io (s) e l
a+% 4' 6em84mano or.
- en en.N IemiDs.*
g 4e m-4-*Oh c
t o uc L u s tonu-l-PL F) ur
',HIy A P
Jcen a rio Jee tra flipbly Impre Loste.
t Lo co fion o f Prina ry Coo fahl / 4/ east 9
failure uncerfain. (,/o fea Hean Gen.
}yie faitue4 ce n Ac Pre crukJ // 4 betiro We to lelect Loco 6cn Ly
.e f u e " W betipn i
e m are be faited Eea lorce}fon o f T disfriLuffoyj
~
more peictifral Efoboroficy af early scen ario o
74 e r.W- //y drn u ticJ j lo op feal Makt /(43 i.e-Poleniial (o,-
mucl 1n created cort re dflanlc fu o
'Ffou
(&lu d Rolfoo n /by,12ustit y) co re efe),
bi{leren+ial 7'le rn a.t Exponcio o, l
\\
fa.icu rej ?
o Aicle m ufufor. - & C tiq. fior) : Defay or Preven f Co ve neu ?
o Cu eren/
Ct fin a /c> o{
& re McL}
auJaf
& nfadn m en f faita rc " Proic4ifiWm m cy.
be tlynifilon fly Erduced
- l l
l l
^
,,# o g.
- l.,' '
Wmet& (o Sandia National Laboratories Albuquerque. New Mexico 87185 December 24, 1986 Professor T..G. Theophanous Dept. of Chemical and Nuclear Engineering University of California, Santa BarDara Santa Barbara, California 93106
Dear Theo,
Thank you for sending me copies of your letters to Dr. speis of 4
November 21, December 8, and December 11, 1986, in which you discuss PWR depressurization.
I wish to express-my agreement with your view that depressurization of pressurized water
{'
reactors under accident conditions could be desirable.
You note, accurately, that there is an escalating concern about direct containment heating as a result of pressure-driven expul-sion of core melt. -Though it is true many have calculated that natural convection within the primary circuit might cause system l
depressurization, the experimental evidence of the Three Mile Island experience makes it difficult--actually impossible--to conclude pressure-driven melt expulsion from the primary circuit I do not believe calculational efforts have yet will not occur
.given sufficient credence to phenomena that may disrupt the distribution of heat to the primary pressure boundary.
But, even if these calculations are entirely accurate and the TMI i
experience can be shown to be a fluke, there is the problem you l
note of the inherent unpredictability of the rupture point.
l
[
Exchanging direct containment heating accidents for V sequence accidents is a poor trade-off.
I clearly, it would be preferable to avoid, altogether, the Certainly, with the exception of pressurized accident sequence.
some station blackout sequences, BWRs have benefitted from a severe accident perspective because of assured depressurization.
You note in your letters several benefits for depressurization in I believe there are other benefits from a source term FWRs.
perspective.
Current modeling in accident analysis codes of fission product release is quite crude.
It only considers time and temperature as important variables.
As a result, there have been difficulties predicting results of tests.
In connection with both the MELCOR and the MELPROG codes, we at Sandia have been One of the upgrading the fission product release modeling.
upgrades has been to include rate limitations caused by gas phase When applied to pressurized accident sequences, mass transport.
gas phase mass transport processes are found to sharply limit the release of the more volatile fission products--Cs, I, and Te.
This is not good.
If these volatile fission products are not l
e-----
y
-wr-ewe--ey.wm--,wy.-y-w,, w wwen----
e-a+--%--es=-----
C
' Professor T. G. Theophanous December 24,'1986 released in-vessel, they will be released ex-vessel, and ex-vessel release is not attenuated by deposition in the primary piping system.
The situation is even worse if the ex-vessel release is the result of direct containment heating processes that also fail the containment.
We have also been upgrading the models of the chem,istry of radionuclides during core degradation.
In this work, we have found high pressure steam will augment the volatility of.many of the less volatile radionuclides:
H moo (gas) + 3H2
[Mo]go
+ 4H O 2
4 2
2
[B40]go
+HO --- Ba(OH)2(gas) 2 2
H UO (gas) + H2 U02 + 2H O 2
4 2
The steam augmentation of the release of'these radionuclides,will In place additional heat loads on the primary piping system.
particular, thermal loads on steam generator tubes may be enhanced--again threatening sequence V type events.
We suspect, but have not proved that high pressure hydrogen will also enhance the volatility of radionuclides now thought not to be released during the in-vessel phases of an accident.
Assuredly, the complexities of both the chemistry and the release of radionuclides caused by high pressures would be obviated if the PWR were depressurized by some design system.
The existence i
of a known flow path from the primary system would permit useful engineered mitigation of in-vessel fission product release.
Were 4
the vent into a containment sump, then release would be attenuated by the same mechanisms that attenuate release during blowdown through a BWR suppression pool.
May I add, then, to your suqqestion of PWR venting that this venting be done under water because of source term considerations.
(
I cannot comment on the ease of introducing assured depressuriza-tion of PWRs.
It does appear, however, that the idea was merit and does deserve some attention.
Sincerely yours, m
W Dana Powers, Supervisor 3
Severe Accident source Terms Division 6422 j
l l
Copy to:
USNRC - T. Speis USNRC - D. Ross USNRC - F. Eltavila e
e
/kamed7 EPRI NP 4455 eactor accidents 89raded core accidents Pro eedings Electric Power r
March 1986 Research Institute I
O Proceedings: The Sixth Information Exchange Meeting on Debris Coolability Meehny hte.1 Nov. 7 c3, (9g4 cd O cLR.
I Prepared by University of California at Los Angeles Los Angeles, California
s
, g....
s t
Section 24 l
NATURAL CIRCULATION PHENOMENA AND PRIMARY SYSTEM FAILURE IN STATION BLACKOUT ACCIDENTS I
by H. P. Nourbakhsh, Chien-Hsiung Lee and T. G. Theofanous School of Nuclear Engineering Purdue University West Lafayette, IN 47907 1
s I
l
____.._.-r-y.-
i -,. v.
INTRODUCTION The potency of high pressure steam natural circulation phenomena to redistribute decay heat within the primary system components during the core post-dryout/ pre-melt period has not been fully appreciated in the past. The original suggestions of their possible significance were made by Denny and Sehgal [1] and by Winters [2].
Denny and Sehgal built an idealized natural circulation flow regime, coupling the core with the upper internals, into the code CORMLT. From calculations for a Station Blackout accident that are emphasized to be preliminary, they draw signifi-cant implications concerning: delays in core degradation, reductions'in thermal energy released to containment, and higher hydrogen generation rates. Winters extended Station Blackout calculations with RELAp5 into the post-dryout regime.
These calculations produced a natural circulation loop between the reactor vessel and the steam generators and an associated energy redistribution of such magnttude that the possibility of primary system boundary failure prior to fuel cladding
~
failure became evident.
Theofanous and Lee [3] also considered this mechanism in an assessment of the likelihood of the so-called "high pressure scenario," i.e., vessel failure and core release to containment from.a high primary system pressure. This simplified analysis will be summarized first as a way of developing a feel for the order of magnitude of various effects. Some simple experiments of natural circulation within a partially volumetrically heated porous medium will be presented to more concretely demonstrate th' efficacy of natural circulation cells to penetrate e
deeply into a heated porous bed. This paper will then conclude with a brief de-scription of a numerical model including applications to this demonstration experi-ment as well as the reactor conditions of interest.
e I
I e
24-1
.~. _.
s-7 ro ORDER OF MAGNITUDE ANALYSIS We consider a simple 3-mass lumped parameter model as illustrated in Figure 1.
Each mass is characterized by the same temperature and the fluid leaving is sup-posed to be in thermal equilibrium with it. This is a reasonable assumption for this approximate analysis given the highly distributed (high interfacial areas)-
core, upper internals, and steam generator masses. The coolant flow through each mass h, is controlled by a loss coefficient (permeability) and the available head e
due to the temperature and associated density differences.
3
-- w w----
I2 h
I_as a
2 m;
h,,
o It a
3 Figure 1.
Schematic illustration of the 3-mass lumped parameter model.
1
{
l The mathematical model may be expressed as:
dT' R,
=, Q' + T, - T, (1) ag l'
dT*
T, - T, (2)
R,
=
dt dT R
at T, - T, (3)
=
A*
l ny g(p,-0)+g6(o,-p,)
(4)
=
3 l
pA where (5) and Q'
R.
=
=
OC hC c pe e pe i
24-2
{
_,_.y--,
_,,...----,.y#,-
~~ -
.~.
~
J-The solution for this system is:
(6)
T, = C,e*1# cosz,t + C,e*8# sinz,t + C,t + C,,
-l e*1# cosz,t T,=fC,+R(Cz,+CZ) 3 3 22 (7)
C, + R,(C,z, - C,z,) e*1# sinz,t + C,t + R,C, + C,,
+
C, R,
R, (3)
T, -
T, O't T, =
7-
+
g, g,
where C, = - C,z,/z, - C,/2, C, = T, - C, C,=T,(R,+R,+R,)
2 C, = C,/c R, - bQ'/c R, C, = Q'/cR
+1 c = R,/R, + R,/R a = R,R,,
b = R, + R, + R,R,/R, (4ac - b )*/*/2a 2
z,= - b/2a z, =
For the computations Eq. (4) was coupled to Eq. (6) through the fluid equat
~
Several numerical examples were state and solved by successive approximations.
considered to determine the effect of the assumed natural circulation flow patte and of the core pemeability (which would be relattid to the degree of degrad
~
Geometric and mass data used are typical of a Westinghouse 4-loop plan The assumed).
In all cases primary steam was at 2300 psia and decay power was set at 42 M results are sumarized in Figures 2 and 3.
d For the " loop circulation" case it is assumed that the loop seals at the pump a lower plenum locations are broken and a steam flow path is open all around The masses m,, m,, and m, are identified with the core, upper 3
No heat primary system.
" internals including upper head, and steam generator masses respectively.
Frictional losses losses to the outside of this three mass system are considered.
f in the core region dominate and they were represented by a pemeability v l
which corresponds to a rubbled core with a characteristic partic e 1.5 x 10"' m 8
h dimension of 8 m and includes a factor of 1/2x for deviations from spheric A parametric using a higher pemeability by a factor of 5x was also co The calculations yield a nearly steady represent a less disrupted core geometry.
This flow is sufficient to redis-natural circulation flow of 25 kg/s per loop.
initial tribute the decay heat such that the three mass system heats up after an* 'C/s (Fi transient of -1 hr., essentially with a unifom rate of -1.7 x 10*
The upper internals follow closely the core temperature while the steam l
24-3
_ _ ~
. 1
~
lags by -150, 70'C for the two cases considered.
1 is e
so -
Y, Ge.
y T,
.e
'S T,
=4 a
T a
To
'g "e
T, T,\\
\\
k
'e a
e a
to t a 10-e(,)
4.
, [j{,
T, -
)
Figure 2.
Predicted temperature tran-i sients for the " loop circulation" case.
ie so; t a 10-e(,)
Figure 3.
Predicted temperature tran-sients for the "in-vessel recirculation" case.
Although some code calculations [2] do yield opening of the loop seals it is generally recognized that significant uncertainties in predicted behavior exist in To complement the behavior depicted in the previous paragraph in this this area.
The flow pattern regard the "in-vessel recirculation." case was also considered.
envisioned is shown as an insert to Figure 3.
Now the three masses of our model were identified as the core, one-half of the upper internals (central region), and the other one-half of the upper internals (outer region). The same two cases of For the nominal core permeability as in the loop circulation case were considered.
case a gradual increase in natural circulation flow to a steady value of 45 kg/s over a period of 2000 s was calculated. Again, these flows provide substantial themal coupling between the core and upper internals masses. However, as seen in That
. Figure 3, this coupling is not as efficient as in the loop circulation case.
is the rate of heatup is different for the three masses and at 10,000 s the outer portion of the upper internals lags by 120 and 60'C behind the core temperature for the two car,es considered. Furthermore, as illustrated in Figure 3 these high 24-4 4
V*f *.'
- "*2.,
__.,;'"[*^",,,
,l[
[.Y
.,,A,
,__.,,_l,_.,_,,,,,,,n___,.',
_,_,,,m.,
. e.
I.
l
.t-#*
~
- i. ',
-_e-i y
temperatures would propagate into the hot leg, towards the steam generator, and pressurizer lines, by net flow caused by relief valve cycling as well as by
. counter-current steam f ow in the hot leg (4}.
l These scoping studies indicate that primary system temperatures would follow the core temperatures during heatup within "a few hundred degrees Celsius" such that primary system boundary failure prior to core melting and slumping would be Such failure would be due to loss of strength of structural material expected.
expected at around 650*C, i.e., would produce leak areas sufficient to depressurize Furthermore if conditions for the loop circulation case were to be the system.
present, concerns about steam generator tube failures (by the same mechanisms) at high pressure should be addressed. On the other hand, conditions leading to clad ballooning to such an extent that blocking of the flow paths in the core occurs could also significantly alter these predictions. In such event phenomena of clad oxidation, melting, and relocation (which will reopen the paths), and fission product relocation within the primary system (and of associated thermal loading) must also be considered together with the natural circulation flows discussed here.
AN EXPERIMENTAL ILLUSTRATION Purdue's Large Scale Debris Bed Coolability Simulation Facility [5,6] was adopted to illustrate the energy redistribution mechanisms for the "in-vessel recircula-The present bed measures 21 cm in diameter and 100 cm in height and is tion" case.
packed by alternating layers of 1.25 cm aluminum spheres and a layer of similar thickness of 0.8 cm irregularly shaped stone fragments (gravel). It has a porosity 2 which is close to the of 3.86 and a turbulence permeability of 1.4 x 10" m pemuoility of a bed made up only with gravel and about one-half that of a bed of equal size spheres and the same porosity. The power to each layer of aluminum 3
spheres is individually controlled such that any vertical portion of the bed can be volumetrically heated at will. For the experiment described herein a total of
]
5.7 kw were applied over a bed height of 30 cm extending from an elevation of 23 cm The bed was flooded with (from the bottom of the bed) to the 53 cm elevation.
The tempera-20*C water and the experiment lasted until the bed reached boiling.
ture transients were measured by means of 250 thermoe:wples appropriately distri-ll buted throughout the bed. These thermocouples were scanned every 5 s with the p
help of a PDP-11 minicomputer.
y The results are shown in Figures 4 and 5.
From Figure 4 we can clearly observe the conduction regime and a rapid transition to convection which is of sufficient The predictien of the strength to quickly homogenize (thema11y) the whole bed.
simple model already discussed for which convection was taken to initiate at a i
Observi:d spatial critical Rayleigh number of 30 is also shown for comparison.
24-5
I s. '-.
s temperature distributions at selected times are shown in Figure 5.
This figure contains also predictions based on the numerical model described in the next section.
See 7
I re 7
G e
ee.
z es
'i e
4 e
e se se t a to*8 e Comparison of 2. mass lumped Figure 4.
parameter model predictions with experi-mental results.
19 0
.o gen f*
- i. eve.
g ee.
t i.....
n g,,
E O 's e.se e.e e.r e i.e ZlH Figure 5.
Comparison of numerical model predictions with experimental results.
24-6
~
e j
<o j
i A NUMERICAL MODEL by means of a distributed The "in-vessel recirculation" case was further examinedThe who heated while that corresponding parameter model as follows.
medium, the region corresponding to the core beingThe pore space is oc The vessel, as a whole to the upper internals being inert.
free to convect as buoyancy forces develop from heating.
Friction is governed by Darcy's law, and density varia is adiabatic.
handled in the Bussinesque approximation.
ticles is commonly includes themal nonequilibrium between the fluid and the par ssed as:
used for natural convection in porous media and may be expre(9) 8" 3_v-+1=0
-- + - ay r
ar (10) 8E-- + 1 u = 0 ar c
(11) h+og-S(T-T)go+p=0 g
g 3
g BT -. = 6 - h S(T -T) p g
(1 - c) p c p
3T 1
BT (12)
+k(1-c)f3T-P- + - E - + r.dk 8
2 ar J
gy:
2 P
1 Br BT BT cogg[T B
=
e c
gf
+ vog g
+ up c
{13)
L+
+
= hS(T -T)+K*
p g
6 i
lls. This model The boundary conditions are obtained for impermeable adiabat c wa i
was solved in nondimensional form by the fin teat each time step the energy equatio differencing and successive over-relaxation.f the exchange tems), while were solved explicitly (with implicit coupling o a marker and cell technique.
i momentum and continuity were solved simultaneously us ng i
step, and con-For all calculations reported here a number of nodalization, t methe r vergence criteria, studies were perfomed to assuretions may be used in i
Clearly, appropriate porosity, pemeability, and power distr bu Two cases are considered here.
detailed evaluation.
24-7 e
,S**
f*','
~
s0
,.i-Calculations for.the demonstration experiment discussed in the previous section were carried out using the measured bed porosity of 0.386 and a permeability of 1.5 x 10~' m'.
From the~ comparison of Figure 5 we deduce that all essential l
features, including the conduction heatup regime, the time for the onset of convec-I tion, and the rapid transition to uniform heating following the onset of convection, are adequately predicted. In addition, for a short period following the onset of convection the model predicts a turnover similar to that observed in thermals with sufficient cold fluid penetrating the heated region to temporarily invert the ver-tical temperature gradient. Such behavior was qualitatively observes also in the experiments. However, due to non-axisyninetric behavior in the experiments precise
}
interpretation of this aspect must await further investigations.
i Reactor calculations were carried out with a uniform porosity of 0.5, a core power of 42 MW and a radial power distribution given by:
2.32 J.(4.81h)
(14)
=
A low permeability value of 1.5 x 10~' m2 was chosen to maximize the thermal gradi-ents calculated. An intact core geometry would be characterized by significantly
~
higher permeabilities. Selected results of one such calculation are shown in Figures 6 and 7.
In figure 6 we ~ observe the unicellular flow pattern envisioned in our simple model presented earlier. In fact the predicted flow rate is -50'kg/s which is in gcod agreement with the recirculation rate of 45 kg/s produced by the I
3-mass model. The time-wise variation of the spattal temperature distribution may be visualized as in Figure 7.
The isotherms are nearly vertical and seem to be continuously (in time) displaced outwards as new higher temperature isotherms are generated near the centerline. However, the maximum temperature difference at i
10,000 s is only 240*C, and most of the mass is within 150*C which is in good agreement with the 120*C obtained in the 3-mass model.
CONCt.USIONS High pressure steam natural circulation phenomena can be responsible for redistri-buting the core decay power to all primary systems components that are accessible to form circulating loops with the core region. As a result the primary system boundary is expected to fait prior to core melting, and the so-called "high pressure scenario" would appear unlikely. The possible role of clad ballooning, I
and oxidation remains to be assessed before these conclusions can be applied to all possible variations of a Station Blackout (or other similar) accident.
24-8
'de-
.Ws *
.z '
p..
o.
't
- \\
Iin----..,
u r
h Iss-...'
- a.
2 --
i lI//s..,I e
t ti.u 4 /<.,g1 e
fe.,g o
-6//.s\\
=
fp s}
a st r
,s.g.
j.,j l.l i
f, iqis.,'j
)\\\\.
\\ s.,,
a
\\\\s.,j t/
11\\\\s
>a t
% n -.. !
A Velocity vector plot at time = 1416 s.
Figure 6.
Figure 7.
Isoterms at 1416, 2832 and 9912 s.
ACKNOWLEDGMENTS The financial support by the U.S. Nuclear Regulatory Comission under Contra NRC-03-83-093 is gratefully acknowledged.
24-9 a
Qy M
p._
4
- g. 9 NOMENCLATURE A
= flow area
= specific heats of coolant and of structure Cg.Cg D
= vessel diameter
~
g
= acceleration of gravity h
= heat transfer coefficient between fluid and particles k,k
= themal conductivity of fluid and effective thermal conductivity of fF porous medium
= coolant flow rate 0
m
= mars of 1 compartment g
p
= gressure Q
= volumetric heat generation rate radial coordinate r
r S
= the surface area between the fluid and porous medium per unit volume 0
T
= temperature of 1 ' compartment g
t
= time u
= velocity component in r-direction v
= velocity component in y-direction y
= coordinate in vertical direction Greek Symbols 8
= themal expansion coefficient i
= average density of fluid u
= viscosity n
= frictional length / elevation length 6
= h:3/h33, where h,and h: are shown in Figure 1 a
e
= permeability of the porous medium i
REFERENCES r
1.
V.E. Denny and B.R. Sehgal, " Analytical Prediction of Core Heatup/ Liquefaction /
Slumping." Paper TS-5.4, Proceedings Intl. Meeting on LWR Severe Accident Evaluation, Cambridge, MA Aug. 28-Sept.1 (1983).
2.
L. Winters, "RELAPS Station Balck-out Transient Analysis in a PWR." ECN Memo No. 8.904.00-GR17. July 1982.
3.
T.G. Theofanous and Chien-Hsiung Lee, "The Direct Heating Problem," Presentation to the Containment Loads Working Group Meeting, Rockville, MD, March 1984.
4.
T.G. Theofanous et al., " Decay of Buoyancy Driven Stratified Layers with Applications to PTS," NUREG/CR-3700, May 1984.
5.
K. Hu, P. Gherson and T.G. Theofanous. "The Large Scale Simulation of Debris Bed Coolability " A!ChE Symposium Series 236. Vol. 80, pp. 380-384, 1984 6.
K. Nu and T.G. Theofanous, " Scale Effects and Structure of Dryout Zone in Debris Bed Coolability Experiments " These Proceedings.
24-10
.