ML20235M397

Summarizes NRC Concerns W/Idcor Analysis of Mark I Response to Direct Contact W/Molten Core Debris,Including Need for Sensitivity Study on Melt Temp & Exam of Effects of Larger Pool Depths.Response to Items in Idcor Encl

ML20235M397
Person / Time
Issue date:	10/02/1987
From:	Sheron B NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
To:	Buhl A INDUSTRY DEGRADED CORE RULEMAKING PROGRAM, INTERNATIONAL TECHNOLOGY CORP.
References
NUDOCS 8710060176
Download: ML20235M397 (4)

Text

- - . ._ . .

OCT 2'1987' e

Dr. A. R. Buhl IDCOR Program !!anager c/o Internatienel Technology Corporation 575 Oak Ridoe Turnpike

. Oak Ridge, Tennessee 37830

Reference:

. Letter dated July 9, 1987 from A.'R. Buhl to Eric S. Beckjord;

, .Resrense to Brookhaven Naticnal Laboratory's Comments on IDCOR's ,

fiodel of the Mark I containment Response to Direct Contact with I Molten Core Debris..

We have reviewed the content of the reference-letter on BNL's evaluation that was ettached to the 11 arch 27, 1987 meeting minutes, regarding the adequacy of IDCOR's model of a tierk I response to direct contact with molten core debris.

Responses to each item in your letter are outlined in the attachment to this j letter.

In summary, let me reiterate the staff concern with the IDCOR analysis.

(1) The melt. temperature is assumed to be the solidification temperature without sufficient justification. A' sensitivity study on the effect of this on the IDCOR result is needed.

1 (2) The molten core composition was essumed to be U0,3 with 1cw thermal '

conductivity; however, it should contain a substantial frection of metal which has substantially higher thermal conductivity than the oxide. j l' (3) The molten pool depth was considered in the sensitivity enelysis for depths below 12 cm but not above. The effect of larger depths should be examined.

(4) A consistent set of-bourdary conditions for heat transfer from the molten pool needs to be applied in the analysis; i.e., decomposition gas egita- 1 tion of the pool, heet transfer coefficients to'the pool boundary. con-sidering gas egitetion and radiation. j l

84* Repon s e endjp "

"ER

4 The staff believes thet a rance of initial end boundary conditions should be arployed in a sensitivity analysis prior to dismissing the containment shell meltthrcech as a non-issue.

If needed, we could discuss your concerns further in the technical exchange meetino, ,

l Sincerely, 1 Offb'$$308f Bilen W. $bgran Brian W. Sheron, Director Division of Reector and Plant Systems i Office of Nuclear Regulatory Research l

Attachment:

As steted l I

Distribution l PDR . >

RES Chron/ Circ File DRPS Chron File RFFB Reading File B. SFeron ,

U. Minners l F. Coffman l F. Eltawila i T. Speis  !

D. Ross W. Houston l

l l

• Please see previous concurrence.

n J OFC :PsHFB :DRPS :RHFB:DRPS :FD:D :D:DRP  :  : : )

.....:...__.......:..........__: ,g.....:. .. . .....:.......... _:._.. __ ...:......_...

NAME :*FEltawile/as:*FCoffnan :W nners :BSh 'on  :  : :

DATE :09/ /87 :09/ /87 :10/I/87 :10/,2./87  :  : :

OFFICIAL RECORD COPY l

L___ _

Attachment Resonse to IDCOR Letter dated July 9, 1987 Item (a): ,

In the reference letter you stated that 6 cm corresponds to almost 50% of the core debris whereas in the original analysis, you stated that 50% of the Peach Bottom core debris was deposited on the drywell floor and this resulted in a liquid depth of 12 cm. Not withstanding this inconsistency, other' analysis indicates that when 50% of the core debris deposit on the drywell floor 12 cm debris depth and not 6 cm would result. In addition, it is not clear to us why 50% of core is the upper bound, particularly since the analysis assumes that the molten core was uniformly dispersed and does not account for porosity of the debris, additional molten reactor vessel structures and the potential for large accumulation outside the pedestal door. Without parametric analysis to illustrate the effect of layer thickness, one can not assume the analysis are done conservatively. l It was also stated that the debris is assumed to be U02 with a low thermal

,i conductivity, fully molten and at its freezing point. The staff believe that the molten debris would be a mixture of oxide and metallic components with substantially higher conductivity and temperature might be above or below the freezing point. It is the staff's assessment that the IDCOR analysis can only model U0 at its freezing point because it neglects the pool side heat transfer l 2 I i coefficient without an adequate justification and, therefore, is too simple to l l be credible for all other possible conditions, j Item (b):

The IDCOR model does not allow the debris to be molten at the surface. This prevents convection of heat to the steel liner, a heat transfer mechanism far

'g more effective than conduction in a solidified mass. In addition, the energy 4 balance does not account for the latent heat of UO2 -

t Item (c):

I The staff disagrees that melt can exist on the concrete for tens of seconds without decomposition gas generation and hence no consideration of chemical

! reaction. It may be that the chemical reaction energy is small compared to  ;

decay heat, but that depends on the erosion rate and the thickness of the  ;

debris layer, which have not been addressed by IDCOR. 1 Item (d):

1 The SWISS tests suggest that the heat flux was in gas-flux enhanced film boiling, far below the critical heat flux. However, experiments at BNL clearly demonstrate the validity of a regime where water does not quench a molten pool.

Also, these experiments show no crust formation until T . Thus, boiling is sustained for the duration of the BNL tests. pool = Tfreezing.

1 (1) R.E. Blose et al., " SWISS: Sustatined Heated Metallic Melt Corferete Inter-action with an Overlying Water Pool Experiment and Analysis," SAND 85-1546, September 1986.

(2) G. A. Green, " Heat Transfer in Core-Concrete Interactions: Liquid-Liquid Film Boiling," Brookhaven National Laboratory, Quarterly Progress Report, NUREG/CR-2331, May 1987.

l

\ - -

2

. Item (e):

The staff believes that the outer shield wall boundary condition should be more appropriately modeled as adiabatic. It was claimed by IDCOR that this assumption does not matter but no justification was given for t411s conclusion.

Axial conduction away from the hot region of the steel liner is effective in the IDCOR analyses because of the relatively small amount of heat conducted to the liner from the U02 crust because of the assumed low thermal conductivity.

Item (f):

The staff believes that it is inappropriate to neglect thermal radiation heat transfer effects. In the IDCOR model, radiation from the pool of UO2 to the liner is neglected even for the case with no water.

i In addition the staff believes that the IDCOR model may overestimate the cool-ing effect of an overlying water pool. The liner exposed to water transfers heat at nucleate boiling limits. If the debris were to accumulate locally l along the liner, having flowed out of the pedestal in a lava-like manner, to a foot or more, water could not remain over the debris: it would ficw into the wetwell vent pipes. Furthermore, nucleate boiling would only occur if the heat transfer axially in the liner were less than film boiling could handle,  ;

thus resulting in a wall quench, limited by U02 conduction.

Item (g): l l

The staff believes that the analyses sponsored by the NRC are conservative since they require the liner to reach its ablation temperature before failure is assumed. It is possible that steel shells at elevated temperatures may fail by mechanical creep at 1500 C, this failure mechanism was not considered by IDCOR.  :

The staff agrees that the IDCOR analysis points out a valid qualitative physical h effect of axial conduction along the steel liner wall to the water pool above.

+

However, the quantitative effect is largely affected by the initial melt pool

! temperature, the melt pool composition, and the boundary heat transfer coefficients. Without parametric analysis, we can not conclude on the adequacy of the IDCOR treatment of this issue.

The final IDCOR comment pointed out the conservatism of instantaneous and uniform deposition of the molten core on the pedestal floor. However, this may not be as conservative as one would first imagine. The geometry of the Peach Bottom drywell indicates that there is only one exit from the pedestal region through a doorway. Therefore, once the molten core begins to be discharged from the vessel into the drywell, there will be a preferential directional flow out the doorway of the pedestal region. This means one would not need 50% of the core to be deposited before the steel liner wall near this doorway would be subjected to this thermal attack. So one must be careful to portray this as a bounding calculation without sufficient sensitivity results to demonstrate this. .

l l

~ '

- _ - _ . - - _ _ . _ _ _ . - -