ML20248F053

Review of In-Vessel Coolability & Retention of Core Melt DOE/ID-10460, Jul 1995

ML20248F053
Person / Time
Site:	05200003
Issue date:	05/19/1997
From:	Behbahani A, Yen-Ju Chen, Rubin A NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
To:
Shared Package
ML20248F024	List: ML20248F021 ML20248F053
References
NUDOCS 9806040084
Download: ML20248F053 (50)

Text

- - - - - - _ _ _ - _ _

s REVIEW OF "IN-VESSEL COOLABILITY AND RETENTION OF A CORE MELT" DOE /ID-10460, July 1995 l

by:

A. Rubin Y. Chen A. Behbahani S. Basu i

Accident Evaluation Branch Division of Systems Technology Office of Nuclear Regulatory Research i

May 19,1997 h N$K ob 3

Enclosure

-. n b

l l

Table.of Contents Introduction and Statement of Problem 1

Technical Review 3

Accident Sequence 3

Internal Heat Load on Vessel Wall 3

Amount. Com)osition, and Stratification of 3

l Melt in tie Lower Plenum Decay Heat Level 7

Molten Pool Natural Convection 10 Heat Transfer in the Metallic Layer 11 External Heat Transfer via Ex-Vessel Flooding 12 Margin of Safety 13 Reactor Pressure Vessel Lower Head Structural Integrity 14 Baseline Calculations 14 Uncertainly end Sensitivity Studies 16 Authors' Responses to Reviewers' Comments 19 Potential Open Issues 20 Structural Integrity of Insulation 20 Melt-Vessel Interactions 20 Stratification of Metallic Layer 20 Conclusions 21 References 23 Appendix A. Sensitivity Study of Melt Quenching During Slumping Appendix B. Amount and Composition of Relocated Melt Calculated hy SCDAP/RELAP5 for the Case 3BE Transient Appendix C. Decay Power Calculations for AP600 Case 3BE Transient

INTRODUCTION AND STATEMENT OF PROBLEM As part of its accident management strategy for the AP600 reactor design.

Westinghouse Electric Corporation has provided the capability to flood the reactor cavity, thereby, submerging the reactor pressure vessel (RPV) and providing ex-vessel cooling to the RPV in the event of a severe accident.

This strategy is intended to remove decay heat from the high temperature melt

(-3000K for oxidic melt and -1800K for metallic melt) that could relocate to the RPV lower plenum and keep the lower head sufficiently cool so that the j

vessel does not fail.

Maintaining RPV integrity would prevent core debris from relocating to the reactor cavity, thereby, avoiding subsequent ex-vessel phenomena that could challenge the containment (e.g., ex-vessel fuel-coolant interactions, core-concrete interactions).

As part of the certification process for the Westinghouse AP600 design, the Department of Energy (DOE). through Westinghouse. has submitted report DOE /ID-l 10460. "In-Vessel Coolability and Retention of a Core Melt" A major I

objective of this report was to demonstrate the effectiveness of the concept of ex-vessel cooling for an AP600-like reactor design.

The Accident Evaluation Branch (AEB) of the Division of Systems Technology in the Office of Nuclear Regulatory Research has reviewed this report.

This review focused on the factors discussed in the report that impact lower head integrity. These include:

(1) thermal load distribution on the inner surface of the lower head. (2) critical heat flux (CHF) and heat removal outside the vessel from ex-vessel flooding. (3) mechanical loads on the lower head, and (4) vessel wall thickness.

Based on this information, one can determine the integrity of l

the lower head by comparing the mechanical and thermal loads and vessel wall thickness to the CHF distribution and the minimum wall thickness needed to ensure the structural integrity of the reactor pressure vessel.

The results l

of this technical review are presented in the following section of this l

report.

The discussion includes related analytical and experimental results (including scaling issues) and the range of uncertainty / sensitivity for key l

parameters. The overall conclusions regarding the likelihood of success of ex-vessel cooling for the AP600 are presented at the end of this report.

1

In addition to thermal loads, there are other factors that may impact lower head integrity but were beyond the scope of the DOE report.

For example.

loads on the RPV from in-vessel steam explosions were not addressed in DOE /ID-10460.

(They are addressed in a separate DOE report.) Also, if the structural integrity of the insulation surrounding the AP600 reactor vessel were not maintained, ex-vessel heat removal could be affected. These and other potential open iss'ues that may imoact lower head integrity are discussed later in this report.

As described in DOE /ID-10460, three basic assumptions are made regarding the definition of the problem being addressed.

First, if core degradation takes pl

e. the rimary systen is acumed to be depressurized.

Seconi it is assumed that the lower head will be fully submerged with water in the cavity prior to the arrival of core debris in the lower head; and third, the water level in the cavity will be maintained indefinitely.

However, consideration of these important aspects of the AP600 design is beyond the scope of the DOE report, and therefore is also beyond the scope of this review.

In particular, with regard to the second assumption we found that MAAP and SCDAP/RELAP5 calculations indicated that the first relocation of core debris to the lower plenum could occur between 2 to 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> after initiation of a double-ended break of one of the direct vessel injection lines (case 3BE accident).

The DOE report uses a lower bound value of just over 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> for the first relocation event.

Therefore. the capability of assuring that'the RPV lower head can be submerged with water in the cavity prior to the arrival of core debris in the lower head should be verified.

We understand that NRR is reviewing core melt progression operator actions and response times to initiate cavity flooding, and reactor cavity flooding rates and tines for a spectrum of sequences and system failures. This review will confirm that operator actions to initiate cavity flooding will be taken early enough in the event to assure that the cavity will be flooded to a sufficient level (above the expected level of the debris pool) in advance of l

core debris relocation into the lower head. This review will also assure that external reactor vessel cooling is not credited in sequences where timely flooding of the cavity cannot be assured.

2

It should be noted that even if in-vessel retention of core melt were successful, other severe accident challenges, such as hydrogen combustion and in-vessel melt-coolant interactions, could still occur even if the RPV lower l

head does not fail due to thermal loads.

l TECHNICAL REVIEW ACCIDENT SEQUENCE l

Among the accidents contributing to the core damage frequency (CDF) of the l

AP600 as developed in the Level 1 PRA. the case 3BE accident transient, which l

has been of main interest to DP. was selected in one DOE study. The accident was postulated by initiation of a doubie-ended off-set break in one of two direct vessel injection (DVI) lines. Gravity injection from the in-containment refueling water storage tank (IRWST) was also disabled.

One train of the automatic depressurization system (ADS) and the passive residual heat l

removal (PRHR) system were also assumed to fail.

As a consequence, depletion of the in-vessel coolant inventory and core heatup occurred followed by core degradation and in-core molten pool formation.

The molten pool continued to j

grow until the crust surrounding the pool failed at the side. The molten l

material then melted through the reflector and core barrel and eventually relocated to the lower plenum.

INTERNAL HEAT LOAD ON VESSEL WALL Amount, composition, and stratification of melt in the lower plenum. Various l

molten-pool and/or debris-bed configurations could be formed in the lower plenum depending on the melt mass, composition and natural circulation in the pool.

In the DOE report two important configurations have been identified.

One configuration was dominated by natural convection in the final steady state (Figure 1), and the other was dominated by forced convection and jet impingement effects (Figure 2).

The second configuration, shown in Figure 2 (i.e.. Figure 2.3 of Reference 1), illustrates a " wall jet" released near the top of the molten core pool, through the downcomer, and into the lower plenum.

3

t Significant forced convection effects can arise as the core melts and relocates into the lower plenum, primarily during the major relocation of melt.

Based on the underlying assumptions in the DOE report. 1t was concluded that local wall failures will not occur due to jet impingement.

The authors of the report calculate the jet impingement heat transfer under rather conservative assumptions such as ignoring the subsequent relocation of melt into an existing corium pool.

The report utilizes correlations for heat transfer from an impinging turbulent jet by Saito (Reference 2).

It should be mentioned that more recent experimental studies on jet impingement at the Swedish Royal institute of Technology by Sehgal, et al. (Reference o> confirm the accuracy of Saito's heat transfer correlations.

The report presents results of calculations of the vessel wall's ablation depth as a function of the impinging jet diameter for a relocated melt volume of 2.5 m (about 1/3 of the 3

total fuel volume) (see Figure 8.1 of Reference 1).

Even this amount of melt pour with a jet pour diameter of 10 cm. would not result in vessel penetration through jet impingement.

We also agree with this conclusion.

The end-state (i.e.

final steady state) configuration (Figure 1) was then considered to be of primary significance in the DOE study to assess in-vessel core melt retention. The key elements of this configuration are:

(1) all the core debris (either a solid oxidic crust or a molten oxidic pool) is at steady state in the lower plenum, and (2) a molten steel layer (consisting of material from the reflector and core support plate, a portion of the core barrel. and all lower plenum internal structures) is on top of the oxidic crust / molten pool.

Table 1 (i.e.. Table 7.2 of Reference 1) shows the material inventories in the reactor vessel that were used in the. study.

For this configuration, the authors' main concern was on the " focusing effect" which can cause high heat loads on the RPV wall from a thin metallic layer on top of the oxidic pool.

The authors performed a parametric study on the

" focusing effect" for various oxidic pool heights, thickness of the metallic layer, and volumetric power density of the molten pool.

The steady state configuration for the " extreme case" in the DOE report consists of a 17-ton 4

ie I

metallic layer with a height. H. of 0.22 m on top of a maximum oxidic pool i

height. H. of 1.18 m.

The total molten pool height is just at the bottom face of the lower core support plate (i.e.,1.4 m from the bottom of the lower plenum).

Using these heights, together with the value for the upoer bound of 3

the decay power of 1.4 MW/m used in the DOE report. the authors obtained the results shown in Figure 3 (i.e.

Figure 7.16 of Reference 1). This figure shows the ratio of the heat flux on the vessel wall to the external critical heat flux as a function of angular position from the bottom of the lower plenum. The maximum heat flux ratio of nearly 1.0 is reached at an angle of 65 degrees which is the level of the metallic layer for the case with an emissivity of 0.45.

However, the authors concluded that "it is doubtful that even this extreme specification could in reality produce failure.

This is because in addition to ignoring zirconium in the metal layer and elongation (and possible failure) of the core barrel, as just mentioned, this calculation has ignored 2-D effects due to (a) eddy diffusion in the bulk of the metal layer. and (b) conduction in the vessel wall.

In addition, the effect of highly localized thermal loads on the critical heat flux, also expected to be mitigative, has been ignored."

An independent SCDAP/RELAPS analysis of case 3BE transient for the AP600 design was performed by INEL (Reference 4).

Results of this study indicate that throughout the entire transient the thermally induced vessel wall failure of an externally flooded AP600 reactor will not occur. This INEL study was based on two assumptions.

First, the molten pool in the lower plenum was assumed to be homogeneously mixed and includes U0. Zr0. Zr stainless steel, 2

2 and the control rod absorber material (Ag-In-Cd).

(The limitations of the study with this assumption are discussed below.) Second, no heat transfer (i.e., no quenching) between relocating melt and water during slumping was l

assumed.

[To address the melt quenching phenomena during slumping, two l

additional sensitivity studies were also completed by INEL (see Appendix A)..]

2 The 17 tons of metallic material in the lower plenum consists of -10 tons of melted reflector. -5 tons of melted core barrel, and -2 tons of lower plenum support structures.

(Note that the estimated amount of the melted reflector and core barrel corresponds uo 25% of the core height.)

5

+-

The results of the INEL study in Reference 4 do not provide direct answers to address the thermal loads due to stratification of the metallic layer I

stratification (i.e., " focusing effect") in the molten pool.

However, the results do provide useful information concerning the behavior of in-vessel

{

melt progression, including the amount and composition of the melt relocation during the course of the accident. Table 2 shows a summary of the calculated amount composition and timing of melts that relocated to the lower plenum l

during the course of the accident. The amount of reflector and core barrel material that slumps to the lower head during the transient is also included l

in this table.

(Appendix B gives additional details on the calculation of l

this slumping material.) Six relocations containing molten fuel and five relocations containing control rod absorber material (Ag-In-Cd) were l

calculated.

Five of the six molten fuel relocations occurred as a result of l

side failures through the stainless steel reflector and core barrel. The sixth (and final) relocation was calculated to occur when the molten fuel (and j

l other core components) drained into the lower plenum through the bottom of the core.

l Based on the results shown in Table 2. it can be seen that at the time before the final melt relocation (i.e.

13.252 seconds), the molten pool in the lower plenum contained a total of about 42.000 kg of oxidic melt and 15.000 kg of metallic melt.

Based on the total of 57.000 kg of molten material, the j

estimated inolten pool height at this time is about 1.19 m.

At the beginning of the transient, the lower plenum was full of water.

But, because of melt I

slumping into the lower plenum, the water boiled off.

By the time of the fifth relocation, the estimated height of water above the molten pool was 0.21 m.

Boiloff continued until the lower plenum dryout occurred before the final melt relocation.

It is possible that prior to the final melt relocation, there could be sufficient time for natural convection to develop l

in the lower-plenum molten pool, and during this time the lower density metallic components could rise up and separate from the oxidic pool in a l

stratified layer. This stratified molten pool is a possible intermediate state that could impose high thermal loads to challenge the vessel wall.

Based on the results in Table 2. for this intermediate stratified state, the estimated oxidic pool height is about 1.01 m and the thickness of the l

6

~

(

metallic layer is about 0.18 m.2 This metallic layer of 0.18 m is larger than the threshold value of 0.15 m given in Reference 1 as the thickness necessary to insure vessel integrity under " focused" heat loads.

Thus, it can be concluded that the thermal loads to the pool boundaries throughout the core melt transient are bounded by the thermal loads in the final steady state provided that the decay power density of the molten pool in the lower plenum is calculated correctly.

(See the discussion below on " Decay heat level" for the threshold value and decay power density.)

Decay heatlevel. In this review we found that the total decay power. shown as a solid curve in Figure 4 (i.e.

Figure 7.1 of Reference 1) correctly l

includes a 2a uncertainty.

(See Appendix C for additional discussion.)

However, we have two concerns regarding the " Decay Power in the Pool" shown as a dotted line in Figure 4.

One relates to the shape of the curve. and the l

other to the timing for initiation of fission product release.

The "In-pool" curve takes into account the loss of volatile fission products during the course of core melt accidents. Thus. the "In-Pool" curve was derived from the fractional contribution of the non-volatile components to the total decay power as a function of time after shutdown (see Figure 5. i.e..

Figure 7.2 of Reference 1).

Note that Figure 5 was obtained from Schnitzler's report (Reference 5). According to Reference 5. Figure 5 was based on the modeling of a small burst of fission product release at the time of cladding integrity loss and a major release due to fuel fracturing at quench.

In the Conclusions Section of Reference 5. the statement was made that "the available data cannot be applied with confidence in cases where the fuel failure mode (and resulting volatile release) differs substantially from the fuel heatup and quench failure mode assumed in this evaluation." Strictly speaking.

Figure 5 is only applicable for a scenario with fuel cladding failure 2

The thickness of the metallic layer is calculated assuming that the control rod absorber material (Ag-In-Cd) is not part of this layer.

(The Ag-In-Cd is denser than the rest of the metallic material and could remain in the oxidic pool.) Since a thinner metallic layer imposes higher heat loads on the l

RPV wall (i.e.. " focusing effect"), neglecting t1e control rod absorber material in the metallic layer is conservative.

7

occurring at decay time of 2 minutes and with fuel rod quench occurring at decay time of 12 minutes.

Therefore the decay power fraction curve shown in I

Figure 5 is not applicable to the slow boilcff, core melt progression transient which is the case 3BE scenario in the DOE report.

With regard to the timing for initiation of fission product release, the results of the INEL study indicate that the first fuel clad failure occurred at 4.455 seconds instead of 220 seconds as shown in Figure 4.

Thus, the entire "In-Pool" curve should be shifted to the right, and the magnitude should also be revised accordingly.

Since Figure 5 cannot be applied to the Decay Power curve (i.e., solid line) of Figure 4. the exact magnitude of this "In-Pool" curve needs to be determined based on the available data.

In this review, we obtained new decay power fraction curves (see Figure C-3 of Appendix C) from INEL. The calculated molten pool volumetric power density based on these curves is shown in Table 3.

For Case 1. no fission product release after fuel melting. the volumetric power density is about 1.16 MW/m 3

(i.e.. ANSS.1 + 10% case), which is less than 1.4 MW/m.

(Note that the value 3

of 1.4 MW/m is significant since it was used to derive the threshold value of the metallic layer thickness (i.e.

0.15 m) necessary to insure vessel integrity under " focused" heat loads.) The power density for Case 2. assuming fission product release after fuel melting (as if fuel remains intact.) is also shown in Table 3.

The difference in the calculated volumetric power density between Case 1 and Case 2 is only about 6%. A realistic value will be somewhere between 1.16 MW/m and 1.09 MW/m'. Thus, a mean value of 1.13 MW/m 3

3 power density will be used for the following discussion.

As discussed previously, the estimated amount of molten U0 just before the 2

final melt relocation at 13.252 seconds into the transient is about 37.000 kg.

which is about 50% of the total UO in the core.

The total relocat'd molten 2

material is about 57.000 kg. an equivalent volume of 7.04 m'.

If stratification of the AP600 lower head molten pool is assumed (i.e. the I

metallic layer is on top of the oxidic melt) the oxidic volume is about 5.28 3

3 m

Also, based on a mean value of 1.13 MW/m for the entire pool, the l

calculated volumetric power density for the oxidic pool only is about 1,51 8

l

3 3

MW/m. which is higher than 1.4 MW/m The implications of tt'is finding are:

l (1) Figures 6 and 7 (i.e.. Figures 0.1 and 0.2 in Appendix 0 of Reference 1) l are no longer applicable to the subject stratification case (i.e.,

intermediate state), and (2) the metallic layer of 0.18 m, which was calculated under this configuration in the previous section, can no longer be used to compare the threshold value of 0.15 m.

With 1.51 MW/m power density, the new threshold value will be higher than 0.15 m.

Thus, the question of whether the threshold value would be higher than 0.18 m should be addressed.

This example and any other applicable cases need to be re-investigated by the authors of the DOE report.

l Table 4 lists the time for the first relocation event into the los head (as 4

calculated with the MAAP code in 1994).

For Case 3BE (i.e.. AP600 DVI line l

break accident), the value of this timing is shown as -4+ hours.

On page 7-9 of the DOE report, the authors also indicated that "An independent assessment j

of the timing of all these phenomena (Sienicki, et al.,1995) are in general agreement with the MAAP result, in Table 7-3. of 4+ hours. Thus. [the authors 0

of the DOE report] assign the main probability interval (~10 ) between 4 and 5 4

i hours, an unlikely range (~10 ) between 5 and 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />, and a very unlikely l

range of (~10 2) between 6 and 7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br />." As a result. Figure 8 (i.e.. Fig. 7-l 7 of Reference 1) was used together with Figures 4 and 9 (i.e.

Figs. 7.1 and 7.4 of Reference 1) to derive the decay power density probability distribution l

function as shown in Figure 10.

In this figure, the " extreme decay power density" of 1.4 MW/m' was obtained and used to develop the failure envelope l

for the " focusing" mechanism (see Figure 6).

However, results of MAAP calculations for Case 3BE (Reference 6) indicated that the first relocation event into the lower head occurred at 2.1 hours1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />.

The difference between 2.1 hours1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> and 4-5 hours is substantial, and accordingly, the impact on the decay power density probability distribution could be significant.

Furthermore.

SCDAP/RELAPS calculations indicated that the first relocation event occurred at 3.0 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />.

It can be concluded that a large uncertainty on the timing of the first relocation event exists.

Therefore, the validity of Figures 8. 9.

and 10 together with the failure envelope for the " focusing" mechanism needs r

' to be re-evaluated.

9 l

Molten poo/ nature / convection. The report assumes the primary system is depressurized during core degradation. The lower head of the RPV is filled witn molten ceramic and metallic material as a result of the downward relocation of the degraded core.

The molten ceramic material fills the lower head up to an angle of about 75 measured from bottom of the lower head (see Figure 5.1 of the report), and the space above the ceramic pool is filled with a metallic layer up to an angle of about 90.

The molten ceramic pool containing the volumetric heat source (i.e., decay heat) undergoes natural convection, thereby imposing a heat flux distribution

-on the boundary of the molten pool As a result of this heat flux discributici part of total hea.. generated in the molten pool is transferred to the inside surface of the lower head next to the ceramic pool (this part is called downward heat transfer, denoted by go), and the rest is transferred upward through the crust that is formed on the upper boundary of the molten pool (this is referred to as upward heat transfer, denoted by q,).

The upward heat flux, q,. is then transferred through the oxidic crust to the metallic layer.

Again, as a result of natural convection in the metallic layer, part of q, is transferred to the vessel wall and the rest is transferred to the upper boundary of the metallic layer, which is then radiated to the upper internals of the RPV.

To determine the local heat flux distribution on the boundary of the molten j

pool (i.e.. q, and q,) which dictates the thermal load on the inner surface of the lower head, one must rely on the available heat transfer data base for molten pool natural convection. The report provides a comprehensive summary of the experimental and numerical information available on this subject.

Various investigators have performed numerical simulations of molten pool natural convection in different geometries (i.e., hemispherical, semicircular and slice).

However, due to a lack of proper turbulent models for natural convection, these simulations underpredict q,.

Experimental results on molten pool natural convection in a hemispherical geometry are limited. Most experiments on the molten pool natural convection have been carried out in a 10

horizontal fluid layer or in a slice semicircular geometry at Rayleigh 23 numbers. Ra, of up to 10 with Prandtl numbers Pr. of about 7.

However, the Ra of interest for the lower head of the AP600 RPV is on the order of 10 or 25 ha e bee c rr ed o a

m her a geo tr e

ener n

the large Ra of interest.

One was performed at UCLA using Freon (Pr-8) as the fluid medium: the other was performed in the mini-ACOP0 test facility at UCSB for 2.6 < Pr < 10.8.

Recently, experiments have also been carried out in the ACOP0 test facility by Theofanous, et al. (Reference 8) for Ra up to 2 x 10 with water as the fluid medium.

Figures 11 and 12 present the 8

downward and upward Nusselt number as a function Ra as measured in this large scale test facility (4 scale).

The dotted lines c. these figures show a range of +/- 10% margins on the correlation for these experimental results.

1 The authors of the IVR report calculate the heat flux distribution imposed on the boundary of the molten ceramic pool based on their own, as well as other experimental results.

We believe the correlations used in the IVR report to calculate q, and q, are reasonable.3 Note that recent data from the ACOP0 experiments (See Fi 0res 11 and 12) reveal that some variation from previous 2

heat transfer correlations exist (i.e.. Nu, = 0.3Ra.22.

Nu,= 1.95Rac 22),

o t

However, this should not affect the margin to failure found in the IVR report.

It should be mentioned that sensitivity studies have been performed in the report using heat transfer correlations from other investigators. These i

sensitivity studies show no significant impact on the lower head integrity.

Further experimental results on molten pool natural convection will be available in the near future from other experimental test facilities (i.e..

OECD RASPLAV salt and corium experiments). These results can then be compared to the molten pool heat transfer correlation used by the authors for confirmatory purposes.

Heat transferIn the mete /#clayer. Heat transfer correlations in a fluid layer 2

at low Pr (-0.1) and typical Rayleigh number of about 10 for the metallic The correlations are Nu = 0.0035 Ra 35. Nu,= 0.345Ra and 0

o 233 equations 5.30a and 5.30b in th*e IVR report.

11

4.

layer are also limited.

One available heat transfer correlation is that of Globe and Dropkin (Reference 9).

lhe authors use this correlation for natural convection in a fluid layer and that of Churchill and Chu (Reference 10) for a vertical wall to calculate the heat flux to the vessel wall adjacent to the liquid metal layer.

The validity of the above formulations are supported by data obtained from the MELAD experimental facility at UCSB.

It should be noted that numerical simulation of natural convection in a fluid layer heated from below and cooled from the top and sides was carried out recently by Dinh.

l et al. (Reference 11) at the Swedish Royal Institute of technology. This

- study also agrees reasonably well with the above formulation. The thickness of the metallic layer, the upward heat flux. and the emissivity on the upper coundary of the metallic layer are important parameter s that ir. fluence the i

heat flux imposed by this layer on the vessel wall. The IVR report includes extensive calculations to satisfactorily determine the' heat flux to the side wall as a function of these parameters.

l EXTERNAL HEAT TRANSFER VIA EX VESSEL FLOODING An essential element needed to assess the limiting failure mechanism of the l

lower head is knowledge of the critical heat flux (CHF) around the external l

surface of the lower head.

Experiments have been performed in the UPLU test l

facility at UCSB to determine the CHF distribution on the external surface of i

the lower head. These experiments were carried out under pool and natural l

convection boiling conditions.

Figure 13 presents dimensionless CHF results of NRC-supported work at Penn State University (Reference 12) and results from l

the ULPU experiments. This figure shows that the results of ULPU tests, which are used in the DOE IVR report compare well with those of Penn State.

To clarify this point further. Table 5 shows the CHF values in both small-scale l

(i.e.. Penn State and SNL) and full-scale experiments (i.e.. UPLU: note appendices El-E4 in IVR report) at various angles of inclination. As this

)

L table shows, the results of both small and full scale experiments ~ at locations away from bottom center are close to one another.

However, a higher CHF is attained for the natural convection boiling case (which is the case for AP600 with insulation) as compared to that of pool boiling. This effect is more pronounced for angular positions above 70. thereby. compensating for higher 12

l

~

l internaT heat loads at these locations.

We believe the appropriate

]

correlations for evaluation of CHF for the AP600 design are those which take l

into account the effect of insulation.

Assuming that the structural integrity j

of the insulation remains intact, correlations that do not take into account the effect of the presence of insulation underestimate the CHF distribution on the outer surface of the RPV lower head.

MARGIN OF SAFETY The margin of safety for RPV integrity can be determined by comparing the internal heat load distribution to the external CHF distribution (i.e., local internal heat flux compared to the corresponding local external CHF).

Both the end and intermediate states in the accident progression must be considered l

to determine if a sufficient margin of safety exits for the AP600 RPV.

In either of these states, a homogenous pool mixture or stratified state may exist and should be considered.

For the stratified case. the metallic thickness must be above a certain threshold value in order to show that RPV integrity can be maintained.

The stratified state is potentially an area of concern since a thin metallic

(

layer on the top of the oxidic melt pool leads to a large " focusing" effect from this layer.

The metallic layer thickness depends on the melt progression scenario considered.

As discussed earlier in this report, an intermediate state could potentially exist with a metallic layer thickness of 0.18 m and a decay power density of 1.51 MW/m.

The threshold metallic thickness of 0.15 m (above which focusing effect diminishes) can be determined from Figure 6 (Figure 0.1 in the DOE IVR report) with an oxidic pool height of 1.01 m. a 3

3 power density of 1.4 MW/m. and an emissivity of 0.45.

However.1.4 MW/m may not be an " extreme" value for the intermediate stratified state.

Therefore.

in order to show that a sufficient margin of safety exists for this case.

additional analysis using the methodology in the DOE report is needed.

Fcr the end state, a stratified metallic layer (consisting of steel from the core support plate, portions of the reflector and core barrel and lower plenum support structures) would be several times thicker than that of the 13

threshold value used in the DOE study.

Even with a somewhat higher decay heat power density in the molten pool (e.g.. as a result of an earlier relocation of melt than the lower bound value of 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> used in the DOE report), we believe that for the end state. sufficient margin to failure exists and that RPV integrity can be maintained.

l REACTOR PRESSURE VESSEL (RPV) LOWER HEAD STRUCTURAL INTEGRITY l

Sese#ne Calculations. This section of the review focusses on structural l

analysis (Chapter 4 and related appendices and/or additional information) performed by the authors in support of their demonstration that absent a boiling crisis the lower vessel head is not going to fail structurally under depressurized conditions even if the vessel wall is ablated to a critical heat flux-limited thickness.' If there is a boiling crisis (i.e., the heat flux through the vessel wall exceeds the ex-vessel critical heat flux). structural issues become moot since the feasibility of in-vessel retention (IVR) cannot be demonstrated.

The authors took a simplified approach to structural analysis initially I

whereby they considered the combined effect of hydrostatic and thermal loads 5

on the vessel wall (no pressure load since the vessel is depressurized ) and demonstrated that the minimum wall thickness required to maintain structural integrity is well below the critical heat flux-limited thickness.

For reference, the critical heat flux-limited thickness ranges from about 7.5 cm at 0 degree (bottom of lower head) to 2.5 cm at 90 degree (transition from 2

l hemispherical to cylindrical section) for a heat flux of 1.5 MW/m. and for a temperature variation between 1573 K (inside wall in contact with molten pool) and 400 K (outside wall in contact with saturated water).

A linear l

Figure 4.1 of the DOE report shows the vessel wall thickness to accommodate the CHF is reduced to 2.5 cm (at an angle of about 90 from the bottom center of the lower head) due to thermal attack.

5 As noted later in this section, additional sensitivity analysis is included in the DOE report to demonstrate the structural integrity of the Rf'V at elevated internal pressure.

14 l

t temperature distribution through the wall thickness is assumed so that only 1.1 cm outer skin of the 2.5 cm thick wall in the cylindrical section remains below 900 K - the implication being that effectively this 1.1 cm thick section acts as the load carrying member.

When compared to the minimum thickness required to maintain structural integrity which is 0.15 mm (considering hydrostatic loads only, i.e.

134 tons of melt, additional 29 tons of lower head steel, and 92 tons of buoyancy load corresponding to a 6.33 m of hydrostatic head).. there is a considerable margin (two orders of magnitude) to failure.

Even when the more dominant thermal stress is superimposed on hydrostatic loads, the margin remains substantial (i.e.

0.15 mm required to maintain structural integrity compared to 5 mm available as the load carrying thickness (see analysis on p. 4-3 of the IVR report), hence the margin is still over one order of. magnitude).

The simplified approach was supplemented by detailed finite element calculations using the ABAQUS structural code.

The details of ABAQUS calculations were not provided in the original text in Chapter 4.

A later addendum to the chapter was provided to address ductile tearing, localized large deformation and related issues brought up by the peer reviewers.

The addendum contains an expanded discussion on ABA0VS calculations, and provides further confirmation of the substantial safety (structural) margins calculated in the simplified analysis.

A further supplement to structural analysis was provided by the authors in Appendix G dealing with creep considerations for the lower head. The premise for this supplement is recognition of the fact that at high temperatures, the vessel wall material will undergo creep deformation and may lead to creep rupture as a mode of failure.

The analysis in Appendix G is presented as a demonstration that the vessel wall will not fail by rupture (damage index in j

Figure G.11 less than 0.1 for the outer half of the wall) for the given loading conditions.

Note the creep analysis is done with a 5 cm thick wall under an internal pressure of 2.76 MPa (400 psi) to merely illustrate the

. point that there is margin in the load carrying capacity of the outer skin even at an elevated pressure.

The actual as-built thickness of the reactor vessel wall is around 15 cm.

15

l '.

1

\\

l l.

Uncertainty and Sensitivity Studies. For structural analysis. the important parameters for which uncertainties need to be considered are melt mass. melt temperature (alternatively, melt composition defining the eutectic temperature). system pressure, and material properties.

To compare the structural analysis results with the critical heat flux-limited thickness for j

all bounding scenarios, uncertainties in heat flux distribution need to be considered as well.

In Section 7.3 of the report, sensitivity studies are presented in terms of melt temperature and heat flux.

It suffices to note here that the baseline structural calculations used an imposed heat flux of i

2 2

1.5 MW/m, which is in excess of the maximum value (1.45 MW/m ) of the critical heat flux considered in thermal load calculations.

The implication 2

here is that a heat flux value exceeding 1.5 MW/m may lead to CHF-induced thermal failure in which case structural consideration would be a moot point.

Variations in material properties are discussed in Appendix L. and indications j

are given in Table 7.1 that these variations are accounted for in the integrated R0AAM approach to IVR.

Finally, additional sensitivity analysis, carried out in response to the first round of peer review comments, is presented in Appendix P.

The baseline structural analysis considered a total melt mass of 134 tons (see page 4-1) including steel inventory from reactor vessel internals.

From the discussion of uncertainties in the total melt mass (page 7-5 and Figure 7-5).

it appears that the quantity of steel was varied between 67 and 77 tons.

Appendix L provides another perspective on melt mass considered in the study.

According to Table L.2. the core mass (excluding metallic Zr and steel) estimate ranges from 80 to 92 tons. Therefore, with Zr and steel added, the total melt mass would be between 147 and 169 tons at a minimum.

Consideration of this variation will not change the overall conclusions of structural analysis presented in the report.

In other words. structural integrity of the vessel and a substantial margin of safety will still be maintained after accounting for the range of uncertainties in melt mass discussed here.

In the extreme case (not considered in the report), if the entire core and all reactor vessel internals, including the core barrel and reflector plates form the total melt pool mass (approximately 252 tons) in the lower plenum, the 16 l

-a

minimum thickness required to withstand the load would be 0.3 mm - still considerably lower than 5 mm available for carrying the load.

The baseline structural analysis considered a temperature of 1573 K (melt pool temperature) at the inner surface of the wall and 400K (nucleate boiling) at the outer surface.

Looking at the iron-zirconium phase diagram (Figure 6.1),

the inner surface temperature corresponds to the liquidus of an Fe-Zr eutectic f

mixture containing about 8.8% Zr. Appendix P (Tables P.1 and P.2) indicates

]

that sensitivity studies were performed, though not in the context of structural analysis, with inner surface temperature of 1700 and 1750 K corresponding to 8e-Zr eutectic containing 12% and 15% Zr. respectively.

Additional sensitivity analysis (Table P.1) also looked at shiftit. Zr oxidation to the left by 10%. which is equivalent to about 15% Zr in the melt pool keeping steel inventory the same as in the baseline case.

The range of metallic Zr considered in these sensitivity studies is consistent with the ROAAM-type quantification in Figure 7.3.

From the structural analysis standpoint. 15% Zr content in the melt pool corresponds to a load-carrying outer wall thickness of 9 mm (following the analysis in Chapter 4. Figure 4.2) as opposed to 11 mm for the baseline case.

As an extreme case (again not considered in the report), if the maximum attainable metallic melt pool temperature of 1950 K (corresponding to a Zr-Fe eutectic containing 33% Zr) is considered in the analysis the overall conclusion will remain the same (i.e..

the available wall thickness will have substantial margin to carry the load).

i The system pressure was not considered a variable in the integrated approach

)

for the simple reason that the report focussed on a specific accident scenario (i.e., that of large or medium LOCA with full system depressurization).

Nevertheless, the structural analysis considered, in addition to the depressurized case, a modestly higher system pressure (see Appendix G) representative of accident scenarios involving a small LOCA with partial depressurization and SGTR. and concluded that a substantial margin of safety would still remain in these cases.

l 17

Variations in melt properties of importance are discussed in detail in Appendix L.

In Chapter 7. uncertainties in thermal load calculations are ccasidered in terms of parametric variations in Zr fractions.

Table 7.1 shows I

)

the n.elting point of the mctallic phase to be 1600 K which corresponds to approximately 10% mole fraction Zr in the Fe-Zr phase diagram (see Figure 6.1 of the DOE-report). Appendix P (Tables P.1 and P.2). on the other hand, indicates that the sensitivity studies were performed with inner surface 1

temperatures of 1700K and 1750K corresponding to Fe-Zr eutectic containing 12%

'and 15%'Zr. respectively.

It is not clear from the text in Section 7.3 if the variations in melt temperatures were considered in the sensitivity studies l

. reported in this section. Likewise. it is not clear if variations in melt vis csities ' Table L.1) were co sidered in the sensitivity studies in Section 7.3.

Note that the viscosity values in Table 7.1 do not correspond to those for binary, pseudo-binary or ternary mixtures.

In addition it is not clear if variations in melt thermal conductivities were considered in the sensitivity studies.

Much higher values of metallic phase thermal conductivity than those shown in Table 7.1 are reported in the literature j

(see, for example. Reference 13).

If, indeed, the variations in the above melt properties (as reflected in Appendix L) were considered implicitly through _the variations in Zr mole fractions, the text in Section 7.3 should state so clearly.

i In summary, the baseline structural calculations, as well as uncertainty and i

sensitivity studies, presented in the report (with the exception of some inconsistencies and lack of clarity in material properties discussion) provide adequate demonstration that absent a boiling crisis, the structural integrity l

of the vessel wall will be maintained for loading conditions representative of a large or medium LOCA with full depressurization.

l 18

_._.__.______._...___.____________.__._____-___________________________.______._________.___d

\\

AUTHORS' RESPONSES TO PEER REVIEWERS' COMMENTS I

1 The most significant peer review comments on the draft report related to the j

structural analysis were: (1) consideration (or lack thereof) of various failure modes, and (2) consideration of uncertainties in initial and boundary conditions and in material properties for structural analysis. The authors initially considered longitudinal bending and creep rupture as the failure modes of interest to IVR.

Some peer reviewers pointed out that ductile tearing / rupture, brittle fracture, fatigue, and thermal shock are other l

possible failure modes of importance to IVR.

Subsequently, the authors analyzed ductile tearing and concluded that such a failure mode was not feasible (see Addendum to Chapt'r 4).

This conclua on is supported by ABAQUS calculations of principal stresses and strains.

We concur with the conclusion noting that: (1) the inner wall temperature used in the ABAQUS calculations is 1573 K (used in baseline structural analysis), and (3) strains in the inner wall are somewhat lower than those observed in the recent Sandia lower head failure experiments.

The possibility of brittk. fracture was raised by one reviewer, but the authors and another reviewer dismissed the possibility on the ground that the nil ductility transition (NDT) temperature of the vessel l

material is about 20 C (well below the nucleate boiling temperature).

We concur with this assessment.

We also note that fatigue failure is a longer-term issue, not germane to the IVR concept.

Finally, we note that consideration of thermal shock is essential in the context'of FCI loading, but that is beyond the scope of this report.

In response to peer review comments, the authors performed above-mentioned sensitivity studies to address uncertainties in initial and boundary conditions They also performed extreme parametric variations in a few cases.

By doing so, we believe they addressed all'the significant peer review i

comments on structural analysis (with the exception of the thermal shock issue) satisfactorily, as evidenced from the revised text and several l

appendices.

1 In general, the peer reviewers appeared to be satisfied with the IVR report, and the responses to their comments.

Regarding the internal heat load at the 19 1

l

~

~

end state (due to an oxidic molten pool and natural convection in the metallic layer) and the external CHF distribution most peer reviewers were satisfied with the authors' responses to their comments. Some questions were raised with regard to the melt progression and potential intermediate states that may give higher thermal loading than the end state.

In response to these peer reviewers' comments, the authors of the IVR report prepared Appendix 0 to address this issue.

Our review and comments on this subject are given in the Technical Review section of this report.

POTENTIAL OPEN ISSUES Jtructura/ Integrity ofInsulation. High pressure pulses Ltween the vessel and the insulation, as a result of vapor generation, may induce vibrations that could damage the insulation structure.

This may cause obstructions to the ex-vessel liquid coolant flow and vapor accumulation that could significantly degrade heat transfer on the outer surface of the RPV lower head. This issue is beyond the scope of this report but was alluded to by the authors in Chapter 9.

The subject of integrity of the insulation during severe accidents remains an open issue.

Me/t-Vesse/ Interactions. Due to ex-vessel flooding of the cavity below the lower head of the RPV. the initial relocated melt will most probably freeze upon contact with the vessel wall.

However in order to deal with the possibility of interactions between melt and the vessel wall, a separate effects experimental study will be carried out as a part of the OECD RASPLAV project.

Stratification of Meta ///c Layer. According to Reference 14. the presence of about five atoin percent uranium in the metal phase is sufficient to make the metallic phase of the core debris denser than the oxidic phase.

This would l

result in an inversion of the densities, therefore, causing the metallic layer to be below the oxidic layer.

Such a configuration might result in different partitioning of the heat fluxes and, thereby increased thermal loads on the bottom part of the lower head.

This density inversion may be addressed in a separate effects test as apart of the OECD RASPLAV project.

20

\\

CONCLUSIONS j

i The DOE report provides a comprehensive treatment of the concept of retaining l

the degraded core in-vessel through external cooling of the vessel wall.

f Through scaled experiments and R0AAM-type quantification of important.

phenomena, such as natural circulation in the melt pool in-vessel and boiling from the vessel outer wall (critical heat flux). the authors assessed the j

feasibility of the IVR concept. All the conclusions in the DOE IVR report are based on the assumptions that (1) the primary system is depressurized. (2) the l

RPV lower head is submerged with water in the cavity prior to the arrival of core debris in the lower head, and (3) the water level in the cavity is j

maintained indefinitely.

The margin of safety for RPV integrity can be determined by comparing the internal heat load distribution to the external CHF distribution. The internal heat load on the vessel wall is a function of the amount, composition and stratification of melt on the lower plenum, as well as the decay heat level. Molten pool natural convection and heat transfer in the metallic layer (for the stratified case) are other very important factors that affect the internal heat load distribution. This RES review considered each of these factors in detail in performing this review of the feasibility of the IVR for the AP600 design.

Overall conclusions of this review are summarized below.

The authors demonstrated that absent a boiling crisis, the structural integrity of the vessel wall would be maintained.

The margin to structural failure was shown to be substantial permitting large uncertainties in wall loading (i.e.. thermal and hydrostatic loads).

The authors demonstrated that for a homogenous well-mixed molten pool, sufficient margin to failure exists to prevent RPV failure.

The stratified state is potentially an area of concern since a thin metallic layer on top of the oxidic melt pool can lead to a large " focusing effect" from this layer.

The results of this review have not been able to confirm that a " stratified intermediate state" (i.e.. before final relocation of melt to the lower l

plenum) has sufficient margin to safety to prevent RPV failure.

Therefore, we recommend that additional analysis using the methodology in the DOE IVR report 21

7 _.

~

be performed to demonstrate that sufficient margin of safety exists for this case. A decay power density greater than the 1.4 MW/m used as an " extreme 3

value" in the DOE IVR report should be considered, as well as an earlier relocation of melt than the lower bound value of 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> used in the report.

l l

3 22

-J

REFERENCES 1.

T.G. Theofanous et al.. "In-Vessel Coolability and Retention of a Core Melt." DOE /ID-10460. July 1995.

2.

M. Saito, et al. " Melting Attack on Solid Plates By a High Temperature Liquid Jet - Effect of Crust Formation. " Nuclear Enaineerina and Desian Vol. 121, pp. 11-23. 1990.

3.

B.R. Sehgal. J.A. Green. T.N. Dinh. and W.G. Dong. " Experiments and Analyses of Melt Jet Impingement During Severe Accidents." Proceedings of the Fifth International Topical Meeting on Nuclear Thermal-Hydraulics. Operations, and Safety (NuTHOS-5). Beijing. China. April 1997.

4.

D.L. Knudson. "A SCDAP/RELAPS Analysis of an AP600 3BE Transient with Ex-Vessel Flooding." Lockheed Idaho Technologies Company Letter Report Under JCN L2230-DLK-2-96. April 1996.

5.

B.G. Schnitzler " Fission Product Decay Heat Modeling for Disrupted Fuel Regions (FDECAY)." EGG-PHYS-5698. December 1987.

6.

B.A. McIntyre. " Updated AP600 MAAP 4 Parameter File and PRA Information on the Description of DVI Line Break Cases." Westinghouse Electric Corporation. AW-96-953. April 1996.

7.

M.T. Leonard et al., "MELCOR Calculations Supporting the NRC/NRR AP600 Design Certification Review." Sandia National Laboratories Preliminary Draft Report. ITS/SNL-95-006. August 1995.

8.

T. G. Theofanous, et al., " The First Results From The ACOP0 e

Experiment." Proceedings. PSA'96. Park City. UT Vol III. pp. 1343-1350. September 29-October 3. 1996.

9.

S. Globe and D. Dropkin, " Natural Convection Heat Transfer in Liquid Confined by Two Horizontal Plates and Heated from Below." J. Heat Transfer. Vol. 81, pp. 24-28. Feb. 1959.

10.

S.W. Churchill and H.S. Chu. " Laminar and Turbulent-Free Convection from a Vertical Plate." International Journal Heat and Mass Transfer.

Vol. 18. pp. 1323-1329. 1975.

11.

T.N. Dinh. V.A. Cui. R.R. Nourgaliev, and B.R. Sehgal. "Modeling of Heat Transfer Processes in Reactor Vessel Lower Plenum During a Late Phase of In-Vessel Core Melt Progression." Proceedings of the Eighth International Topical Meeting on Nuclear Reactor Thermal-Hydraulics (NuReTH-8). Kyoto Japan. September 1997 (submitted).

12.

F.B. Cheung. "A Scaling Law for the Local CHF on the Outer Surface of the Hemispherical Lower Reactor Head." CSARP Presentation. May 1996.

13.

H. Esmaili, et al.. "An Assessment of Steam Explosions in the AP600 Advanced Pressurized Water Reactor." ERI/NRC 95-211. September 1996.

14.

D. Powers " Chemical Phenomena and Fission Product Behavior During Core Debris / Concrete Interactions." Proceeding of the Committee on the Safety of Nuclear Installations (CSNI) Specialists Meeting on Core Debris-Concrete Interactions. EPRI NP-5054-SR. February 1987.

23

)

3 m

8863 0690 874 4

(

5

'e 6 5 5 4 7527 351 1 0

1 m8200 3305 301 u

W l

le s

s d

a o

O W

r I

W r

0 I

o 1

e tc x

h t

5 a

98.49 05 0

3 4a8 f

e

)

2 R

g55 32 4 2 4

2 3

o k71 d

(

n e

e h

s r

t s

e n

w o

i l

se w

i G

m ro m

t u

d d

n i

e e

s e

mt t

te v

d a a

le l

ms n

s s s s yy ams s s s s s si s s s s s p

I la ooCis e ee et e eeee

/

t l

sl l l l l g

a n

sl s

l mE l

l l

l i

i in n nEn nnnn n

l eOlaa i

n ii i iiii o

re aU ccu a a a a a a aaa l

t t

t t

t t t

t t

r r i t

SSSS S SSSS n

id a

M ZZn i

c M

)

i

)

/P P

5 r

I 7

eI vL U

s 1

a l

(

(

i S e q

e e

g r

c a

l a

=

m ith k

i r

MM N 'l e

Ee s

l 2

s G

t g

e a

7e b s n

e is r

P sFn ue i

o d

s n

Ti s

l e

t d

t ga e

nn enm eP e e 0

nl s r oo neu sh 0 N r

sd eut pp e plo ei r i

Cd el id ot epp nuCdu od I

p a$

t t

n dlaRn r uu A

e a

ASS o SnuiGCepl d r

rI y

i r

r l

t t

r r m s t

e CeloeB r

t r

r r c er ogee I

i o C

h ee eg ep p

r wpiMIn n

pt we wwnp p pd ml ein ou oe oL oooeU pu oop r

er r o

CFCPC CLLR USRLU'2

i )

er rW o

eM t(

C a

0625 6

9 d

Mr 3189 1

7 n

e 7192 4

7 dw a

eo 1202 1

r tP

~

5 nnnnnn oooooo o

a iiiiii t

cy tttttt c

oa aaaaa a ccccc c l c ooooo o e

ee llllll l

f RD eeeee e rrrrr r e

R sssss s iiiii i hhhhhh g

tttttt n

i gggggg l

d nnnnn n a

iiiii u

rrrrr i i

r l

r uuuuu u c

f e dddddd n

ot 6555 33233 39 dddddd I

eeeee e a

7914 77877 79 gggggg eM 0011 00100 01 rrrrr r t

r 3333 11311 13 eeeee e l

ud mmmmm m e

t e bbbbbb M

at uuuuu u sssss s ra d

ec sssss s e

po

.aaaaa a t

ml) ewwwww w a

eeK bKKKKK K l

c TR(

a o

t368770 l

913211 e

e344444 h

R d

tffffff C

ooooo o f

n ieeeee e o

n 1

rrrrr r I

7 duuuuu u n

67 44 8

7 etttttt o

g 52 57 5

daaaaa a urrrrr r A

96 91 2

i eeeeee l

t cpppppp i

nmmmmmm s

ieeeeee o

l tttttt p

e t

otttttt m

e d

naaaaaa o

t b

3 1

C g S

'53 8

7

)

5 ellllll d

4 4'54 2

'6 2

35 reeeeee k

aeeeeee 4

8 tttttt

(

n S

1161 1

3 6 essssss r

a s

ussssss s

tssssss t

a ceeeeee nl M

r rnnnnnn ullllll ua Z

6 2 7 tiiiiii oi d

4 3 6 saaaaaa mr e

2330 8

7 tttttt Ae t

266.

3

. 5 l ssssss e

t a

8841 6

1 5

ellllll da c

t aaaaaa eM o

srrrrrr t

uuuuuu l

mtttttt al e

ucccccc me R

2 nuuuuuu ir 0

6 7

1 errrrrr tr r

l tttttt 2

7 9

pssssss sa Z

6442 2

2 5 EB 062 9

rmmmmmm 8951 7

3 7 euuuuuu wnnnnnn oeeeeee lllllllpppppp g

2 2

8 8 5 nrrrrrr O

5897 6

8 9 wwwwwww i eeeeee e

U 4254 9

7 9 ooooooo 145.0 5

l lllllll b

6 3 l

offffff a

7841 6

2 6 f oooooo T

egggggg hkkkkkk d

t

)

e 60.1480 0580 t0.04 s

t a

(

84019557192 a

s h32 191 73429959735 c

l t521906 e

35692345682 osa 419113 m

l st e t

11112222223 eao i

o.....

T 11111111111 RMT N ab cd ef.

1i ll1l; ll l

l s

00 5,

2 1

fo e

4 m

i t

ta E

)

8 M

3 W

0

'0 es

(

1 1

a n.

c 5

X X

0 0

1 8

9 0

0 1

6

+

PA l_

l l

r y

5 o

t f

i S

s M

l n

A o

e o

D p

r n

e e

w t

o

)

l P #

om c

W

'0

'0 i

d r

(

1 1

a t

e e

e x

X h

m t

u a

5 9

r l

m 0

9 e

o i

w V

t 1

0 o

s l

E n

t i

se y

B t

isne s

d

)

e t

as e

s c

(n r

d a

ur i

e e

e dega w

N otnm l

o nel rfie p

e ore patr g

i u

l c

a stfe neel) r sc n

osmet ir e

iuri ia uc t

v fdet selfa e

A otl sl e t

m orfo ieufn

(

u N pa m Frfti e

l s

ov a

C e

h 1

2 T

3 e

lba T

I 1

i Table 4. Arche Contritmting to Core Damage Frcquency (CDF) of the AP600 as Developed in the 1.xvel 2 PRA with Update Level 1 Quanti 6cadon (AP600,1995) l l

Case

% of Taming of first

/rcht Descriptica and enm-enre l

1.D.

CDF' Debris l

Relocamos (br)

. Core damage fouowing summent or 2

small I.OCA at high peessare; maanual 1AC 1.9

~5.5 br ADS 8 failure.

~

. Some relevaner to IVR' if hot leg nozzle e and '----

-- p

. IAc 1 AC. cacept PRHR aperaung..

IAPC 0.4

~10.5 br

. Some e.w to IVR: less decay heat than 3BE: partsally pressemand (<.5 * ).

. Core darnage followns large 14CA or w'm other event with In8Il '----

u l

3BE 50

~a+ hr (snrein= failwe); nay re8ood through l

break as cavny Goods

. Of main maeres to IVR.

l

. Core damage foltoumag 14CA wish early h and CMT* imilmes or l

3BRC

~29 N/A any other w what samuual j

depressmaison leads to ndbood iscen IRWST* med wy m-mael.

. Typically, not relevant to IVR 4.1 N/A

. Vessel rp I

3C l

. Not relevaan no IVR.

. Non-large LOCA with perual 3DC 13

~22+ hr

. Of innees to IVR. but less decay best than 3BE.

e Cort damage sher meam guaramar udse 6E 23

~22+ hr nignure.

. Of imeren to IVR, be less decay hem l

l shan 3BE l

t

. hmty case for 3BE without cavity 3BE.cc0 l N/A

~3 hr floodsag_

. Bounds IVR tuning uncenaintv I

l

'8 CDF = Core Daenage ~.wy LOCA = Loss of Coolant Accident

' ADS = Ataarnanc Depressacnawei Synese

' IVR = in Vessel Rreensman

• CMT = Core Makeup Tank j

[

• IRWST = In-Contamment RefuelinF Waner Swaec Tant 1

e+

l g on 0

9 8 0

r oi 5 8

1 5

ep r,

cl e

B bi 0 0 1

1 p

d s

uo u

n C sB e

r U

U h

e s

t t

c t

n it i

a o

e l

i t

c c

U a

F P d L

eg g

m s

U tn n

o t

8 5 i

t e

ai T

2 rl 2 6 d

t n

o s

m ui 0 0 o

b u

/

to io W

aB p

a M

s s

e r

e h

V r

t a

x r

t d

u o

t e

l c

a F L ru N

d+

0 c

C sa t S e g 9

e a

t n 0 4

M.

le t

ai a

r 5

=

1 s s l

l e n u i 0 0

f x o l

L t

ui o

o i

a B a

t B

d lFd c Y s

n g

t n

i C a

an a o t

eC ti H

i r

al g

r e

do l

n ai C

t C

o cli n

gc it o l

d*

e nb i

B a

e*

r c

0i u C

c l g

)

d 9

6 5 5 1hs e

o on l

c o L

oi 5 8 0 3 m

c al o

f na o o cl 0 0 1

1 t

oe Lc bi b

U

(

t uh uo e

S o

gqt u

h sB S

P b

n i

t iew f

d o n t

e ll na a

h oag od t

ocn s e B

csi i

t a a L

o bl r

r d

p B

t uli u

eg mt S

sl o

)

t n 0

1 5 0 g

ab a

o S ai 4 G 8 0 n

am C

rl il sn ui 0 0 0 1

deh o

5 t o nstni

(

le aB osiot s

pew c

ba sv ye T

e gl v rtnen r si g o oelrc e

n ctia l

o oll g

i eb a

n t

0h dr A f a tleu 0

0 0 0 on 0

=

ost l

3 6 9 i

f oaa a

l 0oPBN c

c o

n L

I

+

• ++

l l

1 i

\\

I l

/

\\

, m ~

~~.,

Mendhe 9r Case 9 W Figure f. Materials ccanguranon in the nansal convecrian, sacady stant, extreme i

4,p**

-~

m A

g

)

l q.

m.

9

.w

___ n ~.- -,

j f..k

-M

-.. a

' L - semma skutage i

te== saw== '

lhY rhae waar

_m WomarBeihag q

i M 2. M -

. inn stuc flouced Coirvecies, jet @ cm assoc -

ased wish she senior case % euss.

".se ira dnu gU o

7'y b

e r n 8

6 4

2 o

u li i 1

0 0

0 0

O at f

a 0

e n 8

h a t

l p

ts 0

e x s-

' 7 e

t oro t

f i

0 tu

' 6 ot

, m x

c qq d e et j/

)

0 s

i i qq

' 5 e

a e r

e r e r

c S g

y i

' 0 e d

8 d

u

/m 4

t

(

s qq cW e

ir i

' 3 eM 0

l t

g n

m4 A

a 1

r a

i : 5:

0 p=

5

' 2 4 6 eQ

- 0 t m

ty y ed i

it i inn

' 0 r n v iv t

a 1

x d

e imn m

E E na 2

0 0

0 0

0 0

0 0

o O

f 2

0 0

0 0

0 0

0 0

s 4

2 0

8 6

4 2

t

=

l 1

1 1

u s,

,"Es56 M3{E eH R

m 8

3 1 e1 ru =

g FH i

s0 q i 7 J J. J J.

o w;I.

,g 1

l o

P

=

1 e

0 n

...X.

h 0

t 6

P i

A

...xi

.j' m

r r e e w w o o x.O P P d

y y

,... x e. 'i 5 "

r0-t l

a a

'..x..*,,"1 c c h

e e x._~*-

I D D

)

t s

f g

(

il

. X's, e

m l

mi t

i f

<,..*.^.t l

T x

x s

. X

_ n 'i 8

i t

.a.

i 0

_x f

1 N.,,-i 1

w

.s-x r

_3, -

....,.,_'i,a 3

..;i.....-

2 0

1 0

0 0

0 0

0 0

0 0

0 0

0 0

0 6

5 4

3 2

1

^g 2 gg*e A,v.Q c

lll

<\\l l

l l

no i

tcnu f

sa r

7 c

)

... 0 w

h i

i.:

1 o

3.:.

i f

p y

a

[.

i c

e d

s e

0 i

i 1

h

_ r l

... ~

t f-

't i.

t j

I-t o

s t

n

)

n e

s s

o

, ~ -

H l

i 0(

p rw3 i

1 e l.

l r

...~..,D[.!

t.

m l

i

. 'I-i l

4 1

I i

1 t :--

. i m

1 o

1

.j i

c e

i T

l i

8

^

t a

4 y

lo

_ i. I:i-NI 0a

't i y,:

i j-_

c v.

m....

M i

1 e n

~

l

~

.j e

D o

n h

t

\\

f 3

o k

0 p

i.r-

. - y:.

\\

i i

f q_r U

1 on

]

A li '!

il I

o 1.dg 4

i t

u

~

b

. ~ _..

i t

r

~

n 2

o

~

~

0 c n 61 aw l

0 I

no U

0 od 1

i t

u t

ch cC,.ucbg 5y Ob gI gA-e s

a

'(

r s F r e

t

.f a

5 e

e m

r ui t

g ifF o

22 2

H' 1*8.-

• ..*,,<*a Scm,*

i

~

1.6-

i.4-iocm u

• ..-4 cr 1.2,,.**.*,*,,,*_..,.=,,.-

cr 15cm

~~

,,,.. ~ ~ ~

_~

0.e O.,.,,,,..-....

.8 0.9 1

1.1 1.2 Height of oxidic pool [m]

Figurc 6.

The failure envelope for the Tocasing" =:=im, in the H, x H space. Decay 3

power density is 1.4 MW/m and emissivity 0.45 both at the exueme conservative values.

U f

l 1'

s eI y

/

/

G

/

o.is - /

1 1

4 1.s

/

A o.as 0.s 0.3 1

1.1 1.2 Height of oaksc pool [m]

Figure 7.. The ide in Figure 0.1 replaned in the B, x R space. The avicaory shown is so suustrase the simuh=ence increase in both R, and R through the mew scenano

2 O

.23 AU

- t i

i e

=

O

~

s c.a. -

v

~

c

~

s.

-1 x

o.s s

U 4

s c

o4 x2 E.

I

.c o.2 o

W

~

g

i'\\-

i

-o 0 3

4 5

6 7

Time (hr)

Specincation of the shuedown time (for use in esimatmg &'my po=ed, as probabi Figme8.

density f==rew=.

i i

is I

e l

o

~

u 5

gi

.o i.

=

~

t

-a 1

u w

s 2

v%

\\

.o o

b 1

l C

j 1

1.5 1.6 17 i4 Height of Oxide Pool (m) mg h a W2 dedW First 9.

Tne oxide pool height the @ of Figwe 7.3 of #4.J.

m'. a hemisphe:e radius of 2 m.

s fo no itan i

b moc m.

o 4,

r f

l g

gn

)

i t

'm lu s

/

e r

W n

M o

(

it u

y b

t i

i r

t s

s n

i d

I l

2 e D

y 1

t i

r li e

b w

a b

o o

P rp y

y a

t i

c s

e n

D e

d r

0 ew, I

l 1

o p9 y

a J.

ce

~

d e

g 4

2 o.

h 8

1 g3$w87E.3:m nh12 e

c 4

0 1

s ee r r uu gg iiFF

6

'0 I

9

/

1 8

2

/

n u

R e

5

' '0 1

'a R

L F

f.-

r a

ta n

1 a

' 0 1

D" o

g

' 1 O 'a R

P R8 s

O5 4 C5 0 v

s A0 0 v

v

./.

'0 0

01 0

0 0

1 1

.m s*

l i

4 I

1 I

i a

1 1

i I

4 Q

g g

i i

6 t

7 O

s '. '.

e=-

40 l

~

i s

• 5 3

i t

t *.

60 s

e=

i O

l N

1 M

4*

~

es 1

M Q

.g 1

.,\\

v n

e e

o eQ 9""

n e

s t

M

\\

e e 3

O C4 s

~

.s 4

• g e

O 0%

4 4

4

• g 4

s

.g a

1

. \\.

get

-.s e

o o

o 1

O o

1 o

sn e=.

deON e

1 i

l l

l l

I Comparison of the Scaling Law for Subcooled Boiling with CHF Dats l

0.20 0.16 F,(0) = 0.0473 (1 + 1.78 0 0.5102)

C

/'

0.12

,/,

"g J

A

".c

./ a I

%#g a_.

0.08 p' /

5

/

• er

/

A/

9 SBLE Saeedy. State Boilms m/

/

A SBLS Tramient t

/

0.04 -/

UtfU M soning

--- utsu mmed c

si e d s om 0

m/6 m/3 x/2 0

1

'O I

13 g

l.

L l

APPENDIX A Sensitivity Study of Melt Quenching During Slumping l

l The purpose of this appendix is to provide results of additional sensitivity studies based on the extent of heat transfer between relocating molten material and water during slumping.

As discussed in the main part of this report, the original INEL Case 3BE transient study (i.e., no quench case) was based on the conservative l

assumption of no heat transfer between relocating melt and water during slumping.

As a result of this assumption, the temperature of the melt in the lower plenum will remain high. and there would be higher heat loads imposed on the RPV wall compared to that of the cases with quenching, j

i To ensure that the no quench case is conservative, even for intermediate states, two additior.1 sens~itivity studies involving different assumptions for melt quenching were completed by INEL.

One case assumed 100% quenching (i.e..

100% breakup of the slumping material) and the other assumed 50% quenching (i.e.

50% breakup of the slumping material).

The results of these two cases indicate that during the accident progression and slumping, the bottom of the core plate is continually covered with water until the debris is in direct contact with the core plate.

Thus, the core plate does not heat up by thermal radiation from the upper surface of the debris in the lower plenum until the debris is in direct contact with the core plate.

(Note that a value of 0.5 porosit' is assigned to the debris bed in these quench cases.) Thus, for the quench case the top of the debris bed reaches the bottom face of the lower core plate sooner than for the no quench case.

This would result in melting of the lower core plate and therefore, more steel in the stratified metallic layer for the quench cases compared to the no quench case.

For the quench cases, the temperature of the slumping molten material at the time it settles in the lower plenum is about 500K for 100% quench and about 2000K for 50% quench.

For the no quench case, the temperature at the time of settling is about 3000K.

A-1 m____

__.___.___________.___m.________..__._______________.m__

Based on the above discussions, it can be concluded that the no quench case is indeed conservative, and that cases with quenching do not lead to more severe

. intermediate states than the no quench case.

A-2

APPENDIX B Amount and Composition of Relocated Melt Calculated by SCDAP/RELAP5 for the Case 3BE Transient The purpose of this appendix is to describe how the amount of reflector and l

core barrel structures that melt and slump to the lower plenum during the course of the accident was determined.

From the INEL study (Reference 4). a summary of the calculated amounts.

I composition. and timing of melts that relocate to the lower plenum during the course of the accident is shown in Table B-1.

The amount of relocated stainless s.2el material in TaL'.e B-1 does not include the melteJ reflector and core barrel material.

l The amount of reflector and core barrel material that slumps to the lower head was determined based upon the results of SCDAP/RELAP5 t.lculations and additional hand calculations considering post-accident observations of the TMI-2 accident and OECD LOFT FP-2 experimental results.

This information indicates that the molten p^ol growth is not axisymmetric.

The SCDAP/RELAP5 calculations determined the height of the reflector and core barrel that come into contact with the molten pool and thus melt and slump to the lower head.

The calculational results also showed that the molten pool penetrates thr'ough the reflector within minutes of initial contact by the molten pool.

However.

the molten pool at its average radius would not contact the reflector until much later in time (-10 minutes). After penetration of the reflector., the molten pool material would drain through the lowest elevation of meltthrough of the reflector in less than ten minutes. A comparison of the timing of these events indicates that it is unlikely that more than 1/8 or so of the reflector would be melted by contact with the molten pool. Therefore due to asymmetrical growth of the molten pool, meltthrough of 1/8 of the circumference of the reflector is assumed in this study.

B-1

a i

)

n r W o

e M i

t

(

ss a

0 6 2 5 6

9 e

M r 3

1 8 9 1

7 r

e 7

1 9 2 4

7 g

d w o

e o 1

2 0 2 5

nnnnnn r

t P oooooo P

a iiiiii.

c y tttttt t

o a aaaaaa cccccc l

c oooooo le e e llllll M

R D eeeeee rrrrrr e

sssss s r

iiiii i hhhhhh o

ttttt t C

l ggggg g a

g nnnnn n i

iiiii n

r rrrrr ir i

f e

uuuuu u r

o t dddddd u

a 6 5 5 5 3 3 2 3 3 3 9 dddddd D

e M 7 9 1

4 7 7 8 7 7 7 9 eeeee e r

0 0 1

1 7 0 1

0 0 0 1

ggggg g 3

1 1

1 3

rrrrr r

t u d 3 3 3 3 '

1 eeeee e l

t e e

mmmmm m a t bbbbbb M

r a uuuuu u sssss s e c d

p o sssss s e

ml

)

.aaaaa a t

e e g ewwwww w

a T R

(

l c

bKKKKKK a

o t36877 0

l 913211 e

d e344444 h

R C

tfffff f

ooooo o f

n 1

n eeeee e o

i I

4 4 7

7 rrrrr r

6 7 n

g 5 2 5 7 8 duuuuu u ettttt t o

A 9 6 9 1 5 2

daaaaa a2 i

urrrrr r -

eeeee eB t

lcpppppp i

nmmmmm m

s l

i eeeee e o

e tttttt p

e t

otttttt m

t

)

naaaaa a o

g S

6 d

'5

'6 4

ellllll

'3 C

k 0 2

'5 5

9 8

2 reeeee e

1 1

(

aeeeee e d

S 1 1 6 1

3 9

ttttt t

n s

esssss s

r a

s ussssss a

t ssssss t

M r

ceeeeee n

ullllll Z

6 2

7 rnnnnnn u

d 4

3 6

tiiiiii o

e saaaaaa m

2 3 3 0 P

7 2 6 6 3

5 t

tttttt A

a 8 8 4 1

6 1

5 e

l ssssss c

ellllll d

o t aaaaaa e E srrrrrr l

t 8 e

uuuuuu mtttttt a 3 R

2 ucccccc m

0 6

7 1

nuuuuuu i e r

2 7

9 errrrrr t s Z

6 4 4 2 2

2 5

l tttttt pssssss s a 0 6 2 9

E C 8 9 5 1 7

3 7

rmmmmmm euuuuuu wnnnnnn oeeeeee lllllll 1

pppppp 2

8 8

5 g

B O

5 8 9 7 6

8 9

inrrrrrr U

eeeeee 4 2 5 4 9

7 9

wwwwwww e

1 l

. 4 5 5

ooooooo 0

6 3

lllllll l

b 7 8 4 1 6

2 6

a offffff f oooooo T

egggggg hkkkkkk d

t

)

e 60.1480 0580 t0.04 s

t a

(

0 6 2 3 1

7 9 3 1 4 a

s h32

.191 7 2 3 1 9 8 4 8 6 3 4 c

l t521906 e

9 2 5 8 9 0

. 2 4 8 o s a 419113 1

1 l

s t e

m t

i 0

1 1

1 1 1 2 2 2 2 2 e a o o.....

T 1 1 1 1 1 1

1 1 1 1 1

R M T Nabcdef.

APPENDIX C Decay Power Calculations for AP600 Case 3BE Transient The purposes of this appendix are:

(1) to provice the results of the SCDAP/RELAP5 calculated decay power in comparison with the ANSS.1-1979 and the decay power used in the DOE report (i.e.. Figure 7.1 of Reference 1), and (2)

~

to provide the results of a study on the effect of fission product release on decay power fraction for a generic PWR compared to the AP600 design.

The SCDAP/RELAP5 calculated decay power is in agreement with the ANSS.1-1979 Standard (see rigure C-1).

However the calculated d9 cay power by aCDAP/RELAPS does not include the estimated 10% uncertainty.

Since the ANS5.1' Standard indicates a la uncertainty in decay power of -5% (i.e.. assuming knowledge of the core power within ~2% and knowledge of each of the 2asU. 23sU.

239Pu power fractions within -1%). to capture -95% of the uncertainty in decay power would require one to place error bands of 10% ( 20) on the decay power curve as shown in Figure C-2.

For the SCDAP/RELAPS calculation of the best-estimate decay heat for the AP600. a series of ORIGEN2 calculations was performed for' an AP600 equilibrium 2-year cycle.

The ORIGEN2 results were then used to generate the'following information needed by the SCDAP/RELAP5 decay heat model:

l 1.

power fractions for 2asU. 23aU. and 239Pu during the equilibrium

)

cycle.

l 2.

the 23sU capture rate per fission (the R value in SCDAP/RELAPS),

and 3.

the fission per initial atoms (the Y value in SCDAP/RELAP5)

Thus, the input for the SCDAP/RELAP5 decay heat model are: (1) 634 days for the reactor operating time. (2) 0.623 for actinide multiplier. (3) G-value.

-which is expressed as G(t).- 1.0 + (3.24 x 104 + 5.23 x 10'" x t) T W. where T is the operating time in days and Y is equal to 0.796, and (4) 2350, 238U, and C-1 l

l

e 239Pu for fissile isotopes, and the power distribution for these isotopes is 235 238 239 56.7 % of

0. 6.5% of U. and 36.8 % of Pu.

Figure C-3 shows the predicted effect of fission product release on decay' i

power fraction for a generic PWR compared to aifferent nodes (i.e.. control volume) in the AP600 design.

Note that the SCDAP results of a fast heatup curve for a generic PWR is shown in the figure; this is a replot of a portion

.i of Figure 5 in the main part of this report.

The other five curves shown in Figure C-3 are the calculated results recently performed by Bruce Schnitzler of INEL using the ORIGEN code and the temperature history from the results of the SCDAP/RELAP5 AP600 Case 3BE study.

The temperature history was obtained for typical high power, average power, and low power nodes. The fraction of the core mass associated with high average, and low power nodes is 0.11 (i.e.

16 of 145 fuel assemblies). 0.67 (i.e., 97 of 145 fuel assemblies), and 0.22 (i.e.

32 of 145 fuel assemblies), respectively.

In Figure C-3. the decay power fraction for the average node in AP600 with no fission product release after fuel melting is about 0.6 at 12.500 seconds.

This value of 0.6 is less than 0.77 obtained from the generic PWR fast heatup curve. Based on these new decay power fraction curves, we calcPated the volumetric power density in the lower head molten pool under the homogeneous configuration as shown in Table 3 in this report.

For Case 1. no fission product release after fuel melting, the volumetric power density is about 1.16 MW/m (i.e.. ANS5.1 +

3 3

10% case), which is less than 1.4 MW/m.

(Note that the value of 1.4 MW/m is l

significant since it was used to derive the threshold value of 0.15 m metallic-layer thickness necessary to insure vessel integrity under " focussed" heat loads.) The power density for Case 2. fission product release after fuel melting (as if fuel remains intact) is also shown in Table 3.

The difference in the calculated volumetric power density between Case 1 and Case 2 is only about 6%. A realistic value will be somewhere between 1.16 MW/M' and 1.09 MW/m.

Thus, a mean value of 1.13 MW/m' power density is used in this report.

3 C-2

0 0

0 0

]

h s

t l

u

)

s d

er r

)

o a t

a ed e

h

+n h

d pf A p a

y

(

pr t

a i inI laS c

't

- k et 7 wto e

s d

on 9

la

- - 29 t

6SN1 o

PPE -

t 1

A A.G

^

5 l

I P

LEER5 i,

A.

S

- RRO s

)

OOH (N i

A

(

E R

e d

0G m

n a

ji 0

^

T 2N

^

E

^

G I

0

^

R O

f o

n o

s i

r ap m

o C

1 C

m u

g i

2 3

F 1

0 1

4 G

\\

\\

1y\\\\

g

~

{.

g

,y w

es

.s 3

\\

5 5'

g 0

p 5'

g 4,

0 Sg M

r.

D S'"

9, {e T >-

O 6

y I

\\

h',a.

S*

O-

.M..

's:e s

d

,o s

S c

e, o

e g-g*.

o.#

'o

't g

"!"s; i

tj g

i e

4 f

e o

l s

'o s

s p

a*

'6 tj g

h

+

O 4

0 3

g 1

19 g

e o

v AP600 3BE Scenario

+ First Node To Meit - Cononuous Release 9 Average Node No Release After Melt A First Node To Mett. No Release After Mett

+ t.ast Node To Melt

+ Average Node - Continuous Release

• SCDAP Generic PWR for Fast Heatup 1.0 l

o

-"'*"'""*~"""""*"'"A*"'*"*""N"*"~^""""~

0.9 -

\\

~

i i

" h' * '

\\

l

(

0.8 3

u u

?

/

w 0.7 -

dt j

Ave. Nede Wo.

8 Relesse A(/he

~

Mee 0.6 -

l Ave. Ade -

0.5 _

.can & ur..hk/c s.......

4:

i l

O.4 j

l l

l l

l 0

2000 4000 6000 8000 10000 12000 14000 Time After Scram (Seconds)

Figure C-3. Effects of fission product release on decay power; SCDAP generic PWR versus AP600 hot and average nodos.