ML20005A939
| ML20005A939 | |
| Person / Time | |
|---|---|
| Site: | Yankee Rowe |
| Issue date: | 06/19/1981 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20005A937 | List: |
| References | |
| NUDOCS 8107060060 | |
| Download: ML20005A939 (8) | |
Text
'
ENCLOSURE
~
YANKEE R0WE ECCS MODEL CHANGE
,AFETY EVALUATION REPORT BACKGROUND The NRC has determined that the swelling and rupture curves utilized by the nuclear industry have the potential for b,eing nonconservative, when compared to all applicable experimental data. The NRC position, along with the data comparisons are documented in NUREG-0630.
Yankee Atomic conducted a detailed assessment of the new data for the Yankee Rowe Power Plant and concluded that the new swelling snd rupture data would penalize the operating limits for that plant.
It is for this reason that Yankee Atomic submitted a proposal for modifying their large break LOCA model (Ref.1, 2,12, and 13). Approval of this model change would offset the penalty impo-ed by the new swelling and rupture curves. The modifications of the large break LOCA model is the subject of this SER.
MODIFICATION TO IHE YANKEE R0WE LOCA MODEL t
the Yankee Atomic Electric Company (YAEC) submitted, for NRC approval, a modifica-tion to their Yankee Rowe large break LOCA model. This modification incorporates a phase separation aodel into the lower plenum of the. reactor vessel. Without this modification, the lower plenum would be modeled as two homogeneous volumes.
The effect of this model change provides greater lower plenum coolant inventory 8107060060 81061 7 PDR ADOCK 050000 P
P_
_. ~ _ _
1
. at the end-of-bypass (E0B) period of a large break LOCA.
End of bypass is de-fined as when "...The expulsion or entrainment me'chanisms responsible for the bypassing are calculated not to be effective."(I) Prior to E03, the injected ECCS " bypasses" the lower plenum as it travels around the reactor vessel down-comer and exits the postulated break.
With a greater coolant inventory in the lower plenum at E0B, the lower plenum refill period is reduced. The refill per'iod is defined as the time to refill the lower plenum with ECC, including the reactor vessel downcomer to the cold leg elevation, and the intact loop cold legs to one-half of their capacity. Dur-ing the refill period, the core heat transfer is assumed to be adiabatic (zero heattransfer).
Reducing the period of adiabatic hectup reduces the final peak clauding temperature (PCT) predicted during reflood.
The old method for evaluating lower plenum inventory utilized two (2) homogeneous lower plenum nodes or volumes. Homogeneous volumes imply that the density of the fluid exiting those volumes is a volume-averaged density.
In contrast, a non-homogeneous volume (e.g., containing a phase separation model) would permit steam flow to exit whenever the flow path becomes uncovered by the calculated mixture level. Thus, less inventory would be depleted from a volume whenever the exiting flow path is located near the top of the volume and becomes uncovered by the mixture level.
(1) Appendix X to 10CFR50 (I)(B)(1)(c)
- To take credit for this physical behavior, the licensee requested to combine the two lower plenum homogeneous volumes into one volume and incorporate the Wil-son bubble rise model with a zero bubble gradient within the volume. A bubble gradient signifies the bubble distribution within the mixture level and corres-pondingly the bubble density at the surface level. This, in turn, affects the mass flow rate of steam bubbles escaping the surface of the mixture. A zero gradient signifies a homogeneous bubble distribution and extends the period of lower plenum level decrease.
MODEL EVALUATION Considerable experimental evidence is available to support the conclusion that water remains in a deep lower plenum following the blowdown, phase of a PWR LOCA. YAEC presented data (Ref. 2 and 13) from General Electric, Battelle North-west Laboratory, Creare, and Idaho National Enginnring Laboratory to support their position that not a'.1 of the water becomes depleted from the lower plenum l
during a postulated cold leg guillotine break. Additional information is avail-able to demonstrate this conclusion for lower plenums of geometry similar to that of the Yankee Rowe lower pler.um(i.e., large plenum depth (L) over effective diameter ratio).
For example, tests conducted at INEL with the Bettis Flask and with a simulated reactor vessel showed that water remained wi'.hin the lower plenum for a break near the top of the vessel -(Ref. 8). Both the LOFT and the Semiscale experiments have shown incomplete voiding of the lower plenum (it should be noted that Semiscale has a lower plenum L/D ratio much greater than that of the Yankee Rowe lower plenum.
See Ref. 9). The most typical data for assessing the Yar.kee Rowe model change are the LOFT L2-2 and L2-3 experiments.
Both LOFT
I
-4 and Yankee Rowe have similar lower plenum L/D ratios and design.
Whereas the L2-2 and the L2-3 experiments never depleted the lower plenum inventory, the Licensing Evaluation Model calculations for these tests, utilizing the two homogeneous lower plenum nodes, showed complete depletion (Ref. 5). Utilizing the proposed single node lower olenum with a bubble rise model, the YAEC licensing code conservatively predicted inventory remaining in the lower plenum at E0B.
From the data presented, including RES supporting conclusions (Ref. 9), YAEC stated that DWRs with LOFT geometric loweh plenums will also exhibit incomplete voiding.
The Yankee-Rowe downcomer to lower plenum confP aration differs from many other PWRs. Flow from the downcomer to the lower pienum during normal operation occurs via two annular paths, an outer path between the pressure vessel and thermal shield and an inner path between the thermal shield and the core former region structure.
The flow areas ma approximately the same for both paths. The outer path flow enters the lower plenum a short distance below the core support plate. The inner path enters the icwer plenum essentially at the bottom of the core support plate.
Thus, during reverse flow, a direct path is available from the top of the lower plenum inta the downcomer.
(Many lower plena designs direct flow from the downcomer into the lower plenum at an elevation significantly below the core support plate.
No.e of the Yankee Rowe data comparisons were evaluated from the view of anything except a connection at the top of the lower plenum.)
i Reference 2 contains a number of comparisons between data and calculations per-formed with the approved version of the Yankee Atomic large-break LOCA EM Code, modified to include the Wilson bubble rise model. Some comparisons are also made with a fixed bubble rise velocity to obtain an intercomparison to the data and
. to Wilson. The approach was as follows:
1.
Select a break flow coefficient that provides a good approximation between pressure data and calculation over the transient for each comparison.
2.
Compare the effect of bubble rise velocity to a GE blowdown test. Select Wilson as the best overall predictor.
3.
Select a buble gradient parameter C, based upon three GE tests.
4.
Compare the behavior with the above selections to other experimental data.
The compariscos (including applicable information used for parameter selection) are sunmarized in Table 1 for the behavior of mass in the applicable volume.
Generally, best estimate type behavior is obtained, with some calculations showing more mass than supported by the data, and some with less.
For the most part, devia-tions are not of large magnitude. The most applicable esperiments are the L2-2 and L2-3 tests, which most closely simulates the Yankee-Rowe configuration and which are conservatively predicted.
Examination of the Yankee Rowe heat source modeling within the lower plen showed 7500 lbs. of internal structures were neglected. YAEC,therefore,revisedtheir model such that the reactor vese21 wall, including the lower plenum internal heat sources are modeled in compliance with the requirements of 10CFR50, Appendix K.I.A.6, " Reactor Internals Heat Transfer."
l i
t
Table 1.
Comparison of Results, Wilson to Data Fig. (Ref. 2),
Data Source liass 2.3, 2.4, 3.2, 3.3 GE B-3 high 3.5 GE B-4 3.7 GE B-5 3.9, 3.10 BINL CSE 9 low (sBE) 3.12, 3.13 BfML CSE 10 low 3.15, 3.16 BtNL CSE 128 BE 3.18, 3.19 BfML CSE 51 BE 3.21, 3.22 BfML CSE 52 BE to high 3.24, 3.25 BINL CSE 53 BE, limited calc.
4.4 Creare 13.0060 low to high
~
4.11 Creare 13.0066 low to slightly low Table 4.1 Creare 13.0060 7% high 13.0063 10% high 13.0061
< 1% error 13.0004
< 1% error 13.0070 3% high 13.0071 3% high 13,0066 4% low 13.0068 8% low 13.0067 4% high, 5.6 Loft L1-4 vessel low to high to low 5.7 Loft L1-4 low (volume) 5.13 Loft L2-3 low to slightly low
, volume)
(
.n
"*"-"-'e*
-*-u-ew'*q r, -,m m
yw.,
CONCLUSION We have reviewed the licensee's proposed model change to the lower plenum of the large break ECCS model for t:1e Yankee Rowe plant. Our review concentrated on the ability of the licensee's cede to model the lower plenum inventory through blowdown.
Based on comparative analyses with diverse applicable data, we conclude that the YAEC modification to the Yankee Rowe ECCS LOCA model provides best-estimate lower planum predictions. When performing licensing evaluations, Appendix K of 10CFR50 requires all of the emergency cooling water injected into the reactor vessel during the bypass period be subtracted from the lower plenum inventory.
The remaining reactor vessel coolant inventory is, therefore, still conservatively predicted for licensing analyses. As such, we find the YAEC modification to the iankee Rowe ECCS model acceptable.
e
-y 9--e e
--+--1 e
m 1
g a
w--e----
ry-y1-e=tw--'ge=r teqq-r-
REFERENCES 1.
Kay, J. A..." Incorporation of a Lower Plenum Phase Separation Model in Yankee Rowe Large Break LOCA Methodology," Letter to NRC (NRR), Yankee Atomic Electric Co., December 10, 1980.
2.
Brown, G.
J., et al., " Application of a Lower Plenum Phase Separation Model to Yankee Rowe Large Break LOCA Analysis " YAEC-1231, Yankee Atomic Electric Co., November 1980.
3.
Robinson, H. C.,. " Quick-Look Report on Loft Nonnuclear Experiment L1-4,"
QLR-L1-4, EG&E Idaho, June 1977; 4.
Reeder, D. L., " Quick-Look Report on LOFT Nuclear Experiment L2-3,"
QLR-L2-3, EG&E Idaho, May 1979.
5.
- Levine, S., "Research Information Letter - #63 Loft Reactor Safety Program Research Results from Nuclear loss-of-Coolant Experiments L2-2 and LP-3," NRC memo to H. R. Denton, November 1,1979.
6.
.Kay, J. A., " Clarification on Use of Heat Slabs in large Break LOCA Modeling for Yankee-P. owe," Letter to NRG (NRR), Yankee Atonic Electric Co., May 5,1981.
7.
Code of Federal Regulations, Title 10 Part 50.46 and Appendix K.
8.
Heiselmann, H. W., et al, "Semiscale Blowdown and Emergency Core Cooling (ECC) Project Test Report--Tests 803 through 820," IN-1404, Idaho Nuclear Corp., October 1970.
9.
Levine, S.:
Research Information Letter- #37 LOFT Reactor Safety Program Research Results through October 1,1978," NRC memo to H. R. Denton, September 29, 1978.
-10.
Batt, D.
L., " Experimental Data Report for LOFT Nonnuclear Test L1-4,
" TREE-NUREG-1084, EG&G Idaho, July 1977.
- 11. Prassinos, P.
G., "Fxperiment Data Report for LOFT Power Ascension Experiment L2-3," NUREG/CR-0792, EG&E Icaho, July 1979.
L
- 12. Kay, J. A., " Revision to Report on a Lower Plenum Phase Separation Model in Yankee Rown large Break LOCA Methodology," Letter to NRC (NRR),
Yankee Atomic Electric Co, April 1,1981.
l
- 13. Braum, G.J., et al, " Application of a Lower Plenum Phase Separation Model to Yankee Rowe Large Break LCCA Analysis," YAEC-1231, Revision 1 Yankee Atomic Electric Co., March 1981.
l' s
ee a
ry.
-y w
.--mr
-v<--w.
.-we--,
+
e--
-- ~