ML20235E021
| ML20235E021 | |
| Person / Time | |
|---|---|
| Issue date: | 06/28/1977 |
| From: | NRC |
| To: | |
| Shared Package | |
| ML20234E460 | List:
|
| References | |
| FOIA-87-40 NUDOCS 8707100386 | |
| Download: ML20235E021 (6) | |
Text
e i
n ENCLOSURE SAFETY EVALUATION FOR GENERAL ELECTRIC ECCS EVALUATION MODEL CORRECTIONS I
Introduction-The NRC was infomed by General Electric at meetings on December 15, 1976, and January 12, 1977, of several errors in-the plant calculations perfomed in accordance with the approved GE Evaluation Model. Four errors were discussed which can b'e categorized as:
(1) pressure rule, (2) vapor-
'izationca.lculation,(3)frictioncreditindischargebreak,and(4) structural absorption.
In addition, the NRC was informed that a logic error l
existed in the REFLOOD 03 computer program under certain conditions. The.
.l staff evaluation of the four model corrections and the REFLOOD code l
l modification is presented in the following sections.
{
II. Pressure Rule 1
- The reactor pressure assumed in the REFLOOD computer program is derived L-from the SAFE code calculation since REFLOOD does not incorporate an integral pressure calculation. This pressure is significant in t' hat it determines the steam velocity at the upper tie plate, and consequently limits the core spray water accumulation deterinined by the countercurrent i-flow limiting (CCFL) correlation. A lower pressure is conservative since it increases the steam velocity and reduces liquid penetration. Since the REFLOOD code considers the CCFL phenomena, its calculated reflood times are
. longer than those calculated by SAFE. This difference can influence the manner in which the SAFE pressures are applied to REFLOOD since a pressure increase is calculated by SAFE during core recovery. The use of this pressure gj71f B70623 THOMAsg@7-40
~
~ ~
~
~
l I
- ~
.s
! i rise during core recovery is not justified.
s.
'Section 2.2 on page II-7 of NEDO-20566 states the following: "At the time reflooding is calculated to occur, the system pressure and ECCS flow rates are essentially constant; thus, if the REFLOOD code indicates reflood times longer than the SAFE calculations, the system pressure and ECCS flows will be assumed to be constant".
G.E. has interpreted the above to mean that for all cases the pressure should be held constant at the value calculated by SAFE at the time core recovery begins. This value would be conservative for those cases,where the SAFE pressure is calculated to increase at the start of reflood and remains above the assumed constant value. However, if the SAFE pressure drops below the constant value determined at the time core recovery begins, it would be non-conservative to assume the higher constant pressure.
'The' staff finds that the GE interpretation of the referenced statement can result in a non-conservative evaluation of CCFL effects under certain conditions as stated a'bove. To avoid that situa' tion, the ' pressure input to the REFLOOD code should always be the lesser of the. SAFE transient pressure or the constant value computed at the time core recovery begins in the SAFE calculation.
l
.III. Vaporization Calculation The vaporization of spray water in the fuel channels is one of the steam sources considered in the REFLOOD computer program. This steam source is part of the total steam upflow in the countercurrent flow limiting I
9
..e
. = -
==.ue m e
- e-
.__.m.
5 calculation. To determine the amount of.; ray water vaporized, a vaporization correlation based on full length test data is used in REFLOOD. This correlation is given on page II-20 of NEDO-20566.
We have been informed by GE that this correlation was incorrectly input 1
.to the REFLOOD code. GE proposes to correct this error by incorporating the correlation referenced above. The staff finds this acceptable.
IV. Discharge Break Calculations i
i To calculate the critical flow for postulated discharge line breaks in
]
the recirculation line, the GE LOCA Evaluation Model allows' credit for the minimum design f1/D values.between the vessel and the break. This value is typically about 0.3, but GE has erroneously assumed a value of f
zero in plant calculations. The use of this line friction would reduce
'the flow rate from the postulated break, and result in somewhat lower calculated peak cladding temperatures.
GE has proposed to correct the LOCA calculations for discharge line breaks by taking credit for the minimum f1/D as specified o'n page I-187 of NEDO-20566. The use of this value is acceptable to the staff.
i V.
Structural Absorption i
Section I.A.4 of Appendix K in 10CFR50 states that the fraction of locally l
generated gamma energy deposited in the fuel pin may be less than 1.0 if 1
justified by calculations. Consequently, the GE LOCA Evaluation Model l
i i
W
'. * ~
(
i l,
allows a credit of 4% for ganna absorption in structural components.
This credit is normally taken by reducing the power generation by 4% in the CHASTE. calculation. GE has reported that for several reactors the 4% credit was erroneously subtracted a second time 'during the implementation ofthefuelpowerlimits(MAPLHGRS).
l i
GE has corrected this error in the affected plants. This correction is accepte a to the staff.
VI. Reflood Calculation I
The GE REFLOOD computer program is used to calculate the reactor vessel
~
refill transient during a postulated LOCA. This calculation' starts at the time core spray is initiated and continues until the core is reflooded.
, The REFLOOD calculation considers the countercurrent flow limiting (CCFL) phenomena resulting from steam generated in the reactor vessel, and provides an improved refill calculation over that calculated by the SAFE computer program. CCFL tends to limit the accumulation of water from the core spray l
system.
The steam produced in the reactor vessel during refill can flow upward through the fuel channels and the jet pumps. However, only the steam flow through the fuel channel limits the accumulation of water from the core spray system. When the liquid level is below the level of the jet pump discharge, the steam exits through both the jet pumps and the fuel channels.
When the liquid level is above the jet pump discharge, all steam is assumed L
l
j
-0 a
a to' exit throusti the fuel channels.
l References 1 and 2 describe an error in the REFLOOD 03 logic which determines the flow paths available for steam venting. The' code correctly vents the steam through the appropriate flow paths for all cases except when the liquid level' is initially above the jet pump discharge, but subsequently drops below this level. While above this level, the code logic correctly assumes that all steam vents through the fuel channels.
However, when the level drops below the level of the jet pump discharge, the logic error still assumes that all steam vents through the fuel channels rather than assuming that a porti,on vents through the jet pumps. This assumption is'ove'rly conservative and is unnecessary. This error in code logic was corrected in REFLOOD 04. This' modification is in agreement with the REFLOOD code description in Section 3.6 on page II-21 of NEDO-20566.
.Thus, the staff finds the proposed 1,ogic change in REFLOOD'04 acceptable.-
VII. Conclusion The staff finds that the' G.E. Evaluation Model corr,ections discussed in the previous sections are acceptable and should be used for application to appropriate plant calculations. The REFLOOD 03 computer code and use j
of a zero f1/D value for discharge breaks are still acceptable since these errors produce conservative results. The errors in the pressure rule, vaporization calculation, and structural absorption were non-conservative.
Thus, these features of the model are only acceptabic as corrected.
(
References 1.
Letter A.J. Levine (GE) to Dr. D.F. Ross (NRC), dated October 13, 1976, subject "BWR ECCS LOCA Evaluation Model".
2.
Letter, A.J. Levine (GE) to Dr. D.F. Ross (NRC). dated January 20
-1977, subject " Differences Between ECCS Codes REFLOOD 03 and 04".
4 l
I l
I e
- 3 e
4 e
9 l
l l
l l
t t
i l.
)