A06421, Forwards Response to 870427 Request for Addl Info Re NULAP5. Lack of Water Slug Formation Due to HPI Would Be Independent of Nodalization in Sys Representation
| ML20214S762 | |
| Person / Time | |
|---|---|
| Site: | Haddam Neck File:Connecticut Yankee Atomic Power Co icon.png |
| Issue date: | 06/02/1987 |
| From: | Mroczka E CONNECTICUT YANKEE ATOMIC POWER CO. |
| To: | NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM) |
| References | |
| A06421, A6421, NUDOCS 8706090405 | |
| Download: ML20214S762 (42) | |
Text
{{#Wiki_filter:~ - - _ _ - _ - . O CONNECTICUT YANKEE ATOMIC POWER COMPANY B E R L I N. CONNECTICUT P o Box 270 e HARTFORD, CONNECTICUT 061414270 m e m o,a June 2,1987 mus-sooo Docket No. 50-213 A06421 Re: 10CFR50.46 U.S. Nuclear Regulatory Commission Attn: Document Control Desk Washington, D.C. 20555
References:
(1) F. M. Akstulewicz letter to E. 3. Mroczka, Small Break LOCA Code--NULAP3, dated April 27,1987. (2) F. M. Akstulewicz letter to E. 3. Mroczka, Small Break LOCA Code-NULAP3, dated March 12,1987. (3) F. M. Akstulewicz letter to E. 3. Mroczka, Small Break LOCA Code--NULAP5, dated February 12,1987. Gentlemen: Haddam Neck Plant Response to Request for Additional Information on NULAPS In Reference (1),11 requests for additional information were made concerning the NULAP5 manual. These requests were in addition to questions forwarded by References (2) and (3). Attachment I contains a response to each of these requests except for Question 10. A response to this question will be forwarded by June 30,1987. It is Connecticut Yankee Atomic Power Company's understanding that satis-factory and timely response to this additional question will enable the Staff to complete their review of the NULAPS code prior to the start of the 1987 refueling outage. Should you have any further concerns, please contact us. Very truly yours, CONNECTICUT YANKEE ATOMIC POWER COMPANY hl6 GW E. J.'Mgr czka p Senior Vice President cc: W. T. Russell, Region I Administrator g F. M. Akstulewicz, NRC Project Manager, Haddam Neck Plant P. D. Swetland, Resident Inspector, Haddam Neck Plant '00t 8706090405 870602 I PDR ADOCK 05000213 F PDR
Attachment 1 Response to Requests for Additional Information on NULAP5 June 1987
t
- 1. Northeast Utilities response to question Q.II.5 did not discuss water slug formation due to HPI at the Haddam Neck Plant. The question was intended to clarify how well NULAP5 calculates water slug formation (filling of a volume with volumes essentially steam filled up and downstream of the filled volume) in Haddam Neck SBLOCA analyses. To resolve the question, clarify whether NULAPS is capable of modeling water slugs. How sensitive is this modeling to the noding used? Also, clarify whether water slugs were ever calculated in Haddam Neck SBLOCA analyses. If no water slugs were calculated, clarify whether or not this could be due to the nodaliza-tion used in the cold legs.
The water slug behavior wherein a slug of water in the cold leg piping is completely surrounded with steam was not predicted to occur for any of the breaks analyzed in the Haddam Neck SBLOCA break spectrum analyses. Since liquid was present in the cold legs during initiation of the HPSI, there is no potential for this phenomenon to occur and, as such, the lack of such phenomenon would be independent of the nodalization in the system representation. 0001.0.0
- 2. It was stated in response to question Q.IV.2, that the Semiscale 1-1/2 loop pump data was used for the two phase homologous curves in the NULAP5 model for the Haddam Neck pump because it was the only data available for all the various pump operating regions. An additional source of two-phase pump data is the CE-EPRI 1/5 scale pump data.1 The pump used in this test program had a specific speed of 4209 rpm as compared to a specific speed of 926 rpm for the Semiscale 1-1/2 loop pump. Discuss whether the data from the CE-EPRI tests was considered for use in the Haddam Neck pump model.
If this data was considered but not used, clarify why it was not used. At the time of the NULAPS model development and Haddam Neck Plant small break LOCA spectrum analyses, the CE-EPRI 1/5 scale pump data was unavailable; therefore, this data was not considered for use as input to the Haddam Neck pump model. l 0002.0.0
)
I l
' I i
- 3. Northeast Utilities indicated in response to question Q.IV.3 that the pump model in NULAPS was assessed against a calculated pump coastdown from the Haddam Neck FDSA and a number of assessment calculations for LOFT and Semiscale tests. A review of the LOFT and Semiscale assessment calculation results did not find any code / data comparisons (such as loop flow rates, pump speed, pump differential pressure, etc.) that would verify the adequacy of the pump model in NULAP5. Provide this type of code / data comparisons for the LOFT and Semiscale assessment calculations, using a time scale that allows a meaningful comparison between the code and the data to be made, so that the adequacy of the NULAPS pump model can be verified.
The attached figures (No. I and 2) provide a comparison with the data of the calculated differential pressure across the pump and a comparison of the pump speeds for LOFT test L3-6. In the test the reactor coolant pumps were left running until approximately 2400 seconds after initiation of the break. Figure 1 shows the plot of the pump differential pressure. NULAPS calculated results compare favorably with the data. The code steady-state calculation at time zero shows a somewhat smaller AP than the recorded data (440 kPa vs. 475 kPa). The degradation in the pressure for the first 300 seconds of the transient compares favorably with the data. Between 350 and 400 seconds the code calculated a small increase in the voiding of the suction side. The pump head decreased, but the differential pressure increased. After 400 seconds the pressure across the pump decreased at a somewhat faster rate than the data. The rate of pump pressure degradation for both the calculated and the recorded data becomes essentially equal after 1000 seconds. Figure 2 shows the comparison of the pump speeds. Both the data and the calculated results show essentially constant pump speed throughout the first 2400 seconds of the transient. The pump model runs at a constant speed of 3139 rpm while the data shows the speed of pump rotation equal to approximately 3215 rpm. 0003.0.0
Further comparisons of the pump performance during coastdown are shown in Figures 3, 4, and 5 which are enclosed in the response to Question 4. The calculated coastdown rates for the LOFT L3-7 and the Semiscale S-07-10D tests compare favorably with the data. The calculated coastdown rate for the LOFT L3-5 test indicates that the code calculated a somewhat faster coastdown rate as compared with the data. 0004.0.0
FIGURE 1 LOFT L3-6 DIFFERENTIAL PRESSURE IN INTACT LOOP ACROSS PUMP 2 I 600-4 500-
)
0 F \ l F I h,400- g E g q T \ I \ A \ L \ F 300-R E s R
\ '~\
E \ 200-
\\
K
\
A \
\
\
N 100- \
\
\.
s s 0, , , 0 400 800 1200' 1600 2000 2400 TIME (SEC) CURVE- DATA --- NUL APS
FIGURE 2 LOFT L3-6 PUMP SPEED FOR INTACT LOOP-PUMP 2
-5000 J 4000-P U _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ . _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _
M p 3000-S P E E D
~
R P M 1000-0- ' ' ' I 4 & 6 6 4 6 0 400 800 1200 1600 2000 -2400 TIME (SEC) CURVE ORTA --- NUL AP5
s
- 4. In response to question Q.V.1, Northeast Utilities stated that NULAP5's ability to calculate single- and two phase natural circula-tion was assessed through comparison to the integral system tests from .T and Semiscale. A review of the LOFT and Semiscale assessment calct iation results did not find any code / data comparisons (such as loop flow rates) that would verify the adequacy of NULAPS to calculate natural circulation. Provide this type of code / data comparisons for the LOFT and Semiscale assessment calculations, in a format that allows a meaningful comparison between the code and the data to be made, so that the adequacy of the NULAP5 natural circulation models can be verified.
The attached figures (3 through 5) provide the comparison of the calculated loop flow rates with the data. Figure 3 is a plot of the loop flow rate for the LOFT L3-7 test. The L3-7 Experimental Data Report encompasses only the first 2000 seconds of the transient. The natural circulation starts at approximately 100 seconds after the transient initiation. The plotted data is from two different flow sensors, both located in the hot leg flow venturi, but at two different elevations on the inside circumference of the venturi. The two sensors measured unequal flow rate, possibly due to some flow stratification. The NULAP5 calculated flow rate unpredicts the data. Figure 4 is a comparison of the intact loop hot leg flow rates for the LOFT test L3-5. The plotted data is again from the two sensors located in the hot leg venturi, but at two different elevations in the pipe. The data encompasses the first 200 seconds of the transient. No additional flow data was available in the LOFT L3-5 Experimental Data Report. As can be seen in the plot, the NULAPS calculated results fall, for the most part, between the two data plots. The code calculates a faster initial reactor coolant pump flow coastdowu after the pumps are tripped 0.8 seconds after the break (17 seconds of coastdown time in the experiment vs. 10 seconds in the NULAPS calculation). 0007.0.0
Figure 5 shows a comparison of the intact loop hot leg flow rate for the Semiscale S-07-10D test. The data is taken from Figure 357 of the Semiscale Experimental Data Report (the momentary increase in the flow rate at 70 seconds could possibly be due to instrumentation error). The plot shows that NULAP5 - calculated flow rate is underpredicted by the code from 50 to 460 seconds in the transient. The calculated pump coastdown rate, however, is well predicted by NULAP5. After 460 seconds the agreement in the calculated and the experimental results is very good. 0008.0.0
4 FIGURE 3 LOFT L3-7 FLOW RATE IN THE INTACT LOOP HOT LEG 500l
. =
400-l l 1 F l L 0 I W I 300- l R l A T ! E l I I K ! G 200 - I
/ I s l E
C ! I I I i 100- l l I ' l,
\{._____
s 0-0 500 1000 1500 2000
. TIME (SEC)
CURVE ORTA-BOTTOM ------- DATA-TOP
--- NUL AP5
FIGURE 4 LOFT L3-5 FLOW RATE IN THE INTACT LOOP HOT LEG 500-l
- I l
l 400-l l i F l I O W I 300- l R l 9 i T E I I I I K G 200 - I
/ \
s i E , l : C g ' I i l 100- l
\
\ .
\
\
\ j! ' N ~ ~ ~
_ _w - t, ,,............ ~* ....,,,,,
'w~% _
s....,,,,........ 0-TIME (SEC) CURVE DATA-BOTTOM ------- DATR-TOP
--- NULRP5
FIGURE 5 SEM! SCALE S-07-100 FLOW RATE IN THE INTACT LOOP HOT LEG 8-1 l 7-I ag I l s _ l1 I F \ L j 05- 1 i R I A l T E I q_ l K c I
/_ 1 s$~ l E
{ c ; I a- 1 I I I i 1- I l l l
- o. 1 0 100 200 300 400 500 600 700 800 TIME (SEC)
CURVE DATR --- NUL APS
4
- 5. Additional information on the clad swell and rupture model was requested in question Q.VI.11. In response, Northeast Utilities stated that the full description of these models was found in the NULAPS manual. The section in the NULAP5 manual on the fuel models (Section 3.2.4) was reviewed again and information on the clad oxidation model, the gap conductance model and the fuel rod internal pressure model was found. Also, there was a general description of how modifications were made to the geometry of the heat structures and hydrodynamic components, but no information was found describing how NULAP5 calculated the geometry changes. That is, no information was found describing how the clad swell and rupture models work.
Therefore, provide for review a description of the clad swell and rupture models. Following is a discussion of the calculation of 6, the change in radius of a heat structure. 6 is applied to the original radius to increase the radius to account for the enlargement of the heat structure when there is a pressure and temperature transient. The calculation of 6, the change in length (or radius) is based upon AP and a stress and strain calculation. The rod internal pressure is calculated by a method discussed in previous responses. The AP is calculated by subtracting the heat structure's hydrodynamic boundary volume's pressure from the internal pressure. This is performed at all axial locations. Once the AP has been obtained the calculation of swell and rupture are as follows. A check is made to find the inner clad surface temperature. All further results are based on this temperature. A check is then made to determine if AP is positive; if not, the calculation ends. If it is positive a hoop stress is calculated by assuming a thin wall type formulation where:
)
P. X. - P X l H= AX 0012.0.0
m . . __ __ -_ . . _ . _ _ _ . __ _ _ _ _ _ _ _ - - - - - - - - l l
, I i
1 where: i H = hoop stress (Pa) 1 P. = internal pressure (Pa) 1 P, = external pressure (Pa) Xf , X, = internal, external radii (m) i AX = clai thickness (Xo -X.) i (m) The AX is updated every time step to account for oxidation. The hoop stress is then used to determine if rupture has occurred and the amount of elongation. Aftsr the actual hoop stress is calculated a tensile strength (in this case hoop stress) is found based upon the inner clad temperature from a property table of tensile strength vs. temperature. This . table is the maximum hoop stress that occurs before rupture at a given temperature. It is related to the percent elongation, from a percent elongation vs. temperature table, by the temperature. This table is the maximum elongation that would occur before rupture. A check is made to_see if the hoop stress calculated is greater than the one from the table. If so, then rupture has occurred. The percent elongation is then calculated. The maximum percent' elongation, E, is known from the table, so to determine the actual i amount of elongation for each step: Hoop stress actual
%E = %Emax , Hoop stress max As this is a linear deformation, it is converted to a radial displace-ment.
i i
; 0013.0.0 i,
4
-- - - - - , - - , + , . . ,,-e e- , - ~ _ , - .--, , ,. , - - , ,- , _ , . _,n- , . , . . ,,,m w+-n-.
1 . . . A
'This calculation is carried out from the beginning of the simulation so that the' channel flow area can be changing continually, although the flow area can ne.er increase over the previous time step's value. This allows for blockage to affect the entire course of the transient.
4 Following are the two tables used in the calculation.' They are , based upon Reference 7. I 4 STRAIN 26 DATA PCTES /273.0DO,0.55DO,366.0DO,0.53DO,477.0DO,0.51DO, STRAIN 27 ! *588.0DO,0.49DO,700.0DO,0.45DO,810.0DO,0.45DO, STRAIN 28
*921.0DO,0.48DO,1033.0DO,0.50DO,144.0DO,0.55DO,- STRAIN 29
*1255.0DO,0.65DO,1366.0DO,0.75DO,1477.0DO,0.60DO, STRAIN 30 STRAIN 32.
DATA TINSLS /273.0DO,4.48D8,366.0DO,4.27D8,477.0DO,4.14D8, STRAIN 33 )
*588.0DO,4.07D8,700.0DO,3.86D8,810.0DO,3.45D8, ' STRAIN 34 I *921.0DO,3.16D8,1033.0DO,2.21D8,1144.0DO,1.45D8, STRAIN 35
*1255.0DO,1.03D8,1366.0DO,7.6D7,1477.0DO,5.5D7, STRAIN 36 STRAIN 38 i
r ) l 4 0014.0.0 l
- 6. The response to question Q.VI.27 provided information on the new critical heat flux model in NULAPS. This model used the Biasi correlation to calculate critical heat flux at high mass fluxes and the modified Zuber correlation for low mass fluxes. This model was taken from RELAP5/ MOD 2.2 Because the critical heat flux model used in all previous NULAP5 assessment calculations was replaced, provide comparisons of NULAPS results to at least one set of critical heat flux data applicable to SBLOCAs to ensure the new model has been properly implemented in NULAPS.
In order to assess the effect of the new CHF logic three THTF transient film boiling tests were run. The tests are reported in Reference 9, Section 4.1.2. The three tests run are 3.03.6AR, 3.06.6B, and 3.08.6C. These tests were previously reported in Reference 10. Following is a discussior of test results and temperature plots of the top-most node. This node was arbitrarily picked, as all nodes eventually experienced CHF. It should be noted that all runs are preceded by a 20s steady-state. The discussion of the three plots follows:
- a. Test 3.06.6B (Figure 6)
In this test CHF occurs at approximately the same time in all cases. Biasi gives a somewhat (9 F) higher PCT and later in the transient shows a better agreement with the data.
- b. 3.03.6AR (Figure 7)
In this case the new CHF logic calculated CHF to occur one second later than the original calculation. This later CHF gave a PCT of 1206 F as compared to 1238 F in the original case. The new CHF shows at least as good agreement with the data as the unmodified code and is within the data experimental l error band. l l l 0015.0.0 ,
- c. Test 3.08.6C (Figure 8)
This test again showed a later CHF, about 2 seconds later than , the data and about 0.5 seconds later than the unmodified NULAP5. Both calculations show a peak temperature higher than the data with the results in agreement at the end of the run. 0016.0.0
FIGURE 6. FRS INNER-SHEATH TEMPERATURES AT LEVEL G vs TIME - TEST 3.06.6B 9 a-
=
9 g- ' -8
~
o (- ' 9 o [: g- -h ;7 e_ ', s Bo E 2 N-9$ ee -g 2s
- 8-x D TE-350AG t I p
d o~ O TE-342CG 6
- d A TE-334AG o I d ,,, ' + TE-318AG 0
3 o x TE-316CG O d - ? E. d
?
"o -
- Q 4 NULAP5 o
$~ ~
/
- -E- -S- NULAP 5 (BI ASI) 9 hb ---
0.0 I.0 I.0 [.0 [.0 Id.0 13.0 15.0 Id.0 Id.0 20.0 TIME (s) 1 I l I l l l
FIGURE 7. FRS INNER-SHEATH TEMPERATURES AT LEVEL G vs TIME - TEST 3.03.6AR 8 n N* 'L s g , a f/ -p C o.
/ -
5 E 8- @ h' $ / s $ / ] $ 3 0 9 Nh f 0- i
-g g tj 1 ! 1 8- g >-
?? '
?
$ h- / O TE-3429G O TE-318AG g { 'f C & TE-318CG o
+ TE-3208G D do /
d 3 E~ o
/ ~E 3.a
/ O NULAP5 o / 8 NULAP5(BIASI) 8] b o
[ '
= -
-g H
0.0 2'. 0 4'. 0 6'. 0 8.0 15.0 15.0 14.0 15.0 15.0 20.0 TIME (s)
FIGURE 8. FRS INNER-SHEATH TEMPERATURES AT LEVEL G vs TIME - TEST 3.08.6C 9 b } 9 - -e g-
,6-~1- = -
m-. 4l9
=
'._ s-g.
5-
- o '
/
- a C
~ -
58N & / O g S~ /
-g 5 89 8 3
8 8-5= 6 5 5a 1 -
/
E x
- h- D TE-350AG a.
w - /
/ o TE uaso ,@ =g b a c
b A TE-334AG o I , g_ / + TE-sissa s
/
sw "9 / O NULAPS 9" E' / 8 NULAP5(BIASI) ~! I 8- b A O a4 C.0 2'. 0 4' 0 5'. 0 8'. 0 15.0 15.0 15.0 !$.0 t$.0 25.0 2d.0 2l.0 N.0 N.0 30.0 TIME (il l l
. = - - , .. - - _ ._ . .- - - - .
- 7. - Qestion Q.IX.1. requested Northeast Utilities to explain the differences between a number of equations given in the NULAP5 manual and the 8
RELAPS/ MODI manual . An answer to this question was not included in 1 Reference 4. Provide the answer to Question Q.IX.1. As indicated in Reference 4, the referenced equations in the NULAP5 manual were extracted from the September 1981 version of the RELAP5/ MODI Users' Manual (Reference 2) which was issued in the draft form (NUREG/CR-1826, EGG-2070 DRAFT, Revision 2). Reference 2 itself-is the final version of.the code users' manual which was published in March 1982. The March 1982 edition has undergone a number of error i corrections in the description of the RELAPS/ MODI system of five differential equations that form _the basis for the numerical solution scheme. Below is a list of the errors contained in the NULAP5 7 manual. I
- a. Equation 3.1.1-15: the last term in the equation which represents the virtual mass acceleration term from_the phasic momentum equations contains two typographical errors. The partial derivative of the phasic velocity difference-(v -vj) should not contain the additional interfacial velocity, Vy , term. Also, inside the same square bracket, the second partial derivative should be represented by: vf (Oy /8x). In the RELAPS numerical i solution scheme, the virtual mass term is omitted from the mixture thermal encrgy equation.
- b. Equation 3.1.1-17: the first term on the right-hand side of the equation should carry a minus sign. This equation is the final form of Equation 3.1.1-15 which does contain the negative sign. l The omission of the sign is a typographical error in bott. :tur q RELAP5 September-1981 Manual and in the NULAPS Topical.
+ f J 4 , 0017.0.0 i (
.--- w -r ,- w -v -
-w - - - - - -
r- r - r ,-e,e , - . - - - , e- --~ w-a m* r T- :-*.n r
, - .- . . -. ~ _ . . . - - - -... -- . . . ._ _ .
r
- c. Equation 3.1.1-28: this is the integrated form of Equation 3.1.1-17 (integrated from junction j to 'j + 1). As such, the last term on the right side of the equation should not contain the A term-and, also, the absolute value of the volumetric vapor generation term Fg should be used. The correct form should be: +(l{l/2) l' (v - v Jay, 4 8 f
- d. Equation 3.1.1-29: the last term on the right side of the equation should read: -{(v - vf ) (Xg - XK ). is is a f
typographical error, since the sum momentum equation (Eqn. 13 , in the RELAPS manual) does show the correct subscripts in the i j phasic velocities. Equation 13 is integrated to obtain 3 Equation 3.1.1-29. i
- e. Equation 3.1.1-33: the void fraction terms (a and af) on the right side of the equation should have a dot overscore to indicate that they are donored qua'ntities (d , Ef ). This is consistent with the donored void fractions on the left side t
of the equation and also consistent with the basic assumption in the code that both mass and energy are convected from the same cell.
- f. Equation 3.1.1-36: the last two terms in the square bracket on
- the right side of the equation contain typographical errors.
The first of the two terms should read: -(ap)(v)$*IFWF" ff f
- (the NULAPS manual reads
- ... (v )"+ ...) and the second term should read: - (F 8)j"(v -8 v )J"*I f (the NULAP5 manual reads:
(v -vf )"). The subscripts in the velocity terms should denote the particular phase (either vapor or liquid); the superscripts should denote the new time. level evaluation for the phasic ! velocities in the mass and energy transport terms'(n+1, instead of n). Implicit evaluation is used for the phasic velocities calculations. s 1 I i 0018.0.0
,---,,..,7 , - - . , , . - - - - , , , _ , - _ , , . . , , , , , . ..a_.__, ,- , . _ , . , , - , , , , , - _ . .
. - = . - _ _ . - . . . - . .. .. -.. _ _-
- g. Equation 3.1.1-37: this equation contains a number of errors.
In the first square bracket on the left side of the equation the coefficient C should be multiplied by density, p-(in the equation this is wrongly' represented as C , p being a subscript).
.9 Also, on the same side of the equation the two square brackets containing the "C" coefficient should not have the \ multiplier ,
in front. This is clear in Equation 3.1.1-30 from which '
- Equation.3.1.1-37 is derived in the finite-difference form. On.
the right side of Equation 3.1.1-37, the expression inside the
!- last square bracket should include superscripts indicating a time level index. Thus, the density and the void fraction l terms (p,0) should have an "n" superscript (indicating the old time step value) and the velocity terms (v y , v , and v )fshould f
s have "n+1" superscripts (the new time step value). The interphase mass interchange term, F , should have an "n" superscript. In - addition, a superscript "n" should not be used on the outside of the square bracket.
- h. Equations 3.1.3.2-8 through 3.1.3.2-10: Equation 267 in Reference 3 is the correct version, consistent with the methodology
. contained in Reference 11. The NULAPS equation in the manual is incorrect. Reference 11 is the basis for the RELAPS heat conduction solution.
i,j . Equations 3.1.3.3-1 and 3.1.3.3-2: there is a missing negative ] sign in front of the first term on the' right side of the two 1 equations. Equations 268 and 269 of Reference 3' include the I minus sign and are consistent with the Reference.11 equations. .i
- k,1. Equations 3.1.3.3-3 to 3.1.3.3-5 and 3.1.3.3-6 to 3.1.3.3-8
l these equations contain incorrect superscripts. Equations 270 and 271 of Reference 3 are the correct versions and are consistent with Reference 11. t I d i t 0019.0.0 i
.___. _ _ __ -,- - _._ ._ _ _ ~ . _ , - _ . , _
- m. Equation 3.1.4-10: this equation should have a negative sign on the right side, consistent with Reference 11. Equation 219 of Reference 3 is the correct version.
- n. Equation 3.1.4-11: missing brackets after the first and the second summation terms. Incorrect subscripts on the summation term. Equation 220 of Reference 3 is the correct version, per Reference 11.
0020.0.0
- 8. In the NULAP5 manual, Section 2.0, it was stated that the IBM version of RELAP5/ MODI, cycle 6, was updated to become cycle 14. It was-assumed this process included all the updates from cycle 7 to
- 14. It was then noted in the manual that the code was being updated from cycle 14 to 18. The response to question Q.IX.2 states that one update to RELAP5/ MODI in cycle 18 should not impact the results of the assessment calculations in the NULAPS manual because the original update to RELAP5/ MODI, cycle 14, was not included in NULAPS. Because of this response, it is not clear whether all the updates to RELAPS/ MODI from cycle 7 to 18 have been included in NULAP5. Question Q.IX.2 also asked Northeast Utilities to provide information on the status of each update to RELAP5/ MOD 1 from cycle 19 to cycle 29, the final released version of the code, with respect to its implementation in NULAP5. You stated that the updates from cycle 19 to cycle 25 were included in a test version of NULAPS but a final decision had not been made on the implementation of these updates in NULAPS. You also stated that known code error corrections would be implemented in NULAP5 as they were released by the RELAPS/ MOD 1 developers. Because cycle 29 is the final released version of RELAP5/ MODI, it is clear that all corrections for known code errors have been released by the RELAP5/ MODI developers. These error corrections are included among those listed in the update summary sent with the earlier set of question.5 Clarify the development history of NULAP as it relates to RELAP5/ MOD 1 and the updates to the RELAP5/ MODI code. If any of the cycle 7 to 29 updates to RELAPS/ MODI were not included in NULAP5, identify those updates and state why the update was not included. If an update was used in modified form, provide this information and briefly state why it was modified.
This information is needed to verify that all known errors in the RELAPS/ MODI base code have been corrected in NULAP5. This must be verified before licensing approval of NULAP5 can be given. The Haddam Neck small break LOCA break spectrum analysis was performed with a " frozen" version of NULAPS which was based on RELAP5/ MODI code cycle 14. As such, all the updates to RELAPS/ MODI from cycle 7 to 14 were included in NULAP5. However, during the code qualification efforts, one of the updates was removed. This update is in cycle 14 which specified that donored phase void and phase density be used in calculating the convective terms in the momentum equations. Instead, NULAPS used average convective terms which resulted in a better predictive capability at the time when the code was being benchmarked against the experimental data. This approach was verified as being valid when cycle 18 updates to RELAP5/ MODI later removed the 0021.0.0
_ _m . W6
+ ]
4 , donored properties in the convective terms and specified the average a terms instead. Modifications-to NULAPS and the code's extensive qualification i effort, which are reported in the NULAP5 topical reports, have.shown . .' that the code can satisfactorily predict the phenomena associated
; with small break LOCA transients such-as a core.two-phase level and peak cladding temperatures. Further upgrading.of the code was !
deemed not necessary in view of its satisfactory and conservative, '
~
when compared to a range of separate effects and integral tests,. s ,
. i ~, AU predictive capability. The licensing small break LOCA" analysis for-j Haddam Neck was performed with a " locked" version of NULAPS which , , t i j .could meet the Northeast Utilities quality assurance requirements 1
for code development and code analyses. l j. i Since the submittal of the topical reports, a review of the RELAP5 code updates subsequent to cycle 14 has been performed. The purpose was to determine the impact of each update on the assessment results presented in the NULAPS Manual.
- The attached table lists all the updates to RELAP5 and discusses
; their potential impact. Those updates'that may enhance the code ! capability and correct the code errors will be' considered for a
implementation in NULAP5. This will be done only after a selection . , of the original code assessment tests are reanalyzed with an updated ^ version of NULAP5 and the calculated results warrant a change to the code. i
; In general, the updates can be. classified in three ways: 1) updayes-i
- i to facilitate the input deck preparation (for example, the loop elevation checker), 2) updates to make the code run more efficiently,.
with fewer aborts, and 3) updates to correct.known errors. The . w! first two categories comprise the majority of the updates and as such, do not impact to any significant degree the existing NULAPS assessment results. The'last category does not contain any major , S code corrections (with the exception of the RELAP5Eaccumulator 0022.0.0 t
- . - . . u~ -
,.t' .
)
. 4 Al
.f g
.s del, which was not used in analyzing the Haddam Neck break spectrum
, since the plant does not have the accumulators),and therefore is not perceived to have a major impact on NULAP5. However, the future 1
+
updates to NULAPS will include this.' category for implementation. It should be noted that a majority of the code error corrections concerns RELAPS models that were not used in performing the Haddam ~ Neck break spertrum analysis. 'Ihese include the motor and servo E valve models, the original RELAPS nonco'ndensible option,
,renodalization option, etc.
t
- t. . ,
a i;
-1 i j' >
s f 9
)
's i e
t
; i.
s
\ g
,t
'P 0023.0.0 i( ,7 f
... ~ - , . . - ~_ . - - . .
SEQUENTIAL
- CYCLE NO. UPDATE NO. S110RT DESCRIPTION IMPACT ON NULAPS 15 1 Fix abort in.llTCOND during None, this option not exercised.
S. S. initialization. 16 1 ; Error in renodalization of time- None, no renodalization using T.D.V. or J. dependent volumes or junctions. has been attempted in NULAPS. This correction will be included in the list of future updates to NULAPS. 2 Correct British conversion constant NULAP5 uses a different plotting package, for heat. transfer rates in minor edits and plots. 3 Correct direct moderator heating. NULAP5 does not use this option. 4 Fix incorrect time advance of No impact. This option never was used. reactor. kinetics when time step is repeated. 5 Bypass IJPROP for old junctions No impact. This correction will be to prevent unnecessary changes included in the list of future updates during renodalization. to NULAPS. 6 Fix accumulator momentum equation. No accumulators at lladdam Neck. 17 1 Fix error in DTSS P w$en setting. No restarts from nonstandard edits Courant limit. Error during have been attempted. restart from a nonstandard edit. 2 Fix error in ICOMPN from A minor correction. cycle 16 update. 3 Remove change to IJPROP to 'NULAPS Ifaddam Neck Plant model is prevent initialization of all fixed. .No renodalization is necessary. components during renodalization. This correction will be included in the list of future' updates to NULAP5. 0024.0.0
g. SEQUENTIAL CYCLE NO. UPDATE NO. SIIORT DESCRIPTION IMPACT ON NULAP5 4 Correct problems with noncondensible NULAPS SBLOCA analyses did not use steam table failures beyond the noncondensible gas option. critical point. 18 1 Corrects a hole in the MDOT logic. Change is minor and does not impact the NULAPS nonequilibrium model. This correction will be included in the list of future updates to NULAPS. 2 Changes to the plot package. NULAPS uses a separate plot package. 3 Corrections to the logic and No accumulators at lladdam Neck. momentum solution in ACCUM. NULAPS accumulator model is not used. 4 Corrections to STATE for No noncondensible option used in noncondensibles. NULAPS SBLOCA analyses. 5 Corrections to VEXPLT for the No accumulators at fladdam Neck, accumulator model. 6 Removes cycle 14 update to VOLVEL No impact. and IULVEL when a phase goes to zero or a pha:e disappears. 7 Removes cycle 14 update to VEXPLT NULAPS always used average terms, instead that donored void and density of donored values. for convection terms. 8 Fix reactor kinetics option that NULAPS uses a separate table of decay makes decay heat too low. heat vs. time, llowever, this update will be included in the list of future updates to the code. 9 Temperature convengence criterion NULAPS retained the 0.002 temperature changed from 0.002 to 0.004. criterion in the RELAPS noncondensible model. Ilowever, this option is not used in NULAPS. 0025.0.0
f SEQUENTIAL CYCLE NO. UPDATE NO. SHORT DESCRIPTION IMPACT ON NULAPS 19 1 Fix air state errors NULAP5 does not use RELAPS noncondensible option. 2 Fix regative square root under This correction will be included in the list stratification conditions, of future updates to NULAP5 A minor impact. 3 Fix computation of bias reactivity, No impact on NULAPS calculated results. error when initial reactivity is nonzero. 4 Fix pump renodalization. No impact. Haddam Neck analyses have not used this option. 5 Fix restart error in use of A minor correction, no impact. control variable as power input to a heat structure. 6 Remove test of unit trip component No impact, being initialized to 0.0 or 1.0. 7 Fix printout of moment of inertia A minor change, no impact. in British units in valve input routine. 8 Fix reactor kinetics when no power No impact. This update will be included from gamma decay is specified. in the list of future updates to the code. Error was entered during fix of gamma' decay processing, does not change answers when gamma power was specified. 9 Add missing asterisks to an error A mostly cosmetic change. diagnostic in RTMDJ and RTMDV. 10 Fix bad checking in Doppler No impact on the calculated results. This feedback table update will be included in the list of future updates in NULAP5. 0026.0.0
r e SEQUENTIAL CYCLE NO. UPDATE NO. S110RT DESCRIPTION IMPACT ON NULAPS 11 When time-dependent volume No impact. NULAPS.did'not use the conditions are specified as T, noncondensible option in the time-dependent X, XA, the variable QF=1. - volumes. This update will be included in QUALE (I) was not defined in the list of future updates to the code. TSTATE 12 Fix undefined variable in FIDRAG No impact. This update will be included in. the list of future updates to NULAPS. 13 Fix indefinite in TSTATE for No impact. This update will'be included in restart, the list of future updates to.NULAPS. 14 Corrections to wall friction, This update has a' minimum impact on the NULAPS small break LOCA results. Wall
~
form loss, and dissipation in VEXPLT. friction effects'in SBLOCA transients are not dominant. .However, this update will-be-included in the list of future updates.to NULAP5. 20 1 Fixes a branching problem by 'No branching problems have been noted multiplying viscous terms in during the sensitivity studies.or in momentum equations by ARAT**2. the break spectrum results. Care was taken to use calculated form K-factors, rather than having the code calculate them.' The code-calculated K-factors could have produced oscillations for-large contraction / expansion ratios. This update will be included in the list of future updates-to the code. 2 Correction to RBRNCH.for velocities A minor initial edit correction of the when using English units, branch input. 3 Improvements to reduce mass error. Editorial change, to monitor the~ accumulated. Add mass error and total mass to mass error during the analysis.
- scan request. variables.
0027.0.0
T o SEQUENTIAL-CYCLE NO. UPDATE NO. SHORT DESCRIPTION IMPACT ON NULAPS 4 Mass error into state. Subroutine A printout of the total ~ mass error modified: STATE. during debug print. This is mainly for diagnostic purposes. 5 Updates to remove energy correction NULAP5 does not use the RELAPS when QUALS, QUALA is truncated. noncondensible gas option. There are Does numerical average of junction no junctions during SBLOCA with - properties at zero velocity. zero flow (with the exception of time-dep. junctions). 6 Updates to redo PRESEQ...VFINL An examination of the NULAPS sensitivity and/or cut the time step and studies and the break spectrum results repeat of velocities flip-flop, has not indicated any instances of. velocity flip-flops. These situations may arise in analyzing steady-state conditions which are not characteristics of LOCA transients. 7 Fix subroutine MOVER to reset See the comment above. state for velocity-flip-flop repeat. 8 Put PV work semi-implicity The effect of the PV term has not. in EQFINL. been explicitly determined. However, i NULAPS sensitivity studies indicate 4 that the exclusion of the term in j subroutine EQFINL is not significant. This update will be included in the j list of future updates to the code. 9 Put PV. work semi-implicitly .See the comment above in'(8). in PRESEQ. I 10 See the comr.ent above in (8). Delete the' explicit PVtwork from VEXPLT. 11 Delete the explicit PV work There are no accumulators at Haddam from ACCUM. Neck. NULAPS accumulator model is not used 0028.0.0 i
y SEQUENTIAL CYCLE NO. UPDATE NO. SIIORT DESCRIPTION IMPACT ON NULAPS 12 Update to use heat transfer This option is not used in NULAP5. solution in time-step control. 13 Add coding to call elevation This modification helps in debugging the checker. input for errors. The NULAP5 break spectrum results are'not affected, however, since the input is carefully-checked for elevation closure before beginning the analysis. 21 1 This modification calculates the NULAPS CliF subroutine has been critical heat flux using the modified to-include the ZUBER and ZUBER and BIASI Cl!F correlations. the BIASI correlations. See also response to Question 6. 2 Update to correct heat t rans fer Printing error correction. errors by TIIC, correct heat transfer mode printout. 3 Correct heat. transfer selection A minor modification. No impact. in post DNB. 4 Correct error in llSU's film The only effect is seen at medium flow,.high boiling formulation. pressure of.the transition point of film boiling. This update results in higher heat fluxes. The original version is, therefore, more conservative. 5 Correct transition film boiling See the comment above. slope calculation. 6 Fix heat structure printout with '5ditorial change, no_ impact on English units. calculated results. 7 Valve updates which fix the These updates concern mainly the
. interpolation over form loss -motor and servo valve models. In rather than CSUBV. NULAP5 trip valves are used. This update will be included in the list'of future updates to the code.
0029.0.0
r SEQUENTIAL CYCLE NO. UPDATE NO. SHORT DESCRIPTION IMPACT ON NULAPS 8 Fix typo error in ISTATE and No' impact. TSTATE. 9 Updates to implement VHR stuffer This update reduces-a potential mass for 2 phase mass error. error by recalculating void fractions and qualities of two phase mixture. However, cycle 27 deletes this update. 10 Fix FRICTF for large ROUGHV/DIAMV A minor correction, no impact and Reynold's number, on NULAPS assessment results. 11 Fix FWDRAG for possible negative This update corrects a " residual" friction friction coefficient. coefficient of a phase that has disappeared. There is no1 impact on NULAPS - calculated results since the code'would abort the run if a negative friction coefficient is calculated. This correction will be included in the list of future updates to NULAP5. 12 Change read of water property file - No impact on NULAPS. The code has to use STH2XJ which uses buffer been converted to IBM and uses a instead of read. different core segmentation approach. 13 . Adds control variables to reactor NULAPS does not use reactor kinetics 1 feedback. kinetics feedback in'SBLOCA analyses. 122 .1 Corrects input and initialization No impact. NULAP5 assessment-for check, trip, and swing check and SBLOCA analyses used mainly va lves ~. Also corrects plot scaling- .the trip valves which work error. , correctly. The code also~uses a separate plotting package. 0030.0.0
_ ___ _ _ _ . _ _ _ _ . _ . . _ , _ _ _ _ _ __ __..__ __ - -._ . _ _ _ _ _ _ _ _ _ _ _ _ . . _ .m . -_ . ., m m ,_ k S.sQUENTIAL . CYCLE NO. UPDATE NO. S110RT DESCRIPTION IMPACT ON NULAPS ,
~
23 1 Fix RPUMP to properly give.up No impact. There has been no space when inserting or replacing need to replace or insert a pump a pump component, in the existing NULAP5 analyses. -! However, this correction will be , added to the list of updates to - NULAPS. 2 Fix flow oscillation between choking No flow oscillations have been and nonchoking in JCHOKE. noticed in the NULAP5 analyses. s This update will be however-included in the update list of the code. 24 1 Fix typo error in liTADV. A minor correction. No impact. 25 1 Fix abort when reactor kinetics No impact. This correction will is deleted at restart. be added to the list of future updates to NULAP5. 26 1 Fix error when trying to delete a A minor correction. No impact. nonexistant hydrodynamic cunponent. 2' Fix error' trap in STATE, SUCCESS =2 A modification to a debug print. if RH0/ VISCOSITY =NEG or 0.0. No impact. 3 Fix pressure equation terms in the NULAPS.does not use the accumulator _ accumulator model, model (no accumulators at Haddam Neck). 4 Fix accumulator pressure. solution See the previous comment in (3). terms in subroutine PRESEQ. 5 Fix accumulator final and state See the previous comment in (3). solution in subroutine STATE,. add
-time. step control.
6 'Fix indefinite in the inertial valve A minor correction. NULAPS SBLOCA decks model,. subroutine ICOMPA. have not used this model. 0031.0.0
C SEQUENTIAL CYCLE NO. UPDATE NO. SIIORT DESCRIPTION IMPACT ON NULAPS 27 1 Fixes subroutine IJPROP for negative A minor correction in junction indefinite. input initialization. 2 Deletes the VllR stuffer implementation . Removes a previous update. No impact on for two phase mass error added in NULAPS. cycle 21. 28 1 Skips elevation calculation in An additional improvement to' cycle 20 IELVTN if no volume is connected loop elevation checker. See comments in to one end of a junction. 20-13 above. 2 Corrects errors in comment cards A cosmetic change. in VEXPLT. 29 1 Calculate momentum viscious terms This update basically rearranges the in VOLVEL as derived in the manual. method of calculating the viscious terms. There is no major impact on the code. This update will be added to the list of possible future updates to the code. 2 Uses momentum viscous terms' This update includes the previous calculated in (1) above in modification calculated in-(l) above, subroutine VOLVEL. An additional in deriving the convective terms. There modification minimizes trucation is no major impact on updates (1) and errors in.the convective term (2) on NULAPS momentum equation solution. (VL**2 - VK**2 = (VL-VK)*(VL*VK)) .This update will be included in.the list of possible future updates to the code. 3 Eliminate a possible indefinite for A minor change. NULAPS-does not use the accumulator model in STATE. this model. 4 Fix algebraic error in inertial There are no inertial valves in the valve, subroutine VALVE. Haddam Neck model. 0032.0.0
t
- 9. Northeast Utilities stated in response to question Q.IX.9 that the heat transfer coefficient calculated by the no-return to nucleate boiling logic would be limited to a maximum of 10,000 W/m 2_g (1761 Btu /hr-ft 2- F) based on the maximum post-CHF heat transfer coefficient calculated during an assessment of post-CHF heat transfer regimes. It was also stated that this would result in the calcula-tion of conservative heat transfer coefficients by the no-return logic when compared to nucleate boiling heat transfer coefficients.
However, based on the figure from the next page, which was taken from Reference 6, this does not always seem to be the case. This is because nucleate boiling heat transfer coefficients in the temperature difference range from 10 to 20 F fall below the value selected to limit the heat transfer coefficient calculated by the no-return logic. Provide additional information to demonstrate the no-return logic will yield conservative heat transfer coefficients relative to the nucleate boiling regime. The selection of the maximum film boiling heat transfer coefficient of 10,000 W/m 2 -K was based on the lowest calculated value of the nucleate boiling heat transfer coefficient for the entire range of flow, pressure, quality, temperature conditions experienced for all postulated small break LOCA's. A heat transfer surface was generated, which covers the following ranges: For Pressure: 2.5 MPa 5 P 5 15 MPa For Quality:
-0.1 < X < 1.1 For Mass Flux:
5 kg/m2 -s 5 G $ 4500 kg/m 2 _3 With the no-return logic employed, the heat tranrfer package in NULAP5 will force the heat transfer to remain in the film boiling heat transfer regime and, as such, with the lower heat transfer coefficients for the regime relative to nucleate boiling, the tempera-ture difference that results is of the order 100 to 200*F. A review of the NULAP5 results for a spectrum of break confirms this result. Furthermore, the maximum heat transfer coefficient is set at 10,000 0033.0.0
e W/m 2 K to merely prevent the film boiling heat transfer coefficient from ever exceeding that for nucleate boiling for the same conditions with the no-return logic exercised. l l 0034.0.0 l l u
- 10. Northeast Utilities stated in response to Question Q.X.ll that the fuel rod temperature increase terminated at the same time the core two-phase level went to zero because the code calculated a transition to mist flow in the core and increased steam velocities in the core during the loop seal clearing process enhanced the
.ccoling of the rods. While this may be what NULAPS calculated to occur, it is the staff opinion that these phenomena are not real anc that the rod hot spot at the top of the core should be heating up'^during the period when the two-phase and collapsed levels are at the bottom of the core. Thecalculatedresultsjortheanalyses forming the basis of Question Q.X.ll, the 0.1 ft discharge leg 2
break with the pumps off and the 0.02 ft discharge leg break with the pumps off, should be reviewed in detail to check for the possible calculation of nonphysical phenomena and results. Areas reviewed should, at a minicum, include the interphase drag models, flow . gime maps, and heat transfer logic. The results of this study should be provided to the NRC for review. Northeast Utilities is reviewing the results of the 0.1 ft.2 and 0.02 ft.2 discharge leg break analyses. The results of the review will be provided to the NRC by June 30, 1987. 0035.0.0
4
- 11. Question 4 of the additional questions submitted to Northeast Utilities on February 3, 4, 1987 meeting with the NRC asked how.
application of the return to nucleate boiling lockout and the fuel behavior models to only the hot channel in the-core met Appendix K
' requirements 1.A.5 and 1.B. Northeast Utilities answer to question 4 i
requested clarification from the NRC staff on the Appendix K require-ments concerning the application of these models to the average core , channel in the Haddam Neck plant model.for SBLOCA licensing analyses. The staff opinion is that these models should be applied to both the average and hot core channels in ordet to meet Appendix K requirements ' I.A.5 and I.B unless it can be shown that applying the,se models only to the hot channel would be conservative. Therefore, to resolve this issue, Northeast Utilities must apply the nucleate boiling lockout'and the fuel behavior models to both the average and hot
- core channels or provide for review sufficient information to show that applying these models to only the hot channel is conservative. ,
In order to fully meet Appendix K requirements I.A.5 and I.B. ' Northeast Utilities will apply the return to nucleate boiling lockout and fuel behavior models to both the hot channel and average l core channel in the Haddam Neck plant model for all future SBLOCA { licensing analyses. A t i
- f. ,
i
]
l 0036.0.0 t
,. _,._ ,_. _ _ _ . _ -.,..- - - - - - - - ~ - - - - ~ - ~
( e 4 REFERENCES
- 1. W. G. Kennedy et al. , Pump Two-Phase Performance Program, Volumes 1 through 8, EPRI NP-1556, September 1980.
- 2. V. H. Ransom, et al., RELAP5/ MOD 2 Code Manual, Vol. 1, NUREG/CR-4312, EGG-2396, August 1985.
- 3. V. H. Ransom, et al, RELAP5/ MODI Code Manual, Vol. I and 2, NUREG/CR-1826, EGG-2070, March 1982.
- 4. Letter E. J. Mroczka, CYAPCO to NRC, "Haddam Neck Plant Response to Requests of Additional Information on NULAP5," Docket No. 50-213, March 19, 1987.
- 5. Letter to P. D. Wheatley, INEL, to S. Sun, NRC, " Questions Resulting from the Review of NULAPS, POW-12-86, December 12, 1986.
- 6. Eckert and Drake, Heat and Mass Transfer, 1959.
- 7. A. L. Lowe, Jr., Material Properties and Acceptance Criteria for LOCA Analysis of Stainless Steel PWR Core, BAW-1411, February 1975.
- 8. Calculative Methods for the Northeast Utilities Small Break LOCA ECCS Evaluation Model, Vol. I and II, Docket No. 50-213, Northeast Utilities Service Company, July 1984.
- 9. NULAPS, a FORTRAN IV Digital Computer Program for Nuclear Steam Supply System Blowdown and Fuel Rod Heat Up Analyses, Docket No. 50-213, Northeast Utilities Service Company, April 1983.
- 10. D. G. Morris, C. B. Mullens, and G. L. Yoder, An Analysis of Transient Film Boiling of High-Pressure Water in a Rod Bundle, ORNL/NUREG-85, March 1982.
- 11. R. J. Wagner, Heat-1, A One Dimensional Time Dependent or Steady State Heat Conduction Code for the IBM-650, ID0-16867, April 1963.
0037.0.0}}