ML20098F136
| ML20098F136 | |
| Person / Time | |
|---|---|
| Site: | Arkansas Nuclear  |
| Issue date: | 04/30/1984 |
| From: | Harne R, Jackie Jones BABCOCK & WILCOX CO. |
| To: | |
| Shared Package | |
| ML20098F132 | List: |
| References | |
| BAW-1829, NUDOCS 8410020366 | |
| Download: ML20098F136 (40) | |
Text
BAW-1829 April 1984 9
Thermal-Hydraulic Crossflow hppi; cations by R. L. Harne J. H. Jones Thermal-Hydraulic Engineering BABC0CK & WILC0X Utility Power Generation Division P. O. Box 1260 Lynchburg, Virginia 24505 84100gg(6$$$I3 PDR A o
PDR Babcock &Wilcox a ucoermott company P
Babcock & Wilcox utility Power Generation Division Lynchburg, Virginia Report BAW-1829 April 1984 Thermal-Hydraulic Crossflow Applications Report R. L. Harne, J. H. Jones Keywords: Reactor Core Design, Core Thermal Hydraulics ABSTRACT The themal-hydraulic analysis can now be perfomed using the cross-flow computational tools of LYNX 1, LYNX?., and LYNXT. These thermal-hydrau-lic crossflow codes have demonstrated improvements in departure from nucleate boiling ratto predictions over previous closed-channel analyses.
This report identifies the methods and criteria used in developing the crossflow models and the application of the models in licensing design l
analyses for 177-fuel assembly plants.
I i
- ii - Babcock &Wilcon a McDermott company
CONTENTS Page
- 1. INTRODUCTION .......................... 1-1
- 2. MODEL DESCRIPTIO4 . . . . . . . . . . . . . . . . . . . . . . . . 2-1 2.1. Multi-Pass Model ..................... 2-1 2.2. Single-Pass Model . . . . . . . . . . . . . . . . . . . . . 2-13
- 3. MODEL DEVELOPMENT AND JUSTIFICATION . . . . . . . . . . . . . . . 3-1 3.1. Multi-Pass Model ..................... 3-1 3.1.1. Hot Bundle location . . . . . . . . . . . . . . . . 3-1 3.1.2. Power Peaking Distribution ............ 3-2 3.1.3. Control Component Distribution .......... 3-4 3.1.4. Core Inlet Flow Distribution ........... 3-5 3.1.5. Core Exit Pressure Profile ............ 3-5 3.2. Single-Pass Model . . . . . . . . . . . . . . . . . . . . . 3-5
- 4. MODEL APPLICATION . . . . . . . . . . . . . . . . . . . . . . . . 4-1 4.1. Core Safety Limits .................... 4-1 4.2. Core Operational Limits . . . . . . . . . . . . . . . . . . 4-4 4.3 Accident Analysis ........... ...... . . 4-4
- 5.
SUMMARY
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
- 6. REFERENCES ........................... 6-1 List of Tables Table 3-1. Adjacent Bundle Peaking Impact ................ 3-8 3-2. L3 cal Peaking Distribution Impact . . . . . .......... 3-8 3-3. LYNXT Model DNBR Comparisons ................. 3-9
- iii - Babcock &Wilcox a McDermott company
List of Figures Figure Page 1-1. Crossflow Applications Flowchart . . . . . . . . . . . . . . . 1-2 2-1. Bundle Location Numbering System (LYNX 1) . . . . . . . . . . . 2-4 2-2. Crossflow Gap Numbering System (LYNX 1) . . . . . . . . . . . . 2-5 2-3. Control Component Configuration (LYNX 1) ........... 2-6 2-4. Bundle Radial Peaking Distribution (LYNX 1) . . . . . . . . . . 2-7 2-5. Inlet Flow Factor Distribution (LYNX 1) . . . . . . . . . . . . 2-8 2-6. Pin and Subchannel Numbering System (LYNX 2) ......... 2-9 2-7. Local Peaking Distribution for Hot Bundle (LYNX 2) ...... 2-10 2-8. Location of Hot Channel s . . . . . . . . . . . . . . . . . . . 2-11 2-9. Single-Pass Model Channel Numbering System . . . . . . . . . . 2-12 3-1. Adjacent Bundle Radial Peaking Impact Distributions ..... 3-10 3-2. Single-Pass Versus Multi-Pass DNBR Comparisons . . . . . . . . 3-11 4-1. Mul ti-Pass and Single-Pass Methods . . . . . . . . . . . . . . 4-5 4-2. Core Protection Safety Limits ................ 4-6 4-3. Pressure-Temperature Envelope (Protection System Setpoints) ...................... 4-7
- iv - Babcock & Wilcox a McDermott Company
- 1. INTRODUCTION Reactor core themal-hydraulic design and licensing analyses have tra-ditionally used conservative methods which provide significant real but rel-atively unquantified margins to fuel design limits. The traditional meth-ods use " closed-channel" computer codes in combination with an analysis technique wherein allowances for uncertainties, tolerances, and measurement errors are all considered simultaneously in the most adverse nanner.
" Crossflow" computer codes, which can predict flow redistribution effects within an open lattice reactor core, provide a significant improvement in modeling accuracy, thereby providing . additional departure from nucleate boiling ratio (DNBR) margins relative to the traditional closed-channel modeling. This report describes a revision to the traditional analysis ,
method which, while retaining the inherently conservative treatment of un-certainties, tolerances, and error allowances, incorporates crossflow analy-sis codes in place of the closed-channel codes.
This report describes the crossflow models developed for LYNX 1 , 1 LYNX2 2 , and LYNXT3 computer codes and demonstrates their accuracy and appli-cability for DNB calculations. Figure 1-1 shows the approach taken in de-veloping the crossflow code models for licensing application. The LYNXT code, with its single-pass model, will be -2d for plant licensing analy-ses. The LYNX 1/ LYNX 2 codes, with their more detailed multi-pass model, will be used for benchmarking.
1-1 hod &Mcom a McDermott company
f Figure 1-1. Crossflow Applications Flowchart Analytical Tools Actions Applications LYNX 1 Code Using sensitivity studies, establish the Multi-pass LYNX 2 Code Crossflow Model Multi-pass tiodel Generate Single-pass Model LYNXT Code based on Multi-pass Model P
Verify Single-pass Model DNB predictions with Multi-pass
~
COBRA!!!C Code Model and COBRA!!!C
' Multi-pess Model available for future detailed analyses or benchmarkino Use S. ingle-pass Model to establish DNS based r,afety and operatinq limits 1-2 Babcock &Wilcox a McDermott company
- 2. MODEL DESCRIPTION The implementation of crossflow modeling in thermal-hydraulic evalu-ations is perfonned by using the LYNX 1 and LYNX 2 computer codes for multi-pass modeling and LYNXT for single-pass modeling. These codes cons.ider the mass and energy exchange between adjacent channels to more accurately pre-dict coolant axial flow behavior. The LYNX 1 and LYNX 2 codes are run in ser-fes (multi-pass modeling) to yield steady-state hot bundle and hot subchan-nel predictions, respectively. The LYNXT code can simulate the hot channel performance with a single-pass model for steady-state and transient condi-tions.
The multi-pass model has been established for benchmarking the single-pass model and for providing detailed subchannel hydraulic behavior. The single-pass model has been established for thermal-hydraulic licensing ap-plications.
2.1. LYNX 1/ LYNX 2 Model The LYNX 1 computer program is used to determine the steady-state ther-mal and hydraulic conditions of the bundle coolant flow in a reactor core by modeling the core on an assembly basis. LYNX 1 utilizes one-dimensional conservation equations of mass, momentum, and energy formulated in the axial direction. By using a simplified conservation equation for trans-verse momentum, LYNX 1 can calculate an interbundle crossflow. The forward finite difference numerical solution method is used in converging towards the core exit pressure profile boundary condition. After a converged solu-tion has been obtained, the calculated coolant properties at each axial and transverse location are obtained. The primary output of the LYNX 1 code for DNBR analyses is the interbundle diversion crossflow (IBDCF) for the hot bundle (the fuel bundle containing the hot pin). The IBDCF is provided for the four bundle interfaces at each axial location modeled for the hot bun-die. The resulting hot bundle IBDCF is then used by the LYNX 2 program to 2-1 Babcock &Wilcom a McDermott company
establish the governing mass and energy exchange for the hot bundle for each axial location.
The LYNX 2 program models the hot bundle and calculates the steady-state subchannel conditions. LYNX 2 uses coupled relations for the conservation of mass, energy, and momentum at each axial increment. By incorporating the IBDCF at the periphery of the hot bundle matrix, the edge subchannels transfer mass, energy, and momentum through the periphery of the array. Intersubchannel diversion crossflow is determined from trans-verse pressure differences. After repeated iterations of the conservation relations at each axial location, the IBDCF propagates throughout the hot
. bundle. The subchannel critical heat flux (CHF) ratios may then be calculated using local subchannel conditions.
The LYNX 1 model used for this study consists of 29 whole or partial fuel bundles representing a symmetric one-eighth portion of a 177-fuel assembly (FA) reactor core. The LYNX 1 fuel bundle numbering scheme re-ferred to throughout this report is shown in Figure 2-1. The hot bundle is located in the center fuel location (bundle 1). Interbundle diversion crossflow is provided through the 44 crossflow gaps distributed across the model as shown in Figure 2-2. For design analyses, the axial increment is about 3 inches (as recommended in reference 1).
The control component distribution for the multi-pass model example is shown in Ff gure 2-3. This distribution is applicable for 177-FA reactor cores containing a combination of burnable poison rod assemblies (BPRAs),
control rod assemblies (CRAs), and assemblies with unplugged, or "open,"
guide tubes . The selection of the control component arrangement is dis-cussed in section 3.1.
The bundle radial peaking distribution, established for design analy-ses using the multi-pass model, is shown in Figure 2-4. This distribution provides relatively high radial peaks surrounding the centrally located maximum radial peak of the hot bundle. ,
The remaining fuel bundle peaks de-crease radially from the hot bundle.
The boundary conditions for the LYNX 1 model are a flat exit pressure profile, a flat inlet flow distribution (Figure 2-5), a uni form inlet 2-2 Babcock & Wilcox a McDermott company
enthalpy and density profile, and zero crossflow at the inlet. This re- 4 suits in five boundary conditions for the six mass and energy equations. -'k The hot bundle flow factor, for DNBR calculations, is the statistical mini-mum flow factor (0.95) developed from vessel model flow tests (VMFTs). The selection of these distributions is discussed in section 3.1.
The LYNX 2 model comprises 256 linked subchannels and 208 fuel rods.
Included in the symmetric model of the hot fuel bundle are 16 control rod .- +
guide tubes and 1 instrument guide tube. The subchannel and fuel rod num- 't -
bering used in the LYNX 2 model are provided in Figure 2-6. The power dis- _
tribution within the hot bundle is referred to as the local peaking distri-bution where each fuel rod peaking factor is equivalent to the rod absolute power divided by the hot bundle average power. The hot fuel rod exhibits _
the maximum local peak within the hot bundle with a value of 1.0615. The hot bundle local peaking distribution used in the multi-pass model is shown in Figure 2-7. c The hot channel in LYNX 2 is the subchannel in which the minimum DNBR [
occurs. Since the fuel bundle comprises five tyres of subchannels (unit, ..
control rod, instrument guide tube, wall and corner), hot channels are de-signated for each channel type. Hot channel factors and flow area reduc-tion factors are then applied to each hot channel. Figure 2-8 shows the location of the hot channels throughout the hot bundle. The limiting hot channel , possessing the minimum DNBR with the B&W-2 CHF correlation 3 , is the hot unit channel for the variety of licensing type operating conditions considered in the development of the multi-pass model.
Transient DNB predictions are obtained with the RADAR 9 code for the multi-pass model. RADAR is a transient closed-channel code widely used in ,
traditional DNB analyses. In the multi-pass modeling scheme, RADAR 'is ini-tialized to the LYNX 2 hot channel minimum DNBR at the begirning of the transient by matching the RADAR hot channel flow rate to that predicted by LYNX 2 for the hot subchannel at the axial location of minimum DNBR and by the use of an enthalpy rise factor (FLAH) to obtain the desired DNBR value. '-
After the DNBR is initialized, the inherent conservatism of the RADAR code results in conservative transient DNBR predictions. -
2-3 Babcock & Wilcox a McDermott company
Figure'2-1. Bundle Location Numbering System (LYNX 1)
I' Hot bundle N _
-- 1 ,-- - -- 2 --- -
--- 3 --- - -- 4 --- - -- 5 --- --
-- 6 --- --
-- 7 --- -
-- 8 --- - --
% l
% t 9, 10 11 12 13 14 15 vm 16 , 17 18 19 20 21 g
22 , 23 24 25 s,
26 27 28 1/8 core symmetry \,
s, 29 ,
N%
l l
2-4 Babcock &Wilcom a McDermott to.npany o
Figure 2-2. Crossflow Gap Numbering System (LYNX 1) pHotbundle
's P 5 11 11 11 11 11 11
_ _ _2A;- . 1 _ _ - - _ _ _ 2 . . _ _ _ . ,- . 3 -- _ _ _ _ _ 4 . .. . -- - - - ----- - - ---- -- - -- --- --
11
'lL 8 J L_ 9 JL J J J 13 JL 14 _
s s
's, 15 16 17 18 19 20
'IN L 21 JL 22 JL 23 JL24 _J L 25 JL 26 _
5 s, l l l l l
'N 27 28 29 30 31
'IN L 32 JL 33 JL 34 JL 35_JI 5, l l l
\ 36 37 38 N,
Ib 39
- 40 41 ~
g l l 1/8 core symmetry N 42 43
' 1, L' ,44 JI N
2-5 Babcock &Wilcox a McDermott company
4 Figure 2-3. Control Component Configuration (LYNX 1)
Hot bundle
,'s,
-- CR ,-- -
- OPEN - -
-- CR --- -
- BP RA -- -. . - C R --- -- - BPRA-- --
-- CR -- -
- OPEN -- --
t, s
R, OPEN CR BPRA CR BPRA OPEN s, -
CR ,, BPRA CR BPRA CR OPEN CR , BPRA CR OPEN 1/8 core symmetry ,
CR ', , OPEN OPEN 69 control rod assemblies 48 burnable poison rod assemblies (a) ',
60 open assemblies s,
's, OPEN (a)This arrangement is applicable for ',
cores containing BPRAs.
s d
2-6 Babcock &WHcom i .
a McDermott company l
Figure 2-4. Bundle Radial Peaking Distribution (LYNX 1)
H
's, Totbundle
- 1.554 - - 1.54 -- -< - 1.49 - -
- 1.45 ..
1.36 .. 1.14 -- - 1.01 -- -- - 0.43 - .-
's
, 1.53 , 1.47 1.43 1.35 1.17 0.97 0.32
's, 1.44 1.38 1.28 1.05 0.76 0.36
-s ,
1.'29 1.14 0.83 0.31 .
s, ,
1/8 core symmetry -
's, O 91, 0.79 0.34
-s s
0.37
's
's where hot bundle hot pin radial-local peak 1.65 rad a1 =
pea hot pin local peak " T N IT = 1.554 2-7 Babcock &WIfcom a McDermott company
Figure 2-5. Inlet Flow Factor Distribution (LYNX 1) pHotbundle
'% t
'%, l I. l
- 1.00 - - 1.00 -.
-- - 1.00 -. - 1.00 -. - 1.00 -- -. - 1.00 - -
- 1.00 - -
0.95 ,-- --
1.00 1.00 1.00 1.00 1.00 1.00, 1.00 1.00 1.00 1.00 1.00 1.00 1.00
%g 1.00 1.00 , . 00 1.00
% l 1/8 core symetry ',
1.00 , 1.00 1.00
's 1.00, N
2-8 588 Kock &WHcox a McDermott company
Figure 2-6. Pin and Subchannel Numbering System (LYNX 2) _
241 242 243 254 255 256
" 0 0 0 0 0 0 0 0 0 0 0224 4e 00000000000 24
'" 8 0 0 0 0 0 0 0 0 @ O 0 0 0 e,
'"800@O000000@O0 ,2
'"80000000000000 'e ,e
"'80@OOOOOO@OO@O 18e e
"@O000000000000 44 "0000000@O00000 "e0000000000000 :
"eOeOOeO JOeOOeO "eOOOOOO OOOOOOe:
OOOeObOeOOOO, "eOOeOOOOOOOeOOe:
002.
1 000 OO 1
OO 000 O@14 'O
13 15 16 ..
gu e u guide tube oYmor c p.$f j
Figure 2-7. Local Peaking Distribution for Hot Bundle 9 10 11 12 13 14 15 2 3 4 5 6 7 8 ROW /COU Mi 1
.9734 .9825 .9825 .9762 .9701 .9676 .9518 .9659 9618 .9534 .9568 .9676 .9809 .9875 .9767 1
.9784 .9992 1.0158 1.0042 .9800 .9618 .9601 .9501
.9493 .9593 .9%8 .9784 1.0058 1.0200 .9958 2
1,0349 .9817 1.0158 1.0333 1.0075 .9701 .9559 .9510
.9493 .9551 .9659 1.0067 - 1.0125 -
3
.9859 1.0133 1.0316 1.0540 - 1.0083 .9751 .9593 4 .9584 .9751 1.0049 - 1.0565 1.0366 1.0108
.9933 1.0191 1.0333 1.0466 1.0557 1.0357 1.0017 .9701 5 .9701 1.0017 1.0324 1.0557 1.0490 1.0382 1.0166 1.0100 1.0391 1.0349 1.0349 - 1.0141 .9759 6 .9759 1.0153 - 1.0349 1.0382 - 1.0366 -
1.0141 1.0175 1.0357 1.0183 1.0141 1.0175 .9950 .9709 7 .%76 .9933 1.0125 1.0116 1.0183 1.0391 1.0150 1.0166 1.0100 .9950 .9R84 .9859 .9776 .9667 8 .%43 .9759 .9917 .9867 .9950 1.0125 1.0141 -
1.0166 1.0200 1.0382 1.0208 1.0158 1.0191 .9975 .9726 N 9 .9709 .9967 1.0150 1.0150 1.0216 1.0415 1.0175 1.0124 1.0407 1.0391 1.0391 - 1.0200 .9809 b 10 .9751 1.0141 - 1.0357 1.0382 - 1.0391 -
.9975 1.0233 1.0391 1.0524 1.0615 1.0416 1.0083 .9776 11 .9717 1.0042 1.0357 1.0582 1.0515 1.0399 1.0216
.9917 1.0191 1.0391 1.0615 - 1.0158 .9842 .9684 12 .%34 .9800 1.0108 - 1.0607 1.0407 1.0175 Y
.9892 1.0224 1.0415 1.0158 .9784 .9659 .9618 13 .9576 .9626 .9734 1.0133 1.0407 - 1.0208 -
.9850 1.0050 1.0249 1.0133 .9884 .9709 .9717 9634 14 .9534 .9643 .%43 .9842 1.0100 1.0233 1.0033
.9792 .9850 .9917 .9875 .9784 .9717 .9634 .9762 15 .9692 .9559 .9643 .9726 .9834 .9892 .9834 Hot pin
- i. I 5R
?R g=
E*
8g 2-40M l
... ,, , , = , . . . - . , . . , , _ _ , .
Figure 2-8. Location of Hot Channels (all types) -
000000000000000 i O00000000000000 i O0000@O00@O0000 000@O000000@O00 :
000000000000000 OOOOOOOOOOOOSOO C00000000000000 >
0000000@e000000 i OOOOOOOODOOOOOO ~
000000000000000 0
OOOOOOOOOC3G0 0 0 0 0 0 0 0 0 OOO 0'@0 0 0 0 000000000000000 :
000000000000000 OOOOOOOOOQQOOOOei
- Wall ce 1
Hot channel locations: c 5
- Limiting channel (B&W-2 CHF correlation) i 2-11 Q,y,k[y{jf ,
Figure 2-9. Single-Pass Model Chanr.el Numbering System Symmetry Line 1-: 00000 O Instrument Guide' Tube 4 5 6 7
= pcy,go'o OOO Control Rod Guide Tube Channel 1:
Channel 2:
Hot Unit Channel Hot Control Rod Guide OO Tube Channel s Symmetry Line
/
. . . . . . ....... . ....... . . . . . . . . . . /. . . .. .
Channels 1- 1 -
comprise the hot buncle Channel 12 is composed of the .
remainder of the core -
2-12 Babcock &Wilcox a McDermott company
2.2. Single-Pass Model The single-pass model utilizes the LYNXT code for determining steady-state and transient . flow and temperature distributions within a reactor core. LYNXT is an improved version of the COBRA-IVS coce developed at Battelle Northwest Laboratories.
Single-pass analyses model subchannels, groups of rubchannels, bundles and groups of bundles in one simulation. Historically, core thermal-hydrau-lic calculations have been performed using mul ti-pass analysis methods, such as the LYNX 1/ LYNX 2 models di scussed above, in which bundles are modeled in an initial " pass" and groups of subchannels in another " pass" which yields the minimum DNBR. LYNX 1 and LYNX 2 are the B&W multi-pass crossflow calculational tool s, respectively. However, LYNXT, with its single-pass modeling capabilities, offers the same accuracy at a lower cost as compared to multipass analyses and therefore will be used for licensing applications.
The LYNXT crossflow model selected for licensing applications is a 12-channel model. The variable-scaling feature of LYNXT permits the simul-taneous modeling of the hot subchannel and its surrounding subchannels with the remainder of the core. Figure 2-9 shows the channel modeling scheme.
This specific model is applicable to 177-FA core analyses using the B&W-2 CHF correlation for DNB predictions.
2-13 Babcock &Wilcox a McDermott company
m
- 3. MODEL DEVELOPMENT AND JUSTIFICATION The approach followed in developing the single-pass model which is to be used for core thermal-hydraulic analyses was to use a more complex multi-pass model as a benchmark and as a development tool to evaluate DNBR sensitivity to various modeling considerations such as the core power dis-tribution, hot assembly location, inlet flow profiles, etc. Modeling simp-lifications, desirable for an efficient calculational tool, were made using conservative selections from the options considered. The LYNXT single-pass model was also compared to a similar COBRAIIIC model to provide additional confidence in the final model selected.
3.1. Multi-Pass Model In developing a multi-pass crossflow model, numerous model sensitiv-ities must be understood and quantified to permit the appropriate selection of model characteristics. The following model characteristics required in-vestigation for the multi-pass model:
- Location of the hot fuel bundle
- Peaking distributions (axial, radial, and local)
- Control component configuration
- Inlet flow profile
- Core exit pressure profile The sensitivity studies were perfonned using a basic 112% overpower maximum design case for a 177-FA plant.
3.1.1. Hot Bundle Location In selecting the location of the hot bundle, various possible loca-tions were considered. The hot bundle was moved fran the center location as bundle 1, to numerous locations and ultimately to the peripheral loca-tion of bundle 29. A relatively flat hot bundle-to-adjacent bundle peaking gradient was maintained to eliminate any DNBR impact due to bundle power.
3-1 Babcock &WHcom a McDermott company
The sensitivity study showed that moving the hot bundle throughout the core produces a negligible effect, (0.1% DNB). Since there was no sb ong DNBR dependency on hot bundle location, additional factors were considered in selecting the appropriate hot bundle. A bundle possessing a higher hydrau-lic resistance to coolant flow than adjacent bundles would naturally yield more conservative DNB results. Therefore, a bundle location containing a cc . trol rod assembly (for this example, bundle 1) was designated the hot bun-die since this location could be surrounded by bundles with lower hydraulic resir.tance.
3.1.2. Power Peaking Distribution Another model characteristic requiring investigation was the power peak-ing distribution in the core and the hot bundle. Core thermal-hydraulic analyses use an assumed design peaking distribution which is then imposed on the fuel cycle design as a limiting criterion. The approach followed is to use a " design distribution" for all thermal-hydraulic analyses, both steady-state and transient, then to define combinations of radial and axial peaking factors which provide equivalent DNB prote:: tion. These factors are defined as " maximum allowable peaking" (MAP) limits and are used as one of the bases for the power imbalance trip function in the RPS. Similar MAP limits, which correspond to initial condition peaking conditions assumed for accident anal-yses, are used in the definition of limiting conditions for operation.
The following nomenclature is used throughout the power distribution studies:
The radial peaking factor represents the bundle, or fuel assembly, pow-er. relative to the core average power. The local peaking factor represents the power of an individual fuel rod relative to the bundle average. The radial x local peaking factor (commonly denoted as RxL or FaHN) represents fuel rod power relative to the core average. The axial peaking factor (F z represents the " hot spot," or local, power generation rate relativa to tne N
fuel rod average. The total peaking factor (FQ ) is the product of radial, local, and axial peaks. -
3-2 Babcock & WHcom a McDermott company
The power distribution study presented herein was oriented toward the development of a model with the following peaking factor limits:
FAHN = 1.65; FzN = 1.65; FQ " = 2.72 This represents a reduction in radial x local peaking and increases in -
axial and total peaking factors relative to those typically used (1.71, 1.50, and 2.57, respectively), for design and licensing of B&W 177-FA reac-tor cores. The revised peaking factors were selected to provide a more realistic radial power distribution while, at the same time, accommodating the anticipated need for higher total peaking factors in future very low leakage fuel cycle designs.
Peaking distribution studies were separated into three areas: the bun-die radial distribution, the hot bundle local distributio'n, and the axial distribution. The DNBR sensitivity to the radial peaking distribution was first studied.
Three bundle peaking distributions were considered:
- 1. A base case (conventional distribution) with the hot bundle in bundle location 10).
- 2. The same as 1 above, but with a 5% reduction in peaking of the bundles around the hot bundle.
- 3. The same as 1 above, but with a 15% reduction in peaking of the bundles around the hot bundle.
Figure 3-1 shows the bundle radial distributions for these cases. Note that the peripheral bundle radial peaks were adjusted to maintain peaking normalization. It was observed that the minimum departure from nucleate boiling ratio (MDNBR) was relatively insensitive to the peaking changes of the fuel bundles adjacent to the hot bundle (see Table 4-1). The DNDR dif-ferences between the three cases were less than 0.1%, however, the radial peaking distribution with a nearly flat peaking profile around the hot bundle yielded the slightly more conservative DNBR response. This distribu-tion, used in the crossflow model and shown in Figure 2-4, has a radial peaking gradient around the hot bundle similar to the peaking distribution used in traditional closed-channel kalyses.
l . ..
l 3-3 Babcock &Wilcox a McDermott company ..
i ..
The local peaking distribution was selected after a series of realis-tic distributions were studied. Results from this study revealed that the minimum DNBR decreased negligibly as long as the hot pin radial-local peak remained constant (see Table 3-2). As the local peaking distribution is flattened, the limiting subchannel, at some point, jumps around the hot bun-di e, depending on the characteristics of the CHF correlation being used.
To avert this behavior in design analyses, a realistic local peaking distri-bution is selected which yields the lowest DNBR prediction consistently in one channel type.
This methodology conservatively considers the most DNB-limiting local peaking distribution expected in-reactor. The local peaki ng distibution selected for the model, and shown in Figure 2-7, includes a hot pin local peak of 1.0615. Design analyses, using the design local peaking distribu-tion, will result in the MDNBR occuring in the hot unit channel for the B&W-2 CHF correlation.
The axial peak was selected to achieve an increase in total peak rela-tive to those typically used with closed channel wthods. The axial peak selected was a symmetric 1.65 Pmax/Pavg cosine shape.
+ p.
$ff f 3.1.3. Control Component Distribution
- 2. J.
One consideration in the development of the multi-pass model is the lff, Qig presence or absence of a control component in each fuel assembly since this I., ' $
introduces a difference in resistance to flow at the top of the fuel assem- .([ .~e .I bly. The 177-FA reactor cores typically include control rod assemblies y' , w (CRAs), axial power shaping rod assemblies (APSRAs), burnable poison rod , >51 C assemblies (BPRAs), neutron source assemblies, and "open guide tube" assem- kbf
- blies. Hydraulic resistance factors of all of the control component types (
are the same provided it is assumed that CRAs and APSRAs are fully in- k, m.N;,
serted. Open guide tube assemblies have a lowar resistance to flow since ..
gg they do not have control component 3piders in the flow stream. For the % < ;'K LYNX 1 model, it is assumed that the hot assembly is in a control component ((O.5 v: ,.
location surrounded by open guide tube assemblies. This results in the i fai' A $ %. %
most conservative model since the increased resistance causes flow to be py diverted from the hot assembly into the surrounding assemblies. A represen- 1,,h,';
tative arrangement is shown in Figure 2-3 and is composed of 69 CRAs, 48 >J{
BPRAs and 60 open guide tube assemblies. 2.I"9 v&
3-4 Babcock & WIIcox g.Ml a McDermott company l
,. ;. y @
-- - _ _ _ _ _ _ _ ___ _ _ _ _ _ _ _ _ _ _ _ . . .__ _ _ _ _ _ _ _ _ _ _ {'
I The philosophy followed in developing a specific model is to select a configuration which either represents the actual core arrangement or pro-vides a conservative assessment of thermal-hydraulic perfonnance for a given fuel cycle design. This approach has been followed here to develop a model that can be generically applied to cores with varicus combinations of control components and open guide tube assemblies.
3.1.4. Core Inlet Flow Distribution Studies were performed to determine the sensitivi ty of DNBR to the core inlet flow distribution. Cases ranging from flat inlet flow profiles to realistic profiles with higher core interior flow were studied. Results demonstrated that the hot subchannel MDNBR varied by less than 1% in DNBR for the various cases. The situation yielding the lowest MDNBR prediction was the case assuming a flat inlet flow profile wtth a 5% raduction in flow at the hot assembly location. The 5% reduction is justified in Reference
- 6. Therefore, the flat inlet flow distribution with the hot bundle modeled at the core center, as shown in Figure 2-5, was selected for use in the DNB analyses.
3.1.5. Core Exit Pressure Profile Another consideration for DNBR impact with crossflow modeling was the effect of a non-unifonn exit pressure profile. An analysis with the multi-pass model demonstrated the DNBR impact for realistic core exit pressure gradients (of 1.5 psi) to be about 0.5% in DNBR. This impact was deemed insignificant in light of the conservatism al ready incorporated by the selection of the flat inlet flow profile.
3.2. Single-Pass Model The single-pass model was devel oped to represent the arrangement modeled by the multi-pass crossflow model. The necessary channel modeling detail was first investigated. Models comprising as few as 12 to as many as 56 channels were studied. DNBR results showed that the differences in the model resulted in a negligible impact. Table 3-3 identifies the rela-tively uniform DNBR behavior for the models studied. The 12-channel model was selected for application to 177-FA cores due to its accuracy and econo-mical advantages.
3-5 Babcock &WHcom a McDermott company
J.?> :N!.
{.%M y 3, The 12-channel single-pass model, using LYNXT, was compared to the @*.ta m.
multi-pass (LYNX 1/LNX2) crossflow model for a broad range of operational Q.[f ;f.?
conditions. The DNBR predictions for LYNXT were within 1%, on the average, .7.d.
,. c oy '[.
.. r n in DNBR of the predictions of LYNX 1/ LYNX 2. The results of this comparison .
l f. f. .s< ..
are shown in Figure 3.2. These comparisons covered a range of DNBR predic- T:ge ?
,,c. ;.
tions from 1.30 to 2.60 (B&W-2 CHF correlation). The DNBR predictions near W.- -b the CHF design limit of 1.30 were created with single and multiple varia-tions, from the base design overpower case, of the following system condi- qM tions: -[SQ
- .. p Condition Range g']
,. ,s Power 90-112% 2P;.ts Y
, - . ;*s.
Pump status 3, 4 pump pq d. ,
2r. h. .
- Design flow 80%, 106.5% of design flow mW Inlet temperature -15/+35F variation ;U~'&, S System pressure 1800-2135 psia r%]
L.i.
- . c .w The single-pass model was also found to agree with COBRAIIIC 7 to with- [+;,r;
..,,r d
in 0.25% in DNBR. The COBRAIIIC benchmark model was the same as the single- / .y pass model used in LYNXT. }[ e The active fuel length to be used in single-pass modeling for licens- [. pty ing application is the undensified cold nominal fuel stack height. The Qi selection of this parameter was based on the physical behavior of the fuel SIl ya .
pellet stack within the fuel rod. Irradiation effects comprise both densi- .W, j fication and swelling phenomena. The densification effects are more predom- p.}( .
inant at low exposure levels, while the swelling effects are more predomi-nant at higher exposure level s . Fuel stack densification decreases the h,.J .
, .g :
y%. e active fuel length and increases the surface heat fl ux. Fuel swelli ng, &&
v .r .. . .
which occurs once the fuel pores are filled with fission / backfill gases, ::.-:e -r "
tends to increase the active fuel length. In addition to the irradiation .,::f
- 1. dw(
effects, the active fuel length is affected by thennal expansion. For 95% [ [
TD fuel with a typical densification characteristic, the thermal expansion ifi,y y effects are greater than the irradiation effects as shown below: AW 141.8 in.
k.Di
..yM Cold nominal stack height Hot nominal stack height 143.2 in. k Minimum hot densified stack height 142.2 in. 6:..26 .7 m ,.
kh) v.Y .
3-6 Babcock & WHcox y&gh a McDermott company g.14
,. , 4 -.ly pf'
Therefore, the hot rod fuel stack, while being irradiated, will have a length greater than its cold nominal stack height. It is then conservative to consider the cold nominal stack height in DNBR calculations.
The single-pass LYNXT model will be the primary steady-state and tran-sient analysis tool for licensing analyses. The single-pass model DNBR agreement with the multi-pass (LYNX 1/ LYNX 2) model justifies its use for licensing applications. The demonstrated accuracy and economy are two major advantages of the single-pass model.
1 3-7 Babcock & Wilccex a McDermott company
Table 3-1. Adjacent Bundle Peaking Impact Peaking reduction of bundles around Subchannel Case hot bundle, MDNBR No. (%) (B&W-2) 1 0 2.10 2 5 2.10 3 15 2.10 Table 3-2. Local Peaking Distribution Impact Subchannel Case Hot Bundle Hot Fin Hot Pin MDNBR No. , radial peak Local peak Rxt (B&W-2) 1 1.584 1.042 1.65 2.41 2 1.554 1.062 1.65 2.41 3 1.539 1.072 1.65 2.41 i
l 3-8 Babcock &Wilcox a McDermott company
Table 3-3. LYNXT Model DNBR Comparisons (a)
Minimum DNBR (B&W-2) 112% power / 112% power / 100% power /
LYNXT model 2568 MWt 2772 MWt 2772 MWt 12-channel 2.19 1.91 2.24 14-channel 2.19 1.91 2.24 17-channel 2.20 1.92 2.25
. 23-channel 2.19 1.91 2.24 25-channel 2.19 1.91 2.24 28-channel 2.20 1.92 2.24 30-channel 2.20 1.92 2.24 56-channel 2.20 1.92 2.25 (a)All conditions are maximum design.
3-9 Babcock & Wilcox a McDermott company
s Figure 3-1. Adjacent Bundle Radial Peaking Impact Distributions N
1.520 1.680 1.640 .381 e=
1.444 1.596 1.558 1.47 1.14 -'
.46 ~
.462 _
.92 - -
1.292 -
'~
" 1.428 -
1.394 ~
'- ' ~ ~ - ~ ~
.623 -~
~-
's _
g [
1.640 1.620 -
- 361 I
1.558 1.69 1.539 1.44 1.16 .73
- 442 1.394 , 1.377 .603
' 1.680 1.560 l .341 -
?
1.596 1.482 1.36 .99 .71 .422 .
Hot bundle location 1.428 1.326 .583 for peaking studies N ~
r, 1.33 N,
1.11 .49
.703
[
1/8 core symmetry -
.301
.87 .59 .382 .
.543 Peripheral radial X.XX used in all three cases .
'381 peaks were adjusted 5
- 462
- 623 to maintain peaking 6 g normalization; 2
X.XX Case 1 X.XX Case 2 ,
X.XX Case 3 3-10 Babcock & Wilcox a McDermott company
Figure 3-2. Single-Pass Vs Multi-Pass Steady-State DNBR Comparisons 3 .
/
/ ,
~
2.5 V 7
"ce E
a
- 2 C
E d
0
$ 1.5 -
a 1
1 1.5 2 2.5 3 Multi-Pass (LYNX 1/ LYNX 2) DNBR "
(a)
Steady-state DNBR comparisons (B&W-2 CHF correlation)
J l
~
3-11 Babcock &WHcox a McDermott company , s
- 4. MODEL APPLICATION ,
The application of state-of-the-art crossflow models provides improved DNBR margin over closed-channel models which can translate into increased fuel cycle design flexibility and less restrictive operating and safety lim-its. This benefits both the fuel cycle designer and the plant operator.
Although the calculated DNB margin increases with the use of crossflow model-ing, the application of the margin is consistent with the licensing analysis methods previously established and utilized with the closed-channel codes.
The analysis methods for establishing operating and safety limits will remain the same using crossflow modeling.
As discussed previously, the single-pass crossflow model will be used for DNB analyses with the multi-pass (LYNX 1/ LYNX 2) crossflow model reserved for benchmarking and detailed steady-state analyses.
The single-pass crossflow model will be used for both steady-state and transient calculations including the determination of core protection safety limits. Figure 4-1 shows the comparison of multi-pass and single-pass meth-ods in relation to analysis inputs and results.
4.1. Core Safety Limits The intent of core safety limits is to establish protection for the . .
fuel and reactor system against various hypothetical accidents and operating transients as well as steady-state operation. These limits are incorporated into the reactor protection system (RPS) in the form of setpoints which cause a reactor trip to occur early enough in plant operation to prevent a condition from exceeding the safety limits. Such safety limits, based on thennal-hydraulic considerations, include the following conditions:
4-1 Babcock &Wilcox a McDermott company
E
- Maximum permissible core power level
- Permissible combinations of core outlet pressure and reactor outlet temperature
- Flux / flow limit The maximum permissible core power level for four reactor coolant pumps operating is established by the high flux trip setpoint with appro- -
priate adjustments for measurement allowances. This power level is refer-red to as the design overpower condition (112% full power). The design ,
overpower condition is used to determine steady-state DNBR-based limits for --
four reactor coolant pump operation. The corresponding power levels for two- and three-reactor coolant pump operation are based on the flux / flow -
setpoint with their respective flow measurement error adjustment. Examples of the maximum permissible core power levels for the respective pump opera- ~
tion modes used to define the core protection safety limits are shown in Figure 4-2.
A reactor protection system envelope encompassing pennissible combina-tions of core outlet pressure and reactor outlet temperature, known as a P-T envelope, provides DNBR protection as well as reactor coolant system protection. The DNBR protection is in the form of a limiting safety system setting (LSSS) commonly referred to as the variable low pressure trip func- b--
tion. Figure 4-3 shows a pressure-temperature envelope containing the vari- 1 able low pressure function. This LSSS bounds a DNBR-based relationship of --
reactor outlet temperatures. and core outlet pressures which yield the DNBR D design limit (CHF correlation limit) or exceed the CHF correlation quality limit. These DNB relationships, or P-T limit curves, are typically deter- -
mined for three and two pump operation as well as for four pump operation.
M The LSSS is set to bound all the P-T curves for the different pump opera- g tion modes. The single-pass crossflow model will be used to define the y limiting pressure-temperature curves. g The single-pass model is also used to establish DNB limits for asymme- -
tric axial power distributions. The steady-state power distribution used '%
for determining P-T curves (the " design distribution") utilizes a symmetric F axial power distribution. In order to maintain a basis for DNB protection for axially asymmetric power distributions, a series of maximum allowable peaking (MAP) limits are generated in the form of lines of constant minimum 4-2 Babcock &Wilcom g
a McDermott company
__. ______ e _--
DNBR (1.30 using B&W-2 CHF correlation) for various axial peaks at various axial locations. The MAP limits provide a basis for equivalency between the design symmetric power distributions and asymmetric power distribu-tions. These relationships of allowable peaking are translated into allow-able positive and negative core axial offsets. Axial offset is defined as the difference between the power in the top and bottom halves of the core divided by the sum of the power in the top and bottom halves of the core.
With allowable axial offset ostablished, allowable axial power imbalance is determined and incorporated into the core protection safety limits of the RPS. Axial power imbalance is defined as the axial offset times the frac-tion of power. This DNB based axial power imbalance may define a portion -
of the core power imbalance limits, shown in Figure 4-2, providing there are no other more restrictive safety considerations for the top of the core (Note, in general, only the positive imbalance limit is affected by DNB con-siderations.)
Another DNBR-based safety limit is the flux / flow function. This pro-tection is necessary to assure fuel integrity during transients involving a partial loss-of-coolant flow. The flux / flow trip is used to provide core . . - -
protection from a partial loss of coolant flow transient and also provides overpower protection for three and two pump steady-state operation. The -
flux / flow trip limit is determined by the analysis of either a one or two-pump coastdown, depending on the pump power monitor configt. ration for the specific plant (pump power monitors, which provide an immediate trip signal on loss of electrical power to the pump motors, are used for protection against the more severe loss of coolant flow transients). The analysis method used to determine the flux / flow limit value is the same as that dis-cussed in reference 8. That is, a flow coastdown transient is analyzed with one or more assumed flux / flow trip limit values to determine the DNBR response to the flow transient. A limit value which insures that the mini-mum DNBR will be equal to or greater than the design DNBR limit (1.30, B&W-2) is then selected by the use of a cross plot of minimum DNBR versus flux / flow limit. Appropriate error adjustments are then provided to deter-mine the setpoint value. This trip setpoint value may then be adjusted downward to be consistent with partial pump steady state overpower assump-tions. In practice with crossflow methodology it is anticipated that the .
e*,.' a{u 4-3 Babcock &Wilcox a McDermot company h 4j
- fc .- j.
- g. - a
flux / flow setpoint will be detennined by the three-pump overpower condition rather than by the transient analysis.
The implementation of single-pass crossflow model into the flux / flow setpoint determination is in the DNBR predictions for the coastdown tran-sient. The LYNXT model can determine the transient behavior of the hot channel as well as the remainder of the core in one model simulation.
4.2. Core Operational Limits Thermal-hydraulic DNBR analyses are also performed to determine the combinations of power distributions (axial and radial) which yield DNB per-formance equivalent to that of the design distributions used in the study of thermal-hydraulic transients. A set of peaking criteria, in the fonn of maximum allowable peaking curves, are generated using the single-pass model. These criteria assure the use of the crossflow model design peaking results in conservative DNBR predictions for transient evaluations. These limiting curves are derived in the same fashion as the MAP limits for the RPS.
4.3. Accident Analysis The crossflow methodology discussed in sections 2 and 3 of this report will be used for the determination of core DNBR response and fuel tempera-ture response to anticipated operational occurrences and hypothetical de-sign-based transients.
4-4 Babcock &WHcom A McDermott company
s t
Figure 4-1. Multi-Pass and Single-Pass Methods Vulti-Pass INPUTS Sinale-Pass
!CorePowerDistribution !
! Inlet Flow Distribution l '
dLocal Peakina Distribution l i y 't 4 r 1r St d S eady-State h t te
. ysis Analysis '
1 V if L Steady State ,
Y Hot Bundle l N Analysis iL X !Y 2 !N
.lX ^
!T OUTPUTS '
Core Analysis [-
y DNBR Sensitivity Pressure-Temperature i I Limits i Maximum Allowable i R Transient Peakino Limits A Hot Channel D Analysis A
_R
> .DNBR Versus Time for Accident Analysis Flux / Flow Limit --
4-5 Babcock &Wilcox a McDermott company
Figure 4-2. Core Protection Safety Limits Tneraal Power Laval, 5 FP
- 120 3Maximum Core Power for 4 pump operat on ACCEPTABLE l 4 PUNP l 1 OPERATION -
-100 I I
Maximum Core Power for 3 pump operat" on
' l ACCEPTABLE l l .
80 l 4 ANO 3 PUNP l l OPERATION l
I I I
Maximum Core Power for 2 pump operat on I + ~ l UNACCEPTABLE .
50 l ACCEPTABLE l UNACCEPTABLE OPERATION l
4,3 ANO 2 PUNP l OPERATION l
l OPERATION I I I I -
- 40 I i ,
I l I l
1 I i
l1 I
20 I
l l
I l
ll l l l t i I I I I le 1
-60 -40 -20 0 20 40 60 Reactor Power lanalance, 5 4-6 Babcock &Wilcox a McDermott company
Figure 4-3. Pressure-Temperature Envelope (Protection System Setpoints)
High Outlet WTemperature Pressure-Temperature Trip Envelope
/
Acceptable f Operation j
- - ~ ..
m / High Outlet 5 Variable Low O' Temperature
$ Pressure Set- / Point 2 point (LSSS) /
/
$ /
h ! Unacceptable
/ Operation
/
/
/
/
e sw O' Low Pressure Point Trip /
/
/' DNB Based Relationship
/
Coolant Outlet Temperature 4-7 Embcock &Wilcom a McDermott company 8 @+6
- 5.
SUMMARY
This report describes the models and analysis methods to be used for reactor core thennal-hydraulic analyses. The LYNXT code, with a single-pass core model, will be used for the analysis of core thermal-hydraulic performance for both steady-state and transient conditions. Sensitivity studies have been performed to evaluate various modeling options, such as the number of channels and subchannels to be modeled, the hot assembly lo-cation, the core inlet flow distribution, etc., and the results of these studies have been used to select the final model. Benchmarki ng studies have also been performed with miti-pass LYNX 1/ LYNX 2 analyses and provide justification for use of the single-pass model with LYNXT for core thermal-hydraulic analyses, including the development of core protection safety limits and analyses of response to limiting transients and postulated ac-cidents.
i 5-1 Babcock &WIIcom a McDermott company
. . ,E j
-7
'g e
l 2 : ,
~7 4
- 6. REFERENCES .
- J
,c.y
. - j@
4
- 1. B. R. Hao and J. M. Alcorn, LYNX 1: Reactor Fuel Assembly Thermal _
Hydraulic Analysis Code, EAW-10129, Babcock & Wilcox, Lynchburg, 'h- M Virginia, October 1976. -
w T
3
- 2. LYNX 2: Subchannel Thermal-Hydraulic Analysis Program, BAW-10130, ]
Babcock & Wilcox, Lynchburg, Virginia, October 1976. i ?
t.f
- 3. J. H. Jones et al . , LYNXT: Core Transient Thermal-Hydraulic Program, i [
BAW-10156, Babcock & Wilcox, Lynchburg, Virginia, February 1984. .. g
's
~
- 4. Correlation of Critical Heat Flux in a Bundle Cooled by Pressurized _
Water, Supplement 1, BAW-10000A, Babcock & Wilcox, Lynchburg, Virginia, l' b May 1976. :
E
- 5. C. L. Wheeler, et al . , COBRAIV-1: An Interim Version of COBRA for - ES-Thermal-Hydraulic Analysis of Rod Bundle Nuclear Fuel Elements and 1 Cores, BNWL-1962, UC-32, Battelle Northwest Laboratories, Richland, I Washington, March 1976. -b
- n"
- 6. Reactor Vessel Model Flow Tests, BAW-10037, Rev. 2, Babcock & Wilcox, Lynchburg, Virginia, November 1972. ~t
- 7. D. S. Rowe, COBRAIIIC: A Digital Computer Program for Steady-State and .
Transient Thermal-Hydraulic Analysis of Rod Bundle Nuclear Fuel El e- . },f ments, BWNL-1695, Battelle Northwest Laboratories, Richland, :.
l_
i Washington, March 1976. r
- 8. Arkansas Nuclear One, Unit 1 - Cycle 2 Reload Report, Parts 1 and 2, ((
1 Jr BAW-1433, Babcock & Wilcox, Lynchburg, Virginia, November 1976. - --
- 9. C. D. Morgan, et al . , RADAR, Reactor Thermal and Hydraulic Analysis ..
1 1
r-During Reactor Flow Coastdown, BAW-10069-A, Rev.1, Babcock & Wilcox, -
s Lynchburg, Virginia, October 1974. .-
b.
1 i 6-1 Babcock &Wilcom a McDermott company h -
. F-
--iumi