ML20003F506
| ML20003F506 | |
| Person / Time | |
|---|---|
| Site: | Clinch River |
| Issue date: | 04/13/1981 |
| From: | Minogue R NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Harold Denton Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML20003F507 | List: |
| References | |
| RTR-NUREG-CR-0741, RTR-NUREG-CR-741 RIL-120, NUDOCS 8104210559 | |
| Download: ML20003F506 (10) | |
Text
c/
r p --. #
pm uc UNITED STATES uq'o, NUCLEAR REGULATORY COMMISSION
+
[ ' 3 3, k
WASHINGTON. D. C. 20555 g
.. j
\\,*...+/
APR 131981 MEMORANDUM FOR: Harold Denton, Director Office of Nuclear Reactor Regulation FROM:
Robert B. Minogue, Director Office of Nuclear Regulatory Research
SUBJECT:
RESEARCH INFORMATION LETTER #120 APPLICATIONS OF THE COM4IX 3-DIMENSIONAL THEPfiAL HYDRAULIC COMPUTER CODES 1.
Introduction This Research Information Letter (RIL) describes the COMMIX (Component Mixing) family of computer codes and discusses some of the applications of these codes. Since there has been no licensing activity on Liquid Metal Fast Breeder Reactors by NRC in the past several years, there has been only nominal interest in the NRC (outside RES) in the Three-Dimensional Transient Thermal Hydraulic Computer Code Development Program at the Argonne National Laboratory. Nevertheless the program to develop COMMIX codes has been continued to provide a tool for licensing LMFBR's in the future.
In addition, the code has been applied to a wide variety of LWR design problems.
The initial objective of the Argonne National Laboratory Program, "Three-Dimensional Code Development for Core Thennal Hydraulic Analysis of LMFBR Accidents Under Natural Convection," was to develop a benchmark code for the detailed study of natural convective heat removal in an LMFBR during a loss ofthecodeCOMMIX-1g)forcedcirculationaccident. The original version was slow running and was useful only to solve benchmark problems at great computer expense (hours on an IBM lg.
Since that version, improvements have been made in the numerics of the code and in the initialization scheme so that COR4IX 1A, running more than 10 times faster than COMMIX-1, is a practical code for solving engineering problems. The two-phase operational and has been documented.grsion of the code COMMIX-2 isThe code has impr over COMMIX-1A which makes transient two-phase flow calculations feasible.
In order to analyse experimental data in an endeavor to obtain betf$[6'7) sodium boiling models for COMMIX-2, a specialized code 80DYFIT-IFE has been developed to analyse small hexagonal fuel-rod bundle without the usual subchannel analysis assumptions. Although COMMIX was developed to be a research tool, it has developed into a tool that should be useful when licensing of Clinch River Breeder Reactor resumes.
120555064215 2 L1 US NRC ADM DOCUMENT CONTROL DES K POR 016 WASHINGTON CC 20555 8104g 105 5 9
4 s
4 l
2 Harold Denton Since the COMMIX code is well documented and the primary purpose of this RIL is to discuss application, the details of the code structure willnotg)discussedhere. An excerpt from the COMMIX-2 code des-cription is enclosed. References 1-8 give descriptions of the models and individual computer codes.
As the improvements have been made in the COMMIX codes and the use-fulness of some of its features such as volume porosity, surface per-meability have been understood, the code has been widely accepted by the LMFBR community. The computer program has been obtained by nearly all of the organizations, domestic and foreign, that are working on LMFBR technology. The codes have also been applied to high temp-erature gas cooled reactors and LWR component problems.
- 2..Results Five ag1jgtions of COMMIX-1A are reported in a paper by H. M. Domanus et al These applications use the porous medium approach.
The experiments studied are referenced in the reports. The appli-cations are:
(a) Natural circulation in a 19-pin bundle with a 3:1 radial power skew. As shown in Figure 4 of the paper, the agree-ment between measured and calculated outlet temperature dis-tribution is excellent.
(b) Planar blockage in a 19-pin bundle. as shown on Figure 17 of the paper, thg agreement betweer :alculated outlet tempera-ture is within 5 C.
A wire wrap Rr:e treatment is necessary to obtain reasonable agreement at tha wall.
(c) Loss of piping integrity transient in a 19-pin bundle.
This is a rapid flow and power reduction transient. As shown in Figures 21 through 24 of the paper, the agreement between measured and calculated temperatures vs time is good.
i (d) Flow rundown in a seven pin bundle. This is the Geman 7-pjn flow rundown experiment without rundown of power until boiling starts. Thecalgulatedtimetoincipientboiling assuming the measured 68 C superheat condition is 6.28s which compares favorably with the measured 6.37s. As siiown on Figcres 28 through 32, the agreement between measured and calculated temperatures vs time in t1e bundle is excellent up to the point where the sodium becomes superheated. This experimentwas(
calculated using the BODY-IFE code with good agreement i
l l
g l
e 3
Harold Denton (e) CRBR outlet plenum transient in a 1/10 scale flow model.
This is an example of COMMIX's ability to handle a complicated three dimensional geometry. As shown on Figure 41 of the paper, the agreement between measured and calculated tempera-tures vs time at the exit and at the top of the plenum is good. The relatively poor agreement during part of the transient at themocouple (TC) 12 is due to the coarse mesh used in the calculation. As shown in Figures 27-38 a stratified layer formed at TCl2. The code was calculating the average tempera-ture of a 20 cm thick region, while the thermocouple was meas-uring the. temperature above the stratified layer.
A simulation of a flow and temperature transient in an intemediate pipe loop during a typical CRBR scram, calculated using the COMMIX-1A code, is reported in reference 12. Because of the parabolic inlet velocity distribution and the temperature decrease at the inlet, cooler sodium penetrates into the central portion of the pipe and is relatively cooler than the bottom portion early in the transient. As a result of theflowredistrfbution,aconsiderabletoptobottomtemperaturedifference (in excess of 65 C) due to flow stratification was found in the later part of the transient. Temperature gradients of this magnitude need to be considered in evaluating piping integrity.
Some early results using COMMIX-2 are reported in reference 13.
These are:
(a) Transient single-phase flow with heat flux.
(b) Two-phase flow with heat flux from a wall.
(c) Separation of vapor and liquid by gravity forces.
(d) High-pressure jet impingement on a vertical wall.
The reported results confirm the ability of COMMIX-2 to handle transient, I
three-dimensional, two-phase flow.
Application by non-NRC users The COMMIX-1A computer code has been obtained by 14 non-NRC users.
Although they are not adequately documented to be referenced in this RIL, there are a number of user applications of the code that may be i
useful within NRC.
1 (a) A version of the code COMMIX-IHX has been developed by Foster Wheeler and ANL for analysis of intemediate (sodium to sodium) heat exchangers under DOE sponsorship.
(b) Westinghouse (WARD), General Electric, and Babcock & Wilcox Co. are using the code to analyze flow patterns in LMFBR plena.
(c) The flow pattern in the Donnreay pFR reactor vessel is being analyzed at ANL under the current NRC program in accord with D,
fC8Srecommendations.
e T
1 4
Harold Denton (d) General Atomics Co. has used the code to analyze natural convection in upper plenum of the HTGR during loss of forced flow conditions. They have also used it to analyze the natural convection in the core, upper plenum and lower plenum of the gas cooled fast reactor in a single calculation (see enclosure 2).
(e) ANL is setting up a complete in vessel calculation for a pot-type LMFBR as a carry-on of the Conceptual Design Study.
(f) The Westinghouse Electric Corporation Steam Generator Division and ANL have made a joint proposal to DOE to develop and verify steam generator codes based on the COMMIX codes.
The code would be used to analyze flow patterns, temperature transients, and vibratory forces.
3.
Evaluation The COMMIX-1A code has been well tested against experimental data on sodium and water systems and can be used with confidence to evaluate transient flow patterns, temperature distributions and flow-induced forces in fuel-rod bundles and reactor componcnts where single-phase flow is applicable. The code has been used by General Atomic Company on helium systems. They are enthusiastic about the code but have not done any appreciable testing against experiment. A scaled plenum flow experiment is planned but is several years away. Further eval-uation of the code against experiment is reconmended before relying on it for assessment of gas cooled reactors.
COMMIX-2, the two phase version of the code, is applicable to transient dispersed turbulent flow. Unfortunately, restits are sensitive to the drag coefficient (K) and evaporation-condensation rate constants, as are all two-phase flow computer codes. Also, more work is needed on initiation of boiling models since the degree of superheat before initiation of boiling is unpredictable in sodium systems. A comparison between COMMIX and other thermal hydraulic coces (TH13D-1, COBRA-lllC, and SABRE-1) is given in reference 14.
RES contact:
P. M. Wood ext. 74326 y
Robert B. Mincgue', pirector Office of Nuclear Regulatory Research
Enclosures:
As stated 1.
Excerpt from COMMIX-2 Report 2.
Letter 2/17/81, Del Bene, GA to Kelber, NRC cc:
R. Matson, NRR P. Check, NRR T. Speis, NRR G. W. Knighton, NRR R. Audette, NRR P. Williams, NRR B. Sheron, NRR r
l
r s
References:
1.
W. T. Sha, et al. COMMIX-1, "A Three-Dimensional Transient Single-l Phase Component Computer Program For Thermal Hydraulic Analysis,"
j NUREG-0785, ANL-77-96, January 1978.
2.
V. L. Shah, et al. "A Nunn C Procedure For Calculating Steady /
Unsteady, Single-Phase /Two.
.e Three-Dimensional Fluid Flow.Mith Heat Transfer," NUREG/CR-0782, ANL-CT-79-31, June 1979.
3.
H. M. Domanus, et al. " COMMIX-2: A Steady / Unsteady Single-Phase /
Two Phase Three-Dimensional Computer Program for Themal-Hydraulic Analysis of Reactor Components" NUREG/CR-1807, ANL-81-10, March 1981.
4.
C. C. Miao, et al. " Analytic Rebalance Technique for Prassure Calculation in Two-Phase Flow Systems" NUREG/CR-1422, ANL-CT-80-19, April 1980.
5.
W. T. Sha and J. F. Thomnson, " Rod-Bundle Thermal-Hydraulic Analysis Using Boundary-Fitted Coordinate System," NUREG/CR-0001, ML-78-1, i
January 1979.
i 6.
B. C. J. Chen, et al. "BODYFIT-IFE: A Computer Code for Three-Dimensional Steady-State / Transient Single-Phase Rod-Bundle Thermal-Hydraulic Analysis," NUREG/CR-1874, ANL-80-127, November 1980.
7.
B. C. J. Chen, S. P. Vanke, and W. T. Sha, "Some Recent Computaticns of Rod-Bundle Thermal Hydraulics Using Boundary Fitted Coordinates,"
Nuc. Eng. and Design, Vol. 62, pg.123, Decemiser 1980.
8.
W. T. Sha and S. L. Soo, "Multidomain Fluid Mechanics," ANL-CT-77-3 March 1977.
9.
H. M. Domanus, V. L Shah, and W. T. Sha, " Applications of the COMMIX Code Using the Porous Medium Formulation," Nuc. Eng. and Design, Vol. 62, pg. 81, December 1980.
- 10. H. M. Domanus, M. J. Chen, and W. T. Sha, " Computational Results For a 7-pin Hexagonal Fuel Assembly During a Flow Rundown Transient Using the COMMIX-1A Computer Code," NUREG/CR 1285, January 1980.
- 11. B. C. J. Chen and W. T. Sha, " Simulation of Steady-State and i
Transient Sodium Boiling Experiments in a Seven-Pin Bundle Under Flow Rundown Conditions by Using BODYFIT-IFE Code," NUREG/CR-1814, ANL-CT-81-9, January 1981.
- 12. M. J. Chen, H. M. Domanus, and W. T. Sha, " Simulation of a Therma-hydraulic Transient in a Pipe Using the COMMIX-1A Computer Code,"
NUREG/CR-1323, ANL-CT-80-15, February 1980.
I
13.
V. L. Shah, et al. "Some Namerical Results With the COMMIX-2 Computer Code" NUREG/CR-0741, ANL-CT-30, March 1979.
14.
W. T. Sha, "An Overviev of Rod-Bundle Thermal-Hydraulic Analysis" Nuc. Eng. & Design, Vol. 62 (1-3) page 1, December 1980 and NUREG/
CR/1825 November 1980.
15.
F. H. Harlow and A. A. Amsden, " Flow of Interpenetrating Material Phases," J. Comp. Phys. 18, p. 440 (1975).
16.
S. V. Patankar, " Numerical Heat Transfer and Fluid Flow, Numerical Heat Transfer," Vol. 2, McGraw-Hill, New York (1979).
17.
D. B. Spalding, "A Novel Finite-difference Formulation for Differ-ential Expressions Involving Both First and Second Derivatives,"
Int. J. Num. Methods in Eng., Vol. 4, p. 551 (1972).
18.
D. B. Spalding, "The Calculation of Free-convection Phenomena in Ga-liquid Mixture," Imperial College, Heat Transfer Section Report, HTS /76/ll (1976).
1 l
l t
l l
l r
r~
t
. Exccrpt from COMMIX-2 Report NUREG/CR 0741, Reference 13 I.
INTRODUCTION The present-generation computer speed and stt, rage capacity, coupled with recent advances in numerical-solution techniques for systems of quasi.-
linear Partial differential equations have made possible detailed numerical simulation of many engineering problem,s. With the anticipated improved performance of the next generation of computers and further advances in numerical-solution techniques, use of numerical simulation for solving engineering problems is expected to increase for many years to come.
Basically, numerical simulation in engineering applications can be clas-sified into two categories: the system computer program and the component computer program. Generally, the system computer program consists of a number of components; therefore, it cannot afford to give a detailed numerical modeling of each component.
In contrast, the component com-puter program deals with one component of interest; therefore, it can afford to provide a detailed numerical simulation. The work presented in this report is focused on the component computer program.
During loss of coolant or transient overpower accident situations, boil-ing of liquid coolant in a reactor core is postulated due to high temper-atures in the core. The fluid mixture of liquid and vapor, in such circumstances, is nonhomogeneous with both phases being in nonequilibrium themodynamic states.
It is, therefore, desirable to develop a computer code for obtaining numerical solutions of three-dimensional, transient, two-phase (gas-liquid) flow system with nonequilibrium and nonhomogeneous coiditions.
The COMMIX-2 code is a steady / unsteady, three-dimensional two-phase com-puter code for themal hydraulic analysis of reactor components under nomal and off-normal 5 perating conditions.
It uses the two-fluid model o
of Harlow and Ansden. to describe the conservation equations of mass, momentum and energy. Consequently, we can analyze a wide spectrum of flow conditions; i.e., from Pomogeneous and equilibrium to nonhomogeneous and nonequilibrium condition. The interactions between two fluids are accounted for by incorpo*ating the corresponding tems in all of the conservation equations. The staggered grid system is used to describe the field variables at the cetter of a cell and flow variables at the center of a cell and flow variables at the surface of a cell.
The structure of the code is similar to that of COMMIX-IA.2 The calcu-lation pro procedure,gdureemployedisanextensionofthesingle-phasenumerical known as SIMPLER (Semi-Implicit Method.for Pressure Linked Equation-Revised).
In this procedure, we use the liquid phase continuity equation to obtain the void fractions, and use the combined continuity equation to derive the pressure and pressure correction equations.
The specific features of C0 MIX-2 are the fullowing:
s.,
~
I
t 2
1.
To permit an analysis of a flow domain with solid objects, the vol-ume porosity, surface permeability, distributed resistance and distributed heat source are incorporated in the conservation equations.
I7 2.
An approximate form of Spalding's equation is used to derive the the finite difference formulation of the convective and diffusion terms.
This equation is a function of the Peclet number and it combines the best features of both, the central difference and upwind difference schemes.
3.
The discretization equations are obtained by integrating the conservatior equations over a control volume surrounding a grid point. Thus, the Derivation process and the resulting equations have direct physical meaning, and the consequent solution satisfies the conservation principles.
4.
The convective, diffusion, interfacial friction and interfacial heat transfer terms are made implicit for more stable formulation and to permit larger time steps.
5.
The discretization equations are formulated with time step size appearing only in the denominator of all transient terms. With this arrangement, for a steady state calculation, all of the transient terms can be eliminated from computation by specifying a very large value of time step size.
6.
The general form of all discretization equations is 3
0 + nb a b nb
=b '
n O
where, 4is a dependent variable and subscript NB stands for neighboring a
points. This general form of the discretization equation permits various solution schemes, e.g., cell by cell, line by line, plane by plane, block iterative, direct inversion etc.
1:
7.
The COMMIX-2 code is structured such as to pennit solution of single phase (gas or liquid) as well as two-phase (gas and liquid) flow problems.
In addition, it permits 10, 2D, or 3D calculation in either (X,Y,Z) or (R,Z,0) coordinates.
8.
The COMMIX-2 code has modular structure. This permits rapid c;
implementation of the latest available drag models, heat transfer models, boiling models, etc.
x' 9
The code has also an option permitting use of either sodium p"operty package or water property package.
4
- 10. The program also contains (i) A generalized resistance moael to permit determination of resistance due to internal structures (fuel rods, wire wrap, baffles, grid spacers, etc.)
l*
I l
(
~~
G s
3 (ii) A generalized thermal structure formulation to model thermal interaction between structures (fuel rods, wire wraps, duct wall, baffles, etc.) and surrounding fluid, and (iii) A local regional mass rebalancing scheme, such as plane by i
plant, for improving the convergence rate.
This report describes the COMMIX-2 program for the solution of the governing equations for three-dimensional, single-phase /two-phase, steady / unsteady flow with heat transfer. The description here i
starts with the differential equations and deals with numerical method incorporated into a computer program. Section 2 is devoted R
to the set of governing equations for the situation considered.
In Subsection 2.4, the general form of all the governing equations is recognized; this generalization facilitates a unified devel-i opment of the numerical method and the construction of the computer i.
program.
?
The conservation equations for quasi-continuum regime are presented
'c in Section 3.
We define the quasi-continuum regime as a medium which
.,e i
contains finite, dispersed, stationary heat generating (or absorbing)
{
solid objects. The effects of solid objects in a medium are accounted for by introducing volume porosity surface permeabilities, distri-l buted heat sources. The physical models and constitutive equations y
used in COMMIX-2 for describing the mass, momentum and energy B.
exchange phenomena are presented in Section 4.
f.
In Section 5 we present some preliminary considerations before we start assembling the finite difference equations. The finite differ-ence equations. The finite difference formulation of the general i
equation is presented in Section 6.
As we use a staggered grid system, the control volumes for momentum equations are different u
and require special considerations. The special features of the 7-finite-difference equations for momentum are discussed in Section 7.
g In Section 8 we have presented the finite difference forms of the continuity equations.
5 Section 9 contains the derivation of pressure and pressure correction pv equations.
In the present program we have two alternative forms of i
pressure correction equation leading to two alternative solution numerical procedure,g procedure is an extension of the single-phase procedures. The fir l'&
/
known as SIMPLER (Semi-Implicit Method for Pressure Linked Equatior.-Revised).
In this procedure we use one lY of the two phase continuity equations to determine the liquid volume c.
'. g.
fractions, and use the conbined continuity equation to derive the pressure correctic.i equation. In the second procedure we use both of the phase continuity equations to determine the liquid volume fractions; the difference lies in the derivation of the pressure correction f l' equation. In this procedure we differentiate the phase continuity l
equation}8and momentum equations and then combine them to obtain the t
l 4,
pressure known as Inter Phase Slip Analyzer (IPSA).
t
.. -)!
i l
l e
^
l, Y
4 Section 10 deals with the boundary conditions for the different i
dependent variables. A discussion of the ways of handling irregular geometries is included in Subsection 10.4.
A line-by-line procedure for solving the finite-difference equations is presented in Sectiorf 11.
For most of the problems analyzed, this procedure has been found to be superior to the usual point-by-point procedure without rebalance technique.
In Section 12, we take an overall view of the entire
,8 calculation sequence. The various steps in the iteration scheme are listed in Section 12.1, while the remainder of Section 12 is de-i voted to matters that enhance the chances of obtaining a converged solution. Section 13 describes the flow chart.
c The thermodynamic and transport properties of sodium and water are
.f given in Appendix A.
The thermal structure module is described
?
in Appendix B.
Appendix C contains the descriptions of the resistar,ce
?y, and wire wrap models. The code input description and sample problems are given in Appendices D and E, respectively.
kIs
't
!s).
i Y
f
?.
1 1
L s.
- i
.~.
- i 4
%(;
- f-
. ' ?., '
4, i
s e
f'
.