This paper presents and illustrates some of the key features of the RELAP5-3D^© code being developed at the Idaho National Engineering andEnvironmental Laboratory (INEEL) for the US Department of Energy (DOE). The purpose of thepaper is to inform potential users of the code of its unique capabilities that extend itsrange of applicability beyond that presently available with any other thermal-hydraulicssystems code. The paper also discusses areas of ongoing development, future plans, and theavailability of the code to the international community.

The RELAP5-3D^© code is an outgrowth of the RELAP5/MOD3 codedeveloped at the Idaho National Engineering & Environmental Laboratory (INEEL) for theU.S. Nuclear Regulatory Commission (NRC). Development of the RELAP5 code series began atthe INEEL under NRC sponsorship in 1975 and continued through several released versions,ending in October 1997 with the soon to be released RELAP5/MOD3.3. The U.S. Department ofEnergy (DOE) began sponsoring additional RELAP5 development in the early 1980s to meet itsown reactor safety assessment needs. Following the accident at Chernobyl, DOE undertook are-assessment of the safety of all of its test and production reactors throughout theUnited States. The RELAP5 code was chosen as the thermal-hydraulic analysis tool becauseof its widespread acceptance. Systematic safety analyses were carried out for the DOE thatincluded the N reactor at Hanford, the K and L reactors at Savannah River, the AdvancedTest Reactor (ATR) at INEEL, the High Flux Isotope Reactor (HFIR) and Advanced NeutronSource (ANS) at Oak Ridge, and the High Flux Beam Reactor (HFBR) at Brookhaven. DOE alsochose RELAP5 for the independent safety analysis of the New Production Reactor (NPR)proposed for Savannah River before that program was canceled in the wake of the end of thecold war.

The application of RELAP5 to these various reactor designs created the need for newmodeling capabilities. For example, the analysis of the Savannah River reactorsnecessitated a three-dimensional flow model and heavy water properties be added to thecode. ATR required a new critical heat flux correlation applicable to its unique fueldesign. All together, DOE sponsored improvements and enhancements have amounted to amultimillion-dollar investment in the code.

Toward the end of 1995, it became clear that the efficiencies realized by themaintenance of a single source code for use by both NRC and DOE were being overcome by theextra effort required to accommodate sometimes conflicting requirements. The code wastherefore "split" into two versions, one for NRC and the other for DOE. The DOEversion maintained all of the capabilities and validation history of the predecessor code,plus the added capabilities that had been sponsored by the DOE before and after the split.

At the outset of the decision to split the code into NRC and DOEversions, the INEEL recognized the importance of retaining the pedigree stemming from theextensive validation history of RELAP5/MOD3. Consequently, the developmental activitieswith respect to RELAP5-3D^© since the split have been carefullyintegrated so as not to compromise this legacy validation. In fact, virtually all of theenhancements in RELAP5-3D^© are optional and supplementalto the proven performance of RELAP5/MOD3.2. Consequently, users of RELAP5-3D^© can confidently apply the code using existing, one–dimensional RELAP5/MOD3.2 input decks and expect their results to be the same orimproved.

To ensure that the code remains true to its validation history (or provides improvedresults), developmental versions of the code are periodically tested using a subset ofcases from the RELAP5/MOD3 validation library. Two such cases are the G.E Level Swell andTHTF experiments.

The G.E. Level Swell experiments¹, conducted in the late 1970’sprovided excellent data for assessing flashing and interphase drag models. Two tanks, one1 ft. in diameter and the other 4 ft. in diameter were used. In the experiments, the tankswere partially filled with water and heated to near the saturation temperature. They werethen depressurized through a blowdown valve. Differential pressure measurements made alongthe vertical length of the tank allowed for computing average void fractions as a functionof elevation inside the tank.

Figure 1 shows the calculated and measured void profiles in the 4ft. diameter tank of Test 5801-13 at four times during the blowdown (5, 10, 15, and 20sec.). RELAP5/MOD3.2 and RELAP5-3D^© produced identicalresults, both being in excellent agreement with the data.

The Thermal Hydraulic Test Facility (THTF) at Oak Ridge National Laboratory (ORNL) wasdesigned to simulate conditions in a PWR core. The test section contained 64 electricallyheated rods with internal dimensions typical of a 17 by 17 rod bundle. Differentialpressure measurements positioned along the axial length of the bundle were used tocompute void fractions, and thermocouples were employed to measure steam temperature andheater rod surface temperature.

In Test 3.09.10i², the bundle was maintained in a partially uncoveredcondition at a pressure of 4.5 MPa, an inlet mass flux of 29.8 kg/m²-s, and aninlet subcooling of 57.6 K. The heater rod heat flux (uniform axial and radial) was7.44x10⁴ W/m². Figure 2 compares RELAP5/MOD3.2 and RELAP5-3D^© results with measured data for bundle void profile, heater rodtemperature profile, and vapor temperature profile. Again, the two code versions arevirtually identical and agree well with the data.

The consistency of results between RELAP5/MOD3.2 and RELAP5-3D^©is not surprising, in light of the fact that the constitutive models for one-dimensionalflow have not been significantly altered.

Figure 1. Calculated and Measured Void Profiles During GE Level SwellTest 5801-13

RELAP5-3D^© CODE FEATURES

The most prominent attribute that distinguishes the DOE code from the NRC code is thefully integrated, multi-dimensional thermal-hydraulic and kinetic modeling capability inthe DOE code. This removes any restrictions on the applicability of the code to the fullrange of postulated reactor accidents. Other enhancements include a new matrix solver, newwater properties, and improved time advancement for greater robustness. Together with theexisting modeling capabilities of RELAP5/MOD3.2, these enhancements make the code the mostpowerful tool of its kind available.

The balance of this paper focuses on the capabilities of the three-dimensionalhydrodynamic model, the multi-dimensional kinetics model, and the new BPLU matrix solver.Other features unique to RELAP5-3D^© are briefly described.

The development of the three-dimensional hydrodynamic model in RELAP5 began in 1990under funding from the Savannah River National Laboratory (SRNL). Since then developmentand testing has continued from both planned improvements and responses to user requests.This has resulted in a reliable model. Progress has been documented in the literature andat various meetings (References 3-12).

The multi-dimensional component in RELAP5-3D^© was developedto allow the user to more accurately model the multi-dimensional flow behavior that can beexhibited in any component or region of a LWR system. Typically, this will be the lowerplenum, core, upper plenum and downcomer regions of an LWR. However, the model is general,and is not restricted to use in the reactor vessel. The component defines a one, two, orthree-dimensional array of volumes and the internal junctions connecting them. Thegeometry can be either Cartesian (x, y, z) or cylindrical (r, q, z). An orthogonal, three-dimensional grid is defined by mesh interval input data in eachof the three coordinate directions.

Verification of the model has been performed by using conceptual problems that haveexact solutions. These types of problems are used to demonstrate that the equations havebeen correctly coded, and are a precursor to model validation using experimental data.Three such problems are the "Rigid Body Rotation", "Pure Radial SymmetricFlow", and R-q Symmetric Flow" problems. Each of theproblems is based on a cylindrical, multidimensional component with eight rings, sixsectors, and one axial level. All six sectors are symmetrical, with non-uniform radialspacing. Six time dependent volumes are attached to the six outer sectors by timedependent junctions for inlet flow specification. In addition, six time dependent volumesare attached to the six inner sectors by a multiple junction component. Figure 3 shows thenodalization for 1 sector.

In each of the simulations, losses due to friction and body force terms are deactivatedto create problems with exact solutions. Also, the flow is assumed to be steady andincompressible. The azimuthal flow pattern, where necessary, was imposed by setting theouter ring azimuthal velocities to the desired value. All problems have analytic solutionsfor velocity from the continuity and q -momentum equations, andanalytic solutions for pressure from the r-momentum equations.

The Rigid Body Rotation problem represents a hollow cylinder with a symmetric flowpattern in the azimuthal direction. Flow in the radial direction does not exist. Thisproblem tests only the azimuthal momentum flux terms. The test conditions and boundaryconditions are: The azimuthal velocity at the 6.5 m radius position is 1 m/s, pressure atthe 1 m radius position is 5x10⁵ Pa, and all radial velocities are 0.0 m/s.

Comparisons between the RELAP5-3D^© calculated results andthe analytic solution for the Rigid Body Rotation problem are shown in Figures 4 and 5.The calculated velocity and pressure profiles are seen to exactly match the analyticalsolutions.

The Pure Radial Symmetric Flow problem represents a hollow cylinder with a symmetricflow pattern in the radial direction. This problem tests only the radial momentum terms.The test conditions and boundary conditions are: All azimuthal velocities are 0 m/s,pressure at the 1 m radius position is 5x10⁵ Pa, and radial velocity at the 7.5m radius position is 0.8667 m/s (from outside to inside of the ring). Comparisons betweenthe calculated results and the analytic solution for the Pure Radial Symmetric Flowproblem are shown in Figures 6 and 7. Again, the RELAP5-3D^©results are in agreement with the exact solution.

The R-q Symmetric Flow problem represents a hollow cylinderwith a symmetric flow pattern in both the radial and azimuthal directions. The azimuthalvelocity at the 6.5 m radius position is 1 m/s, pressure at the 1 m radius position is5x10⁵ Pa, and radial velocity at the 7.5 m radius position is 0.8667 m/s (fromoutside to inside of the ring). The calculated and analytic solution results are seen tobe in agreement (Figures 8, 9, and 10).

Both the analytic and calculated velocity profiles come from the continuity and q -momentum equations; they are each a function of radius and outerring velocity. The velocity profile from the Rigid Body Rotation problem is linear (Figure4), while the velocity profiles from the other problems are inversely proportional toradius (Figures 6, 8 and 9). The pressures are calculated from the r-momentum equations(Figures 5, 7, and 10). The pressure profile for the Rigid Body Rotation problem isconcave up (Figure 5), while the others are concave down. Additionally, the pressureprofile from the R-q Symmetric Flow problem is influenced bythe 1-D outlet connections. There, the azimuthal flow contributes to the pressure solutionup to the point of connection.

These simple test problems verify the correct implementation of the three dimensionalcontinuity and momentum equations. This is a necessary, but certainly not sufficient, testof the 3D model. Work has begun on performing a number of validation cases usingexperiments exhibiting multidimensional flow behavior.

One such validation case using experimental data with multidimensional flow behavior isdiscussed in Part 2 of the detailed paper.  The experiment discussed is the LOFT TestL2-5.  This link will take you to Part 2 of thedetailed paper.

Part 3 of the detailed paper discusses the mutlidimensional neutron kinetics model andBorder Profiled Lower Upper (BPLU) sparse matrix solution technique implemented inRELAP5-3D^©.  This link will take you to Part 3 of thedetailed paper.

You may also return to the briefpaper.

Questions or problems regarding this web site should be directed to danidl@inel.gov.

Last modified: Thursday November 12, 1998.