II. BACKGROUND. has been developed for estimating radioactivity release

Transcription

1 Joy L. Rempe Idaho National Engineering Laboratory P.O. Box 1625, MS 3840 Idaho Falls, ID (208) Melinda J. Cebull Idaho National Engineering Laboratory P.O. Box 1625, MS 3779 Idaho Falls, ID (208) ABSTRACT The Parametric Source Term (PST) code has been developed for estimating radioactivity release fractions. The PST code is a framework of equations based on activity transport between volumes in the release pathway from the core, through the vessel, through the containment, and to the environment. The code is fastrunning because it obtains exact solutions to differential equations for activity transport in each volume for each time interval. It has successfully been applied to estimate source terms for the six Pressurized Water Reactors (PWRs) that were selected for initid consideration in the Accident Sequence Precursor (ASP) Level 2 model development effort. This paper describes the PST code and the manner in which it has been applied to estimate radioactivity release fractions for the six PWRs initially considered in the ASP Program. I. INTRODUCTION A new code, the Parametric Source Term (PST)code, has been developed for estimating radioactivity release fractions. This code was created to support the Level 2 and Level 3 'model development effort for the U.S. Nuclear Regulatory Commission's Accident Sequence Precursor (ASP) program. The ASP program was initiated to provide a structured, probabilistic method for reviewing operational experience to determine and assess both known and unrecognized vulnerabilities that could lead to core damage accidents. For plants modeled in NUREG-1 150,' the XSOR codes2 could be applied to estimate radioactivity release fractions. However, the ASP program includes plants not analyzed in W G Thus, the PST code was developed so that radioactivity releases could be estimated for all plants considered in the ASP program. The PST code is a framework of equations based on activity transport between volumes in the release pathway from the core, through the vessel, through the containment, and to the environment. The code is fastrunning because it obtains exact solutions to differential equations for activity transport in each volume for each B. Gay Gilbert Idaho National Engineering Laboratory P.O. Box 1625, MS 3855 Idaho Fails, ID (208) time interval. Hence, PST allows fairly large time intervals to be selected. PST currently performs point estimate calculations. However, PST may easily be modified so that source term uncertainty distributions could also be estimated using Monte Carlo with importance sampling or Latin Hypercube techniques to select from PST input parameter uncertainty distributions. The primary differences between the XSOR parametric code and the PST code are that the PST modeling framework 1) ensures conservation of activity as it is transported across various volumes in its release pathway, 2) allows a limited amount of consideration on the timedependent behavior of input parameter uncertainty distributions, 3) allows more direct application of recent state-of-the-art severe accident analysis code results for estimation of input parameters (and thus reduces the dependence on expert opinion), 4) increases modeling flexibility because linkage between volumes is specified by user input, 5) allows plants not considered in the W G studies to be evaluated, and 6) facilitates input parameter uncertainty and sensitivity studies. PST has been developed to be compatible with subsequent codes used in the NUREG-1150 studies. Although the user may specify parameters, such as volumetric flowrates, deposition rates, and source rates, values for production runs are determined based on a series of 12 characters, the Source Term Vector (STV). II. BACKGROUND For the NUREG program, source terms were estimated using the source term estimation codes, XSOR? These codes were written for three PWRs (Surry, Sequoyah, and Zion). Because of differences between these plant designs, plant-specific versions of the XSOR codes were written for each plant (referred to as SURSOR, SEQSOR, and ZISOR). The XSOR codes are not detailed mechanistic models; they were not designed to be first-principle models of fission product transport, physics, and chemistry. Instead, the XSOR codes integrate the results of many derailed codes and conclusions of many experts by decomposing the source term into several parts.

2 First, the total release is decomposed by considering the time of release from the core and the pathway followed. For example, release fractions may be considered that pertain to releases originating in the vessel (e.g., releases from the fuel to the RCS before vessel failure) and releases originating ex-vessel (e.g., releases from core-concrete interactions after vessel failure). Second, the XSOR models subdivide each release fraction into their constituent parts. Specifically, each release fraction is expressed as a product of factors that correspond to the number of steps in the release path or important processes for the release mode. For example, XSOR decomposes the release from the fuel in the vessel into the passage from the fuel to the RCS, the passage from the RCS to the containment, and the passage from the containment to the environment. XSOR input is not time-dependent. Rather, code input corresponds to integrated release fractions or decontamination factors over the specified time period, early or late. Typically, XSOR release fractions were quantified by expert opinion. However, the ASP program required that plants not considered in the NUREG-1150 program be evaluated. Furthermore, significant advances in severe accident research had been made since the time that the XSOR codes were created, and there was a desire to incorporate these advances into source term estimates for the ASP PWRs. Thus, a method was needed that could be applied to estimate source terms for these plants, which had substantially different design features than the Westinghouse PWRs considered in NUREG Finally, a method was desired that could be easily extrapolated to other plants and modified to incorporate future updates in source terms. These desired improvements in source term estimate methodology motivated the development of the PST code. III. METHODOLOGY Although a larger number of volumes containment subcompartments) may be incorporated into PST, the model developed for the initial ASP PWR source term calculations only considers the volumes illustrated in Figure 1. In the ASP PWR models, input was selected that specifies which flowpaths are active between volumes for each time period. Possible sources and losses from each volume are shown in Figure 1. ASP PWR input also determines which source and loss mechanisms are active for each time period. Resuspended f-----l reiease Oxidation TemperatureResuspended induced plateout + RCS I Leakage Containment F Containment \ Deposition f Deposition Figure 1. Control v included in the initial PST model. 4 I f Environment s,,

3 DISCLAIMER Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

4 DISCLAIMER This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, make any warranty, express or implied, or assumes any legal liability or respolm'bility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commeraal product, pmess, or service by trade name, trademark, manufacturer, or otherwise does not necessan'ly constitute or imply its endorsement, recammendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

5 Similar to the manner in which XSOR groups radionuclides, PST considers nine release classes. These nine release classes encompass a total of 60 isotopes. Representative isotopes have been selected (based on health effects) in order to select a nuclide decay constant and group to which the nuclide decays. The general transport equation for each volume is: where -- the nth volume's fractional activity (dimensionless) the fractional loss rate (s-') of the nth volume's activity the fractional rate at which activity enters the nth volume (s-'). The fractional loss rate may include the following mechanisms: the fractional leakage rate to the next volume (s-*) the deposition rate within a volume (s-l) the decay rate (s-i). Equation (1) is applied to each volume and solved analytically in PST. Specific equations obtained for each of the control volumes may be found in Reference 3. IV. COMPUTATIONAL DESIGN PST has been designed to read ASP Level 2 Source Term Event Tree (STET) output and produce output that interfaces with the NUREG-I 150codes. Specifically,PST imports a file containing a list of 12-character Source Term Vectors (STVs) from the ASP STJ3T and a file containing default input values, such as volumetric ff owrates, deposition rates, and source rates, that are based on the STV. PST output includes a summary file that is formatted similar to the XSOR output file and a detailed output file that lists output for each PST timestep. Typically, PST input values for production runs are selected from the default input file based on the STV. However, the user may perform sensitivity studies by modifying default values. PST can easily be applied to other plants. User input determines the number of volumes to be considered in the release pathway and which Bowpaths are active between each volume. Plant-specific analysis results from more detailed, severe accident analysis codes are used to quantify lower-level PST input. PST is designed to run on an IBM or IBM-compatible PC. Both DOS and Windows versions of the code are available. The DOS version of PST requires 640 k 3 of RAM; the Windows version of PST requires 16 MB of RAM if Windows NT is used, or 8 Ml3 of RAM if Windows 95 is used. Typical plant calculations require approximately 10 M6 of disk space per plant, although this value may vary depending on the number of STVs required. In addition to the above system requirements, a laser printer may be connected to the PC through port LPTl. Figures 2 and 3 illustrate some of the user-friendly windows available in PST. Figure 2 illustrates screens that help the user identify lower-level PST input for a particular STV. If sensitivity calculations are desired, the user can modify any of the input shown in this window. Figure 3 illustrate the screens used to input default values for production runs. Note that PST includes many help screens, such as the screen shown in Figure 3 that defines all PST default input variables. V. APPLICATION PST has been applied to six PWRs. Table 1 identifies the six plants and the type of PWRs encompassed by each plant. For each of these plants, approximately 1,000 unique STVs required PST source term estimates. Ideally, PST default values for estimating source terms should be quantified using output from severe accident analysis codes. However, it was beyond the scope of the initial ASP Program Level 2 model development effort to run such codes. For each PWR considered in this initial effort, a literature search was performed to determine what code calculations were available for quantifying PST input. Sources of information inciuded calculations performed with SCDAE'/RELAPS? MAAP: MELCOR? and STCP.7 Sequences with similar core, vessel, or containment releases were grouped, where possible; and the higher release fraction for the group was selected as the encompassing release fraction. Rules were then encoded into PST, which allows default input values to be selected for each sequence based on the 12-character source term vector. Reference 3 discusses the method

6

7 r Table 1. Plants considered in initial ASP Level 2 models. PWR Classification Zion Unit 1 Westinghouse 4-loop PWRs with large, dry containment Sequoyah Unit 1 Westinghouse 4-loop with ice condenser containment 1 Calvert cliffs Unit I I Palo Verde Unit I I Ocoiee Unity I surry Unit 1 ~ -1 Combustion Engineering PWRs with PORVs I Combustion Engineering PWRs without PORVs ~ ~~ ~- ~~ ~ Babcock & Wilcox PWRs I Westinghouse PWRs with subatmosphericcontainments used to quantify input for each of the six plants initially considered in the ASP Level 2 models. A plant source term calculation (-1000 STVs) typically runs in approximately 10 minutes on a Pentium computer. VI. SUMMARY A new code, the Parametric Source Term (PST) Code, has been developed for estimating radioactivity release fractions. The PST code is a framework of equations based on activity transport between volumes in the release pathway from the core, through the vessel, through the containment, and to the environment. The code is fast-running because it obtains exact solutions to differential equations for activity transport in each volume for each time interval. It has successfully been applied to estimate source terms for the six PWRs that were selected for initial consideration in the ASP Level 2 Model development effort. In these cases, source terms for between 900 and 1200 STVs were required. PST was capable of generating the required number of source terms in times comparable to previous PSA source term methods, such as XSOR. However, PST provides a superior method for source term estimation because the PST modeling framework 1) ensures conservation of activity as it is transported across various volumes in its release pathway, 2) allows a limited amount of consideration on the time-dependent behavior of input parameter uncertainty distributions, 3) allows more direct application of recent state-of-the-art severe accident analysis code results for input quantification (and thus reduces the dependence on expert opinion), 4) increases modeling flexibility because linkage between volumes is specified by user input, 5) allows plants not considered in the NUREG-I150 studies to be evaluated, and 6) facilitates input parameter uncertainty and sensitivity studies. -I I I ACKNOWLEDGEMENTS The U. S. Nuclear Regulatory Commission supported this work under DOE Contract No. DE-AC07941D REFERENCES 1. U. S. Nuclear Regulatory Commission, Severe Accident Risks: An Assessment for Five U.S. Nuclear Power Plants, NUREG-1150, Volumes 1 and 2, December H-N. Jow, W. B. Murfin, and J. D. Johnson, XSOR Codes Users Manual, NUREGICR-5360, November J. Rempe, M. Cebull, and G. Gilbert, PST User s Guide,to be issued. 4. C. M. Allison, et al., SCDAPA?ELAPS/MOD3.1Code Manual, Volume 11: Damage Progression Model Theory, NUREGICR-6150, EGG-2720, October Fauske and Associates, Incorporated, Modular Accident Analysis Program, Version R. M. Summers et. al., MELCOR Computer Code Manuals, NUREG/CR-6119 and SAND , March Source Term Code Package: A Users Guide, NUREGKR-4587, BMI-2138, July 1986.