Benchmarking a new closed-form thermal analysis technique against a traditional lumped parameter, finite-difference method

Huff, K. D. ; Bauer, T. H. (Nuclear Engineering Division)

关键词: AR FACILITIES;    ACCURACY;    ANL;    ENGINES;    GEOLOGY;    HEAT SOURCES;    LAWRENCE LIVERMORE NATIONAL LABORATORY;    RADIOACTIVE WASTES;    SIMULATION;    STORAGE;    THERMAL ANALYSIS;    THERMODYNAMIC PROPERTIES;    TRANSPORT;    UNDERGROUND STORAGE;    WASTES;   

DOI  :  10.2172/1049041
RP-ID  :  FCRD-UFD-2012-000142
PID  :  OSTI ID: 1049041
Others  :  TRN: US1204389

学科分类：核能源与工程

美国|英语

来源: SciTech Connect

【 摘 要 】

A benchmarking effort was conducted to determine the accuracy of a new analytic generic geology thermal repository model developed at LLNL relative to a more traditional, numerical, lumped parameter technique. The fast-running analytical thermal transport model assumes uniform thermal properties throughout a homogenous storage medium. Arrays of time-dependent heat sources are included geometrically as arrays of line segments and points. The solver uses a source-based linear superposition of closed form analytical functions from each contributing point or line to arrive at an estimate of the thermal evolution of a generic geologic repository. Temperature rise throughout the storage medium is computed as a linear superposition of temperature rises. It is modeled using the MathCAD mathematical engine and is parameterized to allow myriad gridded repository geometries and geologic characteristics [4]. It was anticipated that the accuracy and utility of the temperature field calculated with the LLNL analytical model would provide an accurate 'birds-eye' view in regions that are many tunnel radii away from actual storage units; i.e., at distances where tunnels and individual storage units could realistically be approximated as physical lines or points. However, geometrically explicit storage units, waste packages, tunnel walls and close-in rock are not included in the MathCAD model. The present benchmarking effort therefore focuses on the ability of the analytical model to accurately represent the close-in temperature field. Specifically, close-in temperatures computed with the LLNL MathCAD model were benchmarked against temperatures computed using geometrically-explicit lumped-parameter, repository thermal modeling technique developed over several years at ANL using the SINDAG thermal modeling code [5]. Application of this numerical modeling technique to underground storage of heat generating nuclear waste streams within the proposed YMR Site has been widely reported [6]. New SINDAG thermal models presented here share this same basic modeling approach.