Parameter estimation of in-situ thermal response tests for borehole ground heat exchangers

Abstract

This paper discusses some aspects of parameters estimation used in in-situ thermal response tests of ground heat exchangers, including sensitivity analysis and comparison of iterative minimization algorithms. First, several sensitivity coefficients of parameters and uncertainties occurring in in-situ tests are examined in depth. Starting from an analytical heat transfer model of borehole ground heat exchangers, sensitivity coefficients for single U-tube ground heat exchangers are of analytical forms. The sensitivity analysis provides some general and new guidelines for mitigating the influence of testing uncertainties. Next, Monte Carlo simulation is performed to evaluate reliability of two powerful minimization algorithms, the Levenberg–Marquardt method and a trust region method subject to bounds. The Monte Carlo simulation shows that if thermal diffusivity of soil is an estimated parameter, the Levenberg–Marquardt method may result in unreliable results, and its performance depends strongly on the quality of the initial guessed values. In contrast, the interior trust region method considered in this paper can produce more reliable results, and its performance depends on the range of input parameters bounds. Finally, feasibility of the highlighted methodology is illustrated by applying it to two real in-situ thermal response tests.