Uncertainty study and sensitivity analysis of soil-rock thermal response test

All right reserved. In order to solve the uncertainty in data interpretation involved in soil-rock thermal response test(TRT), a parameter estimation program was developed by employing numerical inversion of cylindrical source model combined with a trust region reflective algorithm, which can avoid being trapped in local optimum and provide underlying model for subsequent uncertainty study.Uncertainties of input parameters, including inlet and outlet temperatures of ground heat exchanger, heat transfer rate and undisturbed ground temperature, etc, were reasonably quantified through literature reviews, theoretical considerations and educated guesses.Then, Monte Carlo stochastic simulation was applied to propagate input parameters uncertainties to output parameters of TRT.A mapping of ground thermal conductivity and borehole thermal resistance was then obtained, and their inherent correlation was presented by a bivariate joint distribution constructed by Copula function.In the end, the influence of variable parameters on the mean temperature of heat transfer fluid was analyzed by Lasso method.The result shows that ground thermal conductivity and borehole thermal resistance are the most influential parameters.