Automated FEM-Based Determination of Thermal Diffusivity and Temperature Profile for Rockfall Prediction
摘要
A novel method for the determination of thermal diffusivity and the generation of a continuous temperature profile of a rock formation based on temperature measurements at several depths is presented. The methodology combines finite element modeling and optimization techniques to accurately simulate the heat transfer processes in the rock. The optimization procedure is divided into three successive steps using three different finite element models. The first model uses optimization to calculate the thermal diffusivity by minimizing the difference between simulated and measured temperatures. The second model uses optimization to determine the unknown surface temperatures of the rock. The third model uses finite element analysis to create a continuous temperature profile over the entire rock depth. It uses the optimized thermal diffusivity and surface temperatures determined by the first two models. The optimization procedures use gradient-based methods and consider sensitivity analysis to determine the derivatives of the objective functions. The methodology is able to accurately determine the thermal diffusivity and create a continuous temperature profile even if the measured data has inaccuracies.