Simultaneous reconstruction of temperature-dependent optical and thermophysical parameters of insulation material by the GA-SQP/SQP technique

Abstract

A combination genetic algorithm (GA) with sequential quadratic programming (SQP) algorithm is proposed to simultaneously identify the temperature-dependent thermal conductivity (κc(T)) and extinction coefficient (β(T)) of insulation material without any prior information of parameter distribution function. The GA is applied to obtain the initial values of κc(T) and β(T) because of its lesser dependence on initial guess, then the reconstruction results from GA are used as the initial values of SQP to reckon an optimal solution of κc(T) and β(T) and finally the established κc(T) in previous step are adopted as known parameters to reconstruct the β(T) by SQP algorithm, which is called a GA-SQP/SQP technique. The retrieval results from numerically generated temperature demonstrate that the GA-SQP/SQP technique can significantly improve the accuracy of estimated κc(T) and β(T) compared with the other traditional optimization algorithms. Corresponding experiment is performed to estimate the κc(T) and β(T) of insulation material at high temperatures by the GA-SQP/SQP technique and the results are validated.