Research On Transient Non-linear Inverse Heat Conduction Problems Based On A Genetic Algorithm

The inverse heat conduction problem (IHCP) has a wide range of significant applications in many engineering problems, such as the aerospace, the metallurgy, the chemical industry, the nondestructive testing and so on. Although many achievements and new findings on IHCPs are obtained, there are still a lot of weak points, such as the transient nonlinear inverse heat conduction problems, irregular or complex structures and so on. This thesis presents a new method based on genetic algorithms and commercial finite element software, using FORTRAN language programming, to investigate transient nonlinear inverse heat conduction problem.This thesis includes the following main parts:The effectiveness and the accuracy of the genetic algorithm in the inverse problem is validated. The one-dimensional model of a transient nonlinear heat conduction problem is built, and a finite difference scheme is derived to obtain its numerical solution. Then boundary conditions and temperature-dependent thermal conductivities are inverted. The effects of the measurement error on inversion results are analyzed. Two and three dimensional transient nonlinear heat conduction models are built using the finite element software. Boundary conditions and temperature-dependent thermal conductivities are recovered, using the genetic algorithm.Numerical examples show that the proposed inversion algorithm is accurate and robust for solving transient nonlinear inverse heat conduction problems. For simple problems, one can use the finite difference method and the genetic algorithm to identify the thermophysical parameters and boundary conditions. For complex structures, one can use the ABAQUS and the genetic algorithm to identify the thermophysical parameters and boundary conditions. This work provides a theoretical basis for identifying the heat flux parameters and the thermophysical parameters for the thermal protection system (TPS) materials in the aerospace field.