ABSTRACT
The inversion of thermophysical parameters is a typical representative of inverse heat conduction problems (IHCP), which has a broad application prospect in scientific and engineering problems such as nondestructive testing, materials processing, geometry optimization, biomedical detection, architectural design, food technology and so on. An integrated workflow of utilizing deep learning techniques to retrieve thermophysical parameters is exhibited exhaustively. The experimental approaches to determine the thermal conductivity can be divided into two categories, namely the steady-state and transient ones. The process of obtaining thermal conductivity by experiment generally requires solving the inverse problem of heat transfer. The Gradient-based Methods have the advantages of high accuracy and fast convergence. However, they are computationally cumbersome, and the optimization results are unstable, which is possible to fall into the local minimum. The chapter introduces the fundamental inverse heat conduction models, including the 2D and 3D cases.
