Session: K20-W2: WORKSHOP ON INVERSE PROBLEMS AND PARAMETER ESTIMATION IN HEAT TRANSFER II

Paper Number: 138688

138688 - Inverse Heat Conduction Problem: Filter Form Solution and Machine Learning Techniques 

Abstract: 

The use of digital filter form solution for inverse heat conduction problems (IHCPs) allows fast and accurate calculation of surface heat flux using internal temperature measurements which can facilitate monitoring and control of the surface condition in various industrial applications. In this presentation, the concept of developing filter matrices for expressing various IHCP solution techniques in the form of filter coefficients will be explored. The characteristics of the filter matrix associated with each IHCP solution technique will be assessed. The application of the filter form Tikhonov regularization method for near real-time heat flux estimation in both linear and non-linear IHCPs will be discussed. Examples involving temperature-dependent material properties, multi-layer medium, medium with moving boundary, and multi-dimensional domains will be reviewed through several case studies. In addition, the application of machine-learning algorithms for solving IHCPs will be discussed as efficient and strong tools for addressing complex problems. Particularly, the use of selected architectures of artificial neural networks (ANNs) for developing a filter form-inspired solution for IHCPs in a one-dimensional domain both with and without temperature-dependent material properties and with and without moving boundary will be discussed. The use of MATLAB for implementing these solutions is elucidated through the review of several examples.

Presenting Author: Hamidreza Najafi Florida Institute of Technology

Presenting Author Biography: Hamidreza Najafi is an Associate Professor of Mechanical Engineering and the director of the Heat Transfer Lab at the Florida Institute of Technology. He conducts research and teaches courses in the areas of thermal sciences. His research works have been focused on designing and optimizing thermal/energy systems, computational heat transfer, and inverse heat conduction problems.

Authors:

Hamidreza Najafi Florida Institute of Technology