Where ThermoMesh meets ThermoNet: A machine learning based sensor for heat source localization and peak temperature estimation

Abstract In this paper, we propose a new thermal sensor (ThermoMesh) for experimental heat source localization and peak temperature estimation (HSL&PTE) enabled by Machine Learning algorithms (ThermoNet). The mathematical model of the ThermoMesh sensor is first derived and experimentally validated. Its use for HSL&PTE of a single heat source is then numerically demonstrated with location accuracy of 99% and RMS temperature error of 1.2%. Complementing existing thermal imaging techniques based on radiative heat transfer, the ThermoMesh plus ThermoNet framework sheds new light on high-speed high-resolution heat source sensing via conductive heat transfer.