Physics-informed learning for thermophysical field reconstruction and parameter measurement in a nano-porous insulator's heat transfer problem

Hao-Qiang Pang,Xia Shao,Zi-Tong Zhang,Xin Xie,Ming-Yang Dai,Jiang-Feng Guo,Yan-Bo Zhang,Tian-Yuan Liu,Yan-Feng Gao

DOI: https://doi.org/10.1016/j.icheatmasstransfer.2023.107045

IF: 6.782

2023-10-01

International Communications in Heat and Mass Transfer

Abstract:To deeply explore aerogel's insulation performance, we proposed a thermophysical field reconstruction based on physics-informed learning with limited parameters measurement. It's the first time that the model has been utilized to make high-precision measurements of thermophysical parameters in heat transfer problem . In this work, the monolithic aerogel's center temperature response is measured, and the thermal conductivity is extracted at small (< 15 K), medium and large (≥ 400 K) temperature differences, respectively. By leveraging the physics-informed learning method, the temperature field can be reconstructed and thermophysical parameters are precisely identified collaboratively, yielding the majority error within 5% and 1%, respectively. Additionally, our method solves the inverse problem of nonlinear heat transfer. The heat conduction and the overall transport attenuation coefficient of thermal radiation are precisely identified at large temperature differences (≥ 400 K), and the majority of errors are <2% and 4%, respectively. Moreover, the opacifier-doped aerogel displays the lowest thermal radiation contribution of 6.36% at temperature differences (≈ 700 K), whereas pure aerogel shows the highest contribution of 35.40%. Notably, the temperature field exhibits significant nonlinearity currently. The proposed study of experiments and model provides a novelty method to make high-precision measurements, which make up the gap of the classical numerical methods.

thermodynamics,mechanics