Deep learning or interpolation for inverse modelling of heat and fluid flow problems?

Purpose The purpose of this study is to compare interpolation algorithms and deep neural networks for inverse transfer problems with linear nonlinear behaviour. Design/methodology/approach A series runs were conducted a canonical test problem. These used as databases or “learning sets” both networks. second set was the prediction accuracy approaches. Findings results indicate that outperform in heat conduction, while reverse true conduction problems. For convection problems, methods offer similar levels accuracy. Originality/value This first time such comparison has been made.