A Neural Network Based Method for Estimation of Heat Generation from a Teflon Cylinder

The paper reports the estimation of volumetric heat generation (qv) from a Teflon cylinder. An aluminum heater, which acts as a heat source, is placed at the center of the Teflon cylinder. The problem under consideration is modeled as a three dimensional steady state conjugate heat transfer from the Teflon cylinder. The model is created and simulations are performed using ANSYS FLUENT to obtain temperature data for the known heat generation qv. The numerical model developed using ANSYS acts as a forward model. The inverse model used in this work is Artificial Neural Network (ANN). Estimation of heat generation is carried out by minimizing the error between the simulated temperature and the experimental/surrogated temperature. The efficacy of the ANN method is explored for the estimation of unknown heat generation as both forward model and inverse model. The concept of Asymptotic Computational Fluid Dynamics (ACFD) is introduced as a fast forward model which is obtained by performing CFD simulations. The unknown heat generation is estimated for the surrogated data using ANN. In order to mimic experiments, noise is added to the surrogated data and estimation of heat generation is also carried out for the perturbed/noise added temperature data