This paper provides a comparison of multiple linear regression analysis with artificial neural network (ANN) and adaptive neuro fuzzy inference system (ANFIS), for predicting the emitted radiations from a flat plate heat sink. HFSS simulations were designed using Ansoft version 12 for the L27 orthogonal array and optimised using Taguchi design of experiments method. The heat sink geometry factors considered for the L27 (six factors, three levels) design are length, width, fin height, base height, fin thickness and number of fins and the response studied is the emitted radiation from the heat sink. A meta model is developed based on a multiple linear regression method using HFSS simulations. Also, the results of L27 orthogonal array were used to train the artificial neural network and the ANFIS-based intelligent networks. The accuracy of results for the prediction of emitted radiations using multiple linear regressions, ANN and ANFIS were compared with HFSS simulations. From the results, it is found that the ANFIS outperforms ANN and regression models for the prediction of the emitted radiations from the heat sink.Keywords: emitted radiation; heat sink; multiple linear regression analysis; MLRA; artificial neural networks; ANN; adaptive neuro fuzzy inference system; ANFIS.Reference to this paper should be made as follows: Manivannan, S., Devi, S.P.
Biographical notes: S. Manivannan is currently pursuing his research in AnnaUniversity in the area of thermal and EMI management of electronic packages. He has obtained his BE and ME from Madras University. He has published over seven papers in referred international journals.S. Prasanna Devi is pursuing her research in Anna University. She obtained her BE and ME from the Department of Computer Science from the same university. She has published over ten papers in refereed international journals in the area of operations research.