Artificial neural network (ANN) modelling for the thermal performance of bio fluids

Abstract

By using the concept of Artificial Neural Network modelling, an attempt is made to predict the Thermal characteristics of Bio fluids in this study. The ANN model was trained and tested to compare with five different learning algorithms namely One Step Secant(OSS), Conjugate gradient backpropagation (CGP), Conjugate gradient backpropagation with Fletcher-Reeves updates(CGF), Scaled Conjugate Gradient(SCG)and Levenberg-Marquardt(LM). The chosen input parameters oil types, percentage, and stirrer speed, while the model outputs focus upon thermal performances like thermal conductivity and absolute viscosity. According to the findings, the suggested ANN model may be utilized to accurately predict the thermal performance of bio fluids. Because the results obtained are well within a 95percent accuracy criteria, LM was found to be the best output of all the training methods utilised in all stages of ANN modelling. As a result, the ANN model constructed utilising the LM approach is one of the most accurate for projecting bio fluid thermal performance.