Application of machine learning in the investigation of entropy generation incurred during electroosmotic flows of Casson nanofluid

Abstract

The present research article delves into the application of machine learning techniques in the study of entropy generation arising from the electro‐MHD Casson nanofluid flow in a symmetric microchannel. The inertial effect on fluid flow through a non‐Darcy porous medium is considered where we have taken into account the velocity, thermal, and concentration slips at boundaries. The resulting system of PDEs is converted to system of ODEs using appropriate wave frame transformations. The metamorphosed system is solved using bvp4c routine of MATLAB. The impact of several factors represented by respective parameters on velocity, nanofluid temperature, fraction of nanoparticles, entropy generation, and Bejan number are shown through graphs. This study also employs a machine learning technique in which Particle Swarm Optimization (PSO) algorithm has been used in conjunction with an Artificial Neural Network (ANN) to develop a predictive model capable of optimizing entropy generation within this specific fluidic context. Such an analysis could be useful in determining the optimum entropy generation in pulsatile transport of physiological fluid through intestine. With the use of above mentioned machine learning technique, the considered model offers the minimum entropy generation for specific values of the electroosmotic parameter, dimensionless temperature difference, and velocity slip parameter. It is observed that the flow with minimum electroosmosis effect and maintained at a specific velocity slip gives the optimal entropy generation.