Computational analysis for thermal and mass transfer in pumping flow of Ellis fluid along with solid particles: Morlet wavelet neural networks approach

Abstract

The current study conducts a parametric analysis of thermal transport and mass transfer during peristaltic circulation within a symmetric channel, utilizing numerical methods and artificial intelligence techniques. This study examines a multi-phase flow involving a non-Newtonian Ellis fluid with suspended solid particles. Additionally, the research explores the influences of chemical reactions along with the electroosmotic phenomena on the flow dynamics. Using the approximation of the lubrication hypothesis, the flow problem is modeled with the assumptions of lengthy wavelength and tiny Reynolds coefficient. The combined ordinary differential equations of the fluid and particulate phases are solved by the Morlet-wavelet neural networks integrated with particle swarm optimization and neural networks algorithm. To validate the intelligence-based especially mean square error results are further compared using the numerical NDSolve and physics informed neural networks with Adam optimizer as a reference solution. The values of Morlet-wavelet neural networks with particle swarm, Morlet-wavelet neural networks with neural networks algorithm and hybrid of Morlet-wavelet with particle swarm and neural networks algorithms based fitness functions are ranging from 10 − 03 − 10 − 04 , 10 − 04 − 10 − 05 and 10 − 05 − 10 − 07 respectively. The absolute error between the proposed three approaches and reference solution between the 10 − 02 − 10 − 04 , 10 − 04 − 10 − 05 and 10 − 05 − 10 − 07 . The mean and median values of mean squared error over 70 independent runs are ranging between 10 − 06 − 10 − 07 and 10 − 07 − 10 − 09 for 12 distinct cases visualized in through figures. The velocity function enhances with the gain in the values of electroosmotic factor and Ellis fluid parameter. Further, statistical metrics calculated over multiple runs to check the convergence and efficiency of the proposed approaches. This study can be used in chemical processing, biomedical engineering, and environmental engineering industries, where the transportation of non-Newtonian fluids with solid particles is important. It optimizes heat and mass transfer in reactors, drug delivery, water treatment, and slurry transport, utilizing AI-based simulations for enhanced efficiency and performance.

Topics & Concepts