Heat transport in entropy-optimized flow of viscoelastic fluid due to Riga plate: analysis of artificial neural network

In this paper, we focus to design intelligence numerical computing through artificial neural networks (ANNs) which are backpropagated with Levenberg–Marquard technique (NN-BLMT) for the theoretical approach of physical aspects of heat generation in second-grade fluid (PA-HG-SGF) due to Riga plate. The Riga plate (RP) is known as the physical interaction stimulus, which consists of electrodes and enduring electromagnets that are located on plane surface. The purpose of NN-BLMT is to learn the weight of neural networks on the basis of optimization of the fitness value based on mean-square error between the proposed result and reference numerical solution. The innovation and reliability of NN-BLMT used will be greatly better as compared to traditional numerical techniques that are used to solve commercial and industrial problems. NN-BLMT is fast and easy to apply on nonlinear problems and get best results. The original model PA-HG-SGF in term of PDEs is first converted into system of nonlinear ODEs through suitable transformation and then numerically solved. Through Adam numerical technique (AMT) in Mathematica software, a dataset for PA-HG-SGF is attained for different scenarios of PA-HG-SGF by variation of second-grade parameter, heat-generation parameter, Hartman number, thermal lamination parameter, thermal composure parameter and Prandtl number. Expected solutions are described for PA-HG-SGF through the NN-BLMT testing, training, and validation process. In addition, NN-BLMT relative studies and performance analysis are validated by histogram studies, regression analysis and MSE and then analyzed PA-HG-SGF.