Intelligent computing technique to analyze the two-phase flow of dusty trihybrid nanofluid with Cattaneo-Christov heat flux model using Levenberg-Marquardt Neural-Networks

This study examines the characteristics of activation energy on the two-phase flow of a tri-hybrid nanofluid with variable thermal conductivity, viscous dissipation, and NHCMBM using a stochastic-based Levenberg-Marquardt backpropagated neural network (LMB-NN). The Darcy Forchheimer porous media characteristics is included in the momentum equation. The model of Cattaneo-Christov heat flux is employed to investigate the significance of heat transmission. The sigmoid function is utilized as the activation function in the hidden layer along with 20 neurons. Three different scenarios are covered by the suggested Levenberg-Marquardt scheme, which uses 15 % of the generated dataset for testing and training and 70 % of the data for network training. To confirm that the suggested method for solving the NHCMBM model is valid, comparisons between the outcomes of the LMB-NN approach and reference solutions are given. The efficacy of the method is confirmed by regression analysis, state transitions, MSE, correlation, and error histograms; nonetheless, its accuracy is impacted by absolute error. As the Marangoni convection factor increased, the results showed that the flow field of the dust and fluid phases increased while the solutal and thermal fields in both phases dropped.