A novel reliable deep neural network scheme to investigate thermal properties of engine oil using hybrid nanofluid with irreversibility analysis

This work explores the irreversible behavior of MHD-HNF flow containing nanoparticles in engine oil through a porous elastic surface by considering the cumulative impacts of viscous heating and variable thermal source. The flow behavior of HNF is describe by applying the ANNs on governing model. Due to the electrical conductivity of HNF, an induced magnetic field arises within the flow field. The governing PDE’s are reduced to ODE’s by means of similarity transformations. The entropy generation characteristics influenced by the major parameters are comparatively investigated for HNF in Cartesian geometry by using ANNs. The graphical outputs are obtained by varying different parameters such as; , , , & , for discussions of numerical results on , and profile. By using ANNs total 101 samples are attained from matrix data. The total samples are divides, 81 for training data, 10 for testing and 10 for validation. For each scenario the graphical outcomes are attained for EHA, FFO, TSD, RG-A, soltuion graph, MSE and AE. Observed that profile drops with rising values of . The profile continuously declines when we increase the values of . Moreover, profile increases due to increment in . The profile tends to rise when we increase difference of , and . The MSE penalties (testing, training, validation) MHD flow on a stretching porous surface lies between The gradients values lie around for MHD flow. EHA of MHD flow is recorded around . The AE of MHD flow by using LMBNNs is noted between for all six scenarios.