Artificial Neural Network Prediction on negative and positive activation energy of magnetohydrodynamic nanofluid flow with multiple slips

In this study, the non-Newtonian magnetohydrodynamic stagnation point nanofluid flow with negative and positive activation energy and multiple slip boundary conditions over a slippery surface has been investigated and an artificial neural network model has been developed to predict the Sherwood number (mass transfer rate). This study can be helpful to identify the optimal conditions for heat and mass transfer enhancement in magnetohydrodynamic nanofluid flow with multiple slips. Artificial neural network model can provide real-time predictions, so it can play an important role in process control, optimization and reducing computational cost. All of the previous study has focused on positive activation energy but in our current study we have considered the negative and positive activation energy together. Thus, our study is unique. Through the use of similarity transformations, the system of non-linear partial differential equations that represent the fluid flow has been converted into a system of non-linear ordinary differential equations and then solved numerically with the help of Spectral Quasi-linearization Method. It has been seen that, velocity increases and temperature, concentration decreases for the increasing values of velocity slip parameter. Concentration of the fluid decreases for the rising values of thermal slip parameter and concentration slip parameter. For the rising values of positive activation energy, concentration of the fluid first decreases then increases and opposite behaviour has been seen for the rising values of negative activation energy. It is also seen that, for the rising values of activation energy from 0.5 to 2.5, Skin friction coefficient and Sherwood number are increased by 0.65 % and 4.64 % respectively while Nusselt number is decreased by 3.65 %. When activation energy goes from −0.5−2.5, Skin friction coefficient and Sherwood number are decreased by 1.74 % and 17.47 % respectively while Nusselt number is increased by 12.40 %. This investigation can take a key role in the field of biochemical engineering, medical and thermal management such as heat exchangers, cooling systems, tissue engineering, protein production etc. Another important matter to discuss here that we have used feed-forward back-propagation multilayer perceptron artificial neural network with Levenberg-Marquard algorithm as the training algorithm to predict the Sherwood number for both activation energy values 1 and -1. We have analysed Mean Square Error, Root Mean Square Error, Coefficient of correlation, Mean and Standard deviation of errors to justify the accuracy of the designed artificial neural network model. From our observations, we can conclude that artificial neural network model is an ideal tool which can be employed for the prediction of magnetohydrodynamic nanofluid flow behaviours.