Numerical computing of induction motor model through Levenberg-Marquardt method

The aim of study is to develop a progressive stochastic numerical solver by means of neural networks through Levenberg-Marquardt backpropagation to examine the dynamics of the higher order nonlinear boundary value problem arising in induction motor model. The fifth order ordinary differential equation dataset is constructed by using Hermite numerical solver and determined the target parameters for continuous mapping of neural networks. The training, testing and validation processes are employed in neural network tool through backpropagation of Levenberg-Marquardt method to estimate the numerical solution of nonlinear induction motor model by constructing different scenarios through varying involved continuous functions on the specific real valued interval. Validation and verification of neural network model is endorsed to calculate the solution of nonlinear induction motor model on the cost of achieved accuracy by histograms error, mean squared error and regression studies.