Anticipating Surface Roughness in Inconel Turning: Harnessing the Power of Neural Networks and Bayesian Regularization

In recent years, hybrid machining techniques have gained popularity in processing materials that are notoriously tough to work with. This is due to the inherent benefits that hybrid machining techniques hold over individual machining methods. These operations need the coordinated use of a wide variety of machining equipment in a variety of configurations. This research uses a neural network-based Bayesian regularisation technique to investigate the possibility of transforming the problematic material Inconel into a superalloy. The depth of cut, spindle speed, and feed rate are the control elements examined in this study. The performance metric is the primary focus of the average surface roughness (Ra). The experimental testing and the coefficient of determination were used to evaluate the capabilities of the suggested neural network in modeling and forecasting. The findings indicate that there is only a 1.25 percent discrepancy, on average, between the predictions made by neural network-based modeling and the values that were obtained. The surface roughness (Ra) values discovered in the experimental data are identical to those predicted by these equations.