Evolution of Regression and Neural Network Models on Wire Electrical Discharge Machining of Nickel Based Superalloy

Abstract

In addition to traditional methods, there are also non-traditional techniques that can be used to overcome the challenges of conventional metal working. One such technique is wire electrical discharge (WEDM). This type of advanced manufacturing process involves making complex shapes using materials. Utilizing intelligent tools can help a company meet its goals. Nickel is a hard metal to machine for various applications such as nuclear, automobile and aerospace. Due its high thermal conductivity and strength, traditional methods are not ideal when it comes to producing components using this material. This paper aims to provide a comprehensive analysis of the various steps in the development of a neural network model for the manufacturing of Inconel 625 alloy which is used for specific applications such as exhaust couplings in sports motor vehicle engines. The study was conducted using a combination of computational and experimental methods. It was then used to develop an index that measures the correlation between various process variables. After analyzing the results of the study, a set of equations was then created to forecast the future performance of a manufacturing process. The parameters used in the study were then updated to create a new multi-performance index prediction model. The results of the study were then used to develop an equation set that can be utilized to forecast the future performance. A comparative analysis was performed between the two sets of equations. After analyzing the results, the neural network model was found to perform better than the multi-performance index.