Predictive wear analysis of SS316L fabricated by direct energy deposition using machine learning techniques

Many critical components like turbine blades, and high-speed trains exposed to particle wear in the environment can be produced using metal additive manufacturing. This paper focuses on studying the particle erosion behavior of 316L stainless steel components built by Direct Energy Deposition (DED) and subsequent precision machining using wire Electrical Discharge machining (wire EDM). Focus on critical factors wire EDM speed, current, and wire diameter, the experiment is conducted using the L9 orthogonal array generated and Minitab is used for statistical analysis. This statistical analysis aims to improve the surface finish of the machined component. Further, the study is extended to analyze the material wear resistance using a slurry erosion wear test on specimens cut by wire EDM. Initially wear analysis was performed using Minitab to find the influential parameter on wear rate and then data analysis techniques such as Linear Regression, K Nearest Neighbor Algorithm, and Artificial Neural Network were used to create a model that predicts the wear rate accurately which may reduce lot of experimentation time and cost. This paper successfully analyzed the particle erosion behavior of 316L Stainless steel parts manufactured through the DED technique and refined using the wire EDM machining process. The surface roughness of the samples is improved by performing the statistical analysis using Minitab software. The developed machine learning models demonstrated the potential in terms of reduction of cost and experimentation time by the accurate wear rate prediction.