On the Prediction and Optimisation of Processing Parameters in Directed Energy Deposition of SS316L via Finite Element Simulation and Machine Learning.

In the current study, the integration of finite element simulation and machine learning is used to find the optimal combination of processing parameters in the directed energy deposition of SS316L. To achieve this, the FE simulation was validated against previously implemented research, and a series of simulations were conducted. Three inputs, namely laser power, scanning speed, and laser beam radius, and two outputs, namely residual stress and displacement, were considered. To run the machine learning model, artificial neural networks and a non-dominated sorting genetic algorithm were applied to determine the optimal combination of processing parameters. In addition, the current study underscores the novelty of combining FE simulation and machine learning methods, which provides enhanced precision and efficiency in controlling residual stress and displacement (geometrical deviation) in the Directed Energy Deposition (DED) process. Then, the results obtained via machine learning were validated with confirmatory tests and reported. The findings offer a practical solution for process parameter optimization, contributing to the progression of additive manufacturing technologies.