Applying a Neural Network-Based Machine Learning to Laser-Welded Spark Plasma Sintered Steel: Predicting Vickers Micro-Hardness

Abstract

This paper presents an artificial neural network (ANN) approach to the estimation of the Vickers hardness parameter at the weld zone of laser-welded sintered duplex stainless steel. The sintered welded stainless-steel hardness is primarily determined by the sintering conditions and laser welding processing parameters. In the current investigation, the process parameters for both the sintering and welding processes were trained by ANNs machine learning (ML) model using a TensorFlow framework for the microhardness predictions of laser-welded sintered duplex stainless steel (DSS 2507 grade). A neural network is trained using a thorough dataset. The mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), and R2 for the train and test data were calculated. The predicted values were in good agreement with the measured hardness values. Based on the results obtained, the ANN method can be effectively used to predict the mechanical properties of materials.