Assessment of Porosity Defects in Ingot Using Machine Learning Methods during Electro Slag Remelting Process

The porosity defects in the ingot, which are caused by moisture absorption slag during electroslag remelting process, deserve researcher’s attention summer wet season. prediction of weight gain is critical for developing baking and scheduling strategies can assist workshop managers making informed decisions industrial production electro remelting. under conditions different air humidity, experimental time, particle size, CaO content investigated experiments. purpose this study to predict rate using observed data machine learning (ML) models. observation dataset includes features growth, serve as independent dependent variables, respectively, ML Four models: linear regression, support vector random forest multi-layer perceptron, were employed study. Additionally, parameters models selected 5-fold cross-validation. Support regression outperformed other three terms root-mean-square errors, mean squared coefficients determination. Thus, ML-based model a viable significant method forecasting rate, whereas produce results that competitive satisfying. show increases with increase slag. defect ingot ESR process often appears when exceeds 0.02%. Considering saving electric energy, complexity on-site scheduling, 4 h T3 (CaF2:CaO:Al2O3:MgO = 37:28:30:5) H13 steel winter, T2 48:17:30:5) summer.