The prediction of the ZnNi thickness and Ni % of ZnNi alloy electroplating using a machine learning method

Abstract

ABSTRACT ZnNi alloy coating is commonly used to enhance the corrosion resistance of steel. The percentage of Ni should be maintained between 12% and 14% in the coating for best corrosion performance. The response surface design (RSD), polynomial regression (PR), support vector regression (SVR), XGBoost regression (XGB), K-nearest neighbours regression and Gaussian process regression (GP) algorithms have been used to predict the ZnNi alloy coating thickness and Ni % amount in the coating. As statistical indices mean square error (MSE) and correlation coefficient (R 2) were used to compare the models. The results of the analysis show that the XGB algorithm gives the best estimation for both ZnNi thickness and Ni%. A high correlation was observed between the predicted values and experimental results. R 2 values of 0.87 and 0.81 were acquired for ZnNi thickness and Ni %, respectively, using the XGB algorithm. This study has proved that the machine learning algorithm is a promising method to predict the ZnNi coating thickness and Ni % in the alloy based on the composition of the ZnNi electroplating bath.