Development of Ni-B coating through the aid of artificial intelligence

Abstract

Electroless Ni-B coatings with excellent wear resistance have been reported by widely varying bath compositions in different research works. If a database can be made, trained using artificial intelligence (AI) and optimized, a readily available model with correlation of bath parameters and coating properties would be available for the scientific community and industries. In this work, an effort has been made to list the composition reported in some works, train them using artificial neural network (ANN) and find an optimal condition leading to higher deposition and hardness of the coatings using non-dominated sorting genetic algorithm (NSGA). The predicted bath has good correlation with experimental results (R² ∼ 1). The usual nodular morphology was seen with amorphous structure. A high as-deposited hardness of 950–1075 HV_0.1, ∼10 µm/h deposition rate and first critical load of failure (L_c) > 24 N was detected. Thus, this work serves as a preliminary model for predicting coating with enhanced quality without having to perform numerous experiments thereby saving cost and time with substantial accuracy.