Dilution Optimization of Laser Cladding Assisted by Pulsed Current Based on Genetic Algorithm and Neural Network

Abstract

The dilution ratio of the Ni coating prepared by the laser cladding under the assistance of the follow-up feeding pulsed current was optimized by combining back propagation (BP) neural network and genetic algorithm. The model was trained according to the results of the 6-factor 3-level orthogonal experiments. A BP genetic neural network forecast model between cladding parameters (laser power, scanning speed, powder feeding rate, pulsed current, pulse frequency and pulse width) and dilution ratio of coating was constructed. On this basis, technological parameters under the target dilution ratio of the coating were optimized by a genetic algorithm. Results demonstrated that the predicted results of the model are very close to the experimental results in term of dilution ratio of the coating, with a relative error no higher than 2.63%. This demonstrates that the model is reliable and effective. The optimal technological parameters are gained when the dilution ratio of the coating is 17.5%, including laser power=1926.3 W, laser scanning speed =·s-1, powder feeding rate= ·min-1, average pulsed current =, pulse frequency=445.6 Hz, pulse width= 108.4 μs.