Prediction of Geometric Characteristics of Broad Beam Laser Cladding Layer Based on Neural Network and Genetic Algorithm

Aiming at the problems that the geometric characteristics of broad beam laser cladding layer are affected by many factors and difficult to control, the BP neural network prediction model was established with laser power, scanning speed and powder feeding rate as the input, and the cladding layer width and height as the output. The prediction accuracy was analyzed, and the weights and thresholds of BP neural network prediction model were optimized by genetic algorithm. The results show that the relative error of BP neural network in predicting the geometric characteristics of cladding layer is within 7.434%, and the relative error of GA-BP neural network model in predicting the geometric characteristics of cladding layer is within 5.348%. GA-BP neural network model has high accuracy in predicting the geometric characteristics of broad beam laser cladding layer, which can effectively guide the selection of process parameters of broad beam laser cladding.