RAPID OPTIMIZATION OF LASER QUENCHING PROCESS BASED ON BP NEURAL NETWORK

The rapid optimization of laser quenching process parameters were studied. First of all, the experiment is carried out by orthogonal method, which achieves the purpose of finding the influence law of process parameters with a small number of samples. Then, the neural network modeling is used, the process parameters and the experimental results are functionally fitted, and the model is continuously revised by reducing the feedback error, and finally the prediction model with the smallest error is obtained. Finally, the genetic algorithm is used to quickly search for optimization, and based on the prediction model, population selection, crossover, mutation and iteration are used to obtain the optimal fitness and corresponding variable value.