Prediction of mechanical properties of welded joints based on GA-BP neural network
Welding process parameters are the main factor affecting the mechanical properties of welded joints,and developing welding processes is a very complex task.Introducing machine learning into the field of welding and using neural network algorithms to predict the mechanical properties of welded joints can assist the formula-tion of welding processes.Due to the high accuracy required for predicting the mechanical properties of welded joints,this paper proposes a method for optimizing BP neural networks through genetic algorithms.The model se-lects the welding process parameters as input and the joint tensile strength as output.The experimental results show that the neural network model optimized by genetic algorithm performs better with an error of±0.25％,which can assist in the design and optimization of welding processes.
prediction of mechanical properties of welded jointsgenetic algorithmBP neural networkmodel optimization
NETL
NSTL
万方数据
