Tensile Strength Prediction of Radial-additive Friction Stir Repaired Out-of-tolerance Hole by IPSO-BPNN
The improved particle swarm optimization(IPSO)algorithm is used to optimize the back propagation neural network(BPNN)weights and thresholds.The IPSO-BPNN tensile strength prediction model for the radial-additive friction stir repairing joints of out-of-tolerance hole of 2024-T4 aluminum alloy with 3×6×1 three layers topology is established with rotating velocity,plunge speed and dwelling time as input and tensile strength as output.The results show that the optimized IPSO-BPNN prediction model has high prediction accuracy and convergence speed,and the relative prediction error is 1.01%.
万方数据
维普
