关键词: 碳纤维复合材料, 超声振动钻孔, Hashin-Puck失效准则, 分层因子, 神经网络, MOEA/D遗传算法

Abstract: Carbon fiber reinforced plastics were widely used in aerospace field because of its superior performance,but it was easy to produce burr,delamination and tear when drilling,which seriously affects its development and application.In order to accurately analyze and predict the delamination defects during CFRP drilling,this paper firstly established a CFRP drilling finite element model by using the Hashin-Puck failure criterion of the interlayer material and the Cohesive viscous layer damage failure mechanism.The comparison with the experimental results shows that the maximum errors of the simulation values of thrust force and delamination factor are 8.68% and 5.54%,respectively.Secondly,a genetic optimization neural network prediction model of thrust force and delamination factor was constructed with the 3.60% average error of thrust force and the 7.06% maximum relative error with respect of experimental results.Finally,a multi-objective optimization model with the minimum quality index and efficiency index was established,and MOEA/D genetic algorithm was used to solve the model,which provides a practical theoretical basis for the reasonable se-lection of CFRP drilling process parameters.

Key words: carbon fiber reinforced plastic, ultrasonic vibration drilling, Hashin-Puck failure criterion, delamination factor, neural network, MOEA/D genetic algorithm