Abstract: Carbon fiber reinforced plastics are widely used in aerospace field because of its superior performance,but it is easy to produce burr,delamination and tear when drilling,which seriously affects its development and application.To accurately analyze and predict the delamination defects during Carbon Fiber Reinforced Plastic (CFRP) drilling,a CFRP drilling finite element model was established 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 showed that the maximum errors of the simulation values of thrust force and delamination factor were 8.68% and 5.54% respectively.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 provided a practical theoretical basis for the reasonable selection 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

摘要： 碳纤维复合材料(CFRP)因其优越性能被广泛应用于航空航天领域,但在钻孔加工时,容易产生毛刺、分层和撕裂等缺陷,严重影响其发展应用。为了准确分析与预测CFRP钻孔过程中的分层缺陷,通过层内材料Hashin-Puck失效准则和Cohesive黏性层损伤失效机制,建立了CFRP超声振动钻孔有限元模型。与实验结果对比分析可知,轴向力和分层因子仿真值的最大误差分别为8.68%和5.54%。其次,构建了轴向力和分层因子的遗传优化神经网络预测模型,实验结果表明轴向力平均误差为3.60%,最大相对误差为7.06%。最后,建立了质量指标和效率指标为最小的多目标优化模型,并用MOEA/D遗传算法对模型求解,为CFRP钻孔工艺参数的合理选择提供切实可行的理论基础。

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