基于遗传算法的碳纤维增强树脂复合材料层合板单搭胶接结构的多目标优化

基于遗传算法对碳纤维增强树脂复合材料(CFRP)层合板单搭胶接结构进行了多目标优化，以提高其结构性能。首先，通过三维Hashin准则和三角形内聚力模型建立三维有限元模型来预测CFRP层内损伤过程、层间失效和胶层损伤过程，并通过试验验证其有效性。其次，利用拉丁超立方抽样(LHS)方法和二次多项式响应面法(RSM)，基于搭接长度、胶层厚度和被胶接件宽度等胶接参数建立以拉伸强度和剪切强度为目标函数的多目标优化代理模型。最后，基于遗传算法(GA)对拉伸强度和剪切强度代理模型进行优化，得出一组Pareto解集，并基于理想解排序方法(TOPSIS)对Pareto非劣解集进行折中处理，得到最好的胶接参数设计方案。结果表明:CFRP层合板单搭胶接结构的数值模拟结果与试验结果相比具有很高的吻合度，验证了有限元方法的可靠性；CFRP层合板单搭胶接结构的拉伸强度和剪切强度与搭接长度、胶层厚度和被胶接件宽度具有显著的关联性；二次响应面代理模型结果与数值模拟结果相比误差均小于2.3%；与常规的单搭胶接结构方案进行对比，搭接拉伸强度和剪切强度分别提高了2.65%和17.24%。 

Multi-objective optimization of adhesively bonded single-lap joints of carbon fiber reinforced polymer laminates based on genetic algorithm

To improve its structural performance, the multi-objective optimization of adhesively bonded single-lap joints of carbon fiber reinforced polymer (CFRP) laminates was carried out based on genetic algorithm. Firstly, finite element (FE) models were constructed using 3D Hashin damage criteria and triangle cohesive zone model (CZM), those well capture the intra-laminar, inter-laminar and adhesive damages during the tensile loads, respectively. And its effectiveness was verified through experiments. Secondly, using the Latin hypercube sampling (LHS) method and polynomial response surface method (RSM), a multi-objective optimization agent model with tensile strength and shear strength as the objective function was established based on lap the length, the adhesive thickness and the width of the bonded parts. Finally, the tensile strength and shear strength agent model was optimized to obtain Pareto solution set based on genetic algorithm (GA), and the Pareto solution set was sequenced by technique for order preference by similarity to ideal solution (TOPSIS) method to obtain the optimized single-bonded joint structure design scheme. The result shows that the experimental measurements of tensile load tests concur with the numerical predictions and validate the FE models. The tensile strength and shear strength of the adhesively bonded single-lap joints of CFRP laminates have a significant correlation with the lap length, the thickness of the adhesive layer and the width of the bonded parts. Compared with the numerical simulation results, the error of the quadratic response surface proxy model results is less than 2.3%. Compared with the conventional single lap bonding structure, the tensile strength and shear strength are increased by 2.65% and 17.24%, respectively.