基于改进型支持向量机的胶接强度预测模型  

Prediction model of bonding strength based on improved support vector machine

作　　者：许明阳 殷晨波[1] 陈曦[1] XU Mingyang;YIN Chenbo;CHEN Xi(Institute of Automobile and Construction Machinery, Nanjing Tech University, Nanjing 211816, China)

机构地区：[1]南京工业大学车辆与工程机械研究所,江苏南京211816

出　　处：《合肥工业大学学报（自然科学版）》2021年第2期164-169,共6页Journal of Hefei University of Technology：Natural Science

基　　金：江苏省质量技术监督局科技资助项目(KJ795914)。

摘　　要：文章建立了有限元模型计算复合材料修复的含中心裂纹钢板的裂纹处应力强度因子(stress intensity factor,SIF),把仿真结果作为训练样本;提出通过支持向量机(support vector machine,SVM)对样本集进行训练和预测,建立基于复合材料补片尺寸参数的胶接强度预测模型,综合考虑补片的长度、宽度和厚度之间的关联性,为补片参数选择提供参考;采用遗传算法(genetic algorithm,GA)、粒子群算法(particle swarm optimization,PSO)和改进的粒子群算法(improved particle swarm optimization,IPSO)对SVM的惩罚因子c和核函数的参数g进行寻优,分别建立GA-SVM、PSO-SVM、IPSO-SVM 3种模型对胶接强度进行预测。结果表明:IPSO-SVM模型的预测效果要优于GA-SVM和PSO-SVM模型;IPSO-SVM模型能够准确地预测修复结构的胶接强度。利用该预测模型可以避免重复建模仿真,降低了考虑补片尺寸参数耦合时的复杂性;应用该预测模型所得的最优尺寸参数对结构进行修复,结构的强度得到了有效提高。A finite element model was established to calculate the stress intensity factor(SIF)at the crack location of the steel plate with central crack repaired by composite materials.The simulation results were used as training samples.Support vector machine(SVM)was proposed to train and predict the sample set.A prediction model of bonding strength based on the dimension parameters of composite patches was established.The correlation among the length,width and thickness of patches was considered comprehensively.It can provide reference for the selection of patch parameters.Genetic algorithm(GA),particle swarm optimization(PSO)and improved particle swarm optimization(IPSO)were used to optimize the penalty factor c and the parameter g of the kernel function of SVM.Three models,GA-SVM,PSO-SVM and IPSO-SVM,were established to predict the bonding strength.The results show that the prediction effect of IPSO-SVM model is better than that of GA-SVM model and PSO-SVM model.IPSO-SVM model can accurately predict the bonding strength of repaired structure.The prediction model avoids repeated modeling and simulation and reduces the complexity of considering the coupling of patch size parameters.The optimal size parameters obtained from the prediction model are used for repairing,and the strength of the structure is effectively improved.

关 键 词：应力强度因子 支持向量机(SVM) 预测模型 遗传算法(GA) 粒子群算法(PSO) 

分 类 号：TB332[一般工业技术—材料科学与工程] TG491[金属学及工艺—焊接]