基于代理模型的消声器噪声和背压多目标优化  

Multi-Objective Optimization of Muffler Noise and Back Pressure Based on Surrogate Model

作　　者：黄泽好 黄荆荣[1,3] HUANG Zehao;HUANG Jingrong(School of Vehicle Engineering,Chongqing University of Technology,Chongqing 400054,China;Key Laboratory of Advanced Manufacturing Technology for Automobile Parts,Ministry of Education,Chongqing 400054,China;China National Heavy Duty Truck Group Chongqing Fuel System Co.Ltd.,Chongqing 401120,China)

机构地区：[1]重庆理工大学车辆工程学院,重庆400054 [2]汽车零部件先进制造技术教育部重点实验室,重庆400054 [3]中国重汽集团重庆燃油喷射系统有限公司,重庆401120

出　　处：《西南大学学报（自然科学版）》2022年第11期201-208,共8页Journal of Southwest University(Natural Science Edition)

基　　金：重庆市教委科研项目(KJQN201801101);重庆市研究生导师团队项目(渝教研发[2018]6号).

摘　　要：以汽车尾管噪声和排气背压为目标,以消声器结构参数为变量,以精度较高代理模型进行多目标优化.通过最优拉丁超立方方法采集试验样本,比较不同代理模型精度后确定应用Kriging代理模型并结合带有精英保留策略的非支配排序遗传算法(NSGA-Ⅱ),进行多目标优化得到了最优解集帕累托前沿.在背压小于20 kPa附近选择一个多目标优化解并对优化后的消声器进行仿真和试验验证,各工况下尾管噪声误差在2%以内,额定转速排气背压误差为0.86%,优化后的尾管噪声和排气背压均在目标限值以内,说明代理模型可信,优化方法可行.Taking automobile tailpipe noise and exhaust back pressure as targets,and muffler structure parameters as variables,the multi-objective optimization was carried out with a high-precision proxy model.Test samples were collected by the optimal Latin hypercube method.Kriging proxy model and non-inferior ranking genetic algorithm with elite retention strategy(NSGA-Ⅱ)were adopted after comparing the accuracy of different proxy models.The Pareto frontier of the optimal solution set was obtained by multi-objective optimization.A multi-objective optimization solution was selected near the back pressure less than 20 kPa,and the optimized muffler was verified by simulation and experiment.Under all working conditions,the tailpipe noise error was within 2%,the rated speed exhaust back pressure error was 0.86%,and the optimized tailpipe noise and exhaust back pressure were within the target limits,indicating that the proxy model is credible,and the optimization method is feasible.

关 键 词：消声器 代理模型 尾管噪声 排气背压 多目标优化 

分 类 号：U463.51[机械工程—车辆工程]