基于Kriging模型的高压共轨柴油机标定应用  

Application Study of Kriging Model in Common Rail Diesel Engine Calibration

作　　者：柴智刚[1] 张付军[1] 刘波澜[1] 黄英[1] 付耿[1] 

机构地区：[1]北京理工大学机械与车辆学院,北京100081

出　　处：《北京理工大学学报》2015年第3期242-245,共4页Transactions of Beijing Institute of Technology

基　　金：国家部委预研项目(DEDP-1004)

摘　　要：为解决电控柴油机随着控制参数的增多造成的标定工作量大、工作效率低的问题,将基于模型的标定方法应用于VMR425DOHC高压共轨柴油机的经济性优化过程中.采用LHS-ESE算法设计控制参数的试验空间,根据获得的试验数据建立控制参数与响应数据之间的Ordinary Kriging代理模型,并在此基础上进行优化得到最终的控制参数MAP.试验结果表明,对比常规标定方法,采用基于模型的标定可以使试验样本点选取更加合理,大幅提高试验标定效率,Kriging代理模型的使用使预测结果更为精确,最大误差率在2%以内,满足共轨柴油机建模需要,优化结果使发动机燃油经济性区域得到扩大.With the high development of modern diesel engines,an increasing number of new controlling parameters have been put out to calibrate,which lead to the growth of calibration loads and the decrease of the work efficiency.This paper dealt with those problems by modelbased calibration method which applied the LHS-ESE algorithm to design testing sample,built an ordinary kriging surrogate model based on test data and then optimized these calibration parameters,and this method had been used in an economic optimization process of VMR425 DOHC common rail diesel engine.The result of the study showes that the method made the distribution of the testing samples more reasonable.Moreover,the use of the surrogate kriging model had better prediction accuracy with the error less than 2% as well as a good match with these modeling requirements.The fuel consumption had been improved according to the optimization result.

关 键 词：试验设计 Kriging建模 共轨发动机 标定 

分 类 号：TK421[动力工程及工程热物理—动力机械及工程]