基于Chaos-RS-RBF算法的汽油机油膜动态参数辨识研究  

Study on Estimation of Gasoline Engine Oil Film Dynamic Parameter Based on Chaos-RS-RBF Algorithm

作　　者：徐东辉 徐向阳[1] 徐新仁 Xu Donghui;Xu Xiangyang;Xu Xinren(Department of Mathematics and Computer Science,Nanchang Normal University,Nanchang 330032,China)

机构地区：[1]南昌师范学院数学与计算机科学系,南昌330032

出　　处：《大理大学学报》2020年第6期29-36,共8页Journal of Dali University

基　　金：国家自然科学基金项目(51176014);江西省科技支撑计划项目(20151BBE50108);江西省教育厅科学技术研究项目(GJJ170896,GJJ170893,GJJ170892)。

摘　　要：针对汽油发动机动力学系统的高度复杂的非线性特性,提出了Chaos-RS-RBF(chaos-rough sets-radialbasis function)算法对油膜动态参数进行辨识。在判断发动机动力学系统混沌(chaos)特性的基础上,通过相空间重构技术恢复其固有的高度复杂的非线性特性,获得多维状态空间时间序列,利用粗糙集(rough sets,RS)删除大量冗余数据,最后采用径向基函数(radial basis function,RBF)算法对多维状态空间时间序列进行辨识,获得油膜动态参数辨识值。仿真结果显示,与最小二乘法及RBF神经网络相比较,Chaos-RS-RBF模型具有更高的精度,对实际工程应用具有较好的借鉴意义。In view of the highly complex nonlinear characteristics of the gasoline engine dynamics system,the Chaos-RS-RBF algorithm is proposed to identify the dynamic parameters of the oil film.On the basis of judging the chaos characteristics of the engine dynamics system,the inherently complex nonlinear characteristics are restored through the phase space reconstruction technology to obtain a multi-dimensional state space time series,and a large number of redundancies are removed using rough sets(RS).Finally,the RBF algorithm is used to identify the multi-dimensional state space time series to obtain the oil film dynamic parameter identification value.The simulation results show that compared with the least square method and RBF neural network,the Chaos-RS-RBF model has higher accuracy and has a good reference for practical engineering applications.

关 键 词：汽油机 相空间重构 混沌 RS RBF 估测 

分 类 号：TK411[动力工程及工程热物理—动力机械及工程]