摘要
润滑油在发动机各部件间流动时,不仅发挥其应有的功能,同时也承载了丰富的关于发动机运行状况的信息,能够有效地反映发动机状态.以某型装甲车底盘发动机为对象,提出一种对润滑油信息进行分析以实现发动机状态评估的方法.该方法基于核主成分分析(KPCA)和沙猫群优化(SCSO)算法优化的支持向量机(SVM),使用KPCA对收集的油液数据进行降维处理,得到的降维数据作为SVM的输入.随后,应用SCSO算法优化SVM的关键参数,建立状态评估模型.通过实际数据的实验验证及与其他几种状态评估模型的比较,结果显示该方法准确率达到了97.35%,能有效评估发动机状态,从而为发动机的维护提供重要参考.
As lubricating oil flows between the various components of an engine,it not only performs its intended functions but also carries rich information about the engine′s operating condition,and can effectively reflect the state of the engine.Focusing on the engine of a certain type of armored vehicle chassis,a method is proposed for assessing engine state through the analysis of lubricating oil information.The method is based on kernel principal component analysis(KPCA)and sand cat swarm optimization(SCSO)algorithm-optimized support vector machine(SVM).It utilizes KPCA to reduce the dimensionality of the collected oil data,and the resulting lower-dimensional data serves as input for the SVM.Subsequently,SCSO algorithm is applied to optimize the key parameters of the SVM,and a state assessment model is established.Experimental validation with actual data and comparisons with several other state assessment models demonstrate that this method achieves an accuracy of 97.35%,which can effectively evaluate engine state and thus provide important references for engine maintenance.
作者
李英顺
于昂
姬宏基
李茂
郭占男
LI Yingshun;YU Ang;JI Hongji;LI Mao;GUO Zhannan(School of Information Engineering,Beijing Institute of Petrochemical Technology,Beijing 102617,China;The Army Third Military Delegate Room in Shenyang Area,Shenyang 110000,China;Shenyang Shunyi Technology Company Limited,Shenyang 110000,China)
出处
《大连理工大学学报》
CAS
CSCD
北大核心
2024年第4期426-432,共7页
Journal of Dalian University of Technology
基金
辽宁省“兴辽英才计划”资助项目(XLYC1903015)。
关键词
发动机
润滑油
状态评估
核主成分分析
沙猫群优化算法
支持向量机
engine
lubricating oil
state assessment
kernel principal component analysis
sand cat swarm optimization algorithm
support vector machine
作者简介
李英顺(1971-),女,博士,教授,博士生导师,E-mail:liyingshunbipt@126.com.
