基于R型系统聚类和因子分析的柴油机磨损状态综合评价  

Comprehensive Evaluation for Wear State of Diesel Engines Based on R-type Hierarchical Clustering and Factor Analysis

作　　者：陈涛[1] 王立勇[1] CHEN Tao;WANG Li-yong(Key Laboratory of Modern Measurement&Control Technology Ministry of Education,Beijing Information Science and Technology University,Beijing 100192,China)

机构地区：[1]北京信息科技大学现代测控技术教育部重点实验室,北京100192

出　　处：《组合机床与自动化加工技术》2020年第7期65-68,72,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学青年基金项目(51605035);国家国防科技工业局基础科研项目(JCCPCX201705)。

摘　　要：为有效判别设备的早期磨损状态,利用设备采集油样的光谱分析数据作为原始数据,运用R型系统聚类和因子分析方法综合判别设备的磨损状态。首先,对光谱检测元素进行R型系统聚类,确定磨损特征元素,减少磨损评价指标的数量;其次,对磨损特征元素的浓度指标进行因子分析特征降维,挖掘出磨损主因子;最后,计算磨损综合得分,判断设备的磨损状态。对两台船舶主柴油机进行磨损状态判别实例分析,结果表明R型系统聚类和因子分析方法的评价结果与柴油机实际磨损状态一致,能够为设备的磨损状态综合判别提供一种新方法。In order to evaluate the wear state of the equipment effectively, the spectral analysis data of oil samples are used as the original data, and R-type hierarchical clustering and factor analysis methods are used to comprehensively evaluate the wear state. Firstly, R-type hierarchical clustering is carried out for the spectrum monitored elements to determine wear characteristic elements and reduce the number of wear evaluation indexes. Then, factor analysis is carried out for the concentration index of the wear characteristic elements to reduce characteristic dimension and mine the main wear factor, and finally the wear state of the equipment is judged by the calculated comprehensive wear score. An example shows that the evaluation results of R-type hierarchical clustering and factor analysis method are consistent with the actual wear condition of two marine diesel engines, which can provide a new method for the comprehensive determination of wear condition of equipment.

关 键 词：R型系统聚类 因子分析 特征降维 光谱分析 磨损判别 

分 类 号：TH17[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]