基于相对隶属度的发动机磨损模糊模式识别  

Fuzzy Pattern Recognition of Engine Wear Based on Relative Membership Degree

作　　者：王南兰[1] 李晓峰[1] 

机构地区：[1]湖南文理学院电气工程系,湖南常德415003

出　　处：《车用发动机》2008年第4期28-31,共4页Vehicle Engine

基　　金：湖南省教育厅基金项目(04c414)

摘　　要：针对模糊数学需人为确定隶属函数的缺陷,利用以相对隶属度为基础的模糊理论对发动机磨损模式进行了识别,得到了较为客观的识别结果,证明了该方法对发动机磨损模式识别的有效性。同时,Matlab的仿真结果还证明了利用最优模糊聚类中心矩阵可以消除冗余信息、简化模式特征值的提取,使模式识别的准确率提高。In view of the flaw that the membership function had to be determined in fuzzy mathematic, the fuzzy theory based on the relative membership degree was used to recognize the engine wear patterns, and the more objective recognition results were acquired, which proved that the method was effective for engine wear pattern recognition. Simultaneously the simulation result of Matlab proved that the most superior fuzzy clustering center matrix might eliminate the redundant information, simplify the pattern characteristic value extraction, and acquire the higher pattern recognition accuracy.

关 键 词：发动机 磨损 相对隶属度 模式识别 

分 类 号：TK427[动力工程及工程热物理—动力机械及工程]