基于改进蚁群算法的齿轮故障模式识别方法研究  

Study on Mode Indetification of Gear Fault Based on Ameliorated ACO

作　　者：李琳[1] 张永祥[1] 李军[1] 

机构地区：[1]海军工程大学,武汉430033

出　　处：《煤矿机械》2009年第12期230-232,共3页Coal Mine Machinery

摘　　要：根据蚁群算法优化路径原理,提出一种基于改进蚁群算法的简单的模式识别方法。通过对TSP蚁群模型的适当修改,将模式识别问题转化为路径优化问题。为了尽量避免蚁群算法过早收敛或收敛到局部最优路径,采用了最大-最小和自适应蚁群算法相结合的方法对算法进行改进。通过实验,以齿轮3种不同状态下的关联维数为特征,成功实现了对不同故障状态的识别。According to the deficiency of the common method of mode identification, a new method of mode identification based on ACO is put forward. The problem of mode identification is translated into of best way finding by amending the model of algorithmic. The way of max-min ACO and adaptive ACO is combined in order to minish the probability of early convergence. The correlation dimension is chosen as the characteristic of different states of gear. The new method of mode identification is proved to be valid and simple by experiment. Three different states of gear is classified successfully by this way.

关 键 词：蚁群算法 故障诊断 关联维数 模式识别 

分 类 号：TH165[机械工程—机械制造及自动化] TP306[自动化与计算机技术—计算机系统结构]