汽车变速器磨损轴承振动信号混沌行为分析

Analyzing the Chaotic Feature of Worn Bearing Vibration Signal in Automotive Transmission

本文首先讨论了判断时间序列具有混沌性态的准则,探索了求时间序列奇异吸引子嵌入维数的方法,并用G-P算法求关联维数。在尝试用几个传统的奇异吸引子验证了该方法的正确性之后,将它应用于计算一组变速器磨损轴承加速度振动信号的分形维数。计算表明,变速器振动信号具有分形特征。在其它条件不变时,轴承技术状况与振动信号的分形维数密切相关。它们均在2维与3维之间,随着轴承磨损加剧,它向3维逼近。当样本点达到一定数目时,计算结果的稳定性好。

This paper, first discusses the principle to identity a time series with chaotic character, and explores a method of finding out the embedding dimensions of time series, and calculate correlative dimension using G-P method. After its correctness is verified with several traditional strange attractors, the method has been used to calculate the fractal dimensions of some vibration signals in automotive transmission. The result shows that the vibration signal of automotive transmission has a fractal feature. With other conditions unchanged, the more seriously the bearing is worn, the bigger the fractal dimension of vibration signal become, which are between 2 and 3. With the bearing's condition deteriorated, the fractal dimension is approaching to 3. When the number of sample data points is big enough, the calculating result is rather stable.