基于混沌理论的往复式压缩机故障诊断

实测某往复压缩机气缸在正常、活塞体松动及拉缸工况下的振动信号,计算信号的关联维数、Kolmogorov熵及最大Lyapunov指数,并证明信号具有非线性。用奇异值分解(SVD)降噪法对信号进行降噪,通过奇异熵增量曲线选择降噪阶次,用互信息法求最佳延时,并用假近邻域法求最佳嵌入维数,同时用G-P算法、小数据量法求出关联维数、Kolmogorov熵及最大Lyapunov指数。计算结果表明,3种工况下的往复压缩机气缸振动信号均为非线性混沌信号,故障越严重,信号表现出的混沌特性越明显,可依据混沌指标对往复压缩机气缸的状态进行监测。

Fault diagnosis of reciprocating compressor based on chaos theory

The vibration signals of a cylinder were measured in three work conditions,including normal operation,loosening of piston and piston-liner wear,then chaotic indexes were calculated and the nonlinear characteristics of these signals were proved.Firstly,the original signals were decomposed using singular value decomposition(SVD),and a reasonable order for noise reduction was selected according to the singular entropy of singular spectrum,and the best delay was calculated using mutual information method,and the best embedding dimension was calculated using fault near neighbour method.Then correlation dimensions and Kolmogorov entropies were calculated using G-P algorithm simultaneously,and the largest Lyapunov exponents were calculated using small-data method.The results show that the cylinder vibration signals are nonlinear chaotics signal,and the more serious the fault is,the more apparent the chaotic character is.The vibration condition monitoring for the reciprocating compressor can be implemented using chaotic indexes.