车用发动机运行状态的模糊聚类与识别

Fuzzy Clustering and Identification for Operating State of Vehicle Engine

利用混沌理论对车用发动机多组振动信号进行相空间重构,计算出不同运行状态的关联维数,同时结合各状态的振动平均值、峰峰值、燃料敲缸系数及机械敲缸系数组成状态向量样本。在此基础上,利用模糊C 均值的方法对样本数据进行聚类分析,得到标准状态向量,从而实现发动机当前状态的模糊识别。试验结果表明这一分析方法可为发动机的故障监测和及时维修提供可靠依据。

The phase space reconstruction based on the chaos theory is used to a series of engine vibration signal to calculate the correlation dimension of different operating state. Meanwhile, combine other vibration parameters such as fuel knocking coefficient and mechanical knocking coefficient to form the state estimation vectors, and then using the fuzzy C-means method to get the standard state vector and realize current state fuzzy identification. The experimental result shows that this analysis method can offer reliable reference in fault detection and timely engine repairing.