循环平稳度在发电机组故障趋势分析中的应用

Application of Degree of Cyclostation to Trend Analysis of Fault in Generator Sets

由于发电机组各部件的运动存在明显的周期性,当发电机组发生故障时,机组振动信号具有循环平稳特性,这种非平稳性随故障的变化而变化。因此,提出利用循环平稳度监测分析发电机组的故障趋势。循环平稳度是利用三阶循环累积量构造并利用二元决策方法估计三阶循环矩。它能抑制噪声与工况干扰,反映机组的故障状态。实例分析表明,将其作为发电机组故障特征量,提取故障的发展趋势,效果好于传统的通频值和烈度值趋势分析。

Generator sets are typical rotary machines and the movements of each part of them have cyclic variation. When there are some faults in generator sets ,the vibration signal of sets has a cyclostationary feature. The nonstationarity always changs with the faults. So,the degree of cyclostation was proposed to monitor and analyze the fault trend of generator sets by using the third-order cyclic moment and binary decision method.The method can reflect the state of fault and suppress the noise and interference. It can be used as the fault feature vector to analyze trend of the fault development. The result is better than the analysis result of using traditional vibration amplitude and intensity.