机械信号复杂性快速刻划方法研究

A Rapid Complexity Evaluation Method of Mechanical Signals

针对目前机械信号复杂性定量评估的研究现状 ,充分考虑信号采样频率、转子转速、振幅、高频噪声对指标评估结果的影响 ,提出一种适用范围更广 ,描述方法更加精确 ,在线实现能力更强的信号复杂性刻划方法 ,成功地实现了回转机组运行状态的快速定量评估。试验结果表明 ,该方法有效、可靠 。

In the detection of mechanical faults, we need accurate and rapid methods even if the problem is highly complex. In many cases, such methods are not yet available. We propose a rapid complexity evaluation method of mechanical signals to satisfy a part of this need. In sections 1 and 2, we discuss the effect of sampling frequency, vibration amplitude and high frequency noise on evaluation result and explain how to eliminate this undesirable effect. We applied our method to a certain compressor rotor during its run up process. We measured the orbit of rotor center for six rotating speeds: 2 168, 6 035, 6 970, 9 199, 10 254, 12 246 r/min. Figure 5 shows the orbits. The relative complexities were computed by our software to be 0.975, 1.636, 0.362, 1.942, 1.966, and 1.763 respectively. These results show preliminarily that our method for computing relative complexity is rapid (on line capability) and accurate.