基于分形盒维数的汽轮机转子振动故障诊断的实验研究

汽轮机转子是汽轮机故障预测及诊断问题的一个重要方面.利用振动信号或状态量对设备进行诊断,是设备故障诊断中最有效、最常用的方法.本文在Bently实验台上进行了实验,模拟电站汽轮机转子的不对中、不平衡、碰摩和轴承松动4种故障的振动过程.根据汽轮机故障的分形特征,采用分形盒维数方法进行了不同故障试验数据的分形盒维数研究计算.结果表明:碰摩状况下的轴心轨迹最为混乱且呈现雪花状,其对应的分形盒维数也最大;不平衡时的轴心轨迹较为平滑,其对应的分形盒维数最小;不对中故障及松动故障的分形盒维数居中.因此,分形盒维数在进行汽轮机转子故障类型诊断时有更好的区分度.

Experimental Research on Steam-Turbine Rotor Vibration Fault Diagnosis Based on the Fractal Box Counting Dimension

The rotor is the main component during turbine fault forecasting and diagnosing. Vibration signal and state parameter are usually selected as the monitored signals to equipment diagnosis. The steam-turbine vibrating is experimentally simulated on the Bently facility. Using box counting dimension method the different type of turbine fault with fractal characteristic is numerically researched. The result shows that the axis track under mass unbalance is smooth and its box counting didaension reaches the small- est. The axis track under friction presents disorder and its box counting dimension is the largest. The box counting dimension of un-countershaft and loosing are in the middle. As a result, the box counting dimension possesses better distinguish during steam-turbine rotor fault type diagnosing.