论文部分内容阅读
为实现在工况变化条件下对旋转机械的故障预测,提出使用相空间曲变和平滑正交分解理论在变工况条件下跟踪旋转机械的故障演化过程.首先在对目标系统的观测时间序列相空间重构的基础上,通过量化相空间曲变构建信号损伤演化的跟踪函数,为弥补累积模型误差和相空间点局部分布概率差异造成的误差,将时间序列和相空间进行分割,并以此构建跟踪矩阵;再利用平滑正交分解方法将跟踪矩阵中分别由实际损伤劣化和工况变化造成的演化趋势进行分离,根据平滑正交特征值提取出其中能够反映实际故障演化趋势的平滑正交分量;最后以变转速情况下轴承外环故障退化的仿真信号为例验证算法的有效性.计算结果表明:本文提出的算法能够对旋转机械故障的演化趋势实现有效跟踪,基本排除转速波动造成的工况变化影响.
In order to realize the fault prediction of rotating machinery under the condition changing, a method of tracking the fault evolution of rotating machinery under varying working conditions is proposed by using the theory of phase space warping and smoothing orthogonal decomposition.Firstly, in observing the time series of the target system Based on the reconstruction of phase space, the tracking function of signal damage evolution is constructed by quantizing the phase space distortion. In order to make up the error caused by the difference between the cumulative model error and the local distribution probability, the time series and phase space are partitioned, Then, the tracking matrix is constructed. Then, the evolution trend caused by the actual damage and the change of working conditions in the tracking matrix is separated by using the method of smooth orthogonal decomposition. According to the smoothed orthogonal eigenvalue, the smoothing positive factors that can reflect the evolution trend of the actual fault are extracted Finally, the simulation results show that the algorithm proposed in this paper can effectively track the evolution trend of rotating machinery faults and basically eliminate the fluctuation of rotational speed The change of working conditions.