基于短时峭度法的钢丝绳损伤检测及定位研究

Research on detecting and locating damage of steel wire rope based on short-time kurtosis method

经典的基于漏磁信号的钢丝绳损伤检测方法,过分依赖使用者的先验知识,其准确率和泛化性能不稳定且损伤定位困难。针对上述问题,提出了基于短时峭度法的钢丝绳漏磁信号损伤检测方法,利用峭度对冲击信号的敏感性实现了钢丝绳损伤的识别。在分析信号与峭度之间关系的基础上,定义了短时峭度指标,利用遍历算法得到最优窗宽及步长,结合分段的思想,通过分段峭度实现对短时峭度图的绘制,实现了损伤特征的提取和定位。结果表明,与基于小波变换的阈值判别法相比,所提方法可以有效对钢丝绳损伤进行检测及定位,准确率稳定且具有一定的泛用性。将短时峭度法应用到钢丝绳损伤检测中,为钢丝绳损伤检测领域的研究提供了新的参考。

The classical method of detecting the damage of steel wire rope based on MFL(magnetic flux leakage) signal relies too much on the user’s prior knowledge, its accuracy and generalization performance are unstable and the damage location is difficult. In view of the above problems, a short-time kurtosis based method for detecting the MFL signal damage of steel wire rope was proposed. The sensitivity of kurtosis to the impact signal was used to identify the damage of steel wire rope. After analyzing the relationship between signal and kurtosis, the short-time kurtosis indexes were defined, and the optimal window width and step size were obtained by the ergodic algorithm. Combined with the idea of segmentation, the short-time kurtosis map was drawn by segmented kurtosis, and the damage features were extracted and located. The results showed that compared with the threshold discrimination method based on wavelet transform, the proposed method effectively detected and located the damage of steel wire rope with stable accuracy and certain universality. The application of short-time kurtosis method to the damage detection of steel wire rope provided new reference for the research on the damage detection of steel wire rope.