Structure health monitoring sensing data loss recovery method based on low-rank Hank matrix

Abstract

The invention discloses a structural health monitoring sensing data loss recovery method based on a low-rank Hank matrix, and the method comprises the steps: obtaining the sensor data of multiple channels in a time domain, and forming the sensor data of the Hank matrix arrangement, specifically a low-rank matrix; and complementing lost data in the sensor data arranged in the Hank matrix, and obtaining sensor complemented data of multiple channels on a time domain. According to the method, sensor data of multiple channels in a time domain are expanded on the channels to form sensor data arranged in a Chinese-kick matrix mode, the sensor data arranged in the Chinese-kick matrix mode are low-rank matrixes, lost data in the sensor data arranged in the Chinese-kick matrix mode are complemented, sensor complemented data arranged in the Chinese-kick matrix mode are obtained, and the sensor complemented data in the Chinese-kick matrix mode are obtained. And finally, multi-channel sensor completion data on the time domain are obtained. The method can effectively recover the lost data of the multi-channel sensor data on the time domain in different modes, has a good recovery effect on various data loss modes, and is high in robustness. 本发明公开了一种基于低秩汉克矩阵的结构健康监测传感数据丢失恢复方法，包括：获取时间域上多通道的传感器数据，并形成汉克矩阵排列的传感器数据，具体为低秩矩阵；对汉克矩阵排列的传感器数据中的丢失数据进行补全，并得到时间域上多通道的传感器补全数据。本发明将时间域上多通道的传感器数据在通道上扩展形成汉克矩阵排列的传感器数据，且汉克矩阵排列的传感器数据为低秩矩阵，再对汉克矩阵排列的传感器数据中的丢失数据进行补全，得到汉克矩阵排列的传感器补全数据，最后得到时间域上多通道的传感器补全数据。本发明的方法能有效恢复不同模式下时间域上多通道的传感器数据的丢失数据，对多种数据丢失模式均具有较佳的恢复效果，且鲁棒性强。