Cleaning teratment of dynamic measuring data

摘要

It's inevitable that there are abnormal components in dynamic testing data. When processing data, it is possible to cause unpredictable errors of data processing result if the data are not cleaned previously, that is, the effects of abnormal signals areeliminated. It can even cause the failure of partial or total data analysis methods. A new and effected method is described in this paper. It can automatically identify and separate abnormal components in dynamic measuring data, and reasonably repair the data which are polluted. Its availability is proved by simulation instances.