The Isolation and Restoration of Abnormal Signal under High Temperature Dynamic Measurement

In the practice process of high temperature float glass online measurement, the thickness of the glass, which is the target being measured by system, is changing with the time and space. The input and the experimental result are changing with time. The data input by instrument has not only the data of certain thickness of glass, but also many uncertain data which cannot be express by clear and definite math formula. For example: environment noise, the abnormal value random anomaly. These noise cannot be removed by using simple static measurement during the dynamic measurement process. This article came up a method using moving average method to separate the abnormal signal, and combined with five point one time end effect method to determine the width of the abnormal signal, and then using weighted mean to repair. Combine this experiment and bring out the elimination method of the error and the result of the instrument dynamic measurement.