Study On Condition Monitoring Method Of Unit And Pre-Processing For Signal

The condition monitoring of hydropower generator sets not only makes the missionary mastering knowing the working condition of unit, prolongs overhaul in terval, but also saves the cost of hydropower plant, ensures production security. Condition-based maintenance of hydropower generator sets consists of three components: information acquisition, fault diagnosis methods and maintenance policies. Because of hydropower generating sets's working environment, information acquired often contains all kinds of noise. Therefore, information acquisition plays an important roles of the system. And, it is important to acquire data of operation mode of hydroelectric generation set for making maintenance policies.This paper deals with the way of data measuring and de-noising method. The major tasks covered in this paper include:Firstly, according to various character of hydropower generating sets and various, and characteristics of the various microsensors, corresponding monitoring method is presented. Then, a measurement system using this method is designed.Secondly, after analyzing existing signal processing method, given the nonstationarities of data and abrupt changes in signals, wavelet analysis can be applied as a per-processing method. Based on comparing results of the noise reduction by using various wavelet function and de-noising method, a optimal method is presented.Finally, computational results of measured data shows that wavelet analysis play a very important role in improving of data acquisition accuracy and identification accuracy forspectrum frequencies.