Detection and Identification of Abrupt Changes for On-Line Sensor Output Signal

In order to accurately distinguish the origins bringing about abrupt changes in sensor output,a method is presented for wavelet frequency analysis based on mathematical model.Aimed at measurement and control systems in industrial processes,the relation between frequencies and origins of abrupt changes in sensor output signals are analyzed in detail.The high-low frequency signals are separated using wavelet frequency analysis and the energy of frequency signals is calculated.Based on the energy proportion changes among high-low frequency signals,origins bringing about abrupt changes in sensor output can be estimated.The results through computer simulation for typical control system and experimentation for water supply systems in invariable pressure showed that sensor malfunction in industrial process can be effectively diagnosed using the above-mentioned method.