在线传感器突变信号的检测与区分

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

为了准确区分传感器突变信号产生的原因,提出了基于数学模型的小波频带分析法.针对工业流程中的测控系统,分析了输出突变信号的频率组成与突变原因的关系.用小波频带分析技术,将高低频信号分离,并进行能量统计,根据高低频信号能量比例的变化,判断出突变信号产生的原因.经典型控制系统的计算机仿真和恒压供水系统实验结果表明,该方法能够有效地诊断出传感器是否发生故障.

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.