In the measurement and control system, the sensor output signal is affected by many factors, and mutations often occur. This mutation point value contains important fault information, accurately capturing and distinguishing the factors that cause these mutation points, is the key to sensor fault diagnosis. Most rely on the sensor's output time series to diagnose sensor faults, and the sudden changes in sensor output signals are attributed to sensor faults.

The specific method of detecting faults is to perform wavelet transform on the input and output signals of the control system respectively. When the wavelet function can be regarded as the first derivative of a smooth function, the sudden change point of the signal corresponds to the modulus maximum of the wavelet transform. This detects the abrupt point and produces a residual sequence and analyzes the sensor fault, and considers that the sudden change in the sensor output signal is due to a sensor failure or a sudden change in the system input signal. In fact, there are many reasons for the sudden change of the sensor output signal, in addition to the system input abrupt and the fault of the sensor itself, in addition to process disturbance, actuator failure, controller failure, controlled object and external electromagnetic field interference, etc. In practical applications, The above sensor fault diagnosis method has certain limitations. Often, in industrial process control, the time constant of the controlled object is large and cannot respond to high frequency components in the abrupt signal. Based on the wavelet analysis of the band analysis technology, the causes of the sudden change of the sensor output signal are discussed, which provides a practical analysis method for fault diagnosis and performance evaluation of online sensors.

The abrupt signal of the sensor output contains important fault information. The mutation factors are different, and the frequency of the mutation signal is different. For a controlled object with a large time constant, most of the sensor abrupt signals caused by a given input change, disturbance change, controller failure, and actuator failure are only low-frequency components. Generally only low frequency components. The abrupt signal caused by external electromagnetic field interference is generally a pulse signal, which contains low frequency components and more high frequency components. The abrupt signal of the sensor deviation fault contains a small amount of high frequency components in addition to the low frequency component.