The Research Of Fault Diagnosis Method With Unknown Input Disturbance

In the actual production process, the control system work environment tends to be poor, input signal inevitably contain some interference noise named the unknown input disturbance signals. The control system for the accuracy of fault diagnosis of the challenges arise, the corresponding unknown input observer theory was created. This work mainly includes the following:This article describes the process of development in the diagnosis of Bernard important fault diagnosis observer theory. This method was first developed fault diagnosis. The principles of Bernard detection filter were study, and the observer- based fault diagnosis was given. Bernard mathematical model representation, and fault diagnosis of actuators and sensors of the constant bias, the dead and the constant gain fault are described, and the observer Bernard application in fault diagnosis of conditions can be observed.In the third chapter introduces the feature structure assignment theory of fault diagnosis methods, followed by the given diagnosis based on feature structure assignment design method of the structure, and that the observer system in the original system stability. Finally, simulation design, observer based fault diagnosis Bernard output of feature structure-based fault diagnosis output results were compared.Kalman filter is an optimal estimation theory, which focuses in this article describes the Kalman filtering method in fault diagnosis approach. First, introduce the development process of Kalman filtering and the application of practical engineering situation. Then, the Kalman filter theory of fault diagnosis and the specific implementation method. Finally, according to Bernard observer by comparing the diagnostic methods and diagnostic methods of the Kalman filter, the method based on Kalman filter of fault diagnosis is effective.