Hui Kete set sail from the same angle, and the classification of the defects was not the same. The shortcomings of diagnosis are briefly divided into: methods based on analytical mathematical models and methods not dependent on mathematical models. 1. According to the method of analyzing mathematical models According to the method of analyzing mathematical models, sensor fault diagnosis can be divided into: parameter estimation method, situation estimation method and equivalent space method. According to the shortcomings of the model, the method of diagnosis is a method of diagnosis that has been developed early. It is also a method of research and application of extensive diagnosis. Benefits are model mechanism understanding, simple structure, easy to complete, easy to analyze, and can be diagnosed in real time. The orientation in the field of defect diagnosis will still be the research direction of the sensor defect diagnosis method in the future development. The disadvantage is that the amount of accounting is large, the system is messy; there are modeling errors, the adaptability of the model is poor; the phenomenon of false positives and false negatives is obvious; the robustness of external disturbances, the noise and annoyance of the system are not sensitive. Nowadays, the research results of this diagnosis method still need to focus on the linear system, which has implications for in-depth study of the general defects of the nonlinear system, and the robustness problem has a high research value. Table 1 describes the benefits and shortcomings of some of the shortcomings in the model approach. Table 1 Comparison of the benefits and shortcomings of some shortcomings in the model method 2, the sensor does not depend on the shortcomings of the mathematical model to diagnose the method At that time, the control system becomes more and more messy. Because it is difficult to establish an analytical mathematical model of the control system in practice, when there is modeling fault, the diagnosis method will have false positives and false negatives according to the shortcomings of the model, so The method of diagnosis based on the shortcomings of the model has received much attention. The benefit of a method that does not rely on mathematical models is that it does not require a model of the target and is highly adaptable. The disadvantage is that the structure is messy and difficult to finish. Such a defect diagnosis method that does not depend on the system model can be classified into a defect diagnosis method based on a data-driven method, a diagnosis method based on a shortcoming of knowledge, and a method based on discrete work. 2.1 According to the data-driven approach There are two broad categories of data-driven methods: signal processing methods and accounting methods. Some commonly used methods for diagnosis based on signal processing defects include: value checking and trend checking, using the disadvantages of the Kulbb ack information criterion, according to the shortcomings of the adaptive sliding Lattice filter, according to the shortcomings of the signal modal estimation method. Analytical methods, wavelet analysis methods, and information methods. 2.2 Methods based on knowledge According to the shortcomings of knowledge, the method of diagnosis can be divided into two types according to the shortcomings of the diagnosis and the methods of diagnosis based on the shortcomings of the qualitative model. 2.3 According to the method of discrete work The method of diagnosis based on the shortcomings of discrete work is a new type of defect diagnosis method developed over the years. The basic idea is that the situation of the discrete work model reflects both the normal situation and the shortcomings of the system. Other Other Yilian Smart Manufacturing Co.,Ltd , https://www.ylcrepemaker.com