KPCA-based visual fault diagnosis for nonlinear industrial process

With the increasingly large-scale, continuous, and complicated chemical process, it is particularly important to ensure the stability and safety of the production process. However, in past studies, the accuracy of fault diagnosis and the degree of system visualization are still insufficient. Here, in order to solve these problems, a visual fault diagnosis system based on LabVIEW and Matlab is designed. First, the system uses LabVIEW interface design, applying Matlab to compile the algorithm program, which makes the system has a powerful data calculation and processing functions, as well as a clear visual interface, the system design also optimizes the communication interface. Second, the typical chemical production process TE (Tennessee Eastman) process is the subject of systematic testing. Additionally, because most of the industrial processes are non-linear, the fault diagnosis method based on Kernel Principal Component Analysis (KPCA) is used in the system design, and the implementation process of this method is elaborated. Finally, the system achieves the functions of TE process data acquisition, data preprocessing, and fault diagnosis lamps. A large number of simulation results verify the effectiveness of the proposed method. The system has entered the stage of laboratory application and provides a good application platform for the research of fault diagnosis of complex systems such as chemical process control.