| In the actual industrial production process,fault detection and fault diagnosis of the entire process and key components are the most important to ensure safe production.If you only rely on manual labor to achieve time-consuming and labor-intensive,you need to use some algorithms instead of manual to complete the fault detection and failure.The process of diagnosis.Among many fault detection algorithms,the PCA method has been widely used due to its simple calculation method,easy application,and complete mathematical theory system.The detection principle of the PCA-based fault detection algorithm is to determine whether a fault has occurred by analyzing the statistical regularity of each monitoring signal data in the industrial production process,and by analyzing whether the statistics corresponding to each data sample exceed the corresponding threshold.However,there are a large number of nonlinear links in the industrial production process,and the kernel method is often introduced into the PCA algorithm.The kernel method is used to non-linearly map the original data to a high-dimensional space,and then perform PCA processing in the high-dimensional space.In this paper,fault detection of TE process is carried out based on PCA method and KPCA method.When the fault is detected by the algorithm,the fault detection algorithm only detects the occurrence of the fault,but cannot determine the source of the fault.Therefore,it is necessary to introduce a fault diagnosis algorithm to determine the source of the fault after the fault detection is completed.The fault diagnosis method based on support vector machine can fully mine the hidden statistical information in the data based on the principle of structural risk minimization when the monitoring signal data information is limited,and finally realize the classification of the fault and determine the fault source.Pneumatic control valve has been widely used in the industrial production process due to its practicability,safety and ease of operation.Once the pneumatic control valve fails,it will cause a series of chain problems.Therefore,this article is based on the failure of support vector machines.The diagnosis method has carried on the breakdown diagnosis to the pneumatic control valve.The core of the support vector machine-based fault diagnosis method is the establishment of the support vector machine model,and the most important thing is the selection of the kernel function parameters.Using empirical methods to manually set the size of the kernel function parameters has great uncertainty,so an optimization algorithm is introduced to find suitable kernel function parameters.Because particle swarm algorithm has good convergence and optimization ability,it has been widely used in solving numerical optimization problems.In this paper,particle swarm optimization algorithm is used to optimize the fault diagnosis model based on support vector machine. |
