Industrial fault diagnosis method and application based on self-adaption feature extraction

Abstract

The invention discloses an industrial fault diagnosis method and application based on self-adaption feature extraction, and belongs to the technical field of industrial process monitoring and diagnosis. Firstly, data feature analysis is conducted on industrial acquisition data, and appropriate feature extraction methods are chosen according to different data features; secondly, fault classification is achieved through a Hidden Markov model method. According to the industrial fault diagnosis method and application based on the self-adaption feature extraction, the self-adaption feature extraction method is adopted specific to diversity of the industrial data with features such as linearity, nonlinearity and nongaussianity, the purpose of reserving effective information to a maximum extent is achieved, and classification of the industrial process faults is conducted through extremely strong dynamic procedure time series modeling capability and time-series pattern classification capacity of the Hidden Markov model, so that compared with other existing methods, due to the fact that the data features are adequately considered, by means of the industrial fault diagnosis method based on the self-adaption feature extraction, higher precision rate of industrial fault diagnosis is achieved.

Description

technical field [0001] The invention belongs to the field of industrial process monitoring and fault diagnosis, in particular to an industrial fault classification method based on self-adaptive feature extraction. Background technique [0002] As the complexity of industrial processes grows, the effectiveness of industrial process monitoring and diagnosis becomes increasingly important to ensure production process safety, maintain product quality, and optimize product benefits, and feature extraction is an important step in fault diagnosis. [0003] There are many traditional feature extraction methods, such as Principal Component Analysis (PCA), Kernel Principal Component Analysis (KPCA) and Independent Component Analysis (ICA). PCA can effectively deal with the linear relationship between variables, but it cannot detect the nonlinear structure among variables. However, in industrial processes, non-linear relationships between variables are ubiquitous. To solve this probl...

AI Technical Summary

Problems solved by technology

Therefore, this method has great advantages in solving complex data characteristics such as non-Gaussian and nonlinear in the process industry, and can realize effective fault diagnosis

Images