Chemical process fault detection method based on multi-sampling-probability-kernel principal component analysis model

Abstract

The invention discloses a chemical process fault detection method based on a multi-sampling-probability-kernel principal component analysis model, and the method comprises the steps: firstly constructing a multi-sampling-probability-kernel principal component analysis model through collecting a training sample set, and obtaining the T2 and SPE statistic detection control limits of the training sample; carrying out the online collection of the technological parameters of the to-be-detected chemical process in the actual running process, and obtaining a test sample; calculating the T2 and SPE statistic quantities of the test sample, and finally performing the comparison with the detection control limits, and obtaining an online detection result of the chemical process. The method employs a kernel learning method based on the multi-sampling-probability-kernel principal component analysis model, an effective nonlinear fault detection model is built, and a problem caused by the difference of sampling rates in the chemical production process is solved, thereby improving the online detection efficiency and performance of the process, enabling the chemical production process to be more reliable, and enabling the quality monitoring of the product to be more stable.

Description

Technical field [0001] The invention relates to a fault detection method, in particular to a chemical process fault detection method based on a multi-sampling probability nuclear principal component model. Background technique [0002] In the modern process industry, with the application of distributed control systems (DCS) and the advancement of computer technology, a large number of online and offline measurement data are collected and stored on the industrial site. The process monitoring (MSPM) technology based on multivariate statistical analysis has achieved rapid Development, it has the advantages of data-based, dimensionality reduction, easy visualization and easy practical application, etc., and has been widely used in many industrial fields such as chemical industry, pharmaceuticals, and semiconductor manufacturing. Among them, principal component analysis (PCA) and partial least square estimation (PLS) and their extension methods are representative models of MSPM techno...

AI Technical Summary

Problems solved by technology

However, this type of method cannot judge whether the fluctuation of the production process affects the quality of the final product

Images