Multi-behavior process monitoring method based on pivot analysis and vectorial data description support

Abstract

The invention discloses a multi-operating process monitor method based on principal component analysis and support vectors data. The method establishes a uniform PCA model to various operating mixed data firstly, puts score vectors of principal component space and residual space to high dimension characteristic space. Two new statistics are established in the characteristic space for monitoring the principal component space and residual space. When the process goes wrong, a fault reconstruction method based on SVDD identifies fault. The method establishes two SVDD statistics monitor model to various operating based that the principal analysis method is used for reducing process variable dimension, reduces statistics limit of processing monitor, increases sensitivity of processing monitoring. In addition, the invention provides a fault reconstruction and identifying method aiming at detected process fault which can locate source of fault commendably, is benefit to removing fault as soon as possible, returns process to normal operation.

Description

technical field [0001] The invention belongs to the field of flow industry process monitoring and fault diagnosis, in particular to a multi-working-condition process monitoring, fault reconstruction and identification method based on principal component analysis and support vector data description. Background technique [0002] As a process performance monitoring and fault diagnosis technology based on multivariate statistical projection theory, multivariable statistical process control (MSPC) has received extensive attention from academia and industry. Since the 1990s, the MSPC method represented by principal component analysis (PCA) and partial least squares (PLS) has been successfully applied in industrial process monitoring. However, the traditional MSPC methods assume that the process operates under a single stable condition. In fact, due to the diversification of products and other reasons, most industrial processes do not operate under a single working condition, and...

AI Technical Summary

Problems solved by technology

However, the above methods have certain disadvantages

In addition to the establishment of multiple models (large amount of calculation), the multi-model method is very likely to cause false positives and false positives for the transition part of the working condition, and when the method is implemented online, it is necessary to determine which working condition the current sampling belongs to

The iterative update model method has a strong blindness, to be precise, they cannot distinguish between normal operating conditions and fault conditions of the process

Although a single MSPC model can better model the process of multi-working conditions, because the process runs under multiple working conditions, its process monitoring statistical limit will become very loose compared with the single working condition model, so that resulting in a decrease in sensitivity of the method to process variations

Images