A Fault Diagnosis Method for Chemical Processes Based on Multilayer Optimization pcc-sdg

Abstract

The invention discloses a chemical process fault diagnosis method based on multi-layer optimized PCC-SDG. The present invention uses the network topology structure of the whole process as a reference point, utilizes the Pearson correlation coefficient (PCC) statistical index to initially optimize the selected variables, and then uses the PCA weight idea to select special variables with relatively large weights from the multi-layer correlation coefficient set , combined with the signed directed graph (SDG) to establish the optimal PCC‑SDG network, and finally for the optimal PCC‑SDG to establish the rule of aggregation weight coefficient Q for fault diagnosis. The invention proposes a new fault diagnosis method, perfects the SDG modeling method, improves the efficiency of the staff in detecting multi-variable states, avoids the influence of time lag and other non-information synchronization factors, and more effectively reduces the false alarm rate and Accurately identify the type of failure, greatly reducing the occurrence of production safety accidents.

Description

Technical field [0001] The present invention belongs to the technical field of chemical process fault diagnosis, and specific relates to a process fault diagnosis method based on multi-layer optimization PCC-SDG. Background technique [0002] With the increasing scale of modern industrial scale and the continuous improvement of system complexity, the requirements for identifying and detection of faults on faults are increasing. At present, the symbolic map (SDG) is more difficult, and the fault diagnosis of the whole process is not significant based on the simple correlation law, although simple mathematical model analysis can make a certain diagnosis of the fault but not diagnosis The law is in-depth discussion, a stronger partiality, which is lacking on integrity factors, and cannot guarantee the safety and stability of the system, and it is difficult to apply. [0003] The normal operation of the chemical system process follows the first diamond, thus presenting complex correl...

AI Technical Summary

Problems solved by technology

At present, it is difficult to quantitatively model the symbolic directed graph (SDG), and the fault diagnosis effect of the whole process based on the simple correlation law is not significant. Although the simple mathematical model analysis can diagnose the fault to a certain extent, it has no In-depth discussion of the rules of the system, strong locality, lack of overall factors, can not guarantee the security and stability of the system, and the application is relatively difficult

[0003] The normal operation of the chemical system process follows the first law, which makes the process variables show complex correlation characteristics, while the SDG's schematic abstraction of process technology and equipment can represent the complex causal relationship between the variables of the whole process, and SDG theory has been successfully applied to industrial processes, and the method is relatively mature, but there are few studies on SDG quantitative modeling methods

Images