A Data-Driven Adaptive Multilayer Flow Model Equipment Diagnosis Method

Abstract

The invention discloses a data-driven self-adaptive multi-layer flow model equipment diagnosis method, which combines the better accuracy based on the data-driven method and the better interpretability based on the multi-layer flow model method, and overcomes the traditional The fixed abnormal alarm threshold in the multilayer flow model method is easy to cause missing alarms, which leads to wrong diagnosis results. Based on the data-driven equipment status classification and identification results, the abnormal alarm threshold of the multi-layer flow model is adaptively adjusted, and the current equipment status is further confirmed through the multi-layer flow model anomaly detection and diagnostic reasoning to enhance the reliability of equipment fault diagnosis. The equipment diagnosis method has important scientific research value and engineering application value.

Description

technical field [0001] The invention relates to a fault diagnosis method for complex systems and equipment, in particular to improve the fault diagnosis method based on a multi-layer flow model by using a data-driven classification recognition method. Background technique [0002] Fault diagnosis is a process of signal acquisition, processing, analysis and identification, and decision-making evaluation. Its main task is to monitor the state of the system, judge whether it is normal, predict and diagnose system failures and eliminate them, and guide the management and maintenance of the system. After decades of development, fault diagnosis technology has formed a variety of diagnostic methods. According to the views of experts in the field of fault diagnosis, the existing fault diagnosis methods can be divided into methods based on analytical mathematical models, methods based on signal processing and methods based on knowledge. [0003] The method based on the analytical m...

AI Technical Summary

Problems solved by technology

[0010] The technology of the present invention solves the problem: overcomes the deficiencies of the prior art, and provides a data-driven adaptive multilayer flow model equipment diagnosis method, which is both accurate and interpretable, and overcomes the fixed abnormality in the traditional multilayer flow model method The alarm threshold is easy to cause missing alarms, resulting in the problem of wrong diagnosis results, which improves the reliability of the diagnosis results. At the same time, it is easier to generate alarms in time than the original multi-layer flow model method, thereby speeding up the fault diagnosis speed and leaving room for fault handling. more time

Images