Multi-feature-fused real-time detection method and device for human factor risk

Abstract

The invention discloses a multi-feature-fused real-time detection method and a multi-feature-fused real-time detection device for human factor risk. The method comprises the following steps: S1, collecting human factor test data in the anti-accident process; S2, processing the collected human factor test data; S3, in the data processing process, extracting the features of change in the behavior, audio and physiological signals of a dispatcher during the anti-accident process; S4, by utilizing sampling timestamp of the data, carrying out frame-by-frame time alignment on the various features with different frame rates extracted from video, audio and physiological measurement data, and thus multi-channel synchronous relevance is realized; and S5, carrying out weight allocation on all the feature data indexes, determining the weight of each index by utilizing a principal component analysis method, finally, substituting the plurality of features with the greatest contribution rate into a comprehensive evaluation function, and then evaluating the risk of human errors of the dispatcher. With the method and the device provided by the invention, the problems of hysteretic human factor riskaccident analysis and low-efficiency human factor risk factor detection in the existing electric power system are solved.

Description

technical field [0001] The invention relates to the technical field of human factor risk detection, in particular to a multi-feature fusion human factor risk real-time detection method and device. Background technique [0002] With the advancement of the national smart grid strategy, the deployment of dispatching automation systems, the application of intelligent electronic devices and remote program control technologies have further ensured the reliable operation of equipment. highlight. The prior planning and arrangement of power grid operation, the issuance of dispatching control orders, fault location and elimination, and equipment inspection and maintenance require the participation of operating personnel. However, due to the limitation of their own knowledge level and the influence of external environmental pressure, the operating personnel may not be able to complete the scheduled operations quickly and accurately, resulting in serious consequences such as equipment ...

AI Technical Summary

Problems solved by technology

[0003] According to investigations, the current research on human factor reliability analysis has become an indispensable part of probabilistic safety analysis in nuclear power plants, aerospace and other fields, but it is still in its infancy in the field of electric power, and only simple operational risk training is given to power grid dispatchers Work, the analysis of human error is limited to simple analysis and accident statistics after the event, continuous real-time analysis cannot be achieved, and the effectiveness of human risk prediction in the actual scheduling process is low

Images