Research On Power Operation Illegal Behavior Identification System Based On Deep Learning

With the continuous expansion of the power grid scale,accidents occur frequently in the process of power production and operation of power enterprises,resulting in huge casualties and property losses,and the supervision of power operations is facing great challenges.Aiming at the problem that the existing technology can only identify a single target and cannot identify violations according to specific power operation scenarios,this paper designs and implements a power operation violation identification system based on the power operation scenarios to improve the efficiency of intelligent monitoring and management.The specific research work is as follows:(1)Build a dataset for power operation.Analyze and study the regulations for safe production in power operations,summarize the working characteristics of power operation scenarios,and divide the power operation data into three categories:features that should be present in all operation scenarios,features that should be prohibited in all operation scenarios,and specific behavioral features in specific scenarios.(2)In order to solve the problem of illegal behavior target detection,a detection algorithm based on the YOLOv5 s model is proposed.Aiming at the problems of small targets and low detection accuracy of specific behaviors,the algorithm adopts the strategy of adding small target detection layer and attention mechanism respectively,and makes corresponding adjustments to the initial algorithm,and tests on three types of typical data sets.better results.(3)In order to solve the problem of illegal behavior evaluation,a behavior evaluation model based on job scenarios is proposed.First of all,according to the electric power operation scene,the judgment standard of violation behavior is formulated,and an evaluation function of violation behavior is designed for three typical operation scenarios,and the threshold value of violation behavior evaluation is determined through experiments.Secondly,based on the LEC method of risk assessment of operating conditions,adjust the assessment parameters,design the assessment method in this paper,adjust the differences between the two methods,redefine the indicators and their assignments,and use them for the division of violation levels.(4)Design and implementation of the power operation violation recognition system.The system was analyzed for requirements,and the overall system process was designed.The Pyqt5 framework was used to design and implement various functions of the system.