Research On The Early Warning Of Production Safety Accidents In Power Generation Enterprises Based On Bayesian Network

In the process of electric power production,there are many factors affecting safe production,although the level of power production management has improved,power production accidents still occur from time to time,causing great harm to people and society.The safety of electric power production is always related to the property safety of the state,the vital interests of the people,as well as the life safety of workers.Therefore,it is of great practical significance to find risk sources as early as possible,reduce accident probability,improve the safety management level of power enterprises,and conduct in-depth research on the risk management of power production safety.This paper reviews the research status of Bayesian network and electric safety production accidents by using the literature method,briefly analyzes the characteristics of electric safety production accidents,studies the related theories of risk warning,and expounds the related factors of Heinley's analysis of accident rules.The content explores the advantages of Bayesian network for risk management and common methods of modeling,laying the foundation for the later construction model and simulation analysis.Then the risk of power production accident is identified.Based on the traditional two-dimensional risk model,sensitivity factors are integrated into the risk dimension of seriousness and possibility,and the risk assessment index of power production accidents is established from the perspective of three-dimensional risk.The system uses the multi-level fuzzy evaluation comprehensive evaluation method to empower the evaluation index system,and combines the risk matrix method to formulate the accident risk level standard.After that,the established risk evaluation index system is input into the Bayesian network,and structural learning and parameter learning are used to construct the Bayesian network electric safety production accident risk early warning model.By the reasoning and learning of the risk early warning model,through its reverse reasoning,sensitivity analysis and maximum causal chain analysis,the main risk influencing factors are determined,and control measures for major risk factors were proposed.Finally,DS Power Generation Co.,Ltd.is taken as an example to simulate and verify the operability and practicability of Bayesian network electric safety production accident risk warning model.The research results show that the accident is caused by the interaction of the main factors with other factors that are greatly affected by it.The main risk factors are found through the analysis of the maximum causal chain,prioritizing the risk factors in the largest causal chain can reduce the probability of accidents to a certain extent;Then,through sensitivity analysis,find out other influencing factors that are greatly affected by the main factors,and formulate corresponding control measures for them;Finally achieve the purpose of preventing electric power production accidents and improving the level of power safety risk management.