Research On The Risk Assessment Method And Application Of Power Production Operation Based On Situation Awareness

Power production operations is an important component to ensure the continuous supply of electric energy.However,power production operations involve a wide range of fields,complicated procedures and high risk factors.Serious personal injury accidents often occur during the power production operation.Although the rules and regulations of electric power safety production are gradually improved,the operation management method is gradually optimized,and the low degree of digitalization and intelligence still exposes electric power operations to a higher risk of major accidents.At present,our country’s power industry has gradually formed a power operation risk early warning mechanism based on static risk assessment.By evaluating the possibility of accidents in the operation process before and after operation,corresponding risk management and control strategies are formulated.However,the static evaluation method cannot predict the real-time risk of sudden changes in the operation in real time,and the static control measures formulated in advance are difficult to match with the actual production situation.Therefore,in order to further improve the safety level of power operations,this paper systematically conducts research on the risk situational awareness technology and application of power operations.The specific contents are as follows:(1)In view of the characteristics of various types of power production operations accidents and complex causes,through statistics and analysis of historical cases of power production operations accidents,considering the characteristics of power production operations and accidents,the causes of accidents are analyzed from five aspects: personnel,physical state,environment,operation nature and organization and management.A quantitative characterization method of the association degree between accident causes and accidents based on association rules mining is proposed,and based on the association degree,a practical and reliable power production operations accident cause model is researched and deduced.This lays the foundation for the subsequent construction of a Bayesian network-based power production operations risk situation prediction model.(2)In view of the low utilization rate of multi-source risk information in power production operations,firstly,the data information that can reflect the risk situation of power production operations is screened,and the feature quantities that have a high degree of correlation with power operation accidents are extracted to characterize the risk situation information.Facing massive multi-source heterogeneous risk situation information,using the advantages of deep learning method to process unstructured data,the feature extraction process of unstructured text data and image data is modeled by convolutional recurrent neural network and deep convolutional neural network respectively,thus constructing a digital and intelligent situation element extraction method.This provides a data basis for predicting the risk situation of power production operations.(3)Aiming at the large randomness of power production operation violations,considering the advantages of Bayesian network in the expression and reasoning of uncertain knowledge,this paper proposes a Bayesian network-based power production operation risk prediction model.The node composition of the Bayesian network is determined on the basis of the accident causation model of power production operations,and the Bayesian network structure and parameter distribution are learned based on historical accident data.In the whole process of power production operation,the risk situation feature quantity is used as the input of the constructed risk prediction model,and the probability of occurrence of typical operation accidents is predicted in real time through the Bayesian forward inference method,so as to realize the real-time early warning and control of power production operation risks.(4)Finally,based on the above theoretical analysis and the historical data of real power production operations in Guangxi Power Grid,this paper conducts platform construction and algorithm verification,and deploys and applies it in the provincial power grid system to provide support for the risk management and accident prevention of local power production operations,and effectively improve the safety level of power production.