Research On Dynamic Risk Assessment And Early Warning Method Of Maintenance Based On Intelligent Identification

Maintenance operation is an important means to recover the defects of equipment in the continuous process and improve the reliability of equipment.Using scientific and reasonable safety risk management technology is of great significance to ensure the safety and order of inspection operation.With the development of the intelligent in refining and petrochemical industry,the traditional maintenance static regulation method for risk identification,risk assessment,early warning is difficult to meet the requirements of real-time assessment of risk change,this article obtains from the intelligent point of view,to overhaul process as the application background,puts forward the intelligent risk monitor early warning strategies,build a intelligent risk identification model based on video surveillance.A dynamic risk assessment model based on Bayesian network is established and an intelligent risk management system for maintenance operation process is developed.The specific research contents include:(1)The intelligent risk monitoring and early warning strategy is proposed.For traditional risk management method is a static view of risk identification and risk assessment and did not consider change characteristics introduction to video monitoring and target detection technology based on the intelligent risk recognition technology,combined with the fuzzy Bayesian network,put forward the maintenance process of intelligent monitoring and dynamic risk assessment of the different risk early warning strategies.The dynamic monitoring ability of risk management model is improved.Compared with traditional risk management methods,it can monitor risks in real time,effectively solve the problem of risk dynamic assessment and early warning,and monitor risk changes in real time.(2)The intelligent risk identification model based on video surveillance is constructed.At present video monitoring technology can only be "prison" can’t "control" problem,with the introduction of the target detection technology based on deep learning,using deep convolution characteristics study,the monitoring object recognition in video monitoring and positioning,can build a kind of real-time identification of video monitoring of the observed risk elements of intelligent risk identification model.Through the dynamic monitoring of risk factors in the maintenance process,the real-time display of the status of risk factors,compared with the traditional video surveillance technology,reduces the investment of human resources,improves the accuracy and integrity of risk identification,and enhances the effectiveness of safety management.Taking helmet monitoring in maintenance operation as an example,the effectiveness of the method is verified by helmet identification in different operation processes,and the identification accuracy can reach93%,realizing the function of intelligent risk identification.(3)The dynamic risk assessment method based on fuzzy Bayesian network is established.According to dynamic evaluation of maintenance risk state change problem,analyzes the relationship between risk factors change,because the risk data samples and compare deficient,the fuzzy set theory is introduced to put forward the combination of fuzzy sets and historical data prior probability calculation method,combined with dynamic classification early warning method,risk early warning model for the evaluation of dynamic Bayesian network is established.Taking the Genie software to calculate the maintenance process and the change of risk possibility when risk factors change as an example,the effectiveness of the method is verified.Combined with the intelligent identification model,the intelligent control requirements of risk factors are realized,which solves the problems of delayed evaluation results and low evaluation accuracy,and enhances the safety of maintenance engineering.(4)Developed an intelligent risk management system for maintenance process.Design the system framework according to the system demand analysis,follow the development specifications,and develop modules such as large screen display of data,equipment information,video monitoring information,early warning information,personnel management information and risk assessment information based on the development of front and back end separation technology,which can effectively improve the safety and intelligent operation and maintenance level of refineries.