Research On Dynamic Prediction And Evaluation Method For Major Hazard Installation Systems In Chemical Industry Park

In recent years,China's chemical industry is booming,and the level of safety production has gradually improved.However,there are number of risks to be faced.The production and storage of the major hazard installations system are mainly hidden danger.In order to improve the safety supervision level of major hazard installations,it is urgent to complete the supervision and prevention of potential risks with relevant technologies.The traditional quantitative risk analysis method is a static evaluation system,which impossible realizes dynamic prediction and early warning for dynamic change hazard installations system.Therefore,this paper proposes a dynamic quantitative risk analysis method,which combines dynamic prediction and assessment of major hazard installations,and provides a solution for the dynamic risk assessment of chemical industry park.The specific research work is as follows:(1)The classical quantitative risk analysis method of hazard installation in chemical industry park is widely used in accident scenario modeling and risk quantification,but it's static structure cannot model the risk changes in the operation process,and cannot realize the prediction of accident potential.In this paper,the outdated static evaluation system is transformed into dynamic quantitative evaluation system by integrating correlation analysis,prediction and dynamic evaluation technology.Through qualitative and quantitative correlation analysis of key parameters,the dynamic prediction model of key parameters is constructed,and the fuzzy comprehensive evaluation system is completed to realize a dynamic quantitative risk assessment method for major hazard installations in chemical industry parks.(2)There are many sensitive key industrial parameters in the hazard installations system of chemical industry park,such as liquid level,pressure and temperature.The abnormal fluctuation of these key parameters will cause the change of the risk degree of the whole hazard installations system.In order to monitor the abnormal change and trend of the key parameters in real time and detect the accident symptoms as soon as possible,it is necessary to realize the prediction of the key parameters.Because the chemical process industry has the characteristics of a large time delay,some other industrial parameters may be abnormal or fluctuated in advance before the key industrial parameters are abnormal.These industry parameters and the key industrial parameters of the hazard installations constitute a strong correlation.The method of visual correlation analysis based on mutual information proposed in this paper can be used to screen the variables with strong correlation of key parameters,so that the autoregressive prediction method of key parameters can be transformed into the method of predicting key parameters with strong correlation variables.(3)Prediction of the key parameters of the hazard installations system in the chemical industry park,the traditional prediction method does not consider the time lag phenomenon,it difficult to predict sudden changes in data.And the static structure of the traditional regression analysis method cannot model the risk changes in the process of operation,which leads to the decrease of prediction performance.Therefore,this paper considers a dynamic recurrent neural network prediction method,and combined with time-delay analysis.By eliminating the influence of time-delay,the key parameters and strongly correlated variables are in the same time line.Thus,improving the prediction accuracy.Finally,a two-level fuzzy comprehensive evaluation system is constructed based on the prediction results of key parameters and the attenuation generation algorithm.Dynamic risk prediction and assessment of major hazard installations in chemical industry park are realized.