Study And Application Of Source Intensity Back-calculation Algorithm Based On Monitoring Data

Resources of oil and gas are the national important strategic resources. With the highspeed development of petrochemical and chemical industry, sudden catastrophic chemicalaccidents of fire explosion and poisoning often occur. They cause heavy losses of people'slives and property and statistic data shows ascendant trend. So it is very important to studysuch accidents mechanism and provide emergency rescue decision making basis according tothe forecast for the consequences. Leakage intensity value and position are the importantinfluence parameters of predicting accident consequences and the foundation of emergencyrescue decision with fast and accurately designing escape routes, determining the evacuationzone and escape zone accident, formulating public warning scheme of accident. In emergencycourse, it is an important challenge to estimate the chemicals leakage intensity and location.In order to reduce the harm caused by accident, it is very necessary to quickly and accuratelyestimate the source information of the leak source.Atmospheric diffusion models and common atmospheric diffusion modes weresummarizd and analyzed to choose appropriate positive diffusion model for back calculationof the source intensity. With summarizing back calcilation methods of the source intensity, theadvantages and disadvantages of them are given. The PSO and its hybrid optimizationalgorithm were applied to back calculation of the source intensity problem. The superiorityand inferiority of these algorithms were analyzed and discussed. PSO algorithm is a kind ofnew and simple algorithm of group intelligent which is easily to realize and understand. It canbe applied to estimate leak source location and the source intensity and can quickly andaccurately obtain the source intensity information. But in complex source intensityback calculation problem, PSO algorithm as other intelligence algorithm, is easily to appearthe phenomena of premature convergence and weak local search capability, so thatNelder Mead simplex algorithm in particle swarm algorithm using serial structure, enchasestructure and improved hybrid structure were proposed and applied to back calculating of thesource intensity problem. The results show that the best leake source information could be gotusing improved particle swarm simplex hybrid algorithm by shorter time and smaller error, soimproved particle swarm simplex hybrid algorithm can accurately used to obtain the source intensity information and satisfy the need of emergency rescue decision making. Thefeasibility and applicability of PSO algorithm and its hybrid algorithm were verified in themulti dimensional back calculation problem of source intensity that is the position andstrength of the source of leakage, and they were more suitable for solving multi dimensionalunknown parameters.With VB6.0 and positive model of Gaussian diffusion model and AERMOD model,program PSO and its hybrid algorithm, and compare the veracity in the presence ofmeasurement errors. No matter what kind of diffusion model to select for the forwarddiffusion model, PSO hybrid algorithm more rapidly and accurately than the single PSOalgorithm obtains source intensity value for the emergency rescue support.