Estimation Of Source Position And Intensity In Chemical Releases By A Combined Genetic-Nelder Mead Simplex Algorithm

In the emergency response of chemical accidents, it is necessary to use atmospheric dispersion model to determine the impact of hazardous chemicals in the region in order to provide emergency decision support, and the leakage of dangerous chemicals is a source of strength of the atmospheric dispersion model, the important input parameters which affect the atmosphere diffusion model of the accuracy of the results, thus affecting the accuracy of the emergency decision-making. Thence, Determination of the locations and strength of the source in a chemical leakage is crucial to crowd evacuation and emergency decision.This paper summarizes the existing methods of the back-calculation of the intensity of source, respectively, the contents of the different methods, characteristics and scope of application is analyzed andsummarized.Combined with its practical application in the case of comparison, the thinking and solving process of the methods is comparised and analyzed. The strengths and weaknesses of various methods are given. This paper compares concentrations computed by dispersion model with the measured by receptors and then hybrid genetic-Nelder Mead simplex algorithm model is established. The locations and strength of leakage source are characterized. The feasibility of the algorithm is verified by data of instantaneous release and continuous release. It is indicated that the hybrid genetic-Nelder Mead simplex algorithm is seldom affected by initial values and can get good results even if the initial values are far away from the expected values. This method back-calculates the optimal solution in a smaller errors and a faster pace and is more suitable for the search procedure in multi-dimensional space. Thence, hybrid genetic-Nelder Mead simplex algorithm is able to back-calculate the locations and strength of leakage source quickly and accurately, which will meet the needs of emergency rescue.