Application and improvement of swarm intelligence optimization algorithm in gas emission source identification in atmosphere

Hazardous gas emissions could cause serious consequences for ecology, environment, human life and even society. Thus gas emission source term identification is crucial for emergency response and safety management. Based on experimental data, swarm intelligent optimization (SIO) algorithms including particle swarm optimization (PSO), ant colony optimization algorithm (ACO) and firefly algorithm (FA), are compared to identify the gas emission source parameters including source strength and location parameters. The results show that all three SIO methods used in this work have similar performances in terms of source parameter estimation, and all of them depend slightly on initial range set for individuals in the population. However, PSO method is superior in computational efficiency compared with ACO and FA methods. The convergence rate of FA is faster than that of ACO. PSO method can obtain satisfied estimation results under different boundary constraints, while the estimation results of FA and ACO will become unrealistic under too wide boundary constraints. The impact of atmospheric conditions on estimated results is also discussed. The results under extreme atmospheric conditions are worse than that in other conditions. Finally, SIO method coupled with a new model, correlated matching of concentration distribution (CMCD) model, is applied to the source location estimation. Test results prove that SIO-CMCD model can obtain a satisfied estimation as well as greatly enhanced computational efficiency when only location parameters are required to be determined. Hence, SIO is a useful tool to estimate emission source term for the storage and transportation process of hazardous gas or volatile materials.