Parameter identification for continuous point emission source based on Tikhonov regularization method coupled with particle swarm optimization algorithm

Denglong Ma; Wei Tan; Zaoxiao Zhang; Jun Hu

doi:10.1016/j.jhazmat.2016.11.071

Parameter identification for continuous point emission source based on Tikhonov regularization method coupled with particle swarm optimization algorithm

J Hazard Mater. 2017 Mar 5:325:239-250. doi: 10.1016/j.jhazmat.2016.11.071. Epub 2016 Dec 2.

Authors

Denglong Ma¹, Wei Tan², Zaoxiao Zhang³, Jun Hu⁴

Affiliations

¹ School of Food Equipment Engineering and Science, Xi'an Jiaotong University, No.28 Xianning West Road, Xi'an 710049, PR China; School of Chemical Engineering and Technology, Xi'an Jiaotong University, No.28 Xianning West Road, Xi'an 710049, PR China.
² School of Food Equipment Engineering and Science, Xi'an Jiaotong University, No.28 Xianning West Road, Xi'an 710049, PR China.
³ State Key Laboratory of Multiphase Flow in Power Engineering, Xi'an Jiaotong University, No.28 Xianning West Road, Xi'an 710049, PR China; School of Chemical Engineering and Technology, Xi'an Jiaotong University, No.28 Xianning West Road, Xi'an 710049, PR China. Electronic address: zhangzx@mail.xjtu.edu.cn.
⁴ School of Chemical Engineering, Northwest University, No.229 Taibai North Road, Xi'an 710069, PR China.

PMID: 27940113
DOI: 10.1016/j.jhazmat.2016.11.071

Abstract

In order to identify the parameters of hazardous gas emission source in atmosphere with less previous information and reliable probability estimation, a hybrid algorithm coupling Tikhonov regularization with particle swarm optimization (PSO) was proposed. When the source location is known, the source strength can be estimated successfully by common Tikhonov regularization method, but it is invalid when the information about both source strength and location is absent. Therefore, a hybrid method combining linear Tikhonov regularization and PSO algorithm was designed. With this method, the nonlinear inverse dispersion model was transformed to a linear form under some assumptions, and the source parameters including source strength and location were identified simultaneously by linear Tikhonov-PSO regularization method. The regularization parameters were selected by L-curve method. The estimation results with different regularization matrixes showed that the confidence interval with high-order regularization matrix is narrower than that with zero-order regularization matrix. But the estimation results of different source parameters are close to each other with different regularization matrixes. A nonlinear Tikhonov-PSO hybrid regularization was also designed with primary nonlinear dispersion model to estimate the source parameters. The comparison results of simulation and experiment case showed that the linear Tikhonov-PSO method with transformed linear inverse model has higher computation efficiency than nonlinear Tikhonov-PSO method. The confidence intervals from linear Tikhonov-PSO are more reasonable than that from nonlinear method. The estimation results from linear Tikhonov-PSO method are similar to that from single PSO algorithm, and a reasonable confidence interval with some probability levels can be additionally given by Tikhonov-PSO method. Therefore, the presented linear Tikhonov-PSO regularization method is a good potential method for hazardous emission source parameters identification.

Keywords: Contaminant emission; Inverse problem; Source estimation; Stochastic method.