Exploring ozone formation rules and concentration response to the change of precursors based on artificial neural network simulation in a typical industrial park

Qiaoli Wang; Dongping Sheng; Chengzhi Wu; Jingkai Zhao; Feili Li; Shengdong Yao; Xiaojie Ou; Wei Li; Jianmeng Chen

doi:10.1016/j.heliyon.2023.e20125

Exploring ozone formation rules and concentration response to the change of precursors based on artificial neural network simulation in a typical industrial park

Heliyon. 2023 Sep 14;9(9):e20125. doi: 10.1016/j.heliyon.2023.e20125. eCollection 2023 Sep.

Authors

Qiaoli Wang¹, Dongping Sheng¹, Chengzhi Wu², Jingkai Zhao¹, Feili Li¹, Shengdong Yao¹, Xiaojie Ou¹, Wei Li³, Jianmeng Chen^{1

4}

Affiliations

¹ College of Environment, Zhejiang University of Technology, Hangzhou, 310032, China.
² Trinity Consultants, Inc. (China Office), Hangzhou, 310012, China.
³ Key Laboratory of Biomass Chemical Engineering of the Ministry of Education, Institute of Industrial Ecology and Environment, College of Chemical and Biological Engineering, Zhejiang University (Zijingang Campus), Hangzhou, 310030, China.
⁴ Zhejiang University of Science & Technology, Hangzhou, 310023, China.

Abstract

Industrial parks have more complex O₃ formation mechanisms due to a higher concentration and more dense emission of precursors. This study establishes an artificial neural network (ANN) model with good performance by expanding the moment and concentration changes of pollutants into general variables of meteorological factors and concentrations of pollutants. Finally, the O₃ formation rules and concentration response to the changes of volatile organic compounds (VOCs) and nitrogen oxides (NOx) was explored. The results showed that the studied area belonged to the NOx-sensitive regime and the sensitivity was strongly affected by relative humidity (RH) and pressure (P). The concentration of O₃ tends to decrease with a higher P, lower temperature (Temp), and medium to low RH when nitric oxide (NO) is added. Conversely, at medium P, high Temp, and high RH, the addition of nitrogen dioxide (NO₂) leads to a larger decrease capacity in O₃ concentration. More importantly, there is a local reachable maximum incremental reactivity (MIR_L) at each certain VOCs concentration level which linearly increased with VOCs. The general maximum incremental reactivity (MIR) may lead to a significant overestimation of the attainable O₃ concentration in NOx-sensitive regimes. The results can significantly support the local management strategies for O₃ and the precursors control.

Keywords: Artificial neural network; Maximum incremental reactivity; Ozone formation rules; Sensitivity and response analysis.