Air pollution risk assessment related to fossil fuel-driven vehicles in megacities in China by employing the Bayesian network coupled with the Fault Tree method

journal article in Web of Science

en

Vehicle emissions have become one of the key pollution sources affecting air quality and human health in China's megacities. How to curb excess vehicle emissions has become a key pain point of urban air pollution prevention and control. This work tries to explore an effective integrated approach/framework to quantitatively assess the risk factors of excess vehicle emissions (EVE) and their impact on air quality for China's typical megacities. Bayesian Network is employed as the assessment tool by coupling with the Fault Tree method to curb the above problem for the first time. Four megacities (Beijing, Tianjin, Hangzhou, and Guangzhou) in China are selected as case studies, and the risk factors leading to EVE are identified to construct the Bayesian Network model. At the same time, some accurate quantisation algorithms of the occurrence probability of root nodes were proposed for the target megacities from 2014 to 2019. The analysis results show that the variation trend of the probability of EVE has a good positive correlation with the variation trend of air quality in some megacities. From 2014 to 2019, the no-occurrence probability of EVE in Beijing, Tianjin, and Hangzhou increased from 0.4972, 0.4973, and 0.6314 to 0.6491, 0.6846, and 0.7564; the good air quality rate increased from 47.1%, 47.9%, and 59.2%65.8%, 60%, and 78.6%. Based on the developing trend of the historical data/information and considering the impact of new energy vehicles, the no-occurrence probability of EVE in Beijing and Tianjin is predicted to be increased from 0.6888 to 0.7929 in 2020 to 0.8561 and 0.8645 in 2025. This work may provide a novel approach and perspective that can realise accurate traceability of key risk factors, quantitative risk assessment and prediction function of urban vehicle emissions for sustainable development of China's megacities.

Vehicle emissions have become one of the key pollution sources affecting air quality and human health in China's megacities. How to curb excess vehicle emissions has become a key pain point of urban air pollution prevention and control. This work tries to explore an effective integrated approach/framework to quantitatively assess the risk factors of excess vehicle emissions (EVE) and their impact on air quality for China's typical megacities. Bayesian Network is employed as the assessment tool by coupling with the Fault Tree method to curb the above problem for the first time. Four megacities (Beijing, Tianjin, Hangzhou, and Guangzhou) in China are selected as case studies, and the risk factors leading to EVE are identified to construct the Bayesian Network model. At the same time, some accurate quantisation algorithms of the occurrence probability of root nodes were proposed for the target megacities from 2014 to 2019. The analysis results show that the variation trend of the probability of EVE has a good positive correlation with the variation trend of air quality in some megacities. From 2014 to 2019, the no-occurrence probability of EVE in Beijing, Tianjin, and Hangzhou increased from 0.4972, 0.4973, and 0.6314 to 0.6491, 0.6846, and 0.7564; the good air quality rate increased from 47.1%, 47.9%, and 59.2%65.8%, 60%, and 78.6%. Based on the developing trend of the historical data/information and considering the impact of new energy vehicles, the no-occurrence probability of EVE in Beijing and Tianjin is predicted to be increased from 0.6888 to 0.7929 in 2020 to 0.8561 and 0.8645 in 2025. This work may provide a novel approach and perspective that can realise accurate traceability of key risk factors, quantitative risk assessment and prediction function of urban vehicle emissions for sustainable development of China's megacities.

Air pollution; Bayesian network; Fault tree; Impact assessment; Traffic-related; Air pollution control; Bayesian networks; Fossil fuels; Quality control; Risk assessment; Vehicles; Bayesia n networks; Fault tree method; Fault-trees; Impact assessments; Megacities; Occurrence probability; Risk factors; Tianjin; Traffic-related; Vehicle emission

10.01.2023

ELSEVIER SCI LTDTHE BOULEVARD, LANGFORD LANE, KIDLINGTON, OXFORD OX5 1GB, OXON, ENGLAND

ELSEVIER SCI LTDTHE BOULEVARD, LANGFORD LANE, KIDLINGTON, OXFORD OX5 1GB, OXON, ENGLAND

0959-6526

383

1

13