This study assessed air quality in a semi-enclosed train station using both personal and fixed monitoring and applied eXtreme Gradient Boosting (XGBoost) modelling to identify key factors influencing pollutant levels. The investigation also examined the effects of train idling time and engine speed on fuel consumption and pollutant emissions. Personal monitoring showed the average PM2.5 concentrations of 17-137 μg/m3 during active train operations, with higher levels observed in enclosed platforms and when diesel multiple unit (DMU) trains were present. Continuous fixed monitoring captured 24-hour average concentrations of 3.18-32.56 μg/m3 for PM2.5, with DMU-dominated platforms exceeding WHO PM2.5 limits during 90% of the monitoring period. The 24-hour average NO2 levels ranged from 135.6-169.8 μg/m3, with 11-15% of hourly values in a month exceeding the EU hourly limit. The XGBoost model demonstrated a moderate fit (RMSE: 15.12 μg/m3, R2: 0.61) and identified ambient temperature and average train idling time as the most influential predictors of NO2 levels. Partial dependence analysis indicated that reducing average idling time by two minutes could reduce NO2 concentrations by approximately 2.5 μg/m3 per train. Operational analysis suggests shutting down engines when the next operation is more than 30 minutes, restarting them 15 minutes before departure, and maintaining idle speeds below 600 RPM. These measures could save over 7,000 litres of diesel per month, reduce fuel consumption by 40-45%, and lower NOx and PM emissions by around 54%, offering actionable strategies to improve station air quality.
Priyan et al. (Fri,) studied this question.