Environmental Impact Assessment of Urban Demolition Sites: A Tri-Point Monitoring Framework with Three-Stage Pollutant Quantification and Artificial Neural Network Predictive Modelling

Le résumé fourni par la source

Urban densification in rapidly expanding Indian cities has led to an increase in the construction and demolition of residential buildings, resulting in the release of various pollutants into the ambient atmosphere that adversely impact the health of the workers at the demolition sites and the residents near these sites. An investigation into the emissions from four demolition sites of varying sizes in Bengaluru, India, employed a three-point, three-stage monitoring approach to measure the levels of PM10 and PM2.5, SO2, NO2, CO, and noise pollution for a total period of 108 working hours at each site. The results indicate that the levels of PM10 at two of the four demolition sites exceeded the NAAQS limit of 100 μg/m³ by up to 29.97%, while the noise pollution at three of the four sites exceeded the permissible limit of 75 dB(A) for residential areas. Furthermore, a novel Composite Pollution Index was developed as a means of providing a single metric to represent the pollution levels at each demolition site. An Artificial Neural Network model for estimating the levels of PM10 at demolition sites achieved a coefficient of determination of 0.964 on the test partition of the data set, outperforming other machine learning models for the same estimation task. Additionally, a SHAP analysis of the ANN model revealed that the area of the construction site plan and the phase of demolition were the two most important variables that influenced the levels of PM10 released during demolition activities.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.