MODELING OF PARTICULATE MATTER CONCENTRATIONS ON A CONSTRUCTION SITE BASED ON IoT MONITORING SYSTEM.

An Internet of things (IoT) based system framework integrating a distributed sensor network was implemented to collect real-time data at a construction site. Various sensors were utilized to gather data concerning particulate matter (PM_2.5 and PM₁₀ concentrations) as well as meteorological parameters - wind speed, humidity, pressure, and temperature. The real-time measurements results provide an overview of air pollution levels at the construction site, revealing its association with earth excavation work, the primary construction activity. This connection allows for better management aimed at reducing concentrations of suspended particles. Through on-site monitoring of two pollutant concentrations, this study identified that the dust levels resulting from excavation activities were relatively high. It can be concluded that earth excavation significantly impacts air quality in the construction area. While exploring the primary factors influencing construction dust concentrations, the correlations indicate that these concentrations were not significantly associated with meteorological factors. To predict PM_2.5 and PM₁₀ concentrations in the air using number of working machines and meteorological parameters as predictors, both multiple linear regression and artificial neural network models were applied. The artificial neural network model demonstrates better alignment with the measured air concentrations compared to the multiple linear regression model. The artificial neural network model demonstrated an R2 value of 0.674 for PM₁₀ and 0.618 for PM_2.5, indicating a strong predictive capability. The aim of this research, through modeling PM_2.5 and PM₁₀ concentrations in the air at the construction site is to indicate importance of the topic, especially with respect to the health of the construction site workers. [ABSTRACT FROM AUTHOR]