Sistema de monitoramento da qualidade do ar por meio de micro sensores aplicado ao conceito de cidade inteligente

Abstract

This project aimed to develop a low-cost monitoring system for environmental variables using micro sensors, calibrated with reference instruments and real-time data available online. The system used Arduino® technology and internet communication, with micro sensors for air temperature, relative humidity and fine particles (diameter < 2,5 m). All the calibration methods followed the recommendations of the US Environmental Protection Agency, developed for air quality monitoring using low cost sensors. Performance features such as linearity of response and precision of measurement were evaluated. After validating the data, the platform was installed in field with real-time data streaming for three places in Londrina (PR) on a web platform. The air temperature and relative humidity sensors showed high correlations with the reference sensor (coefficient of determination R2 = 1 and 0.98, respectively) and good stability during operation. These data also showed low standard deviation and variation coefficient values (between 2% and 8%). R2 for the MP2,5 sensor ranged from 0,85 to 0,95 for the different sensors. The variation coefficient dropped as particle concentration increased, showing that the sensors were not very accurate for measurements of low concentrations (<50 g.m-3). All variables showed low bias after the calibration process (<2%). When deployed in the field, the air temperature and relative humidity sensors continued showing excellent behaviour compared with the reference instrument. The MP2.5 data showed a somewhat inferior outcome, as some environmental factors, such as aerosol composition and air humidity, may have affected the sensor response. The communication via wi-fi and cable with the online platform was stable and, thus, feasible for field application. The results showed that this technology is an economic and scientifically robust alternative to provide environmental data to the population and decision makers, with potential for the development of a monitoring network or to complement networks already in place. In any case, the system must always be calibrated against reference sensors following pre-established protocols.