Desenvolvimento de metodologia para monitoramento contínuo da concentração de emissão de PM1 em fontes estacionárias industriais

Abstract

Among the processes that generate emissions of ultrafine particles (UFP), which are known by their high potential to cause brain diseases when inhaled, the prilling towers have a significant contribution, as those processes use flow rates up to 10 times higher than other dryers. From the monitoring perspective, UFP are generally unstable and volatile (as the case of sulfates and nitrates) making its mass concentration determination inaccurate, especially if continuous mass monitoring is intended to be installed. Also, the environmental agencies have demanded that particulate matter from the stacks must be continuously monitored, with the standard procedure for calibration being the gravimetric sampling. In this context, this research has evaluated the performance of two continuous sensors on measuring UFP mass concentration in an industrial fertilizer prilling tower. The sensors tested were electrodynamic diffusion charger (EDA) and optical scattering (OSA). The analyzers were installed at same stack where gravimetric sampling (GS) was performed to find calibration factors. Data was collected over four seasons in Northern Europe, including the particulate matter from the analyzers, meteorological and process parameters. A controlled experiment where flows and meteorological parameters were kept stable was run to test electrical low-pressure cascade impactor (ELPI) performance compared to the other sensors. The results show that OSA follows the process changes and the calibration factors obtained from GS varied on 0.5 standard deviations over products and seasons. EDA represents better sudden variations up to its operating range, but the calibration factors measured were over three standard deviation, independent of seasons and products. EDA also presented an artificially high emission due to intermittent steam injection in the stream. Dust deposition on the sensors affects the reading on both analyzers but only OSA indicates it. While modelling the data collected over the period, it was found that EDA was neither correlated with process nor meteorological parameters (r-squared less than 10%) what can be caused by particles not being charged evenly or droplets read as particles. The OSA concentration model showed r-squared of 45% and strong correlation with meteorological parameters and raw material flow rates. The model presented a standard error of 0.21 mg/Nm3. OSA has potential to be employed for ultrafine particles monitoring if the influence of particle characteristics under industrial operation is considered and meteorological parameters are included, as already in practice for ultrafine particles monitoring outdoors. Finally the linear regression models built based on the methodology here proposed, showed that the commercial sensors did not have enough precision while ELPI analyzer output presented a model with adjusted r-squared of 94.13% and a standard error of 0.02 mg/m3 not including any air stream parameter and detecting changes in the process that allow optimization of those to reduce the UF emission. The ELPI shows potential to be used as calibration device to other in-situ continuous dust monitoring when dealing with UFP, once gravimetric sampling has reached its detection limit for UFP. By applying the methodology here proposed there is potential to install and calibrate continuous sensors at stationary sources emitting UFP with ELPI as well as optimizing the process to reduce the concentration of UFP and by doing so, decreasing the potential impact of those in the environment and on the human health.