Worker exposure to airborne particles in industrial settings: evaluation of exposure assessment and modelling tool

Abstract

[eng] Exposure to particulate matter in work environments has been linked to ischemic heart, cardiovascular and respiratory-related disease risk increase due to inhalation. Increased adverse health effects have been linked to nanoparticles (< 100 nm) due to their ability to reach the deepest sections of the respiratory tract and their longer retention time. Exposure monitoring is widely used method to assess worker exposure to airborne particles. However, other prediction tools have been explored such as the use of the dustiness index, mass-balance models, and health risk assessment tools. Discussions regarding the use and application of the latter tools are ongoing due to their relatively novelty for worker exposure assessment, the need to test their performance under real- world scenarios, and the need to understand the uncertainties related to critical parameters and limitations. The main objectives of this PhD Thesis are to 1) assess worker exposure to particles (4 nm - 35 μm) in ceramic industry real-world workplace scenarios; 2) evaluate currently used exposure assessment metrics and decision-making approaches; 3) understand the relationship between material dustiness and worker exposure; 4) evaluate the performance of mass-balance models, and 5) compare health risk assessment tools. Worker exposure was assessed during mechanical handling of powders in 6 different scenarios and for 15 materials as well as thermal spraying of ceramic coatings. Exposure monitoring was conducted using online and offline instruments which allowed for the characterization of particle mass and number concentrations, particle size and size distribution, particle morphology and chemical composition. In addition, some of these scenarios were also selected to assess relationship between the dustiness index and exposure concentrations as well as the ability of different particle metrics to represent worker exposure. Finally, decision making approaches, and the performance of mass- balance models and risk assessment tools were tested. Results evidenced clear impacts of industrial activities on workplace exposure to coarse, fine and nanoparticles.

Significant increases of inhalable and respirable particle mass concentrations (inhalable mass concentration 80-4000 μg m-3) were observed during mechanical handling of raw materials (d50 2.7-40 µm), when compared to background concentrations. The highest mean inhalable mass concentration (3700 μg m-3) was monitored during packing of ceramic materials, when mitigation strategies were inefficiently implemented.

Conversely, particle number concentrations were not influenced by mechanical handling of powders, but by driving of diesel-powdered forklifts, leading to concentrations up to 70000 cm-3. Thermal spraying, on the other hand, increased particle number concentration up to 105 cm-3 in the worker area. After the application of the ICRP respiratory tract deposition model, airborne particles in the workplaces studied were seen to deposit mainly in the alveolar region (51-64%) during packing of powder materials and (54-70%) during thermal spraying by means of surface area. Source enclosure and modification of the energy settings were pointed out as useful strategies to minimize worker exposure.

The validity, performance and comparability of tools for exposure assessment were evaluated. Several decision-making approaches were tested to determine statistically significant impacts on exposure. Among them, the ARIMA models were seen to be the least conservative while the nanoGEM approach confirmed its usefulness for particle number but slightly underestimated exposure for particle mass concentrations when compared to traditional statistical tests. High degree of correlation was found between dustiness and measured exposure concentrations during mechanical handling of powders in a pilot plant (R2 up to 0.97) and at industrial scale (R2 up to 0.80). This correlation was stronger when material characteristics dominated over process characteristics, and an adequate methodology is applied, using the dustiness method which best mimics the activity under study.

Finally, one- and two-box models were used to model particles under high and low concentrations in terms of mass and particle number concentrations. Ratios between modelled and measured concentrations were 0.82-1.22 when modelling inhalable particle mass in the mechanical handling scenario, whereas ratios of 0.2-0.7 were obtained when modelling thermal spraying particles. Thus, model performance was poorer for the high nanoparticle concentration scenario. The addition of background and outdoor concentrations as input improved model performance. Risk assessment and control banding tools (ART, Stoffenmanager and NanoSafer) were tested for the scenarios under study, and it was concluded that the mechanical processes were estimated with higher accuracy and lower variability by Stoffenmanager (64% of the cases). Conversely, ART and NanoSafer showed higher flexibility for introducing more case-specific input data. A clear need for harmonization between risk assessment tools was evidenced.

[cat] dels materials en planta pilot (R2 fins a 0.97) i en escala industrial (R2 fins a 0.80). Aquesta correlació s’ha vist reforçada quan les característiques dels materials dominen sobre les característiques del procés, i quan una metodologia adequada per a la determinació del poder d’emissió de pols va ser utilitzada, aquella que millor simula el procés d’estudi. Finalment, es van utilitzar els models basats en balanços de massa de una i dos caixes per a modelar dos escenaris diferents, un de baixa exposició a partícules grolleres (manipulació mecànica, ensacat de materials) y un altre d’altes concentracions de numero de partícules (projecció tèrmica). Les proporcions entre les concentracions modelades/mesurades van ser de 0.82-1.22 per a la modelització de la fracció màssica de pols inahalble en el cas de la manipulació mecànica, mentre que les proporcions van ser de 0.2-0.7 per a la modelització de la concentració en numero de partícules durant la projecció tèrmica. Per tant, la bondat dels models basats en balanços de massa va ser menor en l’escenari d’altes concentracions de numero de partícules. Al considerar les concentracions de fons i de l’aire ambient en els models, la bondat i la precisió dels models va augmentar. De les tres eines per a l’avaluació de riscos per exposició a agents químics (ART, Stoffenmanager i NanoSafer) que van ser provades, les concentracions mesurades durant els processos de manipulació mecànica van ser estimats amb una major precisió i menor variabilitat per Stoffenmanager (64% dels cases). Per el contrari, ART i NanoSafer van mostrar tenir una major flexibilitat per a introduir dades d’entrada especifiques per a cada cas.