10th International Aerosol Conference September 2 - September 7, 2018 America's Center Convention Complex St. Louis, Missouri, USA
Abstract View
Mapping Occupational Hazards with a Multi-Hazard Monitor Network in a Heavy-Vehicle Manufacturing Facility
CHRISTOPHER ZUIDEMA, Sinan Sousan, Nima Afshar-Mohajer, Larissa Stebounova, Alyson Gray, Laura Hallett, Xiaoxing Liu, Marcus Tatum, Mitch Fitzpatrick, Oliver Stroh, Campbell Summer, Geb Thomas, Thomas Peters, Kirsten Koehler, Johns Hopkins School of Public Health
Abstract Number: 215 Working Group: Low-Cost and Portable Sensors
Abstract Sensors play an important role in the lower-accuracy/ larger sample measurement paradigm emerging in environmental health. Due to their small size, low power demands and interoperability, low-cost sensors can be deployed in collections that are spatially distributed in the environment, known as sensor or monitor networks. Although examples of these networks in the ambient environment exist in the literature, we have developed and deployed 40 multi-hazard monitors, constructed with low-cost sensors for particulate matter (SHARP GP), carbon monoxide (Alphasense CO-B4), oxidative gases (Alphasense OX-B431) and noise (developed in-house) in a wireless network in a heavy-vehicle manufacturing facility. Monitors were equipped with radio antennae that communicated with a central database and recorded hazard measurements at 5-minute intervals. Here, we report on the temporospatial measurements from the monitor network, precision of network measurements, and accuracy of network measurements with respect to field reference instruments after approximately 5 months of continuous deployment. During ‘first shift’ production periods, 1-hr mean hazard levels across all monitors for PM, CO, oxidative gases and noise ranged from 0.4-0.6 mg/m3, 2-10 ppm, 50-150 ppb, and 78-83 dBA respectively. We examined the influence of major manufacturing processes on the spatial variability of hazards by grouping monitors by the processes that surrounded them. Manufacturing processes associated with the highest hazard levels were flame cutting (PM), staging (CO), machining (oxidative gases) and machining and welding (noise). We observed clear diurnal and weekly temporal patterns in the mean level of each hazard, and daily, hazard-specific spatial patterns attributable to general manufacturing processes in the facility, for example welding, machining and laser cutting. Monitors exhibited varying degrees of measurement precision across the range of hazard levels observed, and the maximum difference between 3 collocated monitors and their mean was equal to 0.4 mg/m3 for PM, 2.5 ppm for CO, 25 ppb for oxidative gases, and 2 dBA for noise. The second-order coefficient of variation (V2) for 3 collocated monitors was equal to 0.17, 0.02, 0.02, and 0.00, for PM, CO, oxidative gases and noise, respectively, at the median hazard level observed. We assessed the accuracy of monitors in the network by conducting side-by-side measurements with field reference instruments and observed the median percent bias for each hazard equal to -17.8%, -0.9%, -25% and 0.3%, for PM, CO, oxidative gases and noise respectively. This study demonstrates the long-term deployment of a multi-hazard monitor network in an industrial manufacturing setting, advances techniques to comprehensively map occupational hazards and sets the stage for using monitor networks to characterize occupational exposures on the individual level.