Estimating Occupational Exposures with a Multi-Hazard Sensor Network

Sensor Network;Exposure Assessment;Environmental Health & Engineering

Zuidema, ChristopherCurriero, Frank ;

Johns Hopkins University

关键词: Sensor Network;    Exposure Assessment;    Environmental Health & Engineering;   

Others  :  https://jscholarship.library.jhu.edu/bitstream/handle/1774.2/60066/ZuidemaDissertation_20180726.docx?sequence=2&isAllowed=y

瑞士|英语

来源: JOHNS HOPKINS DSpace Repository

【 摘 要 】

Problem Statement: Exposure assessment and monitoring of occupational hazards is typically performed to assess regulatory compliance, and almost exclusively relies on personal sampling or measurement. However, personal measurements, primarily conducted by industrial hygienists, can be expensive and burdensome and often suffers from a low number of samples. Motivated to overcome the limitations of personal exposure measurement, this dissertation instead proposed and investigated estimating personal exposure with a multi-hazard sensor network.Methods: In the first of three related manuscripts, we conducted a laboratory evaluation of a low-cost sensor strategy to reduce the measurement error of quantifying ozone and nitrogen dioxide concentrations with electrochemical sensors. Typical sensors for these gases are in actuality ;;oxidizing gas” sensors, detecting both ozone and nitrogen dioxide without discrimination. In the second manuscript, we reported on the long-term deployment of a multi-hazard sensor network designed for this project that included sensors for particulate matter, carbon monoxide, oxidizing gases, and noise. We assessed the space-time variability of the hazards captured by the sensor network, and the accuracy and precision of the sensor network measurements. In the third manuscript we developed a technique to derive personal exposure estimates from the sensor network, simulated facility employees while collecting personal measurements with field reference instruments, and compared the network-derived personal exposure estimates to the personal measurements.Results: In our first study, we observed measurement error for ozone was two to three times higher than for nitrogen dioxide and that ozone was progressively underestimated as the ratio of nitrogen dioxide to ozone increased. In our second study, we demonstrated the first long-term deployment of a sensor network in a manufacturing setting capable of measuring multiple hazards with a high degree of space-time resolution. The accuracy of network measurements differed among the four hazards of interest, with the median percent bias with reference to direct-reading instruments equal to 41%, 7%, 36% and 1%, for particulate matter, carbon monoxide, oxidizing gases and noise respectively. Network sensors exhibited varying degrees of precision with 95% of measurements among 3 collocated nodes within 0.23 mg/m3 for particulate matter, 0.4 ppm for carbon monoxide, 7 ppb for oxidizing gases, and 1 dBA for noise of each other. In our third study, we observed the difference and correlation between personal exposure measurements and network-derived personal exposure estimates varied greatly between the hazard under study. The best correlation was found for noise, with the Pearson correlation coefficient equal to 0.75.Conclusions: Low-cost sensors may be subject to high levels of measurement error, principally related to sensitivity, responsiveness to non-target species, and signal degradation over time. Ultimately, the success of our technique to estimate personal exposures was highly dependent on the accuracy of the sensor network’s underlying measurements. We have demonstrated that estimating personal exposure holds promise as an additional tool to be used with traditional personal measurement due to the ability to frequently and easily collect exposure data on many employees.