Fenceline and Community Sensor Applications and Comparisons
MacDonald, M., C. Smith, E. Thoma, I. George, R. Duvall, P. Deshmukh, AND A. Scott. Fenceline and Community Sensor Applications and Comparisons. To be Presented at Air Sensors International Conference (ASIC) Virtual Meeting Series, Session Title “Fenceline Monitoring with Low-Cost Sensors”, N/A, N/A, May 18, 2021.
This public abstract and presentation (separately cleared) will advance the topic of lower-cost fenceline and community VOC sensors. Background: Energy production operations, refineries, chemical plants, and other industries and waste facilities can emit air pollutants and odorous compounds from fugitive leaks, process malfunctions, and area sources that are hard to detect and manage. From the shared perspective of industries, regulators, and communities, improved understanding of stochastic industrial sources (SISs) can yield many benefits such as safer working environments, cost savings through reduced product loss, lower airshed impacts, and improved community relations. Under its next generation emissions measurement (NGEM) program, ORD CEMM is working with a range of partners to develop and test NGEM tools that can assist facilities in detection and management of SISs. The fenceline sensors described in this work represent one type of these emerging NGEM tools.
The following is an abstract supporting a presentation (separately cleared) to be given at the Air Sensors International Conference (ASIC) Virtual Meeting Series, in the session titled “Fenceline Monitoring with Low-Cost Sensors” on May 18, 2021. In December 2019, the U.S. Environmental Protection Agency (EPA) conducted a 21-day sensor collocation study near a chemical facility in Louisville, KY. The study compared prototype sensors measuring volatile organic compounds (VOCs) for use in either “near-source” fenceline or “away from source” community monitoring applications. A pair of EPA SPod fenceline sensors along with three other pairs of sensors built with 10.6 eV photoionization detectors (PIDs) and one pair based on metal oxide sensors (MOSs) were deployed using manufacturer-applied calibrations and no temperature or relative humidity (RH) drift correction algorithms were utilized. An additional pair of EPA fenceline sensors malfunctioned during the deployment and are not described. The EPA SPods and one other pair of PID sensors used electrical resistance sensor heating strategies to help stabilize temperature and RH-induced baseline drift. The data were processed with and without an EPA baseline drift correction approach used for fenceline applications to separate emission source plumes detected at short time averaging intervals (i.e. seconds) from longer-term trends including drift and air shed signal. The data were corrected for offset to lowest zero and the EPA SPods fenceline counts were normalized to averge peak value of one of the sensor sets for direct comparison to ppb values. Because the “fenceline signal” was not particularly strong during the collocated comparison study, select data from a long-term deployment of EPA SPod fenceline sensors containing two different brand PIDs are also discussed to provide additional context. The collocation study found significant differences in VOC concentrations reported by the units, including those produced by the same manufacturer. Simultaneous meteorological data showed that these differences could not be attributed to the effects of temperature and RH alone. Without EPA fenceline drift correction, the relative percent difference (RPD) in the 21-day integrated mean VOC signal of same-manufacture paris varied from 13% to 95%, with the best unit to unit precision observed among the MOS sensor and the worst from two different prototype variations from the same maker. Across sensors, the mean offset-corrected concentration ranged broadly (36 ppbv to 173 ppbv). Poor unit to unit precision is problematic for community monitoring applications adding to calibration, interference bias, and speciated response uncertainties. Use of the EPA baseline correction algorithm improved unit to unit RPD to within +/- 30% in most cases but may inadvertently remove non-plume air shed VOC signal, as indicated by the overall mean range (7 ppbv to 27 ppbv) for the baseline-corrected set. Examples of effects of data processing will be provided along with an example of a fenceline plume and evacuated canister grab sample acquisition from the long-term study. The role of sensor time resolution in fenceline applications will be discussed.