Files in this item



application/pdfYUEN-DISSERTATION-2018.pdf (5MB)
(no description provided)PDF


Title:Evaluation and improvement of optical remote sensing methods in monitoring particulate matter emissions and plume opacity
Author(s):Yuen, Wang Ki
Director of Research:Rood, Mark J; Koloutsou-Vakakis, Sotiria
Doctoral Committee Chair(s):Rood, Mark J
Doctoral Committee Member(s):Du, Ke; Bond, Tami C; Zhang, Yuanhui
Department / Program:Civil & Environmental Eng
Discipline:Environ Engr in Civil Engr
Degree Granting Institution:University of Illinois at Urbana-Champaign
lidar equation
mass extinction efficiency
emission factor
open burning
open detonation
USEPA Method 9
digital camera
Abstract:Atmospheric particulate matter (PM) affects human health, reduces visibility, and impacts climate. PM also causes plume opacity, which is defined as percentage of light that is attenuated by a plume. This research focuses on two optical methods that have been developed and implemented previously to measure PM mass emission factor (EF) and plume opacity. My contributions to this research are: 1) evaluating and improving the previous approaches used in these two methods, and 2) estimating their associated uncertainties to quantify PM EF and plume opacity values. In the first part of the research, fugitive PM EFs that come from unconfined flow streams are measured by the hybrid-optical remote sensing (hybrid-ORS) method. Fugitive PM EFs pose a challenge for measurements because they are aloft, heterogeneous, have short lifetimes, and can exist within large spatial scales. These challenges can be addressed by the developed hybrid-ORS method. The hybrid-ORS method that I used involves the use of micro-pulse light detection and ranging (lidar; MPL) that measures range-resolved extinction coefficients. Co-located point PM mass concentrations and extinction coefficients are measured to determine PM mass concentration from light extinction measurements provided by the MPL. Fugitive PM EFs are then obtained by integrating PM mass concentrations across the plume’s cross-section with wind data and duration of events. Two field campaigns were completed to measure fugitive PM emissions from mobile vehicles on unpaved roads, as well as open burning or detonation of energetic materials. The fugitive PM EFs from these two sources were also measured by at least one independent and concurrent method for comparison. The results show that PM EFs measured by hybrid-ORS method and other concurrent methods are not significantly different, while hybrid-ORS method offers the advantage of knowing the spatial and temporal distributions of PM mass concentration in a fugitive PM plume. In addition, two lidar equation inversion methods, namely near-end and far-end methods, were compared with respect to its PM EF and its uncertainty. The results show that the far-end method is preferable because it introduces less uncertainty, and the method is mathematically stable. In the second part of the research, plume opacity is measured using digital optical method (DOM) that was previously developed by Du (2007). This method is an improvement over the traditional human observer method in determining plume opacity. DOM was initially developed using compact digital cameras, but the method was extended with this research to smartphone cameras and a camcorder. In DOM, the response curves, which relate exposure to pixel value (PV), are determined for compact cameras, smartphone cameras, and a camcorder. Then, relative exposures of select regions within a plume picture are used to calculate plume opacity using DOM software. There are two DOM models that were previously developed by Du (2007), namely: 1) the contrast model that requires a plume passing in front of and near one of two co-located contrasting backgrounds; and 2) the transmission model that requires a plume passing in front of and near one background, and a diffusive scattering parameter (K) that characterizes the optical property of the plume relative to its background. A field campaign was performed to use compact cameras, smartphone cameras, and a camcorder to measure the opacity of plumes emitted from a smokestack using DOM. The smokestack included a transmissometer inside, which provided independent opacity measurements used as a standard for comparison with the devices using DOM. In this research, a new method was developed for calibration of cameras and a camcorder that uses exposure value (EV) compensation. The results for the compact and smartphone cameras show that: 1) the resulting opacity values are not significantly different from those determined by two previous calibration methods; 2) the color contrast between two backgrounds is the most important variable affecting the uncertainties of opacity measured by the two DOM models; and 3) empirically determined K values for select background and plume color combinations show that K value depends on wavelength of background color. The results for the camcorder show that: 1) the camcorder can accurately measure opacity values in real-time (1 Hz); 2) increasing color contrast between two backgrounds using, DOM contrast model, decreases opacity measurement errors and uncertainties; and 3) background choice is more important than camcorder calibration and number of sampled pixels in determining the opacity measurement uncertainty. This research is novel and significant by providing for improvements and uncertainty analyses of two remote sensing methods for PM. Evaluations of the hybrid-ORS method and DOM are done for method improvements (e.g., increase flexibility and understanding of the applicability of the hybrid-ORS method and the transmission model; demonstrate that DOM can be applied to smartphone cameras and a camcorder aside from compact cameras). The uncertainty analyses identify the major sources of measurement uncertainties and quantify the overall uncertainty for future applications of these methods.
Issue Date:2018-04-16
Rights Information:Copyright 2018 Wangki Yuen
Date Available in IDEALS:2018-09-04
Date Deposited:2018-05

This item appears in the following Collection(s)

Item Statistics