Mobile Sensors: Assessment of Fugitive Methane Emissions from Near and Far-Field Sources

• Main Author: Foster-Wittig, Tierney
• Other Authors: Albertson, John D
• Published: 2015
• Subjects: Civil engineering; landfills; methane; natural gas; oil; sensors
• Online Access: http://hdl.handle.net/10161/9893

<p>The primary focus of this dissertation is on the assessment of fugitive methane emissions from near and far-field sources. Methane is the second most prevalent greenhouse gas (GHG) emitted in the United States from anthropogenic activities. Due to measurement and model limitations, there is not an accurate assessment of how much methane in the atmosphere is due to anthropogenic sources. This dissertation focuses on measuring the methane emissions from two of the three largest anthropogenic sources -- landfills and natural gas systems. All measurements are made with a single fixed or single mobile sensor. Methods are developed to assess the source strength for both near (i.e. natural gas) and far-field (i.e. landfill) sources using either the fixed or mobile sensor. </p><p> </p><p>For far-field measurements, a standardized version of a mobile tracer correlation measurement method was developed and used for assessment of methane emissions from 15 landfills in 56 field deployments from 2009 to 2013. A total of 1876 mobile tracer correlation measurement transects were attempted over 131 field sampling days. </p><p>Transects were analyzed using signal to noise ratio, plume correlation, and emission rate difference method quality indicators. The application of the method quality indicators yield 456 transects (33\%) that pass data acceptance criteria. </p><p>For near-field sources, techniques are developed for 1) fixed sensors sampling through time downwind of a source and 2) mobile sensors passing across plumes downwind of a source. For the fixed sensor, the lateral plume geometry is reconstructed from the fluctuating wind direction using a derived relationship between the wind direction and crosswind plume position. The crosswind plume spread is estimated with two different methods (modeled and observed), and subsequently used a Gaussian plume inversion to estimate the source strengths. For the fixed sensor, the sensor takes measurements for about 20 minutes and we are able to reconstruct the ensemble average of the plume. </p><p>For the mobile sensor, the vehicle drives through the plume in the crosswind direction. </p><p>The measurements show the lateral plume geometry of an instantaneous plume. The instantaneous plume has a narrowed Gaussian structure. </p><p>Two techniques are tested using data from controlled methane release experiments; these two techniques are 1) linear-squares and 2) a probabilistic approach. For the probabilistic approach, Bayesian inference tools are applied and special attention is paid to the relevant likelihood functions for both short time averaged concentrations from a single fixed sensor and spatial transects of instantaneous concentration measurements from a mobile sensor. The two techniques are also tested on measurements downwind of multiple natural gas production facilities in Wyoming for the fixed sensor and in Colorado for the moving sensor. The results for both the fixed and mobile techniques show promise for use with gas sensors on industry work trucks, opportunistically providing surveillance over a region of well pads.</p> === Dissertation