Date of Award

10-30-2020

Publication Type

Master Thesis

Degree Name

M.Sc.

Department

Chemistry and Biochemistry

Keywords

abandoned wells, gas migration, greenhouse gases, Kmeans, oil and gas, principal component analysis

Supervisor

Scott O. C. Mundle

Rights

info:eu-repo/semantics/openAccess

Abstract

Fugitive releases from natural gas wells are a persistent issue in the oil and gas sector and comprises 27% of Canada’s greenhouse gas emissions. Natural gas within this sector accounts for 44% of Canada’s methane releases and 70% of Alberta’s. Releases from wells are documented; however, knowledge gaps persist for abandoned assets. When fugitive gases are suspected, regulatory standards require gas migration testing. This thesis presents the beginnings of developing ‘best practices’ in testing recommendations to better estimate emissions from abandoned wells. Testing requires detection of stray gases utilizing a worker-safety portable handheld multi-gas monitor; however, our work shows this monitor has limited application in gas migration testing. Portable monitors are equipped with non-specific, catalytic combustion sensors that underestimate methane concentrations in the subsurface. To circumvent misleading results, we suggest reporting oxygen levels for subsurface gases or the use of more sophisticated detectors. Additionally, work enclosed addresses single-sample, or sample-to-sample, risk assessments for gas migration testing. A brief commentary on previous testing at an abandoned well site in Western Canada reveals how this approach often produces insufficient evidence of stray gases. In applying a multivariate risk assessment method, using principal component analysis and K-means clustering, we showed sample sizes >20 for reporting gas compositions and >10 for stable isotopes will accurately detect stray gases at an abandoned well. Best practices highlighted in each study can easily be integrated into testing recommendations that will assist Canada in reducing emissions by 45% in 2025.

Share

COinS