Machine Learning Analysis of Hydraulic Fracturing Operations and Susceptibility to Induced Seismicity - Duvernay Formation

Publication Type
Presentation
Topic
Oil and Gas
Publication ID
PRS 2022-002
Publication ID Extended
Presentation 2022-002
Publication
Abstract

A review of hydraulic fracturing operational data submitted to the Alberta Energy Regulator (AER) was undertaken for wells within the Duvernay Formation.  A data science approach was used to investigate whether any operational parameters might influence a well’s seismogenic response, when evaluated in concert with factors that describe the regional geological setting. The ability to understand whether any specific combinations of operational parameters show a relationship to the seismogenic state of an operation and/or resulting earthquake magnitudes is important to allow safe well operations to be planned and mitigations to be incorporated in regions where the geological susceptibility to induced seismicity may be higher. The study area includes operational hydraulic fracture data and well injection information currently residing within AER databases specific to AER Subsurface Order No. 2 region and includes only well drilling and completion activities within the Duvernay Formation.

The analysis results indicate that the geological seismic susceptibility varies spatially within the Subsurface Order No. 2 region, with geologically susceptible areas being delineated by factors that act as proxies for structurally influenced areas. However, within these geologically susceptible areas, the results indicate that operational parameters relating to injection pressures, fluid volumes, and wellbore azimuths are also associated with an increased probability of inducing seismic events. Furthermore, the empirical relationships between the operational parameters and seismogenic state vary spatially and are dependent on the geological subsurface conditions.

The presentation provides a review of the data used in the analysis along with the analytical workflow and presents interactions between operational and subsurface conditions that might influence seismogenic behaviour within the Duvernay Formation and could potentially be informative from a mitigative perspective.

The AER is in a unique data rich situation, and this study demonstrates the use of information to assist in the execution of the AER mandate; the safe efficient, orderly, and environmentally responsible development of oil, oil sands, natural gas, and coal resources over their entire life cycle.

Event
Society of Petroleum Engineers Workshop: Montney and Duvernay: Poised to Capitalize Upon Opportunity
Place Keywords
alberta, canada, fox creek
Place Keywords NTS
83e, 83f, 83g, 83j, 83k, 83l
Theme Keywords
data science, duvernay formation, machine learning, seismicity