Modern machine learning methods in SAGD, PIMS and Energy Datathon Summaries

This month we have three talks. The first is an applications of machine learning in SAGD reservoirs. The Second is a summary of the recent Untapped Energy Datathon. We'll close out with a summary from the PIMS datathon.

Jamal Bajwa will present on:
SAGD (Steam-Assisted Gravity Drainage) is a popular recovery method used in reservoirs containing Bitumen. Majority of the in-situ Oil Sands projects in Alberta use SAGD as their primary recovery method. While significant work has been done in the last two decades to improve first-principles modelling of recovery from SAGD-operated reservoirs, little has been done to systematically leverage operating data to inject steam optimally. In this work, we attempt to use modern machine learning methods to create empirical models for oil production as a function of steam injection. The production models serve as an input to an overarching optimization model that recommends optimal steam distribution across a well pad. The model development process is complete and the models are currently being put into production at Nexen's Long Lake facility.

Jamal Bajwa is a graduate of the Chemical Engineering program at the University of Waterloo. He is currently a member of Nexen Energy's Engineering and Subsurface Analytics group, where his team leverages advance mathematical methods to solve complex, real-world challenges for Nexen.

Marc Boulet will present a summary on The Untapped Energy Datathon, where 5 teams chose to address one of these five topics in the oil and gas industry: social license, ageing infrastructure, health & safety, new business models, and repurposing infrastructure. Samuel Quiroga and Steven Milbradt will discuss the results of the Repurposing Infrastructure project.

Samuel Quiroga is a geophysicist with Cenovus Energy. Steven Milbradt is a technologist with Cenovus Energy. Marc Boulet is a data scientist at Cenovus and a core member of the Calgary R Group.

Ben Reeves and Cliff Sobchuk will present a summary on:
The PIMS datathon created an opportunity for Calgarians to evaluate the available City of Calgary’s Open Data Portal, and discuss ideas to enhance the infrastructure. Results supported how we use water, to improving drop-in sports to online interactive environmental games.

Ben Reeves is a core member of the PyData group and active in the data community.
Cliff Sobchuk is passionate about data literacy and is a core member of the Calgary R group.