Upcoming Division Event
Leveraging Machine Learning for Building Seismically-Constrained Reservoir Models: A Permian Case Study
Speaker: Dennis Ellison, Emerson
Time: 12:00 - 1:00 pm, Thursday, Nov. 28, 2019
Location: Palliser Conference Centre on the 2nd floor, 298, 115 9th Ave SE
*CSPG members can register for free and track their CPD hours!
To evaluate the quality of the reservoir and gain a more realistic measure of its heterogeneity before drilling, geoscientists require accurate, calibrated reservoir models and maps to better understand rock type and fluid distribution and forecast production potential. Typically, these reservoir models are created from wireline and core data. Ideally, these reservoir models are digital representations of the subsurface which include seismic constraints for structure, lithology, and fluids.
An effective tool for meeting this challenge is Democratic Neural-Network Association (DNNA), and Emerson-developed Rock-Type Classification method based on a Supervised Machine Learning approach. This method reconciles seismic data with well data to predict formations, fluids, and structure away from the wellbore. It employs an ensemble of neural-networks running in parallel which simultaneously learn from the multi-resolution wellbore and seismic data using different strategies and associations. This machine learning architecture minimizes the possibility of biasing the results. DNNA includes a secondary training stage where seismic data is introduced away from the wellbore and “voted” on for training set inclusion. This “voting” step stabilizes the learning process while preventing overfitting.
The aim is to generate a seismically constrained probabilistic facies model from electrofacies models, and preferably core data. The strength of this method is its ability to integrate different data types (core, wireline, seismic) and resolutions. This technology reveals new insights into the seismic data reliability for predicting subsurface reservoir models away from well locations.
In his role as Technical Advisor – Geophysics, Dennis Ellison helps producers get more from their data by leveraging next generation technology. His career started in depth imaging of geologically complex land data and transitioned into Reservoir Characterization and Quantitative Interpretation, focusing on unconventional geological and geomechanical property prediction. Dennis actively volunteers with ENERGYminute, CSEG, and Scouts Canada. Earlier this year, the ENERGYminute won the Community Engagement and Educational Outreach Award with the SPE – Calgary Section, and in 2017, he received awards from GeoConvention and the CSEG. He is a member of SEG, CSEG, APEGA, EAGE, and SPE.