Speaking at the MATLAB Energy Conference 2021, Paul Pastusek and Gregory Payette presented the Open Source Drilling Community, an industry and academia coalition that sets out to support an open-source effort for drilling software and encourage the reuse of ‘ever-improving’ models and code. OSDC span out of DSATS, the Society of Petroleum Engineers Drilling Systems Automation Technical Section in 2019. After a covid-delayed start, the web and GitHub sites opened to the industry with a ‘soft launch’ in September 2021.
Today’s commercial models are often opaque while many academic efforts are of limited scope. Models developed under joint industry projects may not be published, and there is a general need for better math and code verification and model validation with case studies. ExxonMobil’s own codebase seeded the initiative and now additional models have been submitted by Scientific Drilling, NORCE, Texas A&M, and the University of Calgary. MathWorks has helped convert the initial ExxonMobil code to MathWorks’ Simulink, improving code stability and execution speed and documenting the models. All models, data, and test cases are freely available for academic and commercial use.
The initiative favors understanding over prediction through full physics modeling. ‘Modeling the drilling process allows us to understand the physics driving our systems’. Proposed tools and procedures can be tested without the time and risks of rig trials. ‘In the near future, it will be inconceivable to put a new tool in the ground or new control system on a rig without fully testing the system for performance and stability’
OSDC provides component models of the complete drilling system, including topsides, coupled drill string dynamics, hydraulics and bit-rock interaction. While simple reduced-order models that ignore couplings may be used in some scenarios, integrated component models are required for comprehensive dynamic modeling of the drilling system. OSDC is working on best practices for the integration of different models with varying time scales and/or spatial sampling.
The open source paradigm has been chosen to enhance collaboration and transparency and to eliminate ‘reinventing the wheel’. There is also an expectation that subject matter experts will be able to create ‘focused models’ for component systems such as the autodriller, bits and motors. It is early days for OSDC, right now at a juvenile V 0.1 release. Currently-available models include a transient torsional drill string model and a yield power law fluid calibration model. The initiative also pulls in prebuilt libraries from the Matlab Drilling Toolbox and PyDrill, the Python Drilling Simulation Library. The University of Calgary is coordinating the organization’s web and GitHub sites. More from OSDC.
© Oil IT Journal - all rights reserved.