Proxy model-based closed loop reservoir management: A data driven approach

• Author / Creator

  Ganesh, Ajay

• Steam assisted gravity drainage (SAGD) is a widely used thermally enhanced oil recovery process in western Canada. Petroleum reservoirs are large scale distributed parameter processes from a systems and control theoretic perspective. Physics-based first principles models pose severe computational challenges in the control and optimization studies; as a solution, proxy/surrogate models which are computationally light are preferred to design the optimization based control strategy. Caprock layer is the outer layer of the reservoir containing the dynamic steam chamber from erupting out to the surrounding. Maintaining the factor of safety (FoS) within the safety limit is crucial in operating the SAGD process. This thesis provides comprehensive analysis of steam chamber and FoS to enable the development of computationally efficient proxy models based closed loop reservoir management strategy. All of the techniques presented are developed from the data collected from a first principles-based commercial reservoir and geomechanical sim-ulator CMG-STARS sequentially coupled with FLAC3D. The first proxy model predicts the caprock pressure and temperature fields using reduced-order dynamic modelling. The second performs dynamic analysis of FoS by modelling the evolution of caprock pressure clusters of high, medium and low pressure regions using graph theory and subspace modelling. System theoretic properties of these proxy models and their practical relevance is also analyzed. Next, a data-driven polynomial chaos expansion (PCE)-based proxy model is developed to provide quick and accurate estimation of caprock FoS along with the intention of propagating the uncertainty in well bottom hole pressure inputs and reservoir petrophysical parameters to the caprock FoS. The next contribution develops a static and dynamic measures for the maximum operating pressure of the SAGD process respectively from the PCE-based FoS proxy model and a model predictive controller (MPC) to achieve FoS-constrained production optimized closed loop control strategy. The final contribution analyzes the system theoretic properties viz. controllability and observability with the intention of actuator and sensor placement and also provides a method of assessment of partial actuation and in-sensor ranges.

• Subjects / Keywords

  • complex systems
  • SAGD process
  • proxy surrogate modelling
  • uncertainty quantification
  • model predictive controller
  • data driven modeling
  • structural controllability observability

• Graduation date

  Spring 2020

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-xb62-0947

• License

  Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.