Multi-Objective Optimization of the Steam Alternating Solvent (SAS) Process using Pareto-based Multi-Objective Evolutionary Algorithms

• Author / Creator

  Mayo Molina, Israel

• The Steam Alternating Solvent (SAS) process is a relatively new auspicious alternative recovery process to produce heavy oil and bitumen resources. This process consists of injecting steam and solvent (i.e. propane) alternatively using the same well configuration as the widely adopted Steam-Assisted Gravity Drainage (SAGD) process. The SAS and other solvent-based processes have gained popularity as they aim to reduce the environmental footprint by reducing water usage and Greenhouse Gas (GHG) emissions. However, to successfully apply these processes in the field, vast knowledge and a proper design of all controllable parameters that intervene in each process and their operational ranges that might conflict with multiple objectives (especially in reservoirs with heterogeneities such as shale barriers) are needed. This study proposes a robust Multi-Objective Optimization (MOO) workflow based on Pareto optimality to determine the optimal operational ranges to implement the SAS process in homogenous and various heterogeneous reservoirs. The MOO is carried out by constructing different simulation models under the following steps. First, a 2-D homogeneous reservoir model is built based on the Fort McMurray formation in the Athabasca region in Alberta, Canada. Then, for the heterogeneous case, multiple model sets superimposing shale barriers at different locations and geometries (shale proportions and lengths) are constructed and subjected to simulation to assess the impacts of heterogeneities according to those characteristics. After, a detailed sensitivity analysis is performed on the most impactful models 1) to determine the controllable operational parameter (decision variables) that impact the most in each model and 2) to select the targets (objective functions) to be optimized. Subsequently, three different Multi-Objective Evolutionary Algorithms (MOEAs) such as Multi-Objective Particle Swarm Optimization (MOPSO), Pareto Envelope-Based Selection Algorithm (PESA-II) and Strength Pareto Evolutionary Algorithm II (SPEA-II) are applied. This is to 1) obtain the Pareto optimal set of decision variables and 2) identify the most suitable algorithm for each problem. Finally, Response Surface Methodology (RSM) to build proxy models is incorporated to estimate each objective function from the chosen decision variables to reduce the computational effort. For the homogenous case, the results indicate that high propane concentration injected over short cycles, coupled with more extended steam injection, is more optimal for the first period. The bottom-hole pressure in the injector and producer should be kept low to reduce the steam and solvent injection and to allow the fluids to be produced, respectively. In contrast, lower solvent concentration and longer cycles are preferred for the second period, and higher steam injection is more optimal to achieve a higher reservoir temperature. In heterogeneous reservoirs was observed that the steam-solvent chamber growth and production profiles are highly impacted by the location and geometry of these heterogeneities. This impact, especially in the area near the wells, is more representative. Conversely, in areas away from the wells pair, just longer and thicker shale barriers are relevant; this conclusion is consistent with other processes studies such as SA-SAGD (Al-Gosayir et al., 2012). The controllable parameters in heterogeneous reservoirs such as solvent composition (i.e. %Propane,%Methane), cycle duration (when either steam or solvent are injected), bottom-hole pressure (BHP) and some production constraints such as steam trap and Bottom-Hole Gas (BHG) have a significant impact on the SAS performance. Since this is a MOO, some trade-offs and relationships among the controllable variables in the process are observed. The robust and detailed optimization workflow presented in this study accounts for multiple targets (objective functions) involving many controllable operational parameters (decision variables). Also, by using different MOEAs to optimize the process, the results might be more accurate and reliable. Thus, this study intends to give a more profound analysis of the SAS process to facilitate field-scale decisions, minimizing the risk that this new technology might have.

• Subjects / Keywords

  • SAS
  • optimization
  • algorithms
  • multi-objective
  • pareto
  • heavy oil

• Graduation date

  Spring 2022

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-yt7h-y277

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.