Loading...

Development of Multi-objective Optimization Framework for Non-vertical Well Placement

Rostamian, Aref | 2017

549 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: English
  3. Document No: 49572 (66)
  4. University: Sharif University of Technology, International Campus, Kish Island
  5. Department: Science and Engineering
  6. Advisor(s): Bazargan, Mohammad; Jamshidi, Saeed
  7. Abstract:
  8. An important step in the early stages of an oilfield development is the accurate identification of the location of production and injection wells which has a significant impact on productivity and profitability of the reservoir. Recently, operation companies in petroleum industry focus on production optimization, recovery enhancement and cost reduction; as a result, application of non-conventional wells (NCW) or non-vertical wells have dramatically increased. Well placement optimization is a complicated task as it is a function of several contributing factors including reservoir heterogeneities and economic constraints. The complexity of this task increases when it is considered as a multi-objective non-vertical problem rather than a single-objective vertical one as it raises the evaluation time. In this work, four mating procedures, as a part of a multi-objective optimization algorithm utilized for this purpose, are examined in order to improve the algorithm’s performance in terms of the computational cost. Here, for the first time, Similarity-Based Tournament selection method is applied in well placement problem which has been proven to be a viable alternative to conventional mating arrangements.In this work, the Non-dominated Sorting Genetic Algorithm-II (NSGA-II), is exploited as the main search engine to optimize Net Present Value and the Recovery Factor as the objective functions. One of the major drawbacks of this algorithm is its population-based nature which leads to higher computational time to evaluate a Pareto front, the primary purpose of this work is to modify the mating scheme in NSGA-II. Four different mating procedures, i.e., (i) Random, (ii) Tournament, (iii) Roulette Wheel and (iv) Similarity-Based Tournament selection mating schemes are implemented and examined to enhance the performance of NSGA-II for well placement problem in terms of the computational efficiency. As a major result of this study, it is revealed that among the mating schemes evaluated, Similarity-Based Tournament selection method, which is utilized here for the first time, demonstrates a substantial reduction in the number of the generation in vertical well placement which is needed to create final solution population (Pareto front). Convergence analysis of this technique indicates that this strategy dramatically decreases the time and number of function evaluations by 40 and 150 percent compared to tournament and random selection methods, respectively.Horizontal well placement and control optimization is investigated in this thesis. Three mating procedure applied to improve NSGA-II efficiency and power to create final Pareto front in this case. This framework is applied on a reservoir case study where the results shows that the optimization approach for each mating techniques have several advantages and disadvantages.Finally, to handle model uncertainty, NSGA-II with a Net Present Value-Recovery Factor approach is implemented to seek robust solutions for well placement, where the robustness is considered for the geological uncertainty. To assist the optimization process, random selection of different mating procedure in each generation is applied in the multi-objective optimization algorithm. This workflow is demonstrated on a reservoir case study where the results indicate that the optimization approach leads to improved decision making capabilities by providing a suite of well planning solutions. Due to convenient modification of the mating arrangement of NSGA-II, the possibility of improving the convergence speed to Pareto front has been the main motive of this work and, in fact, the easy-to-apply nature of NSGA-II makes it an excellent candidate for multi-objective problems, as it has been recently exploited extensively for petroleum industry applications. As a result, Similarity-Based Tournament selection method, implemented in this work, is proposed as an alternative to conventional mating arrangements for NSGA-II. Furthermore, the sensitivity and convergence analysis of NSGA-II which has been done in this work is one of the first published works in this area and, indeed, is very scarcely found in the open literature
  9. Keywords:
  10. Well Placement ; Multiobjective Optimization ; Non-Dominate Sorting Genetic Algorithm (NSGAII) Method ; Genetic Algorithm ; Particles Swarm Optimization (PSO)

 Digital Object List

 Bookmark