No abstract

Determining an optimal location and configuration of well to be drilled is a critical reservoir development decisions as it can cost millions of dollars and determine the volume of hydrocarbons being produced. This is a very challenging task due to large number of decision variables involved. One important factor to consider in field development is scale deposition. It can cause production problem and reduce hydrocarbons recovered. In the extreme case, it can cause a production well to be abandoned as a result of formation damage in the near wellbore or narrowing of the production tubing. Due to the potential high cost of scale management, scale risk deposition would be an important issue. The aim of the paper is to optimize the placement of new wells in a field development accounting for the risk of sulphate scale deposition. This scale is formed when sulphate rich seawater, usually used as injection fluid to maintain reservoir pressure, mixes with barium rich formation brines. Multi-Objective Optimization (MOO) is used to find a range of solutions by maximizing oil recovery and minimizing scale risk. Scale risk deposition is modelled by tracking the injected seawater concentration, which acts as a natural tracer, in the produced water from each well. Penalty parameter is introduced in the optimization workflow and quantified by relating the scale risk impact in the reservoir production to scale treatment (bullhead-squeeze) as the mitigation. The study used a modified version of the benchmark reservoir model, PUNQ-S3; the model was adapted to be developed using seawater flooding, and as a result sulphate scale deposition is of a major concern. Particle Swarm Optimization (PSO), a stochastic population-based and gradient-free optimization algorithm, is applied to search over the new well locations and their completion zones in the presence of potential scale deposition risk. The paper compares a conventional field optimization with objective only to maximize oil production to the one with accounting scale risk as extra information. The paper demonstrates that multi-objective optimization allows us to identify well locations and completion zones with high recovery and yet low scale risk, leading to significant economic benefit in a number of field development scenarios.

Assisted history matching (AHM) can be approached either by single objective (SO) or multi-objective (MO) optimisation. AHM generates multiple matched-models, which are then used to quantify uncertainty of the forecast. The goals in this framework are not only to produce matched-models in an efficient way but also to provide a reliable forecast under uncertainty. Choice of the model paramaterisation is one of the sources of uncertainty in reservoir modelling. This is usually done based on geological information and engineering knowledge.This paper explores the impact of different geological model parameterisations on AHM and its impact on the reservoir forecasting. It compares the performance of SO and MO AHM approaches under different model parameterisation. In the MO approach, different choices of objective grouping were studied. Three performance measures in AHM and forecasting will be discussed, namely: diversity, convergence, and forecast reliability. Diversity is defined as set of models that are different but still match history. Convergence is defined as how fast history matching achieves the desirable match. Reliability, for synthetic case study in the present paper, is defined as encapsulation of a simulated truth case into the probabilistic range of forecasting (P10 -P50 -P90).Study on a standard industry benchmark case in this paper shows that MO optimisation approach provides a more diverse set of matched-models, which leads to a better forecast. It also confirms that a MO approach is more robust and reliable in forecasting than SO under different model parameterisation. Different objective grouping choices in MO approach still give a reliable reservoir forecasting, although variability in convergence rate is seen.

Summary Multiobjective history matching has gained popularity in the last decade. It provides an ensemble of diverse set and good matched models that should lead to improved forecasting. Moreover, in some cases, multiobjective history matching provides faster and more-robust convergence than the single-objective approach. In multiobjective, objective components (usually groups of them) guide the algorithm to different areas of objective space that lead to a diverse set of optimal solutions. These algorithms are widely established and well-developed for problems with two or three objectives. Under an increasing number of objective components, such as in a reservoir model with multiple wells and production data, multiobjective-history-matching performance (convergence speed and match quality) can deteriorate. One effective approach is grouping objective components to reduce the number of objectives. However, the existing literature does not present sufficient information on appropriate grouping techniques and ways of combining objective components. We present a novel technique to group the objective components depending on analysis of the nonparametric-conflict information obtained from a set with a limited number of initial solutions. By grouping the objectives depending on the conflict between them, we aim to achieve better performance in history matching. We apply this framework to history matching of an industry-standard reservoir model and a real-field case study. We also perform history-matching runs of groupings with different degree of conflict, and then analyze the performance among them with the statistical-significance test. Our extensive simulation results show that the proposed conflict-based strategy can be used as a guideline to help select a grouping of the objective components in multiobjective history matching optimally. By calculating the conflict between objectives a priori, we can identify which grouping scheme will result in a better performance. This technique can significantly improve the fitness quality of the matched model given the same number of flow simulations, and can also obtain a diverse set of models.

No abstract