Well placement optimization using imperialist competitive algorithm

An efficient and optimized field development plan is a crucial and primary aspect in maximizing well productivities, improving the recovery factors of oil and gas fields, and thus, increasing profitability most effectively. In this paper, we apply a metaheuristics algorithm known as the Imperialist Competitive Algorithm (ICA) to determine optimal well location for maximum well productivity. The ICA is an evolutionary algorithm that mimics socio-political imperialist competition. This algorithm uses an initial population that consists of colonies and imperialists that are assigned to several empires. The empires then compete with each other, which cause the weak empires to collapse and the powerful empires to dominate and overtake their colonies. We compared the ICA performance with that of the particle swarm optimization (PSO) as well as the genetic algorithm (GA) in the following four optimization scenarios: 1) a vertical well in a channeled reservoir, 2) a horizontal well in a channeled reservoir, 3) placement of multiple vertical wells, and 4) placement of multiple horizontal wells. In all four scenarios, the ICA achieved a better solution than the PSO and GA in a fixed number of simulation runs. In addition, we conducted sensitivity analyses for three important parameters (revolution ratio, assimilation coefficient, and assimilation angle), and the results of these analyses showed that the recommended ICA default parameters generally led to acceptable performances in our examples. However, to obtain optimum performance, we recommend tuning the three main ICA parameters for specific optimization problems.