The Research Of Intelligent Optimization Methods In Oilfield Water Injection System

Water injection flooding is an important exploitation mode during development of the oilfield, water injection system is a continuous and inclosed hydraulic system, it is constituted by nodes that include many water injection stations, water-distributing rooms, wells, and pipelines, every component part restricts and affects each other, and forms one entirety. The optimization of oilfield water injection system is a complex problem of large-scale hybrid optimization design, which involves discrete topological, combination and non-linear parametric optimizations; some of them belong to NP hard problems. It is difficult to achieve good results by use of conventional optimization algorithms, thus at present the design is mainly based on the human experience. With the continuous development of optimization technology, intelligent optimization methods, such as genetic algorithm, simulated annealing etc, with its advantages of high-effective optimization performance, needing no special information for the problem and with its global search capacity, has received extensive attention, and has been widely applied to many domains. It has become a powerful instrument to solve project difficulties, and can be used as an effective alternative method for the optimization design for oilfield water injection system. In the paper, based on careful and profound analysis and research on the oilfield water injection system, mathematical models are established according to the practical needs, solved by use of intelligent optimization methods, and optimization designs of the system are realized by designing proper algorithm and programming corresponding software. The following aspects are researched in this paper:Based on careful analysis on the structure of oilfield water injection system, optimization model of topology optimization for tree type water injection pipe network is established, according to the model characteristics, the problem is divided into distribution and geometry layers, in distribution layer, the genetic algorithm is used to optimize subjection relations between wells and stations, in geometry layer, the nonlinear optimization is used to optimize position of water-distributing rooms.Optimization model of looped pipeline network optimization is established, In the genetic...