基于改进量子遗传算法的油田井位及数量优化

Oilfield well location and quantity optimization based on improvedquantum genetic algorithm

摘要:

布井的数量及位置的选取是油田开发中至关重要的一环。一项最优的布井方案受到地质情况、油藏驱动方式、流体特性、油田设备规格以及多种经济参数指标的影响,是一个具有多决策变量的优化问题,传统的数学优化方法在处理这类问题时,很难找到一个合适的目标函数来满足优化条件。量子算法作为量子计算与智能算法相结合的产物,其优秀的寻优能力以及良好泛化能力,在处理目标函数性态复杂的优化问题时较传统方法有着更好的表现。因此,本文利用MATLAB建立油藏数值模拟模型,将井的数量和井位作为变量,以油田净现值为目标函数结合改进的量子遗传算法(Quantum Genetic Algorithm,QGA)对井位进行优化。通过与传统布井方式的对比,所提出的方法有更好的经济效益,同时摆脱了传统布井方式对于经验的依赖,具有很好的移植性。

Abstract:

The selection of the number and location of well is a vital part of oilfield development. An optimal well placement scheme is affected by geological conditions, reservoir driving methods, fluid characteristics, oilfield equipment specifications, and various economic parameters. It is an optimization problem with multiple decision variables. Traditional mathematical optimization methods are dealing with this type. When it comes to problems, it is difficult to find a suitable objective function to satisfy the optimization conditions. As a product of quantum computing and intelligent algorithms, quantum algorithm has excellent performance and good generalization ability, and it has better performance than traditional methods in dealing with complex optimization problems of objective function. Therefore, this paper uses MATLAB to establish a reservoir numerical simulation model, taking the number of wells and well position as variables, and optimizing the well position with the improved net QGA with the net present value of the oilfield as the objective function. Compared with the traditional well-drilling method, the proposed method has better economic benefits, and at the same time, it has got rid of the dependence of traditional well-placed methods on experience and has good portability.