改进遗传算法在优化油田布井中的应用

Application of improved genetic algorithm to oilfield well array optimization

针对目前中国在油气藏布井方式上常采用规则井网的局限性,提出了一种基于适应值共享机制的小生境改进遗传算法。该算法具有强大的全局搜索性能和不要求目标函数连续、可微等特点。将其应用于油田布井当中,以获得最大累积产量为目标,对油井位置进行了优化,取得了良好效果。结果表明,利用改进遗传算法完全能够满足油田开发过程中多决策变量的选择、组合及优化的需要,是一种行之有效的新方法。

An improved genetic algorithm (GA) of niche based on fitness - sharing mechanism was advanced to overcome the main limitations of regular well pattern in well array way of oil-gas reservoir at home. This algorithm has a strong global - searching performance and has no requests for the continuity and differentiability of object function. This technology was used for well placement optimization with the object of maximizing the cumulative oil production. Good result proved that this technology could completely meet the demands of selection, combination and optimization of multi-decision variables in the process of oilfield development and will be a new available method.