遗传算法在苏里格气田井位优化中的应用

苏里格气田是中国致密砂岩气田的典型代表，储层非均质性及特低的渗透率造成苏里格气田有效泄油面积小。为提高单井控制储量和气田采收率，需要利用井位优化来确定最优井数及其位置。因此，借助Matlab自带的遗传算法工具箱，采用优化理论结合数值模拟的方法进行苏里格气田某区块的井位优化研究。研究中将净现值作为目标函数，将单井控制面积和井位作为变量，优选净现值最大的单井控制面积及井位。结果表明，苏里格气田的最优单井控制面积为0.5km²/井|对于最优单井控制面积，遗传算法得到的最优净现值为14.625 9×10⁸元，大于常规井网的净现值14.337 8 ×10⁸元，净现值增加幅度为2.01%|基于遗传算法的井位优化方法克服了常规布井方式的经验依赖性|最优井位分布与渗透率关系密切，渗透率高的部位井数多，反之，井数少。

#br# Application of Genetic Algorithm for Well Placement Optimization in Sulige Gasfield

Sulige Gasfield is a typical tight sandstone gas field.The effective drainage area is small due to its reservoir heterogeneity and ultra-low permeability.In order to improve single well controlled reserves and gas recovery factor,there is an urgent need to determine the optimum well numbers and location by well placement optimization.Therefore,we perform well location optimization of Sulige Gasfield by using the optimization theory combined with numerical simulation,based on Matlab′s genetic algorithm toolbox.In the study,the objective function is the net present value (NPV) and the variables are well spacing and well location.Then the optimized well spacing and well location which makes the NPV the highest was selected.The optimization results show that the best well spacing of Sulige Gasfield is 0.5km²/well.For the best well spacing case,the NPV of the GA algorithm which is ￥1.462 59 billion is greater than the NPV of the conventional well pattern which is ￥1.433 78 billion.The NPV increases by 2.01%.The well placement optimization based on GA overcomes the disadvantage of the conventional well pattern which is experience-dependent.The optimized well location has a close relationship with the permeability distribution.The high permeability zone has more wells and the low permeability zone has fewer wells.