基于混合遗传算法的油藏井位快速优化  

Rapid optimization of well placement based on hybrid genetic algorithm

作　　者：赵乐义 张蒙 孙成田 李昱东 任雪瑶 胡鳕茹 丁帅伟 ZHAO Leyi;ZHANG Meng;SUN Chengtian;LI Yudong;REN Xueyao;HU Xueru;DING Shuaiwei(PetroChina Yumen Oilfield Company,Jiuquan,Gansu 735019,China;Geology Department,Northwest University,Xi’an,Shaanxi 710069,China)

机构地区：[1]中国石油玉门油田公司,甘肃酒泉735019 [2]西北大学地质学系,陕西西安710069

出　　处：《石油地质与工程》2025年第2期89-94,共6页Petroleum Geology and Engineering

基　　金：陕西省自然科学基础研究计划(2024JC-YBMS-328);国家自然科学基金项目(52474044)

摘　　要：确定合理的井位是一件非常具有挑战性的任务,主要是因为影响油藏生产动态特征的地质因素和工程因素复杂多变,且非线性十分严重。由于深水油藏物性条件比较好,开发策略上通常采用“稀井高产”的方法。目前国内一般采用的是规则井网布井技术,但这种布井思路并不适用于“稀井高产”模式下的不规则布井方式。文中以遗传算法为基础,结合油藏生产潜力图和Gompertz产量预测模型作为辅助技术手段,同时充分考虑井位优化的目标函数、约束条件以及相应的惩罚函数,提出了一种混合遗传算法井位优化流程实现井位快速优化。实例应用表明,该混合遗传算法优化流程优化结果相对于其他方法,在优化效果和优化时间上具有较好的优势,在尽量减少对油藏数值模型调用的条件下,研究成果能够快速完成井位优化,研究成果对于不规则井网的布井方案具有重要指导意义。Determining optimal well locations is a challenging task due to the highly nonlinear correlation and uncertain between engineering and geologic variables impacting reservoir performance.Deep water oil reservoirs have better physical properties due to their favorable conditions,the development strategy mainly adopts the method of"higher yields with fewer wells strategy".Until now,engineers usually adapt regular well pattern technology in daily work in China.Unfortunately,this method can’t meet the need of designing highly yields well pattern with fewer wells.The main idea of this paper is to propose a hybrid genetic algorithm well placement optimization process for rapid well location optimization based on genetic algorithm,combined with productivity potential map and Gompertz production prediction model as auxiliary technology means,and considering the objective function,constraint condition and corresponding penalty function of well location optimization.The example application shows that results of the hybrid genetic algorithm optimization process have better advantages than those of other methods in optimization effect and optimization time.The results of this study make it possible to optimize well location quickly without using reservoir numerical model as much as possible,and have a certain guidance for the layout of irregular well pattern.

关 键 词：井位优化 遗传算法 生产潜力 产量预测模型 混合优化流程 

分 类 号：TE319[石油与天然气工程—油气田开发工程]