基于帝国竞争反向传播神经网络的断块油田开发顺序优化  

Development sequence optimization of fault-block oilfieldsbased on the ICA-BP algorithm

作　　者：徐庆岩[1] 孙晓飞[2] 翟光华 王瑞峰[1] 雷诚 张瑾琳[1] XU Qingyan;SUN Xiaofei;ZHAI Guanghua;WANG Ruifeng;LEI Cheng;ZHANG Jinlin(Research Institute of Petroleum Exploration&Development,PetroChina,Beijing 100083,China;China University of Petroleum(East China),Qingdao 266000,Shandong,China;International Petroleum Exploration and Production Corporation,PetroChina,Beijing 100085,China)

机构地区：[1]中国石油勘探开发研究院,北京100083 [2]中国石油大学(华东),山东青岛266000 [3]中国石油国际勘探开发有限公司,北京100085

出　　处：《石油地质与工程》2024年第3期77-81,89,共6页Petroleum Geology and Engineering

摘　　要：明确断块油田群中断块的开发顺序是进行开发方案设计的前提条件。断块油田数量较少时,可以进行技术经济的组合对比,但是断块数量较多时会形成海量的组合,耗费时间也长。断块油田开发顺序评价的现有方法有权重评价法、层次分析法、综合模糊评判法等,这些方法在选择评价指标和指标权重上带有较强的主观性,无法做到完全客观的评价。因此本文提出一种基于帝国竞争算法改进的反向传播神经网络模型,首先采用Spearman相关系数法确定影响断块油田开发的主控因素,其次使用分段三次Hermite插值方法实现断块油田群开发数据库的扩充,最后在扩充后的大量数据库训练样本的基础上,基于帝国竞争算法改进的反向传播神经网络模型可以确定影响开发效果参数的权重并预测断块油田群中各断块油田的净现值,根据净现值大小可以确定每个断块的开发顺序。该方法以实际断块油田群的地质油藏数据库作为评价依据,断块油田的开发顺序更加的科学合理,项目整体的净现值也明显高于依靠传统方法确定的开发顺序组合,避免了人为主观性,也节省了数值模拟和经济评价的工作量,克服了现有方法的局限性,对于提高断块油田群开发综合效益具有重要意义。Defining the development sequence of each fault block in the fault-block oilfield group is a prerequisite for the development scheme design.When the number of fault block oil fields is few,the combination of technology and economy can be compared,but when the number is large,it will form a massive combination and take a long time.Existing methods include weight evaluation,analytic hierarchy process,comprehensive fuzzy evaluation,etc and they have strong subjectivity in selecting evaluation indicators and indicator weights,and cannot achieve completely objective evaluation.Therefore,ICA-BP algorithm,an improved back propagation neural network model based on the imperial competition algorithm is proposed.Firstly,Spearman correlation coefficient method is used to determine the main control factors affecting the development of fault-block oilfield,and then Hermite interpolation method is used to expand the fault-block oilfield group development database.Finally,based on a large number of training samples of the expanded database,the improved back propagation neural network model based on the imperial competition algorithm can determine the weight of the parameters affecting the development effect and predict the net present value of each fault block oilfield.The development order of each fault block can be determined according to the net present value.This method is based on the geological reservoir database of the actual fault-block oilfield group as the evaluation basis.The development sequence of fault-block oilfield is more scientific and reasonable,and the overall net present value of the project is significantly higher than the development sequence combination determined by traditional methods,avoiding human subjectivity,saving the workload of numerical simulation and economic evaluation,and overcoming the limitations of existing methods.It is meaningful to improve the comprehensive benefit of the development of fault-block oilfield groups.

关 键 词：帝国竞争算法 反向传播神经网络 开发参数权重 投产顺序优化 断块油田群 净现值 

分 类 号：TE319[石油与天然气工程—油气田开发工程]