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 oilfields based on the ICA-BP algorithm
徐庆岩 1孙晓飞 2翟光华 3王瑞峰 1雷诚 1张瑾琳1
作者信息
- 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
- 折叠
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.
Abstract
Defining the development sequence of each fault block in the fault-block oilfield group is a pre-requisite for the development scheme design.When the number of fault block oil fields is few,the combina-tion of technology and economy can be compared,but when the number is large,it will form a massive com-bination and take a long time.Existing methods include weight evaluation,analytic hierarchy process,com-prehensive fuzzy evaluation,etc and they have strong subjectivity in selecting evaluation indicators and indi-cator weights,and cannot achieve completely objective evaluation.Therefore,ICA-BP algorithm,an im-proved 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 ex-panded database,the improved back propagation neural network model based on the imperial competition al-gorithm can determine the weight of the parameters affecting the development effect and predict the net pres-ent value of each fault block oilfield.The development order of each fault block can be determined accord-ing 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 sci-entific and reasonable,and the overall net present value of the project is significantly higher than the devel-opment sequence combination determined by traditional methods,avoiding human subjectivity,saving the workload of numerical simulation and economic evaluation,and overcoming the limitations of existing meth-ods.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Key words
imperialist competition algorithm/back propagation neural network/development parameter weight/production sequence optimization/fault block oilfield group/net present value引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据