|
油藏自动化历史拟合及开发方案智能优化技术的一体化应用研究
|
Abstract:
生产历史拟合和开发方案调整、优化是油藏开发过程的重要环节。通过历史拟合能够进一步明确储层和流体的各类静态参数,从而对油藏特征有更准确的认识。人工历史拟合工作过程繁琐、工作量大,对研究人员经验要求较高。本研究采用集合卡曼滤波方法辅助进行自动化历史拟合,极大降低了油藏工程人员的工作量,简化了历史拟合工作流程。油藏开发方案调整是在历史拟合获得的最佳模型基础上,通过改变注采关系、增加新井、调整射孔段等方法最大化开发效果或利润的过程。人工进行的开发方案优化主要是通过大量试算,逐个对各参数进行优化,过程繁琐且无法一次性获得全局最优解。本研究采用粒子群算法,同时对全部待优化参数进行统一优化,降低了工作量,获得了最佳的调整方案。本文首次将自动化历史拟合技术与开发方案智能调整优化技术结合,应用于实际油藏的生产研究当中。研究结果表明,本文采用的方法和技术流程是合理、高效的。为其他复杂油藏高效开发提供了技术基础和案例参考。
Production history matching and development plan adjustment and optimization are important links in the reservoir development process. History matching can further clarify various static parameters of reservoirs and fluids, so as to have a more accurate understanding of reservoir characteristics. The work process of manual history matching is cumbersome and the workload is large, which requires high researcher experience. In this study, the ensemble Kalman filter method is used to assist in automated history matching, which greatly reduces the workload of reservoir engineers and simplifies the history matching workflow. Reservoir development plan adjustment is a process of maximizing development effects or profits by changing the injection-production relationship, adding new wells, and adjusting the perforation section on the basis of the best model obtained by history matching. The manual optimization of the development plan is mainly through a large number of trial calculations to optimize each parameter one by one. The process is cumber-some and the global optimal solution cannot be obtained at one time. In this study, particle swarm algorithm was used to optimize all parameters to be optimized at the same time, which reduced the workload and obtained the best adjustment scheme. This paper combines the automatic history matching technology with the intelligent adjustment and optimization technology of the development plan for the first time, and applies it to the production research of the actual oil reservoir. The research results show that the method and technical process adopted in this paper are reasonable and efficient. It provides technical foundation and case reference for the efficient development of other complex reservoirs.
[1] | 陈玉雪. 智能井网优化理论与应用[D]: [硕士学位论文]. 青岛: 中国石油大学(华东), 2013. |
[2] | 罗志凡, 荣见华, 杜海珍. 基于遗传算法和梯度算法的一种结构优化混合方法[J]. 计算机工程与应用, 2003, 39(8): 71-73. |
[3] | 姜万录, 刘庆平, 刘涛. 神经网络学习算法存在的问题及对策[J]. 机床与液压, 2003(5): 29-32. |
[4] | 章敬东. 复杂优化问题中智能算法的分析与集成[D]: [博士学位论文]. 广州: 华南理工大学, 2003. |
[5] | 安瑶. 粒子群算法在油田布井优化中的应用研究[D]: [硕士学位论文]. 西安: 西安石油大学, 2013. |
[6] | 丁帅伟, 姜汉桥, 周代余, 等. 基于改进粒子群算法的不规则井网自动优化[J]. 中国海上油气, 2016, 28(1): 80-85. |
[7] | 姜瑞忠, 刘明明, 徐建春, 等. 遗传算法在苏里格气田井位优化中的应用[J]. 天然气地球科学, 2014, 25(10): 1603-1609. |
[8] | Kalman, R.E. (1960) A New Approach to Linear Filtering and Prediction Problems. Journal of basic Engineering, 82, 35-45. https://doi.org/10.1115/1.3662552 |
[9] | Geir, E. (1994) Sequential Data Assimilation with a Nonlinear Qua-si-Geostrophic Model Using Monte Carlo methods to Forecast Error Statistics. Journal of Geophysical Research: Oceans, 99, 10143-10162.
https://doi.org/10.1029/94JC00572 |
[10] | Fernández Martínez, J.L. and Gonzalo, E.G. (2009) The PSO Family: Deduction, Stochastic Analysis and Comparison. Swarm Intelligence, 3, 245. https://doi.org/10.1007/s11721-009-0034-8 |