• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文学位 >Simulation based tools for reservoir management.
【24h】

Simulation based tools for reservoir management.

机译：基于模拟的储层管理工具。

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Reservoir simulation is a tool that offers wide scope for reservoir management. The first part of this research concentrates on developing simulation based tools to maximize the profit in the oil recovery process. The next part focuses on the development of a method to improve reservoir models using real-time measurements. A previously developed two-dimensional, two-phase reservoir simulator that uses a control volume approach to solve the flow equations for an oil-water system was used for the purposes stated above. The simulator has the ability to incorporate fractures and employs a discrete-fracture model to do so. It can also handle any complex field geometry.; The quasi-Newton method using a BFGS update formula provided by the optimization toolkit TAO (toolkit for advanced optimization), developed at the Argonne National Lab, was used to optimize water injection strategies in order to maximize the profit associated with oil recovery. This toolkit was interfaced with the reservoir simulator. Two different cost functions were considered. The first one was a simple profit function in which the cost of oil produced is negated by the costs of water injection and production. This profit is summed over the entire period of operation. The second model for the cost function used here was the net present value (NPV). When using this function, the profit is discounted over the period of operation or in other words, oil's worth is considered to be more at the present time than it is later in the future. The optimization strategies developed for these two cost functions were compared in the studies. Significant improvement was observed in fluid production characteristics when using the optimal strategy as opposed to an uncontrolled case. In this algorithm, the period of operation was divided into more than one stage and the effect of multiple stages on the optima was also studied. Different domains with varying geologic complexities were used in the studies. This method was compared to an earlier algorithm developed and was shown to achieve better multi-stage solutions.; In the next part of this research, a variation of the traditional Kalman filter, known as the implicit Kalman filter (IKF) was used to update the reservoir model by estimating the states of the process using measurements from a reference field. This variation is used to estimate the states of a system cast in an implicit descriptor form. In order to achieve this, the reservoir model was linearized to cast the nonlinear model in an implicit form whenever measurements were available. However, the original nonlinear model was used to carry out the forward simulation. An algorithm employing a first-order Newton's method to estimate the porosity of a homogeneous field was also developed, which was used in conjunction with the IKF. A coarsely gridded domain was used as the model in all the case studies whereas a fine-scale representation of the same domain was used as the reference to generate the measurements. The states themselves were used as the measurements at select locations. The results show excellent agreement of the states of the model with the reference field as time progresses. The porosity is also estimated within a fair degree of accuracy.
机译：储层模拟是为储层管理提供广泛范围的工具。本研究的第一部分集中于开发基于仿真的工具，以使采油过程中的利润最大化。下一部分将重点介绍使用实时测量来改进储层模型的方法。出于上述目的，使用了以前开发的二维，两相油藏模拟器，该模拟器使用控制体积方法来求解油水系统的流量方程。该模拟器具有合并裂缝的能力，并采用离散裂缝模型进行合并。它也可以处理任何复杂的场几何。使用由阿贡国家实验室（Argonne National Lab）开发的优化工具包TAO（高级优化工具包）提供的BFGS更新公式的准牛顿法用于优化注水策略，以使与采油相关的利润最大化。该工具包与油藏模拟器接口。考虑了两个不同的成本函数。第一个是简单的利润函数，其中生产的石油成本被注水和生产的成本所抵消。该利润是整个运营期间的总和。此处使用的成本函数的第二个模型是净现值（NPV）。使用此功能时，利润在运营期内会折现，换言之，当前的石油价值被认为比以后的价格要高。在研究中比较了为这两个成本函数开发的优化策略。与不受控制的情况相比，使用最佳策略可以观察到流体生产特性的显着改善。在该算法中，将运算周期划分为一个以上的阶段，并研究了多个阶段对最佳状态的影响。研究中使用了具有不同地质复杂性的不同区域。将该方法与开发的较早算法进行了比较，结果表明该方法可以实现更好的多阶段解决方案。在本研究的下一部分中，使用传统卡尔曼滤波器的一种变体，即隐式卡尔曼滤波器（IKF），通过使用参考场中的测量值估算过程的状态来更新储层模型。此变化用于估计隐式描述符形式的系统状态。为了实现这一目标，只要测量可用，就将储层模型线性化以隐式形式转换非线性模型。然而，原始的非线性模型被用来进行正向仿真。还开发了一种算法，该算法使用一阶牛顿法估算均匀场的孔隙度，并将其与IKF结合使用。在所有案例研究中，均使用粗糙网格域作为模型，而使用同一域的精细比例表示法作为生成测量的参考。状态本身被用作选定位置的度量。结果表明，随着时间的推移，模​​型状态与参考场之间的一致性极好。孔隙度也可以在相当准确的程度内估算出来。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号