首页> 外文会议>International Conference on Greenhouse Gas Control Technologies >Probabilistic Risk Assessment of CO2 Trapping Mechanisms in a Sandstone CO2-EOR Field in Northern Texas, USA
【24h】

Probabilistic Risk Assessment of CO2 Trapping Mechanisms in a Sandstone CO2-EOR Field in Northern Texas, USA

机译:美国北部德克萨斯州砂岩CO2-EOR场中CO2诱捕机制的概率风险评估

获取原文

摘要

This study employs a polynomial chaos expansion (PCE) technique to develop the reduced order models (ROMs) integrated with the Monte Carlo simulations to quantify uncertainties of CO2 storage associated with enhanced oil recovery (EOR) in a commercial-scale CO2-EOR field. The Morrow B sandstone at the Farnsworth Unit (FWU) EOR field in northern Texas was selected as a case study. A 3-D reservoir model, including simulation of a five-spot injection/production well pattern, was constructed based on the data from well logs and cores, and an updated seismic interpretation in the FWU. Porosity and permeability were considered as sources of uncertainty. The 1000-realization heterogeneous random fields of porosity and permeability were generated based on the variogram analysis on the core measurements of the FWU. Three model outputs of interest include the amount of CO2 trapped by three trapping mechanisms, i.e. hydrodynamic topping, oil dissolution trapping, and aqueous dissolution trapping. The total of 25 runs were conducted with the simulations of a 20-year CO2-EOR period, and a 20-year monitoring period to develop the ROMs between uncertain model input parameters (i.e., porosity and permeability in this study) and model outputs at the end of simulation for each grid cell using PCE approach. The high coefficient of determination (R2) and small normalized root mean square error (NRMSE) between the simulated (using reservoir modeling) and predicted (using ROMs) CO2 storage indicate that the ROMs are acceptable and reliable to be used for the predictions. Given the 1000 Monte Carlo samplings of the model input parameters, cumulative distribution functions (CDFs) and uncertainty bounds (10~(th) and 90~(th) percentiles) of model outputs were estimated based on ROMs. At the end of simulation, the most injected CO2 at the FWU was stored by hydrodynamic trapping (between 121,400 and 166,860 tonnes), followed by oil dissolution trapping (between 55,303 and 76,816 tonnes), and aqueous dissolution trapping (between 1979 and 2751 tonnes).
机译:这项研究使用了多项式混乱膨胀(PCE)技术发展与Monte Carlo模拟积分以量化与在工业规模的CO2-EOR场强化油采收(EOR)相关联的二氧化碳储存的不确定性的降阶模型(ROM)之。莫罗乙砂岩在得克萨斯州北部法恩斯沃思单元(FWU)EOR字段被选为为例。 A 3-d的储层模型,其中包括一个五点注入/生产井图案的模拟中,基于从测井和芯,并在FWU更新地震解释数据构建。孔隙度和渗透率被认为是不确定的来源。孔隙度和渗透率的1000实现异质随机视野中生成基于关于FWU的岩心测量所述变差函数分析。三个感兴趣的模型输出包括CO2的由三个捕获机制捕集的量,即液力摘心,油溶解诱捕,和含水溶解俘获。 25个运行的总用了20年的CO2-EOR周期的模拟进行的,和有20年监视周期开发不确定模型输入参数(即,孔隙度和渗透率在本研究中)和模型输出之间的ROM的在模拟结束使用PCE方法每个网格单元。模拟的(使用储层建模)之间确定(R2)和归一化的小根均方误差(NRMSE)和预测(使用ROM的)CO 2存储的高系数表明ROM是可接受的和可靠的将被用于预测。给定模型的输入参数的1000个蒙特卡罗抽样,累积分布函数(CDF)和误差范围模型输出的(10〜(TH)和90〜(th)的百分)估计基于光盘。在模拟结束,在FWU最注入的二氧化碳被存储在由流体动力学捕获(121,400和166860之间吨),随后油溶解俘获(55303和76816之间吨),和含水溶出俘获(1979和2751吨之间) 。

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