阳光石油网|石油技术交流|石油人论坛

 找回密码
 欢迎注册
查看: 1546|回复: 1

Tempest黑油计算结果的流线显示

[复制链接]
  • TA的每日心情
    开心
    2014-8-7 16:23
  • 签到天数: 7 天

    [LV.3]偶尔看看II

    发表于 2013-3-1 15:04:04 | 显示全部楼层 |阅读模式

    马上注册,下载丰富资料,享用更多功能,让你轻松玩转阳光石油论坛。

    您需要 登录 才可以下载或查看,没有账号?欢迎注册

    x
    目前市场上的模拟器,如eclipse,需要用到流线模型才能做出流线结果。

    有时候我们往往需要黑油计算的结果来模拟油藏,以其得到更准确的拟合结果和修改模型。但是拟合完的模型在转到流线模型时,会出现数据格式、计算等不同的错误,而继续花费时间去作流线结果。

    Tempest在用黑油模型计算完后,就可以直接查看流线结果。
    图片7.png 流线结果



    附Tempest关于流线计算的技术文档:

    Streamline Calculations in Tempest Tempest calculates streamlines for each individual phase. In some cases voidage streamlines are generated by summing the phase velocities, e.g. when calculating the Well Matrix or the Voidage Drainage Regions property.
    Requirements For Tempest to be able to calculate streamlines the following properties must be written to the simulation output:

    • Porosity
    • Transmissibility (X, Y and Z)
    • Pressure
    • Density for each phase
    • Viscosity for each phase
    • Relative Permeability for each phase
    • Capillary Pressure (Pcog for Gas phase and Pcow for Water phase)
    NNC connections and their transmissibility are also required. Streamlines are not supported for radial grids. Calculation Steps

    • Initialize active cell information
    • Load static properties
    • Load dynamic properties for the required report step
    • Pre-process NNCs
    • Set up unit conversion factors - all calculations are in SI units
    • Calculate velocities at each cell face
    • Mark injector/producer cells and source/sink cells (cells with no flow in or no flow out)
    • Lace streamlines
    Pre-process NNCs The NNC data output by MORE is processed as it contains transmissibility data that will be needed to calculate the velocities. This data includes grid to grid NNCs (e.g. Global to LGR) and fault NNCs which allows faster calculation by avoiding the need to scan the whole grid and calculate NNCs within Tempest.
    For each NNC the cell geometry for the two connected cells are loaded. The two connected faces (given by the NNC type) are projected onto the best 3D plane and the connection window stored. The connection windows will be needed for moving between grids and/or faulted cells during the streamline lacing step. User entered NNCs that link two unconnected cells are ignored for streamline calculations.
    Calculate velocities at each cell face Velocities are stored in a three dimensional array. The total number of values calculated = nPhase (1 to 3) * nFace (6) * nCell. A parallel loop over all cells calculates the velocity at each face for all phases. The calculation involves the current cell (A) and the neighbouring cell (B) for the face being calculated. If either cell is inactive then the velocity is zero.
    The velocities are calculated as follows:

    • Calculate the harmonic average of the mobility of the phase in each cell: (2*MobA*MobB) / (MobA+MobB), where the Mobility for each cell is RelPerm / Viscosity for the current phase, e.g. Kro / Viso.
    • Calculate the hydrostatic pressure (pStat) using the depth difference between the two cell centres and the mean of the cell densities: (DensA + DensB)/2.0 * (CenDepthB - CenDepthA) * 9.80665
    • Calculate the pressure difference (pDiff) between the cells: pA + pcA + pStat - (pB + pcB) where pcA and pcB are the capillary pressures for each cell.
    • Calculate the flow volume: Transmissibility * Mobility * pDiff where the transmissibility comes from the Tx/Ty/Tz properties for grid neighbour connections or from the NNC's transmissibility for NNCs.
    • Calculate the area of the connection. This is done by projecting the two cell faces onto the best 3D plane and calculating the area of the intersection.
    • Calculate the flow velocity by dividing the flow volume by the connection area.
    • Calculate the pore velocity by dividing the flow velocity by the porosity.
    • Store the velocity ready for lacing calculations.
    NOTE: If calculating the flow volume for a cell that is connected to multiple cells on a single face then the flow volumes and areas are calculated for each individual connection and summed together before calculating the velocity, i.e. Steps 1-5 may be performed multiple times for cells with NNCs before steps 6-8 are performed.
    Lace Streamlines There are two methods used for seeding streamlines in Tempest:

    • Lace from the cell faces of completed cells - used for streamline visualization and Well Matrix calculation. If a face's cell neighbour is also completed then streamlines are not seeded from that face. The number of streamlines seeded per face depends on the density setting: Low=1, Medium=4, High=9, Highest=16
    • Lace from the centre of every cell - used for calculating streamline properties, e.g. Oil Drainage Regions.
    Streamline lacing uses Pollock's method. Each grid cell is transformed to a unit cube and a linear velocity interpolation applied to calculate the exit point for any given entry point in abg (unit cube) coordinates and a time of flight (TOF) value for the cell traversal. The entry and exit points are transformed from abg coordinates to true xyz coordinates for visualization. No attempt is made to model the trajectory of the streamlines within the grid cells, a straight line is drawn from the entry point to the exit point.
    When a streamline is seeded Tempest first laces from the seed point backwards to a source cell, e.g. an injector completion or an aquifer. The streamline is then completed by lacing forwards to a sink cell, e.g. a producer completion. All lacing is performed in unit cube coordinates.
    The time of flight from a completion to the face of the completed cell is approximated to zero so completions in large grid cells will yield less accurate TOF calculations. However, given that the model's completions are already an approximation of the true completion geometries, this approximation will have little impact on overall accuracy.
    Grid boundaries and faults When a streamline encounters a grid boundary or a fault some extra processing must be performed to find the next cell and entry point. All calculations are in unit cube coordinates.

    • Find the list of connection windows for the exit face.
    • Find the exit window that contains the exit coordinates. If no window contains the exit coordinates then the windows are looped over to find the closest valid exit point so that the streamline can traverse the cell boundary. This may happen if the exit point corresponds to an inactive cell or no cell at all.
    • Transform the exit coordinates to entry coordinates in the new cell.
    • Move to the next cell and continue lacing.
  • TA的每日心情
    开心
    2024-5-10 14:28
  • 签到天数: 2525 天

    [LV.Master]伴坛终老

    发表于 2014-7-2 07:57:14 | 显示全部楼层
    谢谢分享
    回复 支持 反对

    使用道具 举报

    您需要登录后才可以回帖 登录 | 欢迎注册

    本版积分规则

    QQ|Archiver|手机版|小黑屋|阳光石油网 ( 鲁ICP备2021003870号-1 )

    GMT+8, 2024-5-19 23:10 , Processed in 0.067115 second(s), 26 queries .

    Powered by Discuz! X3.4 Licensed

    Copyright © 2001-2021, Tencent Cloud.

    快速回复 返回顶部 返回列表