The Reservoir Engineer Network

The Reservoir Engineer Network

21,152 members
  • Join

    When you join a group, other members will be able to see your profile and message you. The group logo will be visible on your profile unless you change that setting.

  • Information and settings

Have something to say? Join LinkedIn for free to participate in the conversation. When you join, you can comment and post your own discussions.

Hasnain Ali

Falloff pressure transient analysis yields better results when estimating the properties of the reservoir than injection analysis... Is this true?? and why???

Well Testing Engineer at AlMansoori Specialized Engineering

  • Comment (14)
  • July 11, 2012
  • Close viewer


  • Dave P.


    Dave P.

    Consultant, President at AStech Inc

    Mostly true - if you don't mean injection fall-off analysis. Transient analysis is based on constant rate assumption and at no time is the rate more constant than when it is zero.

  • Hasnain Ali A.

    Hasnain Ali

    Hasnain Ali A.

    Well Testing Engineer at AlMansoori Specialized Engineering

    Thanks Dave. But i simulated a simple model using Eclipse, and made a study on both injection and falloff.. During Injection i injected water at a constant rate of 200 bbl/day... I crosschecked and found out that there were no fluctuations during injection i.e my wellbore was injecting water 200 bbl/day throughout. But when i carried out my well test analysis i found falloff gave better results than injection analysis..... Any other suggestions??

  • Tony T.


    Tony T.

    Reservoir Engineer

    Can you pls clarify what you mean by better results. Also is your injection and fall off time periods the same?
    I would like to see your derivatives from both periods.

  • Ram  Tirth J.

    Ram Tirth

    Ram Tirth J.

    Sr. Reservoir Engineer at Cairn India Ltd

    As far as better results are concerned, injection test will give always better results as you don't have to factor in the injection period time which creates shape deviation in some cases esp. when injection time is less. I would like to see the diagnostic plot and what was Injection Constraints in Eclipse model.

  • Walter F.


    Walter F.


    Hasnain, you didn't mention what methods you are using to analyze the data. That might be important.

    However, besides that, there are at least 2 separate issues here. In the field fall-off tests are more representative, as others mentioned, because it is usually very difficult to precisely maintain a constant rate, unless the rate is ZERO. Even small fluctuations in rate will affect the pressure response making the analysis of the injection period data difficult and uncertain. That extends to the fall off portion, but due to the nature of diffusion, the rate fluctuations during injection have a much smaller affect on the fall off data.

    Second, you mentioned you are using Eclipse. There is a problem using Eclipse, and most other general purpose reservoir simulators, for pressure transient work. Basically the equations are optimized to conserve mass in the grid cells by placing the state variables (i.e. pressure, density, etc.) at the center of the cells. That is exactly what you want for most reservoir simulation projects.

    However, for pressure transient analysis, the well should be a boundary condition, not in the center of the cell. For that reason, the skin value for sure is "odd" because it is also compensating for the difference in the well position.

    In addition, most early time analysis methods assume an infinite reservoir - which is OK for realistic time periods. But in most reservoir simulators you cannot specify an infinite reservoir outer boundary, so you are forced to use a constant pressure or a closed boundary. Again, for reservoir simulation that is fine, but for pressure transients, it causes a distortion.

    The exception to the distortion is precisely when the well flow rate is ZERO. Then there is no flow and the distortions are all diminished. As a result, analyzing pressure transient generated from general purpose simulators will generally give slightly erroneous results, but the best results is when the well is shut-in and the distortion is minimized.

  • Fernando T.


    Fernando T.

    President at VYP Consultores SA

    I assume you are using such a small grid size that what Mr Fair said gets minimized, otherwise the error included due to "connecting" a grid to the well can be very big.

  • Hasnain Ali A.

    Hasnain Ali

    Hasnain Ali A.

    Well Testing Engineer at AlMansoori Specialized Engineering

    Thank you all. I have simulated a simple ID homogeneous radial model, with small grids sizes near the wellbore. Wellbore radius is 0.5 ft and first grid in my radial model is also 0.5 ft (central grid). When i carry out analysis for both injection and falloff plots. I obtain better calculated results (like skin, perm, mobility etc) for falloff tests but injection test show me an error of approx +10% (with constant injection rate of 200 bbl/day) when i compare it with the input data, for each parameter calculated.

Have something to say? Join LinkedIn for free to participate in the conversation. When you join, you can comment and post your own discussions.