Abstract
In an unconventional reservoir, the biggest challenge is to know how the natural fractures drain the reservoir as they have the greatest impact on production. But unfortunately, very little information is available about them. Microseismics aid in building a picture of the fracture network, but give no information about fractures where actual fluid flow occurs. Production logging results give information around wellbore area only. Conventional rate transient analysis has major drawbacks, as long shut-in times are not possible and with dimensionless variables multiple results are possible. The method outlined in this paper overcomes these limitations using simplified assumptions.
This simulation modeling method uses dual porosity as an idealization of the fracture network, which is the conventional wisdom, but with constant volume hydraulic fractures. This restricts the possible fracture lengths and the associated geometries of these hydraulic fractures, when modeled in 1D, 2D or 3D orientation. These HF-NF connectivity scenarios, using idealized fracture network of slabs (planar 1D HF-NF), matchstick (non-planar 2D HF-NF) and cubes (non-planar 3D HF-NF), is used to establish those fundamental connectivity scenarios where the fracture spacing can either be 1:1:1 (equidistant) or in the ratio 1:2:3. In order to assign permeability to the fractures, under these six different fundamental scenarios which have the same production performance, we follow the single block approach based on rate transient analysis. It also helps in establishing fracture permeability for other fracture connectivity variants such as 2D HF-3D NF or 3D HF-2D NF and with the two previously specified fracture spacings.
The results of this study, which essentially deals with the reservoir linear flow, are presented in the form of characteristic plots based on the ratio of average dimensionless pressure in the block with the square root of dimensionless time versus the dimensionless time for different fracture pressure declines. In each of fracture connectivity scenarios the solution rises to a discreet 1, 2, 3 value if idealized blocks are used or fall short of these values for non-idealized block combination depending on block geometry of NF. These conclusions are also shown by field models, analyzing actual history matched data.
Basic knowledge of the orientation of NF network gives better history match and prediction results. Also, with the help of a reservoir simulator one can assign physical meaning to different fracture spacings, which could be in the increasing or decreasing form. Rate transient analysis, using dimensionless parameters, fails to illustrate this fact. It helps in a long way to establish the optimum fracture spacing with the same volume of proppant being pumped in the reservoir and with known NF orientation.