Near Well Modeling

Near well modeling has been a niche area for many years. Accurate models for some parts of the problem exist (e.g., well testing semi-analytic solutions) which accurately deal with near-well effects in conventional geometries and simple completions, but support only limited options for stimulation. Numerical well tests are a field of growing interest because of the increased flexibility of these models. However, these models cannot handle new completion and stimulation techniques and lack the flexibility to capture the real-world geometries at work. Additionally, they are only semi-analytic and actually contain basic assumptions such as single-phase flow in a pseudo-steady-state (PSS) flow regime. These assumptions are needed to arrive at a linear model equation, which is amenable to analytic solutions. These assumptions are not strictly valid in real wells, and are increasingly invalid in the flow regimes common in tight sands and shales. Furthermore, productivity index correlations do not apply to stimulated or damaged wells when multiphase effects are considered.

Conventional reservoir simulation tools based on finite-difference or control-volume modeling approaches contain implicit assumptions about the relative importance of the many flow regimes present in porous media flow, each with disparate length and time scales. Examples include the assumption of a fully developed PSS radial pressure profile in the near-well area, in the industry standard Peaceman (1983) well model, and the assumption that the transient flow period during which that pressure profile develops is negligible. These assumptions are acceptable in conventional gas or oil reservoirs but are invalid in UG fields due to the extremely low permeability values, which extend the “transient” flow period from seconds to possibly decades. This lengthening of the time for transient flow is a physical result driven by the extremely low permeability values typical of UG.