Image Well

Key Takeaways

• The log-log pressure derivative plot is the diagnostic tool that reveals boundary effects and the presence of image wells in the theoretical framework — in a well producing from an infinite-acting reservoir with no boundaries, the pressure derivative (the change of pressure with respect to the logarithm of time) plots as a flat horizontal line at late time during a buildup test, indicating radial flow geometry is maintained; when the pressure disturbance radiates outward and reaches a no-flow boundary, the drainage area is effectively halved (the well can only drain the half-space on its side of the boundary) and the pressure drops faster, causing the derivative to double and follow a slope of unity on the log-log plot; multiple boundaries (the well in a rectangle, a wedge, or a channel) create a sequence of derivative behaviors as each boundary is encountered, and the timing and magnitude of each departure from the infinite-acting baseline reveals both the distance to each boundary and the boundary type; the image well framework provides the theoretical underpinning for why these derivative behaviors are what they are: each doubling of the derivative slope corresponds to one additional image well becoming influential in the pressure superposition.
• Sealing faults detected by image well analysis significantly affect reserves estimation and infill drilling planning — when a pressure transient test in an exploration or appraisal well reveals a sealing fault at a distance of 500 meters from the well, the interpreted reservoir size is immediately constrained by that boundary; the well can drain a half-circle of reservoir with radius determined by the productive thickness and porosity, rather than the full circle that would be drained in the absence of the fault; the image well calculation that the analyst performs to match the observed pressure behavior determines the fault distance with uncertainty depending on the duration of the test and the quality of the pressure data; a fault at 500 meters detected with 100-meter uncertainty has significant implications for whether an infill well can be placed between the existing well and the fault, whether additional structure exists beyond the fault (if the fault is a lateral seal of a separate reservoir compartment), and whether the reserves booking for the discovery should include the area beyond the fault; many reservoir compartmentalization surprises in field development — where production decline is faster than predicted — trace back to sealing faults that were either not detected or underestimated in distance by early pressure transient analysis.
• Multiple image wells are required when a producing well is bounded by more than one boundary, and their superposition is what allows interpretation of complex reservoir geometries — a well producing in a channel sand bounded by two parallel no-flow faults requires two image wells (one reflected across each fault) plus additional image wells of the image wells (to satisfy both boundary conditions simultaneously at each fault); the resulting "image well array" for a channel sand extends infinitely in both directions perpendicular to the channel, with alternating sign image wells satisfying both boundary conditions simultaneously; the pressure response of a well in a channel shows the initial radial flow period (no boundary effect), then a characteristic "channel flow" period where the derivative rises to double the radial value and then follows a half-slope on the log-log plot as linear flow dominates along the channel, then eventually approaches a semi-steady state if the channel is also closed at both ends; matching all these regimes to the image well model gives the interpreter the channel width (from the time to first boundary response) and the channel length (from the time to the closed end effect), both of which are critical for assessing the recoverable reserves in the isolated channel reservoir.
• The method of images has limitations that experienced interpreters must recognize when applying it to real reservoirs — the mathematical elegance of the image well approach depends on several idealized assumptions that are only approximated in field conditions: the boundary must be a straight line (or plane in 3D) extending to infinity; the transmissivity and storativity of the reservoir must be uniform on both sides of the well (heterogeneous reservoirs distort the pressure signal in ways that can mimic boundary effects); the reservoir must be single-layer and homogeneous for the classical solution to apply; and the production history must be simple enough that superposition can correctly account for all rate changes; in practice, heterogeneous reservoirs (with natural fractures, permeability streaks, or layering) and complex production histories (with multiple rate changes before the test) introduce noise in the pressure derivative that can either mask real boundary effects or create apparent boundary effects that are artifacts of reservoir heterogeneity rather than geometric boundaries; modern interpretation software allows simultaneous fitting of multiple interpretive models (infinite acting, one fault, two faults, channel) to the data, comparing the goodness of fit and uniqueness of each model before committing to a boundary interpretation that will be used in reserves estimation.
• Constant-pressure image wells model aquifer support and gas cap drive, providing a quantitative framework for assessing reservoir energy — a strong aquifer at the reservoir edge acts as a constant-pressure boundary that maintains reservoir pressure as production withdraws fluid; in the image well framework, this is modeled as an injection image well of equal strength to the producing real well, located at the reflection of the real well across the aquifer boundary; the pressure behavior of a well adjacent to a strong aquifer shows no pressure drawdown at late time (the aquifer image injection well replaces exactly the production of the real well, maintaining constant pressure at the boundary and therefore maintaining reservoir pressure at the well if the aquifer connectivity is good); the distance from the well to the aquifer boundary, determined from when the constant-pressure effect first appears in the buildup data, tells the engineer how much of the pressure transient response is supported by aquifer influx versus reservoir depletion; this distinction is critical for material balance calculations that use pressure history to estimate connected pore volume — without correctly identifying and modeling the aquifer boundary, the material balance will overestimate the reservoir gas or oil volume because it will attribute the pressure support to a larger connected rock volume rather than to aquifer influx at the boundary.