Tightening well spacing while pumping larger hydraulic fracturing jobs seems to be the most practical solution to boost revenue in shale plays in the US, especially under current oil market conditions. However, the outcome of such approach may indeed be the opposite. Drilling infill wells closer than ever to older wells and creating very long hydraulic fractures make the risk of getting fracture hits, or "frac hits", in a multi-well pad more likely. Given its severe consequences, such as temporary or permanent loss of productivity as well as induced blowouts in offset wells, frac hits has become a top concern for shale producers. In response, shale operators and regulatory agencies are commonly working on techniques or procedures that can predict or, at least, mitigate the occurrence of frac hits.
We present a mathematical model aimed to assess the degree of interference in multi-well pads due to frachits. The degree of interference δ is hereby defined as the fraction of fractured stages that are in communication so that 0 ≤ δ ≤ 1. We assume that both wells have the same number of hydraulic fractures. The mathematical model relies on the application of the trilinear flow model to two multistage-fractured horizontal wells interconnected through the inner fracture tips. To emulate the effect of null, partial or total interwell connectivity, we implement a semi-permeable type boundary condition that allows to use the well interference coefficient α as the history-matching parameter between the analytical model and pressure transient data. Once a satisfactory value of α has been determined, the degree of interference δ is estimated from either α vs δ charts or a correlation of the form δ = f(α). Such correlation should be determined using pressure data from either previous interference tests or numerical simulations. In addition, we define the limiting case beyond which the analytical model cannot estimate the exact number of frac hits. In this work, we explore the effect of the stimulated reservoir volume permeability on the relationship α vs δ for 100 ≤ xf ≤ 600 ft.
We use numerical simulations to generate pressure transient data for an interfering dual-well system with 20 fractured stages per well. Reservoir and fluid properties are taken from petrophysical data from the Lower Eagle Ford shale. We artificially impose δ onto the model by manipulating the number and location of the frac hits to generate the α vs δ vs xf curves. For simplification purposes, we assume that both wells have the same properties.
As a result, we demonstrate that the analytical well interference model accurately captures the characteristic pressure transient behavior induced by frac hits, by systematically history-matching α with numerical pressure data for different values of δ, ranging from 0 (no frac hits) to 1 (full frac hits). This procedure allows to obtain δ as a function of α for the cases kSRV = 0.1 and 1 md, and 100 ≤ xf ≤ 600 ft by formulating a nonlinear correlation for each case based on nonlinear regression of history-matched data. Finally, we identify the operational limit beyond which the well interference model will not give precise information about the degree of interference.