Establishing Flow Regimes for Multi-Fractured Horizontal Wells in Low-Permeability Reservoirs

Abstract

This paper presents rigorous theoretical guidelines for durations of flow regimes for multi-fractured horizontal wells in ultra-low permeability reservoirs. Theory and practice lead us to expect four regimes: early ramp-up, transient, transition, and boundary-dominated flow (BDF) in these wells. We must model each of these flow regimes for proper forecasting and for construction of TWPs (aka type wells or type curves), but, without guidance from theory and verification in practice, the durations of these flow regimes are difficult to observe in production histories or to predict when forecasting. These forecasts have significant impact on financial decisions regarding low-permeability reservoir development. We can most readily identify flow regimes using log-log plots of pressure-normalized rate vs. time for wells produced at near-constant bottom-hole pressure. This is adequate to determine the start and end of transient flow, with a straight line whose slope is near −1/2. Diagnosis is enhanced if we add normalized rate vs. material-balance time plots, which transform the well response to an equivalent constant-rate profile, on which we can identify BDF with a straight line with −1 slope. On this plot, the transition flow regime lies between the end of transient flow and the start of BDF. In some wells, with relatively longer production histories, we can readily identify these flow regimes, but many if not most wells in a play will display neither transition nor BDF regimes. To fill this gap in knowledge, we simulated flow histories using analytical solutions, which provide shapes and durations of the flow regimes. Starts and ends of flow regimes depend on arbitrary assumptions about deviations from straight lines, which can be determined in theory using derivatives of the analytical solutions. In practice, wells do not follow theory exactly by any means, but we find in our examination of actual well production histories that theory provides excellent guidance that enhances our understanding of actual production profiles. We present our simulated production histories for wells in terms of dimensionless variables, which generalizes their applicability. For actual situations, with known or estimated reservoir and completion properties, we can use these plots of dimensionless variables to determine approximate durations of flow regimes. Importantly, for the common situation in which no production data are available beyond transient flow, we can estimate the shape of the remaining production profile in a way significantly superior to the common two-segment Arps decline model with an assumed terminal decline rate at an assumed time. Critics of the industry, particularly in the financial community, have suggested that this common approach leads to optimistic production forecasts. Realistic forecasts of production profiles for individual wells, which our workflow based on rigorous theory enhances, can improve the credibility of resource evaluators within and beyond individual companies. This is especially important for TWP construction, on which many important financial decisions are based.