Proppant Holdup, Bridging, and Screenout Behavior in Naturally Fractured Reservoirs

Abstract

Abstract Pressure spikes and early, rapidly developing screenouts are often observed while hydraulically fracturing naturally fissured reservoirs. These occurrences are commonly attributed to fracture width restrictions or slurry dehydration. This paper describes the results of a series of laboratory experiments designed to study proppant transport and distribution in a hydraulic fracture with leakoff through discrete fissures. Mechanisms are described which explain the observed phenomena of proppant induced pressure rises and rapid screenouts. Methods of treatment design to combat these effects are also outlined. Observations made during the experiments provide insight into proppant crushing during fracture closure, poor polymer gel recovery, and final fracture conductivity. The experiments show that lateral flow velocity into natural fissures causes proppant to migrate to the fracture wall and build up a dense proppant pack at the leakoff site. With sufficient leakoff the proppant holdup is severe enough to completely fill the main fracture channel, leading to screenout, even at low injected proppant concentrations. During buildup of the proppant bank the injected fluid is forced to flow in channels of diminishing height at very high shear rate. This unstable channelized flow leads to significantly different proppant and fluid distribution in the fracture than that predicted by uniform flow and transport models. Introduction A successful fracture stimulation treatment requires the creation of a desired fracture geometry filled with a sufficient concentration of conductive propping material. Fracture design simulators are commonly employed in an attempt to achieve this goal economically and with minimal waste of resources. To be successful, a fracture design model must correctly represent both fracture geometry development and fluid and proppant flow. However, most of the effort expended so far in development of these models has centered on fracture growth calculations with relatively little effort on detailed proppant transport modeling. Most of the proppant transport modeling published to date is based on theoretical and experimental work with no fluid loss.1–3 Predictions made with models based on these findings, or based on simplified theories of homogeneous or "perfect" transport are frequently misleading. Especially troubling is the difficulty encountered in predicting fracture screenout conditions when using measured leakoff data from pre-frac fluid efficiency or leakoff tests. This paper presents the results of a series of flow experiments conducted to study proppant transport in the presence of discrete leakoff sites representing natural fractures or fissures intersecting the wall of a propagating hydraulic fracture. Fluid rheology, input solids concentration, leakoff rate, and total pump rate were varied to determine the effects of various parameters on proppant transport. Some results for more uniformly distributed matrix leakoff are also included. Results of the experiments show that leakoff has a much more significant effect on transport than previously believed. While the most comprehensive proppant transport models in current use include the effects of slurry dehydration caused by leakoff, these effects alone are not sufficient to describe the results of leakoff on overall solids transport. The studies reported here show that the particle velocity profile across the fracture width and rate of transport down the fracture are significantly affected by leakoff, even at relatively low rates. Localized high leakoff can result in the complete immobilization of proppant in the vicinity of leakoff sites.