Dual Leakoff Behavior in Hydraulic Fracturing of Tight, Lenticular Gas Sands

Abstract

Summary Stimulation experiments conducted in anisotropic, naturally fractured, tight, lenticular gas sandstones have shown the existence of a dual-leakoff phenomenon. Below a threshold pressure, the leakoff coefficient is very low and fluids are very efficient. Above the threshold, leakoff increases by a factor of 50, slurries dehydrate rapidly, and screenouts occur in minutes. The leakoff has been shown to be controllable to some extent with 100-mesh sand. Results of three stimulation experiments are presented, including a treatment that screened out, a minifracture experiment that showed the effectiveness of 100-mesh sand, and a final successful stimulation. Introduction Hydraulic fracturing has proved difficult in many naturally fractured reservoirs, with such problems as high leakoff, early screenouts, and formation damage seriously degrading the expected production enhancement. Laboratory results indicate that fracture offsets and high fluid-loss rates may occur when a hydraulic fracture intersects a natural fracture. Mineback tests show the complex fracturing that may occur, and field experiments document the likely screenout problems. Miller and Warembourg, Nolte, Nolte and Smith, and Conway et al. discuss methods to recognize or to avoid some problems associated with naturally fractured reservoirs, but each reservoir clearly will have its own special considerations. Tests conducted at the U.S. DOE's Multiwell Experiment (MWX) site in Mesaverde rocks in the Piceance basin near Rifle in western Colorado have shown that the natural fracture system of this reservoir dominates both production and stimulation effectiveness. The matrix rocks are submicrodarcy, but the natural fracture systems increase the effective permeability to tens of microdarcies, just suffi-cient to make production of these reservoirs feasible if optimum stimulations can be obtained. Stimulations at this site, however, tend to screen out within minutes of sand entering the perforations and postfracture well testing has indicated that significant formation damage is created by the hydraulic fracture, presumably because of residual gel/liquid in the natural fractures that intersect the hydraulic fracture. In recent tests, the emphasis has been on understanding the interaction of the hydraulic and natural fractures and on controlling the screenout and damage problems. The results from the last three stimulation experiments, all in len-ticular sandstones, have shown that the problem results primarily from a dual-leakoff phenomenon. At relatively low injection pressures, the base-fluid leakoff coefficient is very low; its magnitude pressures, the base-fluid leakoff coefficient is very low; its magnitude is consistent with effective reservoir permeabilities of tens of microdarcies. Above some critical pressure, however, the leakoff accelerates and appears to become as much as 50 times greater than the base level. When the proppant stage enters the perforations under these high-leakoff conditions, the slurry rapidly dehydrates and screenout occurs. The major difficulties have been recognizing the problem and finding a way to control the accelerated leakoff. These problem and finding a way to control the accelerated leakoff. These are the subjects of this paper. We used pressure analyses, during both injection and pressure decline, to recognize the problem. The characteristic behavior during these MWX stimulations is a rapid flattening of the Nolte-Smith log-log plot above some critical pressure. Diagnostics has never shown much height growth, and the compliances of all the rocks in this interval are similar, so the likely candidate for this behavior is high leakoff. Unfortunately, the pressure-decline behavior invariably yields a low leakoff coefficient, even at early times. The combination of these factors led us to hypothesize an accelerated leakoff mechanism above a threshold pressure. This hypothesis was supported by pressure-history matches of the stimulations that showed that a dual-leakoff mechanism could account for the complex pressure behavior. The mechanism for this accelerated component of the leakoff is clearly the natural fracture system. Once it is recognized that the problem is leakoff and not height growth, complex fracturing, or some other process, leakoff must be controlled through some additive or change in the treatment schedule to keep the injection pressure below the threshold value. In this application, it was impossible to lower the pressure sufficiently and still carry proppant, so we considered leakoff additives. Because the natural fractures are the primary production paths, additives that would entirely plug these systems were not practical. A 100-mesh sand appeared to be an effective additive because it would bridge off excess leakoff but leave a propped, conductive flow path for subsequent production. Two minifractures were conducted in one sandstone lens to investigate the effects of 100-mesh sand, and the results were positive, yielding a reduced leakoff under high-pressure conditions. Finally, a successful stimulation was carried out with these findings. This paper summarizes the results of these three sets of experiments.