The Effect of Stress Sensitive Permeability Reduction on the Evaluation of Post-Fracture Welltests in Tight Gas and Unconventional Reservoirs

Abstract

Abstract The interpretation of well performance in propped fractured tight gas reservoirs is a topic that has been studied for decades, but still presents a number of difficulties. These issues are becoming more important as unconventional reservoirs with even lower permeabilities are developed (micro and nano- Darcy). One of the most troublesome issues in tight-gas fracturing is that post-frac welltests often show very short effective fracture lengths. This is often blamed on gel damage due to poor proppant pack cleanup in the fracture or loss of propped width due to embedment and filtercake in the fracture. It has also been attributed to formation damage, due to “phase trapping” of fracture fluid filtrate along the fracture face. Much work has already been done on these topics, which show that extreme damage to the proppant pack or the fracture face would be needed to explain the short effective fracture lengths that are often seen. Regained permeability testing, when available, in many cases does not indicate that this much damage is expected, either in the proppant pack or on the fracture face, under normal circumstances. In this paper, a large number of high resolution 3D numerical simulations have been performed, based on realistic reservoir parameters for a typical tight-gas reservoir, using stress sensitive permeability data derived from core testing. By modeling both the propped fracture and the reservoir with fine grid refinement, it is possible to simulate drawdown and buildup pressures with and without the effect of stress sensitive permeability in the reservoir matrix. The simulations also include the effect of the effective stress on the proppant pack, which is generally not significant in tight gas, unless the proppant concentration is extremely low. These synthetic drawdown and buildup data sets with known parameters are then analyzed using a standard welltest analysis package, and the results are compared. The results of this work show that for tight and ultra-tight reservoirs, conventional welltest analysis methods will always show a smaller fracture length and a larger permeability, compared to the “correct” values used to generate the test data sets. This behavior is partly due to the rectangular fracture geometry assumed in the welltest solutions. An additional reduction in length can be explained by stress sensitive permeability, without any need for damage to the proppant pack or the fracture face. The results from the incorrect welltest analysis are used for a long-term forecast simulation (with a short fracture length, higher permeability and ignoring stress sensitive permeability), which is compared to a forecast based on the model with the correct fracture length and stress sensitive permeability. This shows a significant overestimation of the ultimate recovery when using the welltest derived parameters.