Low-frequency distributed acoustic sensing shape factors for fracture front detection

Abstract

Accurate knowledge of fracture extents generated in multistage unconventional completions remains elusive. Crosswell low-frequency distributed acoustic sensing (LF-DAS) measurements can determine the time and location of a frac hit. Knowing where and when a frac hit occurs constrains the fracture extent but does not estimate it quantitatively. A recent study on crosswell LF-DAS demonstrated a simple method to rapidly determine the instantaneous fracture propagation rate when a frac hit occurs. This method, the zero strain-rate location method (ZSRLM), is based on laboratory experiments and numerical modeling assuming a radial fracture geometry. The method estimates a fracture propagation velocity that is used to extrapolate the final fracture extent. The propagation rate is calculated based on dynamic estimates of the nearest distance from the fiber to the front of a propagating fracture. The ZSRLM is adapted to estimate the distance to the fracture front based on rectangular fracture geometries. A 3D displacement discontinuity method program generates crosswell LF-DAS strain-rate waterfall plots considering a single, rectangular fracture of constant height. Over 150 different simulations were conducted varying formation mechanical properties, fracture height, and the vertical and horizontal offset between the treatment and monitor well. For each simulated case, the ZSRLM is used to estimate the distance to the fracture front based on the simulated waterfall plots. The difference between the estimated and actual distance to the front is corrected by a shape factor. The relationship among the shape factor, fracture height ratio, and vertical offset ratio is determined. Using a shape factor improves the performance of the ZSRLM by up to a factor of two for rectangular fractures. The updated ZSRLM is applied to extrapolate final fracture extents in two field cases: a single cluster stage in the Montney Formation and a multicluster stage of an Austin Chalk completion.