Assessment of DFIT Analysis Using Laboratory Experiments

Abstract

Abstract Reliable knowledge of the magnitudes and orientations of in-situ stress is crucial for conducting scientific and engineering activities in the subsurface, especially with regards to subsurface energy and storage applications. A commonly used technique for determining the minimum principal stress is to interpret the fracture closure pressure from the pressure decline transient during the shut-in phase of a minifrac test or a diagnostic fracture injection test (DFIT). However, current minifrac or DFIT analysis methods for interpreting fracture closure pressure often yield inconsistent results, leading to large uncertainties in determining the minimum principal stress. This paper presents a series of small-scale laboratory hydraulic fracturing experiments conducted under true-triaxial compression. The injection scheme consists of a hydraulic fracturing cycle followed by a few fracture propagation cycles and several injection/falloff (DFIT) cycles. The wellbore pressure and acoustic emission (AE) activities of each cycle were concurrently measured to monitor fracture initiation, propagation, and closure during fluid injection and shut-in. The pressure data were used to interpret S3 using different hydraulic fracturing-based methods. The results illustrate that the spatial-temporal evolution of AE activities is well associated with fracture propagation. In relatively low permeable rocks (Test 1 and Test 2), fracture reopening pressure generally provides a reliable estimate of the minimum principal stress (S3). ISIP consistently provides a relatively higher estimate of S3 and can be used as an upper limit for constraining S3. Fracture closure was observed using the so-called "tangent" method in all injection/falloff cycles. However, the "tangent" method using a signature close to the peak GdP/dG tends to significantly underestimate S3. The "compliance" method offers a relatively objective (yet still low) estimate of closure pressure. The signature associated with the change in system stiffness or compliance is observed but not consistently in every DFIT cycle. In the relatively high permeable Scioto Sandstone test (Test 3), the G-function plots exhibit a "normal" leak off behavior, and the "tangent" method provides a good stress estimate, and the compliance method lacks a clear signature for determining fracture closure. The stress interpretation results demonstrate that the process of fracture closure is highly impacted by rock properties, such as permeability and elastic moduli, and cyclic injection operations.