Monitoring hydraulic fracture growth: Laboratory experiments

We carry out small-scale hydraulic fracture experiments to investigate the physics of hydraulic fracturing. The laboratory experiments are combined wi th time-lapse ultrasonic measurements with active sources using both compre ssional and shear-wave transducers. For the time-lapse measurements we focu s on ultrasonic measurement changes during fracture growth. As a consequenc e we can detect the hydraulic fracture and characterize its shape and geome try during growth. Hence, this paper deals with fracture characterization u sing time-lapse acoustic data. During fracture growth the acoustic waves generate diffractions at the tip of the fracture. The direct compressional and shear diffractions are used t o locate the position of the tip of the fracture. More detailed analysis of these diffractions can be used to obtain information on the geometry and c onfiguration of the fracture tip, including the creation of a zone that is not penetrated by fluid. Furthermore, it appears that the acoustic diffract ion is generated mainly at the fluid front and only weakly at the dry tip. In addition. the wavefield that has been transmitted through the hydraulic fracture is measured. Shear-wave transmissions are shadowed because the she ar modulus vanishes inside the fluid-filled fracture. From this observation we conclude that the fracture is mechanically open. Tn other words, no fri ction occurs related to the movement of fracture faces that are in mechanic al contact. Compressional transmissions show a distinctive dispersion relative to the m easurement in the unfractured medium. This dispersion can be used to determ ine the width (or aperture) of the fracture by fitting the measured dispers ion with the theoretical prediction as a function of the unknown fracture w idth. We show that the width profile of the fracture can be reconstructed b y using a set of transmission records with different source and receiver lo cations. By performing a validation experiment, we show that the width dete rmination method is reliable, although the estimated fracture width is only a few percent of the incident wavelength. The strength of the method relie s on time-lapse measurements combined with fitting the changes in the measu red waveforms during the experiment. The combination of diffractions and transmissions helps us visualize the dy namic process of hydraulic fracture growth. Hence, acoustic measurements wi th active sources prove their usefulness for fracture characterization.