Experimental Investigation of Low-Frequency Distributed Acoustic Strain-Rate Responses to Propagating Fractures

Abstract

Low-frequency distributed acoustic strain-rate sensors (LF-DAS) experience strain changes due to far-field fracture propagation. To better understand the LF-DAS response to fracture propagation, we performed laboratory-scale hydraulic fracture experiments with embedded optical strain sensors. The objectives of this research are to generate hydraulic fractures of known geometry, measure the strain response along the embedded fiber optic cable comparable to LF-DAS measurements, and use the results to inform interpretation of field-derived LF-DAS data. The experiments were conducted in unconfined transparent cubic blocks with a dimension of 8-inches on each side. The block was made of transparent epoxy in order to visualize the fracture propagation. Fiber optic sensing cables were embedded in the block with different distances to the source of injection. We injected dyed water through an injection tubing to generate a transverse, radial fracture along an initial flaw. An optical interrogator recorded the response of offset fiber Bragg grating strain sensors normal to the plane of the fracture. The strain data was visualized on a waterfall plot, akin to visualizations of field-derived LF-DAS data. Dimensional analysis was used to scale the lab results to field conditions. We compared the evolution of the strain response at the fiber optic cable, injection pressure, and rate with known fracture geometry. The measured strains were compared to Sneddon's (1946) linear elastic solution for a penny-shaped crack and found to follow this behavior. The generated radial fractures in transparent media can be modeled with Sneddon's linear elastic radial fracture model and a mode I critical stress intensity factor. The LF-DAS characteristic response of a narrowing region of extension surrounded by compression was exhibited as a fracture approached and intersected the fiber optic cable. The experimentally derived strain and strain-rate waterfall plots with known fracture geometry, injection rate and pressure response provide insight in understanding LF-DAS responses in the field. Furthermore, we developed a method to estimate fracture geometry evolution from the fiber optic strain data and validated the method against the experimental data.