No abstract

In 2017, distributed acoustic sensing (DAS) technology was deployed in a horizontal well to conduct a vertical seismic profiling survey before and after each of 78 hydraulic fracturing stages. From two vibroseis source locations at the surface, time shifts of P- and S-waves were observed but decayed over days. Some stages also showed waves scattered off the stimulated rock volume. We have used 2D finite difference elastic wavefield modeling to understand these observations and connect them to underlying properties of the stimulated rock. We have developed an effective medium model of vertical fractures that close exponentially with time as fluid leaks off into the formation can match the distribution of P- and S-wave time shifts along the well. This has enabled estimates of the height, normal and tangential fracture compliance values, and decay time of the stimulated rock volume. Additionally, the kinematics of scattered waves observed in the data have been found to be consistent with PS conversion across the stimulated rock volume from an individual stage. With higher quality DAS data, stage-by-stage inversion for height, fracture compliance, and decay time attributes may be possible for characterizing variations in the effectiveness of hydraulic fracturing.

Optimization of well spacings and completions are key topics in research related to the development of unconventional reservoirs. In 2017, a vertical seismic profiling (VSP) survey using fiber-optic-based distributed acoustic sensing (DAS) technology was acquired. The data include a series of VSP surveys taken before and immediately following the hydraulic fracturing of each of 78 stages. Scattered seismic waves associated with hydraulic fractures are observed in the seismic waveforms. Kinematic traveltime analysis and full-wavefield modeling results indicate these scattered events are converted PS-waves. We tested three different models of fracture-induced velocity inhomogeneities that can cause scattering of seismic waves: single hydraulic fracture, low-velocity zone, and tip diffractors. We compare the results with the field observations and conclude that the low-velocity zone model has the best fit for the data. In this model, the low-velocity zone represents a stimulated rock volume (SRV). We propose a new approach that uses PS-waves converted by SRV to estimate the half-height of the SRV and the closure time of hydraulic fractures. This active seismic source approach has the potential for cost-effective real-time monitoring of hydraulic fracturing operations and can provide critical constraints on the optimization of unconventional field development.

No abstract

This study presents a workflow of using a convolutional neural network to automatically classify microseismic events originating from a more productive oil and gasbearing formation (the Eagle Ford Shale), as compared to events originating in less productive formations (the Austin Chalk and Buda Limestone). These microseismic events occur due to hydraulic stimulation and are recorded by fibre optic-distributed acoustic sensing measurements from a horizontal monitoring well. The convolutional neural network is trained to recognize guided wave energy in distributed acoustic sensing seismograms, since microseismic events originating within or close to a lowvelocity reservoir (such as the Eagle Ford) generate significant guided wave energy. The training of convolutional neural network is conducted using synthetic seismograms overlain with real noise profiles from field data. Field events with guided waves are then classified by the convolutional neural network as occurring within or close to the Eagle Ford, while events without guided waves are classified as occurring far outside the Eagle Ford. Noise attenuation steps (including a bandpass filter, median filter and non-local means filter) are implemented to increase the signal-to-noise ratio of the field data and improve the classification accuracy. The accuracy of the convolutional neural network is measured by comparison with labels of the events determined by human inspection of guided wave presence. We also evaluate the impact of different network architectures and noise attenuation methods on classification accuracy. The accuracy and F1-score of the final classification are both 0.85 when tested on a high signal-to-noise ratio subset of the field data. On the complete dataset including low signal-to-noise ratio events, an accuracy and F1-score of 0.80 are achieved. These results demonstrate the high effectiveness of the trained convolutional neural network on guided wave detection and classification of microseismic events inside and outside the Eagle Ford formation.