Loading...
Time-lapse multicomponent geophone and DAS VSP processing and analysis
Schultz, Whitney H.
Schultz, Whitney H.
Citations
Altmetric:
Advisor
Editor
Date
Date Issued
2019
Date Submitted
Keywords
Collections
Research Projects
Organizational Units
Journal Issue
Embargo Expires
Abstract
Various geophone and distributed acoustic sensing (DAS) vertical seismic profile (VSP) datasets acquired in south central Texas are processed for P-wave and PS-wave imaging. The surveys acquired and analyzed in this project were time-lapse zero-offset and walkaway recorded on both geophones and DAS, and a 3D survey recorded only by DAS. This project focuses on reprocessing and analyzing the various datasets to evaluate the differences between the geophone and DAS VSP surveys and to analyze the time-lapse changes from hydraulic fracturing. The DAS performs comparably to the geophones and records PS-waves well enough to be processed and interpreted. DAS fiber has significant advantages over geophones in large VSP surveys due to the larger receiver depth aperture. DAS receiver spacing is also denser than geophone arrays, but DAS records only the particle motion parallel to the fiber whereas geophones can record three-components. A P-wave velocity model is developed from the first-break picks of the zero-offset geophone VSP and used in the processing of various other datasets. Using the Vp/Vs ratio from sonic logs, the corresponding shear wave velocity model is developed. The velocity models are used as an input to the walkaway and 3D VSP processing. The 1-D model has limitations, but is a useful starting point toward development of an integrated 3-D velocity that can be used to processes various surface seismic and microseismic datasets. While converting the zero-offset DAS to geophone response, noise is smeared across the gather and the higher frequencies are lost. Aggressive 2D median filtering is necessary, but the result ties well to the geophone zero-offset gather. Imaging of the time-lapse geophone and DAS walk-away surveys is completed using the 1-D velocity models and ray tracing for the VSP reflection traveltime correction and common reflection point (CRP) binning. Unfortunately, the receiver aperture for the geophone time-lapse survey is limited to 1500 ft. The radial component contains the most P and PS-wave energy and is probably the most useful component for this survey. The limited receiver aperture also limits the offset from the well which can be imaged to $\sim 2500$ ft. When compared to the microseismic data, it becomes apparent that this survey is not imaging a significant portion of the well, and even less for the PS image ($\sim 1000$ ft). Since the hydraulic fracturing occurred in Well A and the geophones are located in Well B, the time-lapse response is expected to be minimal. Currently, the time-lapse response is noisy and further processing is probably necessary to produce a conclusive time-lapse response. The DAS walkaway survey contains a much larger receiver aperture (full length of the well) which improves the ability to cross-equalize the datasets using strong reflections located in the overburden. The fiber images the P-wave reflections very well and also records enough PS energy to image. Overall, the DAS walkaway is comparable to the vertical component for the geophone data, with a much improved receiver aperture. When compared to the microseismic, there is a potential time-lapse response, however, the image is noisy and would benefit from further time-lapse processing. The 3D DAS VSP is imaged using the same ray tracing method and velocity model as the walkaway survey. The PS-wave reflections are recorded on the fiber and produce very good CRP mapping results. Even with potential lateral velocity variations within the survey area, the 1-D velocity model is valid to the first order. The results tie well to the dataset after a few velocity model modifications. Additionally, the 3D DAS VSP is sectioned off into 4 azimuth lines in order to perform both a reverse time migration (RTM) and a least-squares RTM in order to compare the migrations to the stacked results. The migrations are too computationally intensive to run on the full 3D dataset. The least-squares migration performs better than the RTM, but both migrations do not provide as clear an image as the CRP stacking performed on the same azimuth lines. The migrated images are negatively impacted by the irregular source spacing along the azimuth lines and the resulting gaps in reflection points.
Associated Publications
Rights
Copyright of the original work is retained by the author.