Seismic Reflection Imaging using Horizontal DAS Cables for Long Term Reservoir Surveillance
Brian Fuller, PhD
VP Reservoir Geoscience
Sterling Seismic & Reservoir Services

Introduction
The primary technology driving today’s high level of oil and gas production in the United States is hydraulic fracturing in horizontal wells. Horizontal wells provide thousands of feet of exposure to low permeability reservoir rock and hydraulic fracturing provides a method to open production pathways in what would otherwise be non-productive nano-Darcy rock. Continuous improvement in fracing procedures and technology has resulted in ever larger production rates and total recoverable volume from horizontal wells. Even so, decline rate on many wells can result in marginal profitability even when land, drilling, completion, and production costs are tightly controlled.
In this paper we present new technology and ideas that were conceived and designed by Sterling Seismic & Reservoir Services specifically to provide petroleum engineers with numerical measurements of reservoir properties that will lead to better reservoir management decisions. The technology we present capitalizes on the revolutionary new approach to recording seismic data in boreholes provided by fiber optic DAS (Distributed Acoustic Sensing) systems. Additionally, rock mechanics plays an indispensible role in the methods that we promote.

DAS System Basics
Figure 1 below shows a sketch of a wellbore going into the earth and turning horizontal at depth. In a typical DAS installation, a fiber optic cable is strapped to the outside of casing and is cemented in place. Laser light illuminates the interior of the fiber optic strand. As seismic waves from a surface seismic source (as shown in this figure) propagate past the fiber optic cable the cable is elongated and shortened in the long direction of the fiber optic cable. Elongation and shortening of the fiber optic cable causes light interference patterns. These interference patterns are collected and interpreted by a system at the surface called an “Interrogator” that outputs a seismic waveform for specific points on the cable, typically every 5-15 feet. The spacing of “receivers” in a DAS cable leads to a revolutionary capacity of DAS cable to provide low-cost, high-resolution borehole seismic data. If a well is drilled to 10,000 ft depth and the horizontal section of the well is 10,000 ft long a single surface seismic shot will record between 1500 and 2000 data channels. Borehole seismic data has never before had so much data to work with.

Figure 1. Components of a DAS recording system. This sketch depicts the primary components of the DAS recording environment that we promote. There is a surface seismic source, a wellbore with DAS cable strapped to the outside of casing and cemented in place, and surface equipment that includes the laser and interrogator systems.

Seismic Reflection Imaging With Horizontal DAS Cables
Figure 2 below is a sketch that depicts ray paths from a surface seismic source traveling into the earth, reflecting from interfaces below a horizontal DAS cable, followed by the reflections being recorded on the DAS cable system.

Figure 2. Raypath Sketch. This sketch depicts ray paths from a surface seismic source reflecting from an interface below a DAS cable and then being recorded on a horizontal DAS cable.

The raypath geometry of Figure 2 results in the seismic sources and seismic receivers being at very different elevations. This geometry invalidates the common midpoint assumption that is critical to CDP processing and makes creating a reflection image from data recorded in this geometry far more complex than data recorded with sources and receivers at approximately the same elevation. Processing the data for reflection images in the native geometry described above requires special skills and software in order to be successful in generating velocity fields and migration images. Fortunately, we at Sterling Seismic & Reservoir Services were able to develop a novel method of downward continuing surface seismic sources to the elevation of the DAS cable. In other words we are able to create virtual source points on the DAS cable such that the seismic data was recorded with the seismic sources on or near the DAS cable at depth (See Figure 3).

Figure 3. Downward continued virtual sources on DAS cable. This figure shows the concept of virtual source points being placed on or near the DAS cable at depth even though the actual source points were at the surface of the earth. The value of virtual source points being at the same elevation as the DAS receiver cable is that surface-seismic data processing techniques can be applied to generate reflection images rather than requiring specialized skills and software to generate images with the geometry of surface seismic sources and the DAS receiver cable at depth.

The value of the virtual source geometry is that we can now treat the data like a typical surface seismic dataset with sources and receiver near the same elevation. The data can be processed on any seismic data processing system by anyone skilled in seismic data processing. The number of software tools for processing surface seismic data is enormous and have been proven to be valid for decades. Thus, the results with surface seismic sources and a horizontal DAS cable can be produced in a short period of time using software and imaging concepts that are proven.

Figure 4 shows a map view sketch of a DAS seismic data acquisition project in North America. The blue dots depict the projection of the horizontal DAS cable to the surface from a depth of 9,000 ft. The length of the DAS cable in the north-south direction is about 10,000 ft. The red dashes roughly show the source locations that span approximately 18,000 ft in the north-south direction at a nominal source interval of about 150 ft. The sweep frequency of the Vibroseis source was 4-120 Hz. The horizontal well was toe-up to the north with a slight dip from north to south.

Figure 4. Map view of sources and DAS cable for field survey. This North American DAS seismic survey had a DAS cable at a depth of about 9,000 ft, a horizontal DAS cable length of about 10,000 ft, and a surface source interval of 150 ft over approximately 18,000 ft from north to south.

The first arrival and reflection data was relatively high quality in spite of containing severe interrogator noise (a strong DC shift and additional stripes that cross each shot record when the interrogator unit experienced vibrations from the local environment). The dataset was processed via the downward continuation/virtual source method discussed above and then passed through a fairly standard 2D seismic data processing flow including post stack Finite Difference migration. Figure 5 shows the seismic reflection image that was produced.

Figure 5. Seismic reflection image from horizontal DAS cable. The 2D seismic image produced from surface seismic sources and a horizontal DAS cable is shown here. The trace spacing is 10 ft which is possible due to the DAS receiver spacing of 13 ft and the virtual source technology employed which allows source points to be output at arbitrary locations. The datum elevation for the image is less than 10 ft above the elevation of the toe of the well (toe-up well configuration). The image is in the time-domain. The two-way time of 100 ms is about 650 ft below the datum elevation.

Figure 6 shows the amplitude spectrum of the reflection image in Figure 5. The Vibroseis source sweep spanned 4 to 120 Hz.

Figure 6. Amplitude spectrum of seismic image in Figure 5. The amplitude spectrum of the DAS reflection image has a spectrum that covers the spectrum of the sweep of 4-120 Hz.

The small trace spacing of the DAS system allowed for far higher spatial resolution than is typically produced from surface seismic data. The structural features of the DAS data match the structure observed in the PSTM volume of the surface seismic data while the amplitude spectrum of the surface seismic data has a maximum frequency near 85 Hz rather than the maximum frequency of the DAS data being near 110 Hz.

Petroleum Engineering Applications – 1, Pre-frac stimulation
Figure 7a below is a zoomed version of the reflection image in Figure 5. Figure 7b shows a spatial coherence image that was computed from the seismic image. Light shades in 7b indicate high coherence and dark shades indicate lower coherence where faulting, fractures, and disruption of the wavefield is detected. Finally, Figure 7c shows an overlay of the seismic and its spatial coherence image. We could use Figure 7c to alert frac engineers prior to fracing the well that fracture and fault zones are present at specific locations in the well which may change the frac parameters for frac stages in those zones.

Figure 7a. DAS Seismic Reflection Image. This figure is a repeat of the DAS seismic reflection image from Figure 5 above.

Figure 7b. Coherence Image from DAS Seismic Reflection Image. Trace coherence was computed from the image in Figure 7a. Light shades indicate zones of high spatial coherence while dark shades indicate zones of low coherence where faulting and stratigraphic changes occur.

Figure 7c. Overlay of Seismic and Coherence. Figure 7b (coherence) is overlaid on Figure 7a (seismic image) and the coherence image is set to being semi-transparent. This allows the image traces and their relative coherence to be displayed in a single view. Fracture zones, faulting, and stratigraphic changes are indicated where the coherence values are low (dark shades). Frac engineers would often like to know which frac stages in which they will encounter fractured zones so they can optimally plan the sand and fluid pumping schedule.

Figure 7c indicates where there are particularly highly-fractured rock or fault zones that intersect the treatment well. If DAS reflection data is recorded and processed prior to hydraulic fracture completion then frac engineers can know where fracture zones and faults are present along the length of the horizontal well. Hydraulic fracture stages are likely to be designed differently in a fracture zone than in a zone where there are not likely to be fractures. This reflection image would show them where fractures and faults are likely to exist and at far higher spatial resolution than surface seismic data is capable of doing with typical trace spacing of around 100 ft.

Petroleum Engineering Applications – 2, Long-term reservoir surveillance

An additional long-term use of DAS reflection imaging technology is its use in time lapse seismic reflection imaging. We know from Rock Mechanics analysis that seismic velocities change within a reservoir in response to changes in production-induced stress and pore pressure. We also know that production-induced changes in stress and pore-pressure extend to rock units above and below the reservoir due to changes in shifting loads, thus production-induced changes in velocity also extend above and below the reservoir layers. Figure 8 shows 2D finite element modeling results (provided by Dr. Tom Bratton) in which changes in pore pressure and effective stress are computed as a result of 18 months of fluid production from a single layer in the vertical center of the model. Pore pressure and effective stress changes extend both above and below the producing layer, thus changes in velocity in those layers are also likely.

Figure 8. Finite element modeling results. This figure shows the results of pore pressure and effective stress modeling. A model started with uniform pore pressure and effective stress. Production started from the center layer in the model and continued for 18 months. Pore pressure and effective stress changed within the producing layer but changes also extend to the layers above and below the producing layer.

If a DAS seismic reflection survey is repeated every few months for a producing horizontal well then production-induced changes in pore pressure and effective stress will likely be detected by way of measuring changes in velocity from the time-lapse seismic reflection images. Figure 9a and 9b show the difference in modeled DAS reflection images when the velocity model has been affected by production. In Figure 9a the velocity model was layered and completely flat and in Figure 9b the velocity model was decreased in an interval near the center-right of the velocity model. We see that the reflections in Figure 9B sag in two-way time relative to the baseline survey in Figure 9a. We can invert the changes in reflection travel times between time-lapse surveys and determine changes in interval velocity with depth due to production from the reservoir. We can then give the velocity numbers to rock mechanics experts and they can calculate the likely range of pore pressure and effective stress changes that occurred between the time lapse surveys.

Figure 9a and b. Time lapse seismic reflection time changes. Figure 9a shows the results of a synthetic finite difference DAS reflection survey in which the layered interval velocity field was perfectly flat. Figure 9b shows the results of a “time lapse” finite difference survey in which the flat velocity field from 9a was perturbed (reduced velocity) in an interval near the center right of the velocity field. The reflections in 9b obviously sag in time relative to the reflections in 9a. Time lapse changes in travel time can be used to derive changes in interval velocity which rock mechanics experts can invert for changes in pore pressure and effective stress.

We now look at a hypothetical example of how an oil or gas field might be monitored with time-lapse DAS seismic imaging. Figure 10 shows a map-view sketch of a hypothetical oil or gas field in which pressure sinks and current stress estimates are shown on Well A. The pressure sinks and stress fields would have been derived via rock mechanics analysis of time lapse DAS velocity results. Knowledge of where the well has been productive and non-productive and the likely orientation of the local stress direction will be useful in planning refrac operations. Figure 10 also depicts Well B as a new “child well” offset some distance from Well A. It is well-known that hydraulic fracture stimulation of a new well near prior production can lead to fractures traveling toward production-induced pressure sinks. Additionally, pressure sinks in nearby older producing wells inhibit propagation of new fractures into new rock in adjacent child wells. Thus, it would be useful for reservoir and production engineers to maintain ongoing updates of the pressure and stress field status of a reservoir to optimally plan re-fracs, new drilling, and secondary recovery efforts from oil and gas fields.

Figure 10. Sketch of Hypothetical Field. Production-induced pressure sinks and rotation of the stress field from the regional stress direction are expected in a reservoir. The sketch conveys the idea that time-lapse DAS reflection imaging can be used to maintain a current pressure and stress model of a reservoir for use in production decisions.