Finding fractures the hard way

Offshore California, the Monterey shale offers several highly conductive producing zones in the form of swarms of natural fractures. Since the fractures represent the only conduit for commercially viable hydrocarbon production from the Monterey, their identification and characterization is an essential objective of any well evaluation program. The only problem is

Figure 1. 3-D representation of the Monterey formation (6 strata, from top to bottom: blue, green, yellow, magenta, maroon and dark blue) being exploited from the Gail Platform shows the trajectory of the E-6 well redrill (green). (Image courtesy of Venoco Inc.)

that fractures occur in two types — drilling-induced, which are shallow and generally non-conductive, and high-conductivity natural fractures, which are desirable and sought-after. In addition, log data perturbations can be caused by the natural bedding planes resulting from sedimentary deposition or from near-wellbore stress concentration. Accordingly, logging programs used in the area must be able to identify and quantify the fractures intersecting the borehole and characterize them, while discriminating signals from the bedding planes or from near-wellbore stress.

Solving a tough problem

Finding and characterizing the fractures that represent viable completion targets is not easy. However, over the years, operators have come up with a logging technique that works. The optimal openhole logging program calls for an integrated gamma ray, induction, compensated neutron/lithodensity log, augmented by a dipole sonic acoustic log and high-resolution resistivity micro-imaging tool.

Frequently, the operator runs only the dipole sonic/micro-imager combination, because the sole logging objective is fracture identification and characterization. This technique has evolved over the past 7 years’ drilling campaign and was augmented by a Department of Energy (DOE) grant awarded to the operator, Venoco Inc. The tools permit comparison of shear anisotropy data from the dipole sonic with fault and fracture images from the micro-imager. The fracture images are used to “calibrate” the sonic data so fracture width can be quantified.

Venoco has been using the dipole sonic/micro-imager combination successfully in openhole to evaluate its wells and find the fractured zones.

Solving a really tough problem

On Venoco’s Gail platform offshore California, drillers prepared to redrill the E-6 well, a tricky S-shaped well targeting the fractured Monterey as well as deeper formations, and deviating

	Figure 2. The Sonic Scanner tool (top end is on the left) (Image courtesy of The Oilfield Review)

up to 92°— 61° to 78° through the zone of interest (Figure 1). But drilling difficulties caused them to switch to oil-based mud. This caused a few problems, but it was believed that they could be resolved by substituting the Oil-Base Mud Imaging tool (OBMI) for the Fullbore Formation MicroImager (FMI). Although sometimes obtaining logs under difficult drilling conditions can be resolved by using logging-while-drilling (LWD) tools, it was decided that the required high-resolution fracture imaging could not be obtained with currently available LWD tools. Ultimately, the difficult drilling conditions caused cancellation of openhole logging altogether.

Forced to rely on cased-hole logs to evaluate the well, Venoco planned to run an UltraSonic Imaging Tool/ Cement Bond Tool (USIT/CBT) combination on drill pipe to at least
get a cement evaluation. Fortunately, the decision was taken to substitute the new Sonic Scanner acoustic scanning platform from Schlumberger for the CBT tool to possibly identify and characterize fractures behind casing using shear anisotropy.

Matching the model

In a typical well evaluation program, the multifrequency Sonic Scanner tool presents a borehole anisotropy analysis that includes a slowness frequency analysis and slowness time

Figure 3. Dispersion analysis 101—by observing how the fast-shear X dipole (blue) and slow-shear Y dipole (red) dispersion curves relate to each other and to the homogeneous isotropic model (black circles), analysts can resolve the cause of observed anisotropy. Under homogeneous isotropic conditions, flexural wave dispersion follows a predictable increase in slowness with higher frequencies (upper left). Inhomogeneity or anisotropy cause the dispersion curves to digress from the model in recognizable patterns that indicate the cause (upper right, lower left and lower right). (Image courtesy of The Oilfield Review)

coherence projections for both fast inline and slow inline waveforms. Inline in this context refers to the orientation of the tool’s crossed dipole transmitters with respect to the receivers being used. A built-in direction and inclination feature orients the tool in geospace. Low frequency Stoneley data are interpreted with wideband dipole flexural wave data and calibrated by actual fracture images from the FMI tool. Key to the discrimination analyses that reveal the source of the anisotropy are the dipole dispersion curves that result when slowness is plotted vs. frequency for the X (fast) and Y (slow) dipoles and the slowness modeled in an elastic impermeable formation. Four end-possibilities exist: homogeneous/ isotropic, homogeneous/anisotropic, inhomogeneous/isotropic and inhomogeneous /anisotropic (Figure 3).

Using a process of elimination together with a process called normalized differential energies, service company log analysts can discriminate natural fractures from drilling-induced fractures even when anisotropy is as low as 2%. The FMI tool enables detection of sedimentary bedding features as well as quantification of fracture apertures.
The technique has been used successfully in the US Rocky Mountain area as well as the North Sea and, to a lesser extent, the Gulf of Mexico. In the Rockies, the hard rock completion solution facilitated by Sonic Scanner and FMI data is incorporated in the whole earth model that is used in stimulation design. In addition, knowing the location of the fractures allows special cementation techniques to be incorporated to avoid damaging the formation’s natural permeability.

But the company had never been faced with the challenge of fracture identification and characterization in a cased hole environment, where FMI images were unavailable.

Accepting the challenge

Faced with solving the E-6 evaluation problem using only the available logs, scientists at the service company’s/Doll Research lab in Boston developed a sophisticated model of acoustic

Figure 4. Natural fractures in the E-6 well as interpreted from Sonic Scanner data with parallel dispersion curve patterns. (Image courtesy of Schlumberger). Log presentation (left to right) applies to this, and all subsequent figures. Depth Track: Max. offline energy (dashed green); Min. offline energy (solid green). Track I: Spectral Gamma Ray (Thorium-dark green, Uranium-purple and Potassium-light green); Bit size (black dashed); Hole azimuth (blue); Deviation (red dotted); Fast Shear Azimuth (dark red). Track II: Dtc-based anisotropy with color flags (0%-2% grey, 2%-4% light green, 4%-6% dark green, 6%-16% yellow, over 16% red); X-dipole Fast Shear (Blue); Y-dipole Fast Shear (Red); Dtc (black); Time-based anisotropy (yellow). Track IV: Slow shear Variable Density Log (VDL). Track V: Fast shear Variable Density Log (VDL). Track VI: Bond attenuation Db/m (purple). Track VII: Cement Bond Log (CBL/VDL).

velocities and direction of fast shear azimuths. Using all available data, the model can create a homogeneous isotropic solution, which represents the formation being studied as it would appear with no fractures. Then any observed anomalies in the actual recorded data are compared with the homogeneous isotropic case to reveal characterization clues. The model contains forward algorithms that allow “What if?” games to facilitate decision-making. The availability of this new model gave confidence to the idea that at least a qualitative solution to fracture identification, orientation and character could be attempted in cased hole using the Sonic Scanner. Its pure, high-quality data can detect shear splitting (anisotropy) by looking at the dispersion curves (Figs 4-6). A few ambiguities exist, but they could be resolved using logic and careful analysis of the fast shear azimuth. In these cases the model suggests the most likely answer. Where the model based on formation dip data was inconsistent with actual results natural fractures were identified as the cause.

	Figure 5. Stress-induced fractures as interpreted from Sonic Scanner data with characteristic crossover of dispersion curve patterns. (Image courtesy of Schlumberger)

Meanwhile, while the computer model was being developed, Venoco engineers completed the E-6 well in the Monterey using correlations from shows indicated on the logs. The well is presently shut-in, and the engineers are busy designing the completion strategy for the shallower target zones. Typically, the naturally fractured zones are treated with acid to enhance permeability. They are not further stimulated using hydraulic fracturing techniques. Venoco will use the Sonic Scanner results to re-evaluate the zone they previously completed as well as identify the best candidate zones for subsequent completions. The company is confident that the fracture-identification results obtained through casing significantly reduce decision-risk and increase the odds of a high productivity completion.

Asked to comment on the significance of the limited results obtained under difficult conditions in cased hole, a company spokesperson said, “While we obviously prefer to have a full suite of openhole logs, it is encouraging to know that we can obtain satisfactory results and reduce decision risk in situations when they are not available — it’s a good ‘Plan-B,’ and we are pleased with the outcome.”

Applications in other areas

The data from the Sonic Scanner tool should show how to optimize hydraulic fracturing results for maximum productivity in zones with good containment. In addition, it has been

Figure 6. Typical Monterey shale response in an unfractured zone. (Image courtesy of Schlumberger)

postulated that operators can resolve hydraulic fracturing leak-off problems experienced when hydraulic fractures propagate into natural fractures. For example, in Wyoming’s Wamsutter field fractures typically propagate only 300 ft (91 m) or so due to leak-off into natural fractures. Optimum performance is about 1,200 ft (366 m). It is believed that by using Sonic Scanner results and the computer model, pumping service companies can design frac jobs with diverters and modified pumping schedules that prevent the fracs from propagating into the natural fracture swarms, thus losing energy. Optimum results may be achievable.
Most importantly, the successful application of Sonic Scanner technology to evaluate fractures in cased holes opens up the possibility to use the technique to re-evaluate old wells that may still possess commercial production potential. The service could become a key ingredient in pre-abandonment strategies, ensuring that no pay is left behind.