GEOSTEERING THIN BEDS
– Applicability, Challenges, Solutions –
Software assisted geosteering is widely employed for reservoir navigation in unconventional shale plays such as the Duvernay, Montney and the Horn River in Canada, and the Eagleford and Wolfcamp in the USA. It is easily applicable for these thick and laterally continuous stratigraphic sequences that exhibit gamma signatures correlated over vast distances. But is the approach applicable to thin clastic reservoirs?
As long as gamma (or resistivity) signatures are consistent over distances larger than planned lateral length, the method can be successfully implemented. Lateral continuity is key in strat-based geosteering. Shoreface deposits maintain their character over large distances, whereas channel plays tend to be more variable. Mapping depositional environments has the potential to outline areas where remote geosteering is applicable.
Such deposits include the migrating shelf deposits of the Cardium in Alberta, the marine flooding sands of the Viking straddling the Alberta/Saskatchewan border, some McMurray sands, parts of the Glauconitic trend in Alberta, evaporite deposits in Saskatchewan, and lacustrine and marine deposits of the Mannville Group.
Channel plays exhibit extensive lateral variability, making it more difficult to correlate against type wells. Using multiple offset wells alleviates this issue, but in many cases a model based geosteering approach is advantageous, even if it is more laborious.
When gamma profile changes laterally, it becomes very difficult to assess the stratigraphic position of the wellpath. However, it is relatively easy to determine apparent dip along the borehole when encountering duplicate signatures in MWD curves, such as ones registered when drilling up and down shale lenses with limited extent.
The Viking shoreface is relatively continuous on a well scale (1 to 2 km lateral). There are several challenges when it comes to horizontal development of the Viking play. The production rates are relatively low and the decline curves steep. In order to make the play economical, costs are squeezed on all sides.
A potential cost saving solution is replacing traditional wellsite geology with remote geosteering. A team of geosteerers can supervise up to two wells at the same time (not more, due to the extreme speed of drilling).
The remote nature of the supervision makes it impossible to consider cuttings information in the navigation process. While this is certainly a disadvantage, the drill cuttings quality in fast Viking horizontal is dismal to begin with, so the petrographical description input is not very reliable.
There is usually significant variation in gamma values, although the general curve profile is consistent for the most part. Geosteering in the Viking is quick and dirty, but accurate enough to navigate with the same level of precision as petrographical based steering (based on drill cuttings observations) in the field. The addition of azimuthal gamma in the interpretation greatly increases reliability of a geosteering interpretation in the Viking play.
The Cardium Sandstone is usually 4 to 6 meters (12 to 20 feet) thick, often with a layer of conglomerate on top. The lack of Gamma curve character in the upper sand makes it more difficult to pinpoint the stratigraphic position while the bit is in zone, and thus increases the risk of crossing into the hard conglomerate.
There can be significant variation in gamma signature in the upper Cardium Sandstone on a well scale. Good well control and steering using multiple type logs can alleviate the issue to a great extent.
Most operators drill below the main Cardium sand, in the distal ramp siltstone. This makes it a lot easier to geosteer remotely: conclusive gamma signatures, lithological continuity on long stretches, and relatively soft rock make it easier to navigate and stay on target.
The Clearwater sand frequently contains shale lenses and discontinuous shale layers making if more difficult to geosteer remotely. As is the case with the Viking, the Clearwater is drilled at tremendous speeds, rendering drill cuttings almost useless for petrographic analysis. The Clearwater is mostly drilled with multilateral wells at tight spacing. Geosteering findings for one lateral section are easily transferred as learning for subsequent legs and wells. A geosteering software with multi-well capability is extremely valuable when drilling this reservoir.
Other thin reservoirs
A similar approach can be applied in thin reservoirs such as the Peace River Bluesky, sand members of the Upper and Lower Mannville (McLaren, Waseca, Sparky, General Petroleum, Lloydminster, Cummings and Dina) or the Grand Rapids and Wabiskaw sands in the Athabasca region. The Bakken in South-East Saskatchewan and North Dakota is similar in thickness, scope and challenges.
Tools and processes
Azimuthal gamma is of great help when navigating thin reservoirs. By creating image logs along the borehole, it becomes very easy to determine when the borehole is drilled up-structure or down-structure, and thus simplifying the geosteering process. The polarization of curves before a stratigraphic threshold is crossed is also aiding in the reservoir navigation process.
Gas Ratios can give valuable indication when steering. The ratio values are not particularly consistent, there are just too many modifiers linked to rate of penetration, mud weight, tool calibration, level of mud in the trough, flow pressure, etc. However, the change in ratio character is tracking well with wellpath position in the stratigraphic sequence.
Ratios (unlike gamma) are not thrown off by local shale lenses. Gas components, as an addition to total gas detectors, give very cheap information when drilling a well. Ratios produced by fast infrared gas analysis are not very accurate (only components up to C5 are taken into account), but the resulting ratios have enough character to be efficiently used as a geosteering aid.
Resistivity / Induction is measured at different depths of investigation and different frequencies. The combination of bulk resistivity (sometimes combined with azimuthal readings), can be used to determine stand-off to water, stratigraphic layers or lithological contrasts. Resistivity modeling is based on a virtual response of rock layers to current induced at an angle, and the comparison to resistivity tool readings from horizontal wells, acquired with MWD/LWD tools (while drilling). The use of resistivity modeling can solve geosteering puzzles when gamma alone is not sufficient to determine stratigraphic position of the wellbore.
A trove of geomodel information can be used in geosteering reservoirs that exhibit some lateral variability. Seismic surface grids, offset well correlation and mapping based on offsets can all be incorporated into a geosteering model. The resulting 3D modeled stratigraphic surfaces are projected along the wellbore intersection or vertical section plane, and used as guidelines when navigating the reservoir.
A curve correlation approach balances out uncertainty related to directional survey approximation (and even balances out errors in surveys), whereas a geomodel approach does not take cone-of-uncertainty into account. In a typical 1 km long lateral leg placed at a true vertical depth of 2000 meters, the TVD error resulted from survey approximation and cone of uncertainty can reach 3 meters (10 feet). This becomes significant when navigating thin targets. As such, geomodels should be used as guidelines more than diagnostic tools, and a significant vertical deviation must be considered.
Using a variety of techniques and approaches, thin reservoirs can be navigated efficiently using geosteering methods. Strat-based and geomodel-based geosteering can be deployed remotely (from Chinooks Operations Centre or from client real-time operation rooms) or in the field (combined with traditional wellsite geology and drill cuttings petrography observations).