Geosteering CTD, using the drill bit as a sensor

In today’s oil price environment, it’s more important than ever to place the wellbore in the most productive formation. The depth of the formation on a well path is never known with absolute certainty until the well is actually drilled. 

There are two ways in which this problem can be minimized. The first is to rely on seismic data and as-drilled data, combined with a certain amount of linear interpolation and assumptions on the properties of the rock. The second is to use geosteering. The latter is more accurate than the former, but neither offers an optimal solution. Let’s remind ourselves why.

LINEAR INTERPOLATION AND ITS LIMITATIONS

Fig. 1a. A planned sidetrack well, based on linear interpolation of formation tops of a reservoir. Fig. 1b. An actual sidetrack well misses the formation, due to a varied formation top depth between the vertical wells.

Figures 1a and 1b show the problems that can occur when relying on linear interpolation. Knowing the depth and dip of the top of the reservoir, and having knowledge of the performance of the directional drilling equipment proposed for the job, it is perfectly possible to plan a wellbore trajectory. This is based on a casing exit depth and a build rate that will land the lateral at the required distance below the top of the reservoir, combined with a target inclination which will track the top of the reservoir. A modern directional CT drilling BHA will be able to feed positional information (inclination and azimuth) to the surface as drilling proceeds. This information can be combined with bit depth to provide an accurate estimation of the borehole’s position in 3D space.

Seismic data and offset wells provide valuable information when planning a well, but there are inherent uncertainties in TVD, based on the parameters and assessments made by the geophysicist. This is corrected to a certain extent by combining seismic interpretations with formation tops, taken as-drilled on nearby wells. 

However, this still leaves an amount of uncertainty in the exact formation top location. A further problem arises, when the formation is too thin for the trajectory to be planned below the lower uncertainty limit of the formation top, or if the formation comes in at a different depth to that predicted by the subsurface team. If the seal drops below this line, as pictured in Fig. 1b, the “dead reckoning” navigated well risks being drilled entirely in the seal. 

The issue is not only a problem at the top of the reservoir but also at the bottom. If the formation below is unproductive and comes in shallower than the expected depth, the well profile could pass entirely through the unproductive deeper formation. 

What the driller needs—to be able to avoid this non-profitable outcome—is information about what he is drilling through, in addition to pure positional information. This is what geosteering provides. 

GEOSTEERING OPTIONS AND CHALLENGES

In geosteering, properties of the formation being drilled are measured and used to modify the wellbore plan, as it proceeds. This offers a considerable advance on simple interpolation, but there are still drawbacks. We’ll look at these briefly and then consider a new development—RockSense—that overcomes them and offers a greater degree of drilling accuracy, at a reduced cost compared to LWD systems.

Observing cuttings at surface

The most basic (though by no means primitive) geosteering technique is to directly observe cuttings at surface. Cuttings are collected as they pass over the shaker, and subjected to optical, chemical, spectrographic or electronic analysis. There is a delay inherent in this process, because of the time taken for cuttings to travel from the bit to the surface, and the time taken to physically prepare and analyze the sample, once it has arrived at surface. 

In deeper wells, this time can be up to an hour. During this time, the wellbore may have progressed 30 ft or more, something that presents complications, should the cuttings data not match the prognosis. Another significant drawback to this method is that the resolution is low. By the nature of the cuttings transport, the cuttings are dispersed on their journey to the surface, which can make it hard to tell exactly when the bit passed through the formation top. 

Adding Sensors to the CTD BHA

The CTD BHA can be equipped with sensors that measure parameters which characterize the formations being drilled. They may provide information more quickly than observing cuttings, but they also add significant cost to the BHA.

Modern logging while drilling (LWD) tools can accurately measure formation gamma signature, resistivity, density and porosity. The above tool-mounted sensors are mature, reliable and offer a high degree of discrimination between formation layers. However, they do suffer from the disadvantage of being located, by necessity, tens of feet behind the bit. The mud motor needs to be nearest the bit, and above that, near bit survey tools, to give the driller the earliest possible indication of direction change. In a certain slimhole tool, the formation evaluation sensor is approximately 25 ft behind the bit, which reduces the effectiveness of the information considerably.

THE COMMON PROBLEM 

The distance travelled by the bit, between the time when the formation boundary is crossed and the time when the fact of the crossing becomes apparent, is a problem. It is especially significant in layered and thin formations, or in formations that don’t have suitable markers to distinguish between the different layers or boundaries, situations that are very common in re-entry drilling. 

Fig. 2a. A steering decision, based on an up-tool sensor. Fig. 2b. A steering decision, based on at-bit measurement, allows earlier trajectory correction to stay in the zone.

Figure 2a shows a conventional formation evaluation sensor in the drilling BHA. This could be, say, 35 ft from the bit (L_sense). When the bit crosses the boundary from one formation to another, an additional 35 ft will need to be drilled before the change in formation is visible to the sensor. If the thickness of the formation is of the same order, then it will not be possible for the driller to execute a steering action before the drill bit has left the formation of interest. This has a cost, in terms of lost production and time wasted drilling unproductive formation. 

Were formation discrimination available at the bit, then this depth lag could be eliminated. The latest generation of coiled tubing drilling BHAs provides the data to make this new geosteering method possible.

A SOLUTION TO ADDRESS PAST CHALLENGES

To understand how this new geo-steering method works, let’s take an analogy. Picture yourself as a passenger in a moving car. Even with your eyes closed, you know the type of road you’re travelling on (freeway, city street, country road) by the road noise you can hear.

Consider, now, a motor turning a drill bit that drills a hole in a sample of rock. We can measure the power input to the motor, as the hole is drilled, to gain an understanding of the type of rock we’re drilling. If it was an electric motor, we would simply measure the voltage applied and the current flowing during drilling, and multiply them to get the instantaneous power. For a positive displacement mud motor, it’s slightly more complicated, but it can be done. If differential pressure and flowrate can be measured then, given knowledge of principal operating constants for the motor, an expression for power, in terms of pressure and flowrate, can be written. 

If we were then to integrate this power as the hole progresses, we would have a value of energy expended per foot of hole drilled, and would therefore have a relative indicator of the changes in formation being drilled.

UNDERSTANDING THE THEORY BEHIND THE SOLUTION

A drilling bit progresses a hole by breaking the bonds between rock particles. The rate at which the hole progresses is dependent on the rate at which these bonds are broken. The rate at which the bonds are broken depends not only on the rate of penetration, but also on the number of bonds in a specific volume of rock. That is to say, a highly porous sample of rock will have fewer bonds per cubic foot then a less porous sample of the same material.

Similarly, two samples of equal porosity but different materials may have a different energy per volume, due to the different intergranular bonding mechanism of the different materials. For many rock types, there is a clear relationship between porosity and bond energy.

In the idealized case of 100% transmission of power between the input to the motor and the rock face, we can say that the power input to the motor will equal the power used to break inter granular bonds in the rock, as the hole progresses.

In reality, the motor will not be 100% efficient, and neither will the mechanical interaction between the rock and the bit. In the former case, some energy will be wasted in the motor. In the latter case, the bit can bounce in the hole, and heat can be generated as the cutter moves against the rock. We can, however, say that: 

E₁P_m = E₂ P_r             (1)

Where E₁ and E₂ are expressions that encapsulate the efficiency of the motor and cutter, Pm is the power supplied to the motor, and Pr is the power effectively used by the cutter in drilling the hole. (In this formulation, E₁<1 and E₂>1).

Now, let us assume two things. Let us assume that the behavior of the motor is well-understood. It follows that its efficiency under various operating conditions will be predictable, given knowledge of those conditions.

Accordingly, we can modify our raw measurement of power supplied to the motor, to give a more accurate measurement of power supplied to the cutter.

The other assumption is that we have sufficient knowledge of the BHA and bit disposition at the bottom of the hole to enable us to keep the bit in efficient engagement with the rock it is cutting. For modern, latest generation coiled tubing drilling BHAs, equipped with Torque and Weight-On-Bit sensors and high-speed real-time telemetry, this is a very reasonable assumption.

The result of this is that we can establish a relationship between our measured power supplied to the drilling motor and the volume of rock removed. (We know the volume of rock removed, because we know the hole size, and we have a record of depth vs time.)

We can then plot our modified energy per foot vs depth. What this plot will be telling us is the energy required to progress the hole by a unit of length. The efficiency functions described above have taken care of extraneous losses, so any remaining variation in volumetric energy with depth results from variation in the properties of the rock, itself.

What we have explained, so far, will be well known to anyone familiar with Mechanical Specific Energy. MSE was first proposed by Teale in 1965 . Technological limitations have impeded the adoption of MSE as a mechanism for at-bit formation evaluation since that time. These limitations are low bandwidth on mud pulse telemetry, and inability to directly measure downhole torque and Weight On Bit. 

The slow data rates were a critical impediment, since for MSE to deliver the inch-level formation discrimination it is theoretically capable of, requires band width substantially in excess of what is possible when using mud pulse telemetry. 

These problems are solved with modern coiled tubing drilling BHAs. Designed for underbalanced drilling, they use high bandwidth wired telemetry in place of mud pulse, and their integrated sensor package includes Weight On Bit and Torque On Bit sensors. We can now get information on the type of rock being drilled in real time with inch-level resolution.

This disruptive capability is achieved without new hardware, meaning no additional equipment cost and no increase in BHA length. The measurement is made at the bit, not tens of feet behind it, enabling a responsiveness in directional control never seen before. 

THE NEW GEOSTEERING THEORY IN PRACTICE

A coiled tubing drilling BHA was used to sidetrack a well in North America. This was a reasonably densely drilled locality, and open-hole logs for historical vertical wells were made available. No horizontal wells had been drilled in the area previously, and the objective was to increase production through increasing reservoir contact. The operator used 3D seismic to evaluate the formations and identified a subsurface ridge that could be acting as a trap. 

Fig. 3. Compelling similarity in the shape of the RockSense log of a sidetrack, and the neutron porosity log for the vertical motherbore, when plotted against TVD.

The well path was planned to pass approximately 15ft below the formation top, and track the formation by holding inclination. The oil-water contact was believed to be 40ft below the formation top and, if it was entered, would significantly impact the well economics. Gamma ray sensors were used for depth correlation, because relying on seismic depth, alone, would not provide the depth accuracy required. The hole section was drilled using single-phase fluid in the build section. 

The data gathered during this job revealed compelling similarity in the shape of the RockSense log of a sidetrack and the neutron porosity log for the vertical motherbore, when plotted against TVD, as can be seen in Fig. 3. 

When carrying out this analysis, it could clearly be seen that the RockSense trace from the new hole section followed the original well data. The changes in formation could be seen on the graph, where there is a step change in the porosity and RockSense traces. 

These data could be used to identify that the bit was in the target formation, and immediately adjust the trajectory accordingly. As drilling continued, the directional driller and geologist would have been able to watch the RockSense trace, and any major change would have shown that the bit had exited the formation and allowed them to work together to understand whether the bit was starting to exit above or below the formation and how to correct it to stay in
the formation. 

Cuttings were used on the project to confirm the formation tops as they were drilled, and could be combined with RockSense to confirm any directional decision. However, waiting for the cuttings to be transported to surface, and be analyzed before making course corrections, could have cost significant reservoir contact and, therefore, production. 

BENEFITS OF THIS DISRUPTIVE TECHNOLOGY 

In conclusion, RockSense can provide a valuable at-bit formation evaluation service and enable more accurate wellbore placement than ever before. The technology uses the high data rates available over wireline, when directional coiled tubing drilling which gives extremely high resolution to downhole measurements. The value comes from the ability to take this high-resolution data and identify changes in porosity and formation tops immediately after the bit has passed through the formation boundary. 

 With this instantaneous information about the location of the bit, the drilling team can adjust the trajectory to ensure that the wellbore stays in the target formation, maximizing the production from the well. This is incredibly valuable in directional CTD wells, where a limited formation evaluation package is run, but the value can be seen in nearly any horizontal or re-entry well. Utilizing this potentially disruptive technology can provide a step change in drilling performance and cost per barrel. 

Literature Cited 

ⁱ    MSE. Teale, R.: “The concept of specific energy in rock drilling”, Intl. J. Rock Mech. Mining Sci. (1965) 2, 57–73.