Acoustic MWD caliper improves accuracy with digital-signal technology

Christopher A. Maranuk
Sperry-Sun
Houston

A newly developed acoustic-caliper tool, utilizing digital-signal processing technology, has increased accuracy and operational range as compared to previous tools. The tool can be run alone or integrated with other MWD-type sensors. The advantages this tool has over other measurement-while-drilling (MWD) acoustic caliper tools are:

• Digitization of the acoustic pulse-echo signals
• The capability to perform digital-signal processing on the resulting waveforms
• The ability to analyze raw waveforms for detection anomalies.

With these capabilities, many different detection algorithms can be developed and tested, from simple shaped exponentials to algorithms that perform pattern recognition and zero-crossing detection techniques.

Digital-signal processing techniques help correct for the presence of drill cuttings and gas, and can help detect small acoustic signals. A field verifier has been developed to obtain highly repeatable waveforms, providing the field engineer or technician with a simple and reliable surface check.

Measuring hole size while drilling is difficult because the dynamic environment precludes the use of truly mechanical devices. Several MWD caliper tools, utilizing acoustic technology, are available to the industry.

Unfortunately, these tools typically have a limited operating range and are often integrated with other MWD sensors, such as density or sonic tools.^{1 2 4}

Technical issues that can restrict the use of acoustic MWD caliper tools include:

• Noise associated with the downhole drilling environment
• Tool accuracy
• Tool reliability
• Errors associated with mud attenuation and signal detection
• Tool off-centering
• Formation cuttings
• Gas presence.

This acoustic MWD tool has been designed with its own microprocessor, memory, power supply, and surface communications port, so it can be run as a stand alone sensor.

Uses for MWD calipers

The need for an MWD caliper tool has often been cited. ^{3 8} It provides a key measurement that has been missing from the suite of available MWD sensors.

Hole-size information can often be used for environmental correction of MWD measurements, including gamma ray, resistivity, neutron, and other geophysical readings. Hole size information can also be used to determine the presence of, and account for the effects of, washouts on these sensors' responses. Acquiring this information is an important aspect of log analysis, which has often been missing in the MWD environment.

MWD caliper tools are also used to calculate hole volumes. After the final bit run, the acoustic caliper has the capability of collecting data while tripping out of the hole. This is useful for calculating cement volumes.

Other applications for MWD caliper tools might include real-time assessment of casing wear and well bore stability, evaluation of hole cleaning, and determination of tight spots or formation ledges.

With the addition of dynamic directional sensors, borehole geometry (ellipticity) can also be computed. Borehole ellipticity can be used to estimate the maximum horizontal stress field for reservoir calculations.

Market studies identified the need for an accurate, highly reliable standalone MWD caliper tool. The ability to make measurements over a relatively large operating range, with a high degree of confidence, was considered a key requirement for fielding a commercially successful tool.

This drove the need to develop a tool with an internal processor and transceivers that could provide reliable operation under conditions of elevated pressures and temperatures.

Development of this acoustic caliper tool was initiated in April 1996, and field testing began in June 1997.

Operational theory

The tool physics behind the measurement is simple and is based on the pulse-echo technique. The standoff from the borehole wall can be calculated by the following equation:

d = (v_mt)/2 (1)

where:
d = Standoff
v_m = Acoustic velocity of mud
t = Time difference between the pulse emission and echo arrival, or equivalently, the time difference between transmitter firing and echo detection of the borehole wall signal.

In Equation 1, the standoff is assumed to include contributions from delay lines, transducer packaging material, and so on. All of these factors need to be taken into account when the actual standoff-the distance from the tranducer's tool surface to the borehole wall-is calculated.

In a perfect world, where a tool of diameter D_t is centered in a borehole of diameter D_h, the hole diameter can be estimated by the following equation:

D_h = D_t + 2d (2)

Unfortunately, MWD tools are seldom centered in the borehole, and the values of d vary with time. In addition, the borehole is seldom a true circle. Thus, the value of d and D_h will vary with the orientation of the transceiver and the actual size of the borehole.

It is for these reasons that multiple transceivers, geometric calculations, and averaging techniques must be utilized. The resultant value is an average hole diameter, representing the diameter of an equivalent circle with about the same area as the area of the true borehole calculated downhole.

An indication of the relative tool-vs.-borehole sizes that may be encountered is illustrated in Fig. 1 [21,328 bytes]. The potential error in the standoff, as well as the borehole diameter, can be calculated by standard mathematical and statistical means.

When the arrival time can be measured accurately, the dominant error factor in the borehole diameter calculation becomes the fractional error in the mud's acoustic velocity. It is well known that the mud's acoustic velocity changes with mud type, density, salinity, pressure, temperature, and the amount of gas dissolved in the mud.^{4 5}

Recognizing the potentially dominant effect of changes in the mud's acoustic velocity, tests were performed to characterize and measure a mud's acoustic velocity compared to the significant factors that influence it.

The result is a data set, as well as a set of semiempirical equations that allow for on-the-fly corrections. The corrections can either be applied downhole or the downhole data can be reprocessed at the surface.

The error in round-trip time can become significant if the true, zero-crossing equivalent of the echo signal is missed or estimated at the wrong time. Published sampling techniques for downhole acoustic calipers rely on analog-signal manipulation.

The published range of frequencies for downhole ultrasonic transducers used in the MWD caliper tools range from 250 khz to 670 khz. When entirely pure analog techniques are utilized for these timing measurements, the probability of detection error for the first arrival increases dramatically.

The error in the round-trip time can also become another significant error factor in the calculation. For example, it has been noted that detection variations for the first arrival are about 3 msec, based on analog detection techniques. This causes a standoff error of about 0.09 in. in water.⁴ For the acoustic MWD caliper, the detection variation of the first arrival using signal digitization, a digital signal processor (DSP), and true-zero-crossing techniques, is as little as 0.1 msec.

This results in a standoff error of only 0.003 in. in water, or a factor of 30 times less than equivalent analog detection techniques.

The capability to capture digitized signals offers a distinct advantage to this tool. Digital-signal processing techniques are used to enhance the capability of echo detection. Even in the presence of significant signal attenuation-large standoffs and heavy muds-echos can be detected even in heavy muds (16-18 ppg), resulting in a 2-3 in. standoff capability.

In addition, digital signal processing can eliminate proximity effects with no acoustic delay-line required and can be used to calculate an accurate borehole diameter in the presence of drill cuttings and small gas concentrations.

There are two errors associated with any acoustic caliper tool. The first error is the repeatability error, or the capability of the tool to consistently measure the same value. In lab testing, the average repeatability error was on the order of 0.02 in. or less, over most of the operating range of the tool.

In many cases, the repeatability error of this tool was less than 0.005 in. when the tool was centralized. In comparison, published repeatability errors from other acoustic caliper tools range from 0.05 in. to 0.2 in.^{4 5}

The second error concerns the accuracy of the measurement, or the capability of the tool to measure the true diameter of a circle. For lab tests of circular boreholes with known diameters, the average accuracy error was less than 0.1 in., extending over most of the tool's operating range.

Tool design

This MWD acoustic caliper tool uses three ultrasonic transceivers, spaced 120° apart, which transmit and receive acoustic signals for borehole size and ellipticity calculations. Dynamic directional data obtained from the magnetometer and accelerometer sensors help determine ellipticity.

The tool has its own processor, data storage memory, and power supply. The first generation caliper tool is 6 ft long and has an OD of 6 3/4 in. across the majority of the sub body. The transceiver section area has an OD of 7.19 in.

A 7.375 in. carbide-coated wear band is used directly above the transceiver area. It protects the transceivers from abrasion as well as provides a minimal standoff from the wall of the borehole. The transceivers are located about 1.25 ft from the pin end of the caliper sensor.

This tool contains an electronics insert, surface communications port, and a battery pack (Fig. 2 [39,640 bytes]). The tool is designed with a 1.92-in. through bore to allow for the passage of mud, cuttings, and lost circulation material. The battery pack is field-replaceable and supplies dc power to the tool, allowing it to operate in a standalone mode.

The caliper insert contains three electronics boards and a surface communications port. The surface communications port is isolated from the rest of the sub in case of a catastrophic failure caused by a seal leak.

The electronics boards are also isolated from the rest of the insert in case a similar catastrophe occurs near the lower end of the tool.

The transceivers use a pulse-echo technique to emit and receive ultrasonic signals. Signals are produced by piezo-ceramic discs. The transceiver packaging technique allows for pressure compensation in wells with downhole pressures up to 18,000 psi.

The transceiver has been designed such that it is field-replaceable in case of failure due to abrasion or other problems. The transceiver is one of the key elements in the design of an ultrasonic caliper tool.

The performance of the tool depends greatly on the repeatability of the performance of the transceiver. Many transceiver designs suffer from extensive ringing in cases where the firing signal extends over too long a period of time.

This reduces the dynamic range of the caliper tool because the borehole wall echo signals arrive too close in time to the firing pulse, masked by the transducer ringing.

For the tranceivers, the ringing lasts for less than the round-trip echo time when the 7.375 in. wearband is in contact with the borehole. This allows for accurate standoff measurements, even with extremely fast return signals.

The tool has the ability to store raw waveforms. Waveforms from different firings for the same transceiver can be overlaid during post-processing. An example of such an overlay of eight waveforms is shown in Fig. 3 [37,839 bytes].

The eight waveforms represent acquisitions of a nonrotating tool over a period of 45 hr and spanning a temperature range from 20° to 145° C. The repeatability of the transceiver's electronics over time with varying temperatures is a key advantage of this tool's design.

The tool also contains a downhole pressure transducer for determining pump-on/pump-off status. In applications where high sampling rates are desirable, such as tripping out of the hole on a final bit run to gather data for cement volume calculations, the tool can be triggered from the surface into a high sampling mode.

The caliper insert contains three boards: a power-supply board, a transceiver/dynamic directional board, and a digital processor board. All boards have been packaged to facilitate future tool downsizing.

The transceiver/dynamic directional board contains a magnetometer and accelerometer which is used to calculate borehole ellipticity.

The dynamic directional sensors mounted in the tool will not identify the actual direction of the borehole, but do provide the relative direction of the major (or minor) axis of an elliptical borehole.

When combined with the standard directional packages present in most MWD tool strings, the actual direction of the axes of the ellipse can be computed.

The processor board houses the downhole processor, data storage memory, and communications circuits. If required, there is enough nonvolatile memory storage capacity to save all detected arrival times of a bit run for eventual reprocessing at the surface.

The downhole software can process real-time hole size calculations; in addition to pressure, temperature, and mud acoustic velocity corrections. It can also analyze magnetometer or accelerometer data in order to assist in the determination of the relative direction of the axes in elliptical borehole.

Data storage and programming

Raw waveform data are also stored periodically and are available for display and evaluation after the bit run. The processor board acquires and stores several key diagnostic parameters in nonvolatile memory.

These data are then processed at the surface to provide a tool-performance and log-quality report. The tool acquisition system can be fully programmed at the surface before the logging run with key parameters before the logging run.

This allows for greater operating flexibility than most other MWD caliper tools in existence today.

A unique feature of this tool is the capability for pre-log and post-log verification at the well site. A rugged and portable clamp-on verifier can be used to confirm the performance of the transceivers and tool electronics (Fig. 4 [9,275 bytes]). The verifier is easily installed and provides for repeatable measurements.

Fig. 5 [42,038 bytes] shows the theoretical maximum, as well as the expected operating range for the 6 3/4-in. caliper tool. The tool's performance depends on its general centralization.

In most situations, the tool is capable of measuring a 2 to 3-in. standoff, which equates to a maximum borehole diameter of 8 1/2 to 12 1/4 in. In situations where the tool is relatively centralized and used with light mud weights, the maximum measurable borehole diameter can extend up to 15 or 17 in.

Effects of gas and cuttings

One of the key concerns for utilizing ultrasonic caliper tools is its behavior in the presence of drill cuttings and gas. This tool offers the advantage of digitized signals and can perform digital signal processing of the waveform data.

Cuttings may evidence themselves in one of two ways:

First, the density of the cuttings may be high enough so that the signal from the borehole wall will be scattered enough to be very small. Digital filtering and signal enhancement techniques can be used to detect the borehole echo signal, even in the presence of a significant amount of cuttings.

Second, the cuttings may affect performance. If a large cutting masks the return of the signal from the borehole wall, the first arrival will be that of the cutting and not of the wall.

Eliminating such interference is straight forward, utilizing digital signal processing. The presence of gas in the mud is another aspect of concern when dealing with ultrasonic caliper tools.

Gas affects the performance of the tool in many ways because:

1. Gas can increase the attenuation of the signal in the drilling mud.
2. Relatively large gas bubbles may give echoes similar to those of cuttings.
3. A large concentration of gas can significantly lower the mud's acoustic velocity.

For the case of increased attenuation, signal processing techniques are utilized to detect a small signal in the presence of noise. Such techniques are not available when analog signal manipulation is utilized.

For the case of false echoes created by gas bubbles, or where both a gas bubble echo and the true borehole wall echo are present in the received signal, the same techniques employed to eliminate false echoes caused by large cuttings can be utilized.

Although gas influx affects the caliper's performance, in most cases, the magnitude of the downhole pressure will alleviate these effects, especially for smaller concentrations of gas.

Published literature suggests that for the velocity of mud to change dramatically because of gas, both lower pressures and significant gas influx need to be present. Experimental results show that the acoustic velocity of water is 1,440-1,480 m/sec and about 340 m/sec in air, but in an air-water mixture, it falls to as low as 20 m/sec.⁶

Even very small concentrations of gas can dramatically reduce the acoustic velocity. For instance, 1% by volume of air in water can reduce the acoustic velocity by 95%.

Later studies have suggested that at higher pressures, much larger mass fractions are required to cause the velocity to decrease dramatically.⁷ Since this tool will be used in a downhole environment, it is expected that the downhole pressure will eliminate most of these effects.

Field data

The acoustic caliper tool was field tested in the Gulf of Mexico in about 1,300 ft of water. The tool was run in a record only, stand-alone mode. Key goals of the field test included confirming basic tool operation, collecting data to facilitate future tool development, and verifying the basic detection algorithm.

The caliper tool was run twice in an 8 1/2 in. hole section, positioned about 15 ft above an undergauge, stabilizer-mounted MWD density tool approximately 117 ft from the drill bit. No other stabilizers were included in the BHA. The mud weights for both runs ranged from 10.9 to 11.9 ppg.

The hole inclination was about 72° at the beginning of the run, dropping to about 58° at the end. The operator planned to drill the well with one bit run.

The run was anticipated to last for 84 hr with penetration rates between 150 and 200 ft/hr. However, this run was cut short because the rig was unable to obtain a good casing-seat test. Unfortunately, no formation was drilled, and the tool only logged the 9 5/8-in. casing.

Forty feet of the cement plug was drilled during the first run, and a casing leak-off test was performed. The leak-off test failed and the operator decided to pull the string out of the hole.

During the run, the tool measured the casing diameter as 8.75 in. on average with a standard deviation of 0.094 in. During the short period when the cement was being drilled, the tool measured the casing as 8.777 in. with a standard deviation of 0.023 in.

The nominal casing ID for 95/8-in., 43.5 lb/ft casing is 8.755 in. From the analysis of the stored waveforms (Fig. 6 [48,834 bytes]), it is clear that the tool was not centered in the casing, and that the echo signals were significantly attenuated by the reflection angle.

During the entire first run, the detection of the echo was very good with no detection errors.

The second run was made with a second caliper tool after the completion of a good casing leak-off test. The operators drilled out the casing shoe and continued drilling to total depth.

Gas was logged in the mud returns numerous times during this run, ranging from 50 to 550 units of gas. High-viscosity sweeps and significant circulating were done prior to making drillstring connections.

The second run drilled the remainder of the casing shoe assembly over an interval of 107 ft as well as 2,264 ft of open hole. When the results of the two runs were compared, the waveforms for both tools were almost identical.

The caliper data in the open hole were very noisy; in other words, the values showed significant variations from nominal. Upon examination of the open hole waveforms (Fig. 7 [19,417 bytes]), it was noticed that the echo signal levels were very small, smaller that the detection threshold defined in the detection algorithm.

For initiated field testing, these tools were programmed with a simple stair-step, threshold detection scheme.

This scheme is very similar to the detection methods used in other MWD acoustic caliper tools. From the analysis of the recorded open hole waveforms, it became obvious that simple threshold detection techniques could not work in cases like where there are soft formations, gas cut mud, and minimal stabilization.

Fig. 8 [123,028 bytes] shows the time-based data from run No. 1. There are three curves plotted on the graph: hole diameter, pressure, and temperature. The three general modes of drilling-tripping-in, drilling, and tripping-out-were also marked on the graph.

Some key elements to note in this figure are the following:

1. The consistency of the hole diameter over time.
2. The stair-step behavior of the pressure during trip-in, consistent with the use of a float sub and the drill pipe filling operation. The bleed-off between stations was the result of a problem with the pressure transducer electrical circuit which has been subsequently fixed.
3. The pressure spike at 13:00 hr. Although this may seem artificial, it is actually the maximum pressure obtained during the failed leak-off test.
4. The 10° C. temperature drop when circulation was started at 10:00 hr.

The acoustic caliper tool has since been used in over a dozen bit runs in the Gulf of Mexico and the North Sea. Several new detection algorithms have been tested resulting in a robust detection scheme that has allowed the tool to provide caliper logs in all cases.

Acoustic caliper data have been obtained in water and synthetic-based muds with mud weights up to 15 ppg.

Acknowledgments

The author would like to acknowledge the support of the acoustic caliper development team for their valuable contributions to this article. The author would also like to thank the principals and employees of SensorWise Inc. for their valuable contributions with the concept development and the design of this tool, as well as their input to this article.

References

1. Best, D., Evans, M., Holenka, J., Kurkoski, P., and Sloan, W., "Improved Formation Evaluation Using Azimuthal Porosity Data While Drilling," SPE paper 30546, Annual Technical Conference and Exhibition, Dallas, 1995.
2. Birchak, R., Linyaev, E., Mackie, B., Malloy, R., Minear, J., and Young, D., "Compressional Slowness Measurements While Drilling," Paper VV. Transactions of the 36th SPWLA Annual Logging Symposium, Paris, 1995.
3. Norve, K.H., and Saether, H., "Field Experience Using the Full Suite MWD-Combination for Reservoir Logging and Evaluation," Paper A. Transactions of the 30th SPWLA Annual Logging Symposium, Denver, 1989.
4. Moake, G.L., Birchak, J.R., Matthews, R.G., and Schultz, W.E., "Standoff and Caliper Measurements While Drilling Using a New Formation Evaluation Tool With Three Ultrasonic Transducers," SPE paper 26494, Annual Technical Conference and Exhibition, Houston, 1993.
5. Orban, J.J., Dennsion, M.S., Jorion, B.M., and Mayes, J.C., "New Ultrasonic Caliper for MWD Operations," SPE/IADC paper 21947, SPE/IADC Drilling Conference, Amsterdam, 1991.
6. McWilliam, D., and Duggins, R.K., "Speed of Sound in Bubbly Liquids," Proceedings Institute Mechanical Engineers, Vol. 184(3C), pp. 102-7, 1969.
7. Kieffer, S.W., "Sound Speed in Liquid-Gas Mixtures: Water-Air and Water-Steam,", Journal Geophysical Res., Vol. 82(20), pp. 2895-904, 1977.
8. Allen, D.F., Best, D.D., Evans, M., and Holenka, J., "The Effect of Wellbore Conditions on Wireline and MWD (Measurement While Drilling) Neutron Density Logs," SPE paper 20563, 65th Annual Technical Conference and Exhibition, New Orleans, 1990.

The Author