Accelerometer

An accelerometer is a sensor that measures linear acceleration along one or more defined axes. In petroleum drilling, accelerometers are the primary sensors inside measurement-while-drilling (MWD) survey tools, where three orthogonally mounted accelerometers measure the components of Earth's gravitational acceleration along the tool's X, Y, and Z axes. From these three gravity components, the MWD processor calculates the tool's inclination from vertical, which defines how steeply the wellbore is deviating from a plumb line at that survey station. Combined with triaxial magnetometer readings, the accelerometer data yields the full survey: inclination, azimuth, and toolface angle. Accurate accelerometer performance is critical because survey errors propagate along the wellbore and can place the bit tens of metres away from the intended target at depth, causing missed pay zones or wellbore collisions with adjacent wells on a multi-well pad.

Key Takeaways

The physics of inclination measurement from accelerometers: Earth's gravitational acceleration (g = 9.81 m/s²) acts straight down everywhere. A perfectly vertical tool has g entirely on its Z axis (along the tool body) and zero on the transverse X and Y axes. As the tool tilts away from vertical, the gravity vector splits between the Z axis and the transverse plane. The inclination angle is the arctangent of the transverse gravity magnitude divided by the axial gravity: inclination = arctan(sqrt(Gx² + Gy²) / Gz). At 90° horizontal, Gz is essentially zero and all of g lies in the transverse plane. This calculation uses only gravity, so inclination can be determined inside steel casing (which blocks Earth's magnetic field) and works reliably at high magnetic latitudes near the poles where compass-based azimuth readings are unreliable.
Two accelerometer technologies dominate oilfield tools. Quartz-flexure accelerometers use a proof mass suspended on a precision quartz beam; a servo feedback circuit applies a restoring force to keep the mass at null, and the restoring current is proportional to acceleration. These sensors achieve bias stability better than 0.01 m/s² and are standard in high-performance MWD tools and gyroscopic survey instruments. Micro-electromechanical systems (MEMS) accelerometers etch tiny capacitive structures onto a silicon chip; proof-mass deflection changes a capacitance gap that is read electronically. MEMS sensors are smaller, cheaper, and far more shock-resistant than quartz-flexure types, making them common in rotary steerable systems and low-cost continuous inclination tools, though their accuracy is somewhat lower at extreme temperatures.
Accelerometer error sources affect survey accuracy and accumulate through the wellbore position calculation. The main errors are: bias (a constant offset even at zero input, causing a fixed inclination error regardless of angle); scale factor error (the sensor reads a fixed percentage more or less than the true acceleration); misalignment (the three sensor axes are not perfectly orthogonal, causing cross-coupling between channels); and temperature sensitivity (output drifts with temperature as the tool heats from ambient to bottomhole temperature, which can exceed 150°C in deep WCSB wells). Premium MWD tools store factory calibration coefficients and apply real-time temperature corrections. A well-calibrated MWD accelerometer set achieves inclination accuracy of approximately ±0.1° (1-sigma), corresponding to a positional uncertainty of roughly 1 metre per 600 metres of wellbore length.
A second critical function of downhole accelerometers is vibration monitoring. Lateral (transverse) vibration, stick-slip (torsional oscillation), and bit bounce (axial vibration) each produce characteristic acceleration signatures on specific axes. Transverse accelerometers detect lateral whirl of the drill collar; the axial accelerometer detects bit bounce; all three axes register the high-frequency shock bursts associated with stick-slip events. Modern MWD and rotary steerable tools compute a vibration severity index every few seconds and transmit it to surface via mud pulse telemetry. Drillers use this real-time feedback to adjust weight on bit and rotation speed before vibration damages the drill bit, MWD sensors, or drill collars. Tools rate their shock tolerance in g-peak; most premium MWD collars are rated to 250 to 500 g, and accelerometers log the cumulative count of impacts exceeding those limits.
In seismic acquisition, accelerometers are used in downhole geophone strings for vertical seismic profiles (VSP) and in modern broadband ocean-bottom nodes. Traditional surface geophones are velocity sensors (a coil moves through a magnetic field), but modern MEMS-based seismic sensors output acceleration and have superior low-frequency response below 5 Hz, improving the broadband signal needed for full-waveform inversion. Three-component (3C) downhole VSP tools carry one vertical and two horizontal accelerometers, recording both P-wave and S-wave energy independently. In Alberta Montney and Duvernay VSP programmes, 3C accelerometer tools map velocity structure between the wellbore and the surface seismic grid, calibrating depth-to-target predictions before committing to horizontal leg placement. Ambiguities in surface seismic velocity that cause ±20-metre depth uncertainty can often be reduced to ±5 metres with a well-calibrated VSP.

How MWD Accelerometers Build a Wellbore Survey

In a horizontal Montney well, the MWD tool takes a survey roughly every 30 metres of new pipe added. At each survey station, the three accelerometers output their gravity component readings. The tool's Z axis runs along the drill collar; at 90° horizontal inclination, Gz is near zero while the transverse components carry the full gravity signal. At 45° inclination, Gz and the transverse magnitude are equal, and the arctangent calculation returns exactly 45°. The magnetometer readings combine with these gravity data in real time to compute the full survey: inclination, azimuth, and toolface.

The 3D wellbore position is then computed by the minimum curvature method, which treats each survey interval as an arc of a circle and integrates inclination and azimuth changes from surface to total depth to give the north, east, and true vertical depth (TVD) coordinates at every survey station. Errors from individual surveys accumulate along the well. The ISCWSA (Industry Steering Committee on Wellbore Survey Accuracy) error model propagates individual sensor errors through these equations to produce an uncertainty ellipsoid at any point along the wellbore. For a typical 4,000-metre Montney horizontal well, MWD accelerometer-based lateral position uncertainty at total depth is roughly ±10 to 20 metres at 1-sigma confidence, which drives the anti-collision rules on multi-well pads: no two wellbores can be closer than their combined positional uncertainty ellipsoids at any point along their shared approach.

Gyroscopic survey tools, which use spinning gyroscopes or fibre-optic ring-laser gyroscopes instead of accelerometers for azimuth measurement, achieve much smaller positional uncertainty (±3 to 5 metres laterally) but require pulling the drill string to run on wireline. The standard approach in precision horizontal drilling is: MWD accelerometers for continuous real-time guidance, with a gyro confirmation survey when the well approaches a critical target boundary or anti-collision constraint.

Fast Facts

The first directional wellbore surveys were made with mechanical single-shot tools: a weighted compass pendulum photographed by an on-board camera on a timer at target depth. The photograph was retrieved with the tool and the inclination and azimuth were read manually. Electronic MWD with downhole accelerometers became commercial in the late 1970s, with companies including Teleco Oilfield Services and Eastman Whipstock transmitting real-time inclination data by mud pulse telemetry. Early MWD accelerometers were bulky pendulum devices; quartz-flexure sensors replaced them in the 1980s as electronics became reliable enough for 150°C downhole environments. By the early 1990s, three-axis accelerometer MWD surveys were standard on all directional wells in the WCSB. A modern MEMS accelerometer occupies a chip smaller than a fingernail, weighs under 1 gram, tolerates 500 g shock, and measures gravity to 0.01% accuracy, a remarkable combination that makes it well suited to the brutal environment of a rotating drill collar at 3,000 metres depth.

Vibration Monitoring With Downhole Accelerometers

Drill string vibration is a leading cause of premature MWD failure and bit damage. Three vibration modes affect drill strings: lateral whirl (the drill collar bends and orbits inside the borehole, generating centrifugal force that batters the BHA against the borehole wall), stick-slip (the bit stalls on hard formation, torque builds up in the twisted string above, then releases in a violent spin-up to 3 to 5 times normal RPM), and bit bounce (axial compression waves cause the bit to alternately pick up and slam into formation). Each mode is detectable by the appropriate accelerometer axis: transverse axes detect lateral whirl; the Z axis detects bit bounce; all three axes register high-frequency shock bursts from stick-slip events.

Modern rotary steerable systems use real-time accelerometer feedback in closed-loop vibration control: the tool adjusts weight on bit or bit rotation speed in small increments every few seconds, without waiting for the driller to act on telemetry data. In BC Montney horizontal wells where long laterals in hard carbonate-rich intervals generate persistent lateral vibration, automated vibration management has reduced bit damage rates and improved toolface control compared to driller-only WOB adjustments. Service companies report 15 to 30 percent longer bit runs in high-vibration intervals when automated vibration control is active versus passive monitoring only.

The accelerometer is also called a gravity sensor, g-sensor, or triaxial gravity measurement package in MWD documentation. Related terms include measurement while drilling (MWD, the system that transmits downhole sensor data, including accelerometer-based surveys, to surface in real time via mud pulse or electromagnetic telemetry while drilling continues; the primary application for downhole accelerometers in oilfield operations), inclination (the angle between the wellbore axis and a vertical reference line, in degrees from 0° vertical to 90° horizontal; the primary output of the accelerometer gravity calculation in a directional drilling survey), magnetometer (a sensor measuring Earth's magnetic field components; combined with accelerometer gravity data to determine wellbore azimuth and toolface angle in MWD surveys), wellbore survey (the measurement of inclination, azimuth, and depth at regular stations along the wellbore, used to calculate the 3D position of the borehole and detect deviations from the planned trajectory), and drill string vibration (unwanted lateral, torsional, or axial oscillations in the drill string detected by downhole accelerometers; managed in real time to prevent MWD tool failure and premature bit wear).

How an Accelerometer Bias Error Put a BC Montney Well 22 Metres Above Target

An operator was drilling a Montney B landing target in northeast British Columbia with a planned lateral 180 metres north of an existing producer. The planned TVD to enter the Montney B was 2,780 metres, with a stratigraphic tolerance window of ±5 metres TVD. The MWD tool carried standard triaxial quartz-flexure accelerometers and magnetometers.

During the build section from 50° to 88° inclination, the formation evaluation engineer noticed the real-time gamma ray log was picking up Montney B markers approximately 18 metres earlier in measured depth than the correlation predicted. The MWD surveys showed the well tracking 2 to 3 metres above plan, which was initially attributed to natural formation variability.

When the well reached 88° inclination at 2,850 metres measured depth, a wireline gyroscopic survey was run through the open drill string. The gyro-based inclination readings averaged 1.8° higher than the MWD accelerometer readings over the entire build section. Propagating this 1.8° bias through the 800-metre build section using minimum curvature, the actual TVD of the wellbore was 22 metres shallower than the MWD survey indicated. The bit had already passed through the top of the Montney B and was running 22 metres above the intended landing plane.

The bias was traced to the MWD Z-axis accelerometer. During trip-in, the drill string had been inadvertently dropped about 30 centimetres into the slips while making a connection at 1,200 metres. The impact shock partially shifted the sensor's proof mass reference position. The tool passed the standard ambient-temperature surface calibration check before the trip, but the bias developed only at bottomhole temperature (148°C), a condition not simulated in the surface test.