19.4.2 Gyro Error Sources

Due to the number of tool variations and running modes, the gyro error model is more complicated than the MWD model. However, it is based on exactly the same framework for propagation, summation and
transformation as the MWD model.

The gyro model contains 48 error sources. 39 of these are new and 9 are common to the latest revision of the MWD model. As before sensor errors are very prevalent.

Accelerometer biases and scale factors and sensor misalignments are present, and these only effect inclination terms, but different error sources have been identified for tool configurations which have a full orthogonal triaxial set of accelerometers and for tools with only x and y accelerometers, which may be canted relative to each other and not perpendicular.

For the gyros, bias, random noise, gravity dependant errors, scale factor errors and sensor misalignment with the tool have all been identified. With suitable selection of error magnitudes, it is considered that the model is sufficient to model the various types of mechanical spinning wheel gyros in use, both single and dual axes systems, ring-laser gyros and possible future gyro systems.

The gyro sensor errors affect the azimuth measurements, but now different functions apply depending on whether the gyro operates in a stationary, gyro compassing mode at each survey station or is initialised and runs continuously in the hole. In addition, the model accommodates xyz; xy and z-only gyro systems.

For the continuous gyro systems, a new complication arises in that some of the errors will be dependent on elapsed time in the run (gyro drift errors) or the square root of elapsed time (random walk errors). These have been modelled by including a running speed in the tool model and including a term in the weighting functions which will accumulate over time. Terms for errors introduced with an external azimuth reference are also included.

From the outset, the gyro model excludes the need for tool face in any of the terms used. To do this the x and y effects are lumped together. For surveys where the tool is rotated between surveys these terms will propagate randomly, however when the tool slides between stations with fixed tool face, these same terms will propagate systematically.

The gyro error model introduced a new method of calculating the effect of tool misalignment to the borehole.This was later incorporated in to the MWD model and is described in section 19.3.2.