Errors in measurement systems can propagate randomly or systematically. The random errors tend to have a cancelling out effect over multiple observations as their effect can be positive or negative from observation to observation. The systematic errors are the ones that generally expand to dominate the error envelope since their effect is the same from observation to observation.
As an example of a systematic error, the azimuth error in MWD is largely created by the uncertainty of the magnetic field direction (declination). Clearly this does not change from one observation to another and can be taken as a systematic shift in azimuth. A misalignment error on the other hand would be considered random since its effect is toolface dependent and the surveys will have a scatter of tool faces throughout the survey.Some errors, like declination, are referred to as ‘global’ in that they effect every survey in every well in the same field where the declination is uncertain. Others vary from one well to another such as vertical reference and others vary from one survey leg to another such as sensor errors. When we accumulate the uncertainties in a formal error model, the propagation effects of each error source are correctly accumulated so that global errors are always present and random errors are accumulated with RMS values across the boundaries where they become randomized. For a misalignment that is every survey station, for a sensor error, that is every survey leg, for a reference error that is for the whole well and for declination error that would be applied to every well in the field.
The next chapter describes the ISCWSA error models in more detail.
Feedback
Post your comment on this topic.