Erik Nyrnes

High integrity wellbore surveys: methods for eliminating gross errors

□ Although a second independent survey cannot separately identify all potential errors, it is still the most powerful external data QC check. □ Statistical tests are used to ensure that directional survey measurements have an accuracy that is compatible with the applied error model. Internal QC checks, which are based only on single-source downhole data, are not sufficient to fully validate a survey against its error model (see SPE 103734). – John Weston, SPE – Roger Ekseth, SPE – Kazimir Kovalenko, SPE

Positioning and Position Error of Petroleum Wells

Positioning and Position Error of Petroleum Wells We present a new model for the estimation of positional uncertainty of petroleum wells. The model uses a heavy tailed normal inverse Gaussian distribution for the errors in Earths magnetic field reference values. These references are required for calculating the position of a well using magnetic directional surveying. The results show that the normal inverse Gaussian distribution gives a more realistic fit to the Earths magnetic field reference values than the normal distribution, which is the current state of art. Errors in surveying sensors and reference errors are propagated along the well path to obtain the positioning uncertainty. The positional uncertainty is important for the probability of collisions between wells and for drilling accurate relief wells when an eventual collision has resulted in a blow-out situation. We compare results of the normal model used in the petroleum industry with our proposed model. The comparison is made for anti-collision calculations between drilled wells and a planned well. The results indicate a higher risk of collisions for certain well geometries when using the normal distribution, while a collision is avoided in the normal inverse Gaussian situation. We find clear differences between an approximate normal test and a Monte Carlo test.