Field Verification of Model-Derived Natural Frequencies of a Drill String

Abstract

Vibrations generated in a drill string while drilling generally lead to a reduction in drilling efficiency and often cause premature failure of drill string components and bit damage. It is also known that lateral vibrations, in particular, are responsible for most measurement-while-drilling (MWD) tool failures while drilling. One way to increase drilling efficiency and avoid tool damage is to monitor and analyze drilling vibrations so that drilling parameters can be adjusted while drilling to reduce such vibrations. An alternative method is to analyze and determine the natural frequencies of the bottom-hole assembly (BHA) so that resonant conditions caused by various excitation mechanisms in the drill string can be avoided. Even though models have been developed in the past in the drilling industry to determine the natural frequencies of a BHA, few attempts have been made to demonstrate that such models do actually help reduce vibrations or failures. This paper deals with the process of field validation of model-derived frequencies for axial, torsional and lateral vibrations. The results presented in this paper are based on the analysis of drilling data from a field test using downhole vibration measurement sensors. The downhole measurements included X and Y bending moments, axial acceleration, dynamic weight-on-bit, dynamic torque, and X and Y-axis magnetometers mounted in an MWD sub. The data analysis demonstrates that the natural frequencies predicted by the models match well with actual field (measured) values at the locations of interest, particularly for lateral vibrations. This analysis therefore shows that model derived results can be used with a degree of confidence to help avoid resonant conditions in a BHA while drilling and to help reduce failures.