Investigation of torsional drillstring motion using finite element model
Master thesis
Permanent lenke
https://hdl.handle.net/11250/2788236Utgivelsesdato
2021Metadata
Vis full innførselSamlinger
- Studentoppgaver (TN-IER) [147]
Beskrivelse
Full text not available
Sammendrag
In the thesis , a four degree of freedom dynamic drill string model suitable for prediction of self-excited torsional vibrations is modeled using the finite element method. The drill string equation of motion subjected to excitation due to friction forces between bit-rock interactions is simulated using direct numerical integration schemes and the responses are analyzed for torsional vibrations and stick-slip, With change of some parameters, bifurcation and chaos of the system are observed. The intensity of torsional vibrations is represented by a dimensionless parameter stick-slip severity and by sensitivity analysis of input parameters. Increasing the rotation speed reduces the stick-slip severity value and mitigates the torsional vibrations.
Excitation to the drill string is represented as a combination of deterministic and random components to introduce uncertainty to friction modeling due to bit-rock interactions. The response statistics are analyzed using the Monte Carlo method for two distinct spectral intensities of random excitation, and the results are compared to the deterministic case.
The drill string vibration data obtained by the simulations was utilized to train a machine learning model
to detect the pattern in stick-slip vs non stick-slip responses and the trained model has achieved 100% classification accuracy on test data .For the automated drilling sector,physics-based machine learning model developed in this thesis might be one of the solutions for real-time torsional vibration control in drill strings.