RSS Torsional Vibration Severity Prediction in Deviated Drilling: An Application of Artificial Neural Networks

Nowadays, due to the increasing use of Rotary S teerable Sys tems (RSS) in deviated drilling, it has become necessary to know more precisely the dynamic mechanism of their operation, these sys tems are designed to resis t drilling vibrations as much as possible, however and inevitably. They are subject to long-term exposure to these vibrations. Torsional vibrations, in addition to possible failure, reduce the penetration rate of the rotary guided sys tem, so the ability to predict these vibrations can play a significant role in the accuracy of deviation drilling optimization with these sys tems and reduce non-productive time and drilling cos ts, which machine learning algorithms have a high potential in These types of s tudies. The purpose of this s tudy is to demons trate the high accuracy of artificial neural network (ANN) in predicting the intensity of torsional vibration (s tick/slip) of rotary s teerable sys tems, which by choosing appropriate input and output parameters and optimal design, an artificial neural network with an accuracy of 89.9% and an error of less than 0.05 was yielded.