Intelligent real-time drag and torque analysis and sticking trend prediction of drill string

The real-time analysis on the drag and torque of drill string plays an important role in improving drilling efficiency and avoiding sticking risks. At present, the drag and torque analysis is dominated by pre-drilling prediction while the real-time drag and torque analysis in the process of drilling is not developed. Currently, the bottom hole weight on bit (WOB) and torque are not predicted accurately and the friction coefficient of drill string is not determined specifically. To solve these problems, this paper proposed an intelligent real-time analysis method of drag and torque of drill string. In this method, the bottom hole WOB and torque are calculated in real time by using the neural network model and then the friction coefficient is inverted by means of the dichotomy method, combined with the rigid model of drag and torque, so that the force applied on the drill string can be analyzed accurately. When this method is applied to field data, it is shown that the friction coefficient of drill string in one certain well presents the overall trend of increasing gradually in the interval of 6000-6100 m and then increases abruptly from 0.35 to 0.75 near 6100 m, indicating sticking will happen. The drilling log of this well describes that the top-drive drilling tool is stopped and frozen near 6100 m. It is indicated that this method can well predict sticking trend, help field real-time adjustment of drilling parameters, effectively avoid sticking risks and improve drilling efficiency.