Analysis Of The Nonlinear Multi-contact Frictional Drag Of Drill String Based On Statistical Learning

With the development of oil field exploitation and drilling technology, the problem of drill string friction aroused extensive concern. It is extremely urgent to consider the complexity during drilling, in view of how to establish analysis models of drill string mechanics in directional and horizontal well, to solve friction error and utilization ratio on well site.Drilling is a complex systems engineering, it produces mass certain, uncertain and random information. Type and quantity of information limit result of friction, so the information which affecting friction is separated into certain data and uncertain data considering classifying, discovering and collecting.To process certain information, analytical methods of drill string mechanics are investigated and researched and finite element method is selected to build the drill string models, considering the change of deviation and azimuth angle. Drill strings are resolved into discrete dimensional unit limited by well bore, and then the units are compounded into a whole representing the original state, so a lot of algebraic equations are obtained in which nodal displacement is unknown variable. These equations adapt to friction analysis of different well bore trajectories.The article presents analytical models of drill string friction based on statistical learning to deal with uncertain information, in view of the limitation of traditional simplification in drill string modeling and enhancing utilization ratio of drilling information. The basic idea is: we use them to sum up hook load with transverse and longitudinal analysis. In order to find feasible region and rules of hook load distribution, one to one or one to many methods are used, so friction coefficient can be got combining finite element method. The rule of hook load and friction coefficient is applied to enhance predictive accuracy of frictional drag in new area or different BHAs and heighten utilization ratio of frictional drag. Drag theory can be combined with practice, and drill string mechanics analysis system is supplemented.