On the Importance of Boundary Conditions for Real-Time Transient Drill-String Mechanical Estimations

Abstract

Abstract The extraction of mechanical friction, during drilling operations, is an important source of information to estimate whether the downhole conditions are sub-optimal, for instance because of poor cuttings transport or additional tortuosity induced by the directional work. Traditionally, the friction factors are extracted using steady state torque and drag models. But, there are contexts, like when drilling from a floater, where it is difficult to find any periods where stable conditions can be observed. A possible solution to this problem consists in running, in real-time, a transient torque and drag model connected to a high refresh-rate feed of measurements taken by the drilling machines sensors, and to utilize continuous calibration methods to extract information about the unknown or ill-defined parameters that influence the drill-string mechanical system. For those parameters that should stay constant throughout a bottom hole assembly (BHA) run, the estimation method shall account for the totality of the observations and therefore utilizes calibration methods based on statistical global optimization principles, while time-dependent parameters are estimated using a filtering technique. This approach is only valid when utilizing a trustworthy transient drill-string mechanical model. Comparisons between estimated values from transient torque and drag models and actual measurements have highlighted the importance of robust modelling at the boundaries of the mechanical system. For instance, when modelling precisely the top-drive and the hoisting system mechanical behavior with its associated heave compensation system, it is then possible to obtain a good match between estimated and observed top-drive torques and top-of-string forces, while the bit is off bottom. Furthermore, it requires a precise modelling of the bit/formation interaction to get a truthful dynamic response of the model when the bit is on bottom and only passive heave compensation is used. Associated with its continuously updated calibration of the ill-defined parameters of the system, such a transient torque and drag model enables the visualization, in real-time, of the internal displacements of the drill-string even in the most complex drilling conditions. Until recently, drill-string dynamic modelling has mostly focused on the lower part of the drill-string, i.e. just above the BHA, and has been reserved to post-analysis evaluations, as existing solutions have been very computer intensive. This paper describes a solution that runs in real-time and that considers the whole drill-string, allowing for the extension of drilling automation functions to very dynamic conditions like those encountered on a floater.