Uses and Limitations of Drillstring Tension and Torque Models for Monitoring Hole Conditions

Abstract

Summary. This paper presents the results of the application of a tension/torque model to directional wells drilled worldwide. The final inclination of the wells ranged from 25 to 70 degrees, and the tension/torque model was effectively used in these cases to aid in planning the directional program before spudding, to monitor the wells during drilling, and to analyze particular drilling problems after completion. The first set of examples represents the situations to which the tension/torque model can be routinely applied and shows that the model can effectively aid the well planner in selecting the best well path, drillstring designs, casing program, and mud system before spudding. A second set of examples shows how real-time monitoring of drilling conditions can he combined with the model to develop a baseline so that deviations from expected behavior can be analyzed and explained. Appropriate action can then be taken to remedy problems before they become serious. A final set of examples shows how a tension/torque model can be effectively used to determine the actual cause of a particular drilling problem after a well is completed. Experience gained from postanalysis of problems can be used to improve later operations. This paper also compares field data with model predictions for a wide range of different inclinations. Introduction Recent technical drilling engineering literature reveals a large interest in mathematical models that can successfully predict drag and torque in directional wellbores. Johancsik et al. and Shepard et al. published papers dealing with the theory of their models. We have used a similar borehole friction model, capable of predicting drags, torques, nominal forces, and buckling behavior, since early 1982 and have accumulated a large body of useful experience in the practical applications of the model. Our experience shows that routine use of a borehole friction model gives the drilling engineer a reliable tool to prevent directional drilling problems. The model has proved useful in all three stages of a directional well: planning, drilling, and postanalysis. Use of the model on an ongoing basis has provided an accurate data base of field information to verify the model and to provide clues to its shortcomings. This paper provides practical examples of the model's usefulness in the drilling of directional wells and can be used as a guide by those drilling organizations that do not yet routinely apply borehole friction models. It also discusses limitations of the model that require further investigation. During the planning phase, the model was used to optimize the trajectory design to minimize the torques, drags, and contact forces between the drillstring and the hole wall. The thrust of this work was to drill directional wells more efficiently by minimizing drilling problems while creating a wellbore that causes fewer downstream problems during the production phase. In the first example, the model was used to allow a well to be drilled that previously had been considered impossible with the available equipment. The model was used to monitor hole conditions during drilling and was particularly useful in diagnosing hole cleaning problems, impending differential sticking, and severe doglegs. It was also used to determine the possibility of reciprocating casing during cementing operations. In postanalysis, the model successfully determined the true causes of hole problems that previously were unexplained or attributed to mud weight, mud chemistry, or "problem shales." Model Description Drag and torque are estimated with a mathematical model that assumes that the drag or torque acting on any small element of the drillstring is proportional to the nominal force acting on the element. The formulation of the tension/torque model closely agrees with that of Sheppard et al. The tension/torque model is a routine that balances the forces on each element that result from drillstring weight and tension, wall contact force, and apparent friction. Fig. 1 represents the force balance of each element. The proportionality constant, is the apparent friction coefficient in the classic Newtonian friction equation FD =, uFn, where FD = frictional drag force and Fn = normal force. When applied to drilling, however, is not a true sliding friction coefficient. Instead, it acts as a correlation coefficient that lumps together the friction caused by various effects, including Coulomb friction, plowing or gouging of the borehole wall, and differential-pressure sticking forces. Also, the drillstring is modeled as a cable with weight but no stiffness. The model operates by calculating the normal force on the lowest element of the drillstring, computing the incremental tension and torque, and then applying this to the next element above. This continues to the top of the drillstring. Three-dimensional (3D) wellbore geometry is defined by the directional survey file with interpolations made to intermediate points by the minimum-curvature method. Multiple drillstring sections are allowed, with user-defined element lengths (usually 30 ft [9.1 m]). Multiple friction factors are possible, allowing the model to be tailored to particular lithologies. Simultaneous axial motion and rotation can be handled. Torque and weight on bit (WOB) can be applied at the lower end to approximate drilling conditions. Buckling of thin-walled tubulars is modeled with the equations developed by Dawson and Paslay. The output from the model is a table of effective tension, torque, normal force, inclination, direction, effective axial friction factor, and rotational friction factor for each element depth. These data are typically presented in summary or graphic form for easy interpretation. In its present form, the model ignores the stiffness of the individual elements and the pressure-related effects, which may be important in cases where the drillstring clement is large or stiff relative to the hole size and curvature. The model also assumes that the lumped friction coefficient is the same for rotation and axial motion, which may not be correct under some circumstances. Model Uses for Planning The ability to forecast drag, torque, and contact forces correctly allows new flexibility in wellbore trajectory and drillstring design for directional wells. Four criteria are used in trajectory/drillstring design.Rotary torque must be within safe working limits for the rotary-drive system and the drillstring, allowing a suitable safety margin.Similarly, up and down drags must lie within safe limits, with the added condition that the down drag must not place any portion of the drillstring into critical buckling.The contact forces between the tool joints and the borehole wall should lie below safe thresholds, typically 2,500 lbf [11.1 kN] for 5-in. [ 127-mm] drillpipe inside casing and 1,500 to 2,000 lbf [6.7 to 8.9 kN] for the same pipe in open hole. Higher forces have been found to accelerate pipe fatigue, casing wear, and mechanical borehole problems, such as key seating and hole enlargement.The entire drillstring must remain within the stable region with regard to buckling for the entire range of anticipated bit weights. To achieve compliance with all these criteria, the planning engineer can vary the wellbore geometry, mud system (and thus the friction coefficient), and the drillstring design.