Evaluating Drilling Practice in Deviated Wells With Torque and Weight Data

Abstract

Summary. This paper shows how to estimate two friction coefficients on a foot-by-foot basis at the wellsite with both measurement-while-drilling (MWD) and surface values of weight on bit (WOB) and torque. A log of the coefficients with depth can be used to diagnose drilling problems in directional wells. Field examples are given that show how the technique detects incipient sticking and hanging stabilizers. Data are also used to evaluate the time spent on wiper trips and in circulating before making connections. Introduction Torque and drag have always been major concerns in the planning and drilling of a well. Deep, highly deviated wells are becoming commonplace, and the ability to minimize torque losses and drag is often critical for a successful well completion. An accurate estimate of the friction factor(s) during drilling is possible with both downhole and surface torque and WOB sensors. Monitoring the friction factor on a foot-by-foot basis allows such drilling problems as severe doglegs and sticking zones to be diagnosed as they occur. This gives the driller sufficient time to take remedial action--e.g., reaming, adjusting the mud weight, or cleaning the hole before becoming stuck. The technique also allows us to evaluate the benefits of bottoms-up circulations or wiper trips, and to estimate the side forces on the pipe (for both pipe fatigue and wear and casing wear) and the position of the neutral point (to prevent pipe buckling). Technique The equations that we use were originally proposed by Johancsik et al. and have since been used by Sheppard et al. to compare catenary and undersection trajectories. Two kinds of friction are defined: rotating friction in conventional drilling and sliding friction in turbine/downhole-motor drilling or tripping. Threshold forces that might be needed to start the motion are not investigated here. The stiffness of the tubular is neglected in the computation. This is justified as long as the drill collars are subject to only a small degree of bending. In a curved hole, the side forces on the collars are greatly underestimated. These forces can be computed in real time with a simplified version of a model used for the prediction of bottomhole assembly (BHA) tendency. The accuracy of such a computation, however, depends critically on the value of the dogleg severity (hence, the MWD should be located close to the bit) and such unknowns as stabilizers/hole clearances and rock anisotropy. A future measurement of bending moments in the BHA will soon help determine some of these parameters. Values of the friction factor between 0.25 and 0.4 are generally accepted, and most of our field measurements fall within this range. We did not notice large changes in the friction factor with commonly used rotary speeds. But recently published papers stress the importance of rubber protectors and solids content and mud lubricant on the pipe/casing friction factor, and present a fan of contradictory laboratory values from 0.05 to 0.5. Monitoring Friction at the Wellsite Over the past few years, considerable interest has been shown in obtaining a reliable friction factor at the wellsite. The tripping hook loads have been compared with the rotating hook load at selected depths. The torque losses have been computed from surface torque measurements with the assumption of a constant drilling torque. A more accurate value of the mean friction factor during rotation can be obtained, however, by measuring torque and WOB both at the surface and downhole. Because certain sections of a well may be drilled with no surface rotation of the drillstring, a sliding friction factor ("drag") was defined. Its use was then extended to rotating drillstrings. Drag is sensitive to differences in surface and downhole WOB and is normalized for changes in WOB, hole geometry, and mud weight. The term can therefore be used to compare the severity of drilling problems at widely different depths and inclinations. The instantaneous (i.e., foot-by-foot) friction factors are rich in information about the hole conditions, as the examples given here prove. They do not, however, monitor the whole well. The friction factor is computed between the two measurement points, the MWD tool and the surface, not between the bit and the surface. Another parameter, the dimensionless torque, TD (the downhole torque divided by the downhole WOB and bit diameter), watches over the hole between the bit and the MWD tool. Variations in TD have been shown to lead to an accurate computation of bit tooth wear, as well as detection of bearing failure. We use TD to monitor the excess torque created between the bit and the MWD by stabilizers digging or the string traversing a severe dogleg. TD is not required for friction analysis if the MWD tool is placed close to the bit. Computation of the Friction Factors. As shown in Refs. 1 and 2, the rotating friction factor is proportional to the torque losses. Because the side forces and tension profile are not independent, an iterative procedure is used to compute the sliding friction factor. The program stops iterating when the computed surface WOB falls within 100 lbf [445 N] of the measured one. About six iterations are normally required. The computation time is on the order of 1,000 ft/min [305 m/min] on a mainframe system. Field Examples This section discusses four field examples that were used to develop the interpretation technique. None of the interpretations were performed in real time. Rotating-Friction-Factor Field Example. Track 5 of Fig. 1 is a log of the rotating friction factor ("fric") for the 12.25-in. [31.12-cm] hole section of a well drilled onshore in the U.K. Surface and downhole measurements of WOB and torque are plotted in Tracks 1 through 4. The log shows data from sections of the hole where rotary BHA's were used and values of surface torque were available. Over the interval from 2,632 to 3,652 ft [802 to 1113 m], fric was not calculated because downhole motors were used and the drillstring was not rotated (no surface torque). SPEDE P. 248^