Advances in drilling hydraulics and drilling fluid rheology
one to determine more accurate multi-parameterrheological models.
Errors in rheological measurements
Oil-filed rheological measurements are prone toerrors, as with any instruments. Researchershave studied the issue in previous years, but ithas been customary to ignore these errors andutilize the values taken normally of the six (ortwo)-speeds at face value. The most significanterrors in rheological measurements are relateda) to the use of Newtonian shear rates, b) to theend effects of the viscometer, c) to wall slip andd) to yield stress determination.
Newtonian shear rates
The solution to the flow equation of the drillingfluid in the double-cylinder Fann typeviscometer is performed for the oil-field narrowgap viscometer, by assuming that the shear rateswithin the gap are Newtonian. If the fluid isnon-Newtonian, the shear rate on the wall of theinner fixed cylinder depends on the particularfluid model and on the rheological parameters tobe determined (Kelessidis et al. 2006, Guillot1990). Most of the time though, in commercialapplications but also in research, no correctionsare applied and use is made of the Newtonianshear rate instead of the true shear rate, withoutmaking the proper reference both in user guidesof rotational viscometers but also in researchpapers. The fact that the gap between the twocylinders is small for most standard viscometertypes, with the ratio being 067846.1 /
,tends to minimize the error introduced, but theerror is there and the magnitude depends notonly on the rheological model but also on theactual values of its rheological parameters.Successful quantification of these errors hasbeen done recently for Casson and Robertson-Stiff fluids (Kelessidis and Maglione, 2006) andin particular for Herschel-Bulkley fluids(Kelessidis and Maglione, 2008). In Fig. 4, theresults are shown for Newtonian and true shearrates for Herschel-Bulkley fluids. The analysisshows that the errors can be significant but canbe quantified. This can have an impact on thepressure loss estimation, and in Figure 5, theimpact of using Newtonian versus true HB-shear rates for various flow situations in annuliare presented.
Solution of the flow situations in the Fann-typeviscometer involve normally implicitassumptions that the cylinders are infinitelylong, known as the end-effects problem. Thisbeen questioned not only for non-Newtonian[Guguyener et al., 2002] but for Newtonianfluids as well [Barnes et al., 1993].An experimental study has recently beenundertaken (Kelessidis et al., 2010a) toinvestigate end-effects occurrence and theirmagnitude, for Newtonian and non-Newtonianfluids, with standard oil-field, direct indicating-viscometer. Is was shown that such viscometershave already embedded at manufacturing stagean end-effects correction, which was quantifiedfor the first time, of magnitude of 6.69% of thetorque developed. This does not allow end-effect determination in a straight forwardmanner, as it can be done with otherviscometers, and the authors presented aprocedure to resolve this. They furtherdetermined that there is additional end-effectscontribution from the top section of the bob,shown also experimentally in another relatedstudy and depicted in Figure 6 (Rune et al.,2009). Its magnitude ranges from 5 to 6 % forthe high shear rate range, to 12% for the lowshear rate range for the non-Newtonian fluidtested. Non-allowance for the additional end-effects errors could result in over-prediction of pressure losses in the annulus of the order of 11% which could jeopardize well safety indifficult to drill wells, as depicted in Figure 7.
Many times researchers wonder and try toassess whether slippage of the fluid to bemeasured slips, i.e. the fluid particle does nothave the same velocity as the solid wall itadheres. Significant research has gone into thesubject, for oil-drilling situations (e.g.Yoshimura and Prud’homme, 1988) and for