A critical review of the coiled tubing friction-reducing technologies in extended-reach wells. Part 1: Lubricants

1. Introduction

Predictable and successful coiled tubing (CT) operations in extended-reach wells rely on the CT ability to transfer enough weight-on-bit (WOB) along the entire well length and achieve the desired intervention goal. As the horizontal wells have become longer and longer for increasing the reservoir contact, the CT capabilities to entirely service these wells are limited by the performance of the CT extended-reach technologies to reduce the mechanical friction between the CT and well. For instance, there are tens of long horizontal wells, i.e., wells with laterals between 12,000 and 25,000 ft in the North Sea alone that cannot be serviced along their entire lengths with the current CT friction-reducing technologies (Livescu and Craig, 2015). While increasing the CT diameter is a theoretical option to improve reach, practically this is not possible when the larger CT is not available or the completion is too small.

Among the most used and performant CT friction-reducing technologies are lubricants, vibratory tools (Sola and Lund, 2000, Robertson et al., 2004, Newman, 2007, Newman et al., 2009, Newman et al., 2014, Castañeda et al., 2011, Alali and Barton, 2011, Schneider et al., 2011, Schneider et al., 2012, Azike-Akubue et al., 2012, Hilling et al., 2012, Wicks et al., 2012, Wicks et al., 2014, Dhufairi et al., 2013, Guo et al., 2013, Macdonald et al., 2013, Livescu and Watkins, 2014, Livescu et al., 2014b, Parra et al., 2014, Ahn, 2015, Benson et al., 2016, Kolle et al., 2016, Livescu and Misselbrook, 2016, McIntosh et al., 2016, Duthie et al., 2017, Livescu et al., 2017) and tractors (Hallundbæk, 1994, Nasr-El-Din et al., 2004, Hashem et al., 2005, Hashem et al., 2008, Bawaked et al., 2008, Arukhe et al., 2012, Arukhe et al., 2013a, Arukhe et al., 2013b, Al-Buali et al., 2015, Manil et al., 2017, Rajamani and Schwanitz, 2017). There are a few commercial vibratory and tractor technologies globally deployed to the field. For instance, there were six vibratory tools and two CT tractors on display at the SPE/ICoTA Coiled Tubing & Well Intervention Conference and Exhibition held in Houston, Texas, USA on March 21–22, 2017 (the most recent such event when this review was written). However, the vibratory tools are slow beyond the CT lock-up and the CT tractors perform unpredictably depending on the presence of debris in the well (Hallundbæk, 1994, Sola and Lund, 2000, Hashem et al., 2005, Bawaked et al., 2008, Alali and Barton, 2011, Castañeda et al., 2011, Schneider et al., 2011, Schneider et al., 2012, Arukhe et al., 2012, Arukhe et al., 2013a, Arukhe et al., 2013b, Azike-Akubue et al., 2012, Hilling et al., 2012, Guo et al., 2013, Liston et al., 2014, Parra et al., 2014, Wicks et al., 2012, Wicks et al., 2014, Al-Buali et al., 2015, Kolle et al., 2016, Livescu et al., 2017). Note that in addition to lubricants, vibratory tools and tractors, several papers have reported optimized taper designs for extending the lateral reach (Newman et al., 2014, Lee et al., 2016, Galvan et al., 2017). While these have some advantages in extending the CT reach, they have significant technological and operational limitations that are discussed in the second part of this review.

The simplest solution to extend the CT reach is the use of lubricants. Lubricants are used extensively for well intervention operations. Their field performance is still highly misunderstood, despite that several studies on modelling and laboratory and field testing have been reported recently, leading to an in-depth understanding of CT friction phenomena that may help increase the lateral reach significantly (Livescu et al., 2014a, Livescu et al., 2015, Livescu and Craig, 2015, Elrashidi et al., 2016). There are still anecdotal case histories recently published without strong scientific fundamentals. In general, the use of lubricants for CT operations is still based on their cost and marketing information rather than on their field-validated performance (Livescu and Craig, 2015). For instance, there are many commercial lubricants available globally that perform outstandingly in laboratory, but their performance is unpredictable and inconsistent at downhole conditions. Livescu and Craig (2015) reported the laboratory and field results obtained with a commercial lubricant (a blend of a fatty acid and an organic surfactant): in laboratory, the coefficient of friction obtained on a commercial rotational friction instrument was 0.08; however, the field coefficient of friction was 0.24, showing no difference comparing to the unlubricated case when no lubricant was used (Fig. 1).

The science of mechanical friction between two solids, in addition to lubrication, adhesion, and wear, is called tribology. It has been studied since Leonardo da Vinci at the end of the 15th century (Livescu and Craig, 2015). Some of the most important results are summarized in the three well known frictional laws: Amontons' first law (the force of friction is directly proportional to the applied load); Amontons' second law (the force of friction is independent of the apparent area of contact); and Coulomb's law (kinetic friction is independent of the sliding velocity). Assuming that Ff and N are the friction force between the CT and well and the normal force applied to the CT, respectively, then the coefficient of friction between the CT and well is $μ = F_{f} / N$ .

If a fluid film is inserted between the CT and the well, the friction is called wet and the coefficient of friction depends on many factors, such as temperature, pressure, contact surface roughness, sliding speed, humidity and fluid viscosity. If there is no fluid between the CT and well, the friction is called dry and the friction force Ff can take any value between zero (for static friction, when the bodies in contact do not move relative to each other) and μN (for dynamic friction, when the bodies in contact have relative motion to each other). In this case, the Coulomb friction model, $μ \leq F_{f} / N$ , applies.

Excellent studies on friction, buckling and helical lock-up between two tubulars in wellbores (i.e., drill pipe or CT and completion, etc.) have been published extensively in the last three decades. As of July 2017, a search in the Society of Petroleum Engineers (SPE) electronic library of technical papers resulted in 3624 papers with the keywords “pipe buckling”. As the focus of the current review is on CT lubricants and their proven field performance, more details about the pipe buckling and lock-up theory and modelling can be found in other papers (see, for instance, Dawson and Paslay, 1984, Mitchell, 1986, Miska and Cunha, 1995, Qiu et al., 1997, Zheng and Adnan, 1997, Qiu, 1999, Qiu and Miska, 1999, Aasen and Aadnøy, 2002, Zdvizhkov et al., 2007, Mitchell, 2007, Mitchell, 2008a, Mitchell, 2008b, Mitchell, 2009, Gao and Miska, 2010, Mitchell et al., 2015).

Even if usually the contact between the CT and well is not dry, the Coulomb friction model has been traditionally used for CT simulation in the pre- and post-job modelling stages. Until recently, a constant coefficient of friction had been assumed in the CT operations’ modelling, independent of the downhole conditions such as temperature, pressure, fluid type, surface type and roughness, CT speed, etc. Sometimes, the coefficient of friction is assumed dependent on the CT direction of movement (i.e., during running in hole, RIH, or pulling out of hole, POOH), although the scientific reason for this dependence is debatable (Livescu and Craig, 2015). Note that from the downhole parameters above, only the effects of temperature and fluid and surface types on the friction were also investigated experimentally for drilling operations (Kaarstad et al., 2009). It was found that the coefficient of friction depends on the surface types (i.e., steel on steel, steel on cement, etc.) and increases with temperature, but no information was provided on how these results could be extrapolated to CT operations.

Lizcano et al. (2001) presented a case history discussing the predicted and actual weights and the range of coefficients of friction associated with the wellbore geometry. They stated that, in general, the predicted and actual weights vary by 10–15%, while the coefficients of friction vary between 0.20 and 0.25 for RIH and between 0.30 and 0.35 for POOH. However, the actual weights were matched by increasing the field coefficients of friction to 0.31 for RIH and 0.39 for POOH. The authors explained these increases due to the need to apply different coefficients of friction in different sections of the well, such as corners or heels, which have abnormal conditions of azimuth change or continuous dog leg severity.

The unified coefficient of friction theory (i.e., one coefficient of friction for the entire CT) was introduced by Craig (2003) who analysed the recorded versus predicted weigh data for 33 offshore oil producers in the Norwegian sector of the North Sea. Based on his analysis, he concluded that a constant coefficient of friction μ = 0.24 was representative for most of those wells, independent of the well deviation complexity, production rates, and CT sliding direction (i.e., RIH or POOH). Since Craig's study, a default coefficient of friction has been assumed in the pre-planning stage of CT operations in wells with water-like fluids, such as fresh water, seawater, produced water, etc., between the CT and well.

The coefficient of friction may still vary along the CT length due to the presence of scale or debris, contact surface roughness changes, or temperature changes. In this case, several coefficients of friction could be considered along the CT length (Livescu and Craig, 2015). However, these discrete values could be averaged along the entire CT length. The common practice within the CT industry is to communicate one coefficient of friction per well. Based on the field evidence, the authors of this review strongly agree with the findings reported by Craig (2003) that the coefficient of friction does not depend on the well trajectory, CT shape (i.e., straight or helical), or direction of movement (i.e., RIH or POOH) which implicitly affects the CT shape. However, the contact force and the friction force between the CT and well do depend on these parameters (Tailby et al., 1993, He and Kyllingstad, 1995, He, 1995; Newman et al., 2014). Thus, the friction force, contact force and coefficient of friction along the CT should be correctly accounted for in all CT tensile force analysis (TFA) models.

Although the generic coefficient of friction of 0.24 is successfully used in the field for 2-in. CT operations in 5 ½-in. laterals as long as 5000 to 6000 ft, the friction force corresponding to this value is too large to reach longer laterals, assuming the CT ability to transfer at least 500 lbf WOB. To quantify the effect of different coefficients of friction on the potential CT reach, in Fig. 2 are shown several scenarios with lubricants and a fluid hammer vibratory tool.

Outside of the oil and gas industry, the so-called Stribeck diagram (Fig. 3) is well known to represent the friction of hydrodynamic bearings when a continuous or discontinuous thin liquid film is located between the two solid surfaces (Meyer et al., 1998, Livescu et al., 2014a). According to Stribeck's study, the relationship between the coefficient of friction $μ$ and the ratio vη/N, where vis the relative velocity between the two solids and η is the bulk liquid viscosity, can be represented as a single curve displaying three regimes.

In the first regime called boundary lubrication, the liquid film is discontinuous and the two solid surfaces are in contact at asperities (i.e., the average thickness of the liquid film is smaller than the average roughness sum of the solid surfaces); the hydrodynamic effects of the bulk liquid do not significantly influence the mechanical friction and the interactions between the two solid surfaces and liquid dominate the tribological characteristics. In the second regime called elasto-hydrodynamic lubrication, the average thickness of the liquid film is similar to the average roughness sum of the solid surfaces; the liquid viscosity is the most dominant factor on the mechanical friction. In the third regime called hydrodynamic lubrication, the average thickness of the liquid film is larger than the average roughness sum of the solid surfaces; the internal fluid friction alone determines the tribological characteristics.

Despite a broad use of the CT lubricants and their potential benefits for well interventions, reliable and consistent field validation studies of their friction reduction capabilities are rare. As of July 2017, a search in the Society of Petroleum Engineers (SPE) electronic library of technical papers resulted in 217 papers with the keywords “coiled tubing lubricant”. However, field performance studies are mostly anecdotal or proprietary. In addition, there was no available literature review of the knowledge gained to date. Thus, a critical review paper of the existing studies addressing the CT lubricants and scientific demonstrations of their field performance will indicate the challenges and limitations encountered. This will hopefully trigger further research development and best field practices for well intervention operations in extended-reach wells.

The paper proceeds as follows. First, the previous studies on CT lubricants are reviewed, including their field performance and laboratory data. The main two testing methods for predicting their friction-reducing properties are discussed. Then, best field practices for CT lubricants are provided. Finally, the principal findings of this critical review and further recommended work are summarized.

2. Laboratory and field studies on CT lubricants

Among all friction-reducing technologies for CT operations in extended-reach wells, lubricants are by far the most used. This is because of their relatively reduced development cost, broad availability and fast laboratory testing on commercially available rotational friction testers.

2.1. Field requirements for lubricants

Lubricants are used for CT operations for two main reasons. First, lubricants are used after the pre-planning job modelling of a CT operation indicates that they are needed to either reach further in a lateral or lower pulling weights and vibratory tools or tractors are not suitable alternatives (Livescu and Craig, 2015). The inability to reach the target depth is due to either a low margin of safety (i.e., low available weight on the bottom hole assembly (BHA) at the target depth, typically less than 500 lbf) or the actual lock-up is predicted (i.e., any additional injected CT would be placed into a stable spiral with minimal to low additional load transmitted to the BHA). This type of operation will normally utilize significant quantities of lubricant.

The second and by far the most common application for lubricants is to reduce the contact friction to reach slightly farther into the well or to release tubing or tools stuck in hole. These situations, termed operational recovery, typically use small slugs of lubricant to locally reduce frictional drag and achieve the desired goal.

Besides friction reduction, there are other issues related to using lubricants for CT operations. Although there is no standard testing for lubricants before the field operations, the operators are concerned about their thermal stability and shelf life, their chemical compatibility with all other fluids in the well, including any hydrocarbons, and the potential formation damage that these lubricants may induce. Each operating and oilfield services company has their own procedures in testing lubricants.

2.2. Rotational friction instruments

Traditionally, there are two types of friction instruments, rotational (Ke and Foxenberg, 2010, Barraez et al., 2014) and linear (Livescu and Craig, 2015, Livescu and Delorey, 2016), used to analyse the friction-reducing properties of lubricants for CT operations. While the type of movement itself (i.e., rotational or linear) does not affect the coefficient of friction between the CT and well, the rotational friction testers used in all CT lubricant studies available in literature do not have any options to change any parameters for mimicking the downhole conditions during CT operations (Livescu et al., 2014a). These instruments have such technologies as block-on-ring, pin-on-disk or ball-on-disk and consist of two metal pieces with very smooth surfaces that are pressed against each other.

Rotational frictions testers are extensively used for testing lubricants used for drilling and completion operations. A plethora of results from laboratory testing for drilling and completion lubricants have been reported in literature (see, for instance, Holand et al., 2007, Foxenberg et al., 2008, Kaarstad et al., 2009, McCormick et al., 2011, Patel et al., 2013, Amanullah, 2016, Li et al., 2016, Cayeux et al., 2017). However, there is no evidence that any of the lubricants from these studies has been used for CT operations or that it could perform predictably and consistently during CT operations.

Because the rotational instruments do not take into account the effects of downhole conditions on the coefficient of friction, very low coefficients of friction have been reported when testing lubricants using these instruments in laboratory. For instance, coefficients of friction as low as 0.07 were obtained in several laboratory studies (Ke and Foxenberg, 2010). Barraez et al. (2014)reported coefficients of friction reduced in the range of 45 and 78% obtained on a commercial rotational instrument with several undisclosed lubricants, but the absolute coefficients of friction were not reported. Assuming Craig's default coefficient of friction of 0.24 (Craig, 2003), the laboratory coefficients of friction in the study of Barraez et al. (2014) would have been as low as 0.05; assuming a default coefficient of friction of 0.34 for rotational testing, the actual coefficient of friction could have been 0.07. However, none of these low coefficients of friction has been confirmed at downhole conditions in the field. For instance, very few cases with field coefficients of friction as low as 0.17 (i.e., friction reduction of 29% from Craig's default coefficient of friction of 0.24) were reported in literature until 2014 (Livescu and Craig, 2015).

Three reasons have been identified for the significant discrepancies between the rotational testing results and field data. First, the laboratory tests were performed at atmospheric conditions. However, temperature and pressure were identified as the downhole parameters that affect the fluid viscosity and thus the coefficient of friction (Kaarstad et al., 2009, Livescu and Craig, 2015). Second, the laboratory tests should take into account the downhole fluid mixture, as it is used in the field, and not only the lubricant. Usually, the fluids pumped downhole during CT operations are composed from brine or produced water, a lubricant (for mechanical friction reduction), a friction reducer (for fluid pressure drop reduction), surfactants, etc. Hydrocarbons or debris could be present as well. Third, all laboratory tests were performed on rotational testers with smooth contact surfaces that do not mimic the roughness of the CT and well contact surfaces. However, the average roughness of the contact surfaces was identified as one of the parameters affecting the Stribeck curve shown in Fig. 3 (Livescu et al., 2014a). Note that reducing the CT friction is the most important in extended-reach wells, where the pressure drop is also the highest due to their length; friction reducers are usually used in these extended-reach wells for reducing the pumping pressure and increasing the life of the CT string; compatibility of lubricants and friction reducers used for CT operations is thus crucial.

2.3. Linear friction instruments

Based on the discrepancies between the coefficients of friction measured on rotational friction instruments in laboratory and in the field, a linear friction instrument mimicking the CT conditions during well intervention operations (Fig. 4) was developed to measure the coefficients of friction of lubricants mixed in downhole fluids when the following parameters are varied: temperature, pressure, fluid chemistry, and contact surface roughness (Livescu and Craig, 2015, Livescu and Delorey, 2016).

In order to improve the shortcomings of the rotational friction testing discussed above, this instrument was designed to allow using real CT coupons and well casing samples. These could be easily replaced with new samples if significant wear of any of the contact surfaces should be detected. The volume and concentration of fluids could be also changed for each individual measurement (Livescu and Craig, 2015, Livescu and Delorey, 2016).

The entire instrument, including the fluid and the CT and casing samples, could be inserted into a pressurized chamber and heated up. Pressure tests up to 9000 psi were reported with a lubricant in synthetic sea water (Livescu et al., 2014a); however, the coefficient of friction varied by less than 5% comparing to similar tests at atmospheric pressure. This led to the conclusion that pressure does not have a strong effect on the coefficient of friction of lubricants in water-based fluids. Indeed, the viscosity of water varies by less than 1% when the pressure increases from the atmospheric pressure to 9000 psi at 20 °C.

For temperature measurements, a heating pad could be installed under the casing sample. Four thermocouples are used to monitor the temperature of the system: two above the casing sample (at each of its ends), one below the casing sample and one on the CT coupon. These thermocouples are used to assure that the temperature of the fluid and the two contact surfaces are similar during each test. The maximum temperature at which measurements could be performed at atmospheric pressure is 98 °C, with external humidity control. Tests at higher temperatures require increasing the external pressure as well in order to avoid fluid evaporation.

Friction tests with several real CT coupons were reported (Livescu et al., 2014a, Livescu and Craig, 2015). Their sizes and average roughness were between 1 ¾- and 2 ⅜-in and between 0.92 and 10.23 μm, respectively. In addition, several real casing samples and planar steel plates with similar roughness were used. Their average roughness was between 5.68 and 12.44 μm.

2.4. Rotational versus linear testing comparison

Livescu et al. (2014a) compared the laboratory testing results obtained with rotational and linear friction testers and field data (Table 1). This is the only such comparison available in literature. The fluids tested were three lubricants, denoted as Lubricant 1, Lubricant 2, and Lubricant 3, mixed at 2% in 2% Potassium Chloride (KCl) brine and 0.1% fluid friction reducer. Lubricant 1 is a blend of a fatty acid and modified fatty acids; Lubricant 2 is a branched polyoxy ethanediyl blended with surfactants and a phosphate ester; and Lubricant 3 is a blend of a fatty acid and an organic surfactant. The three lubricants are commonly used for CT operations globally and readily disperse in water.

Table 1. The coefficients of friction of 2% KCl brine with 0.1% fluid friction reducer and three lubricants currently used for CT operations, as obtained from laboratory rotational and linear friction tests, at 20 and 98 °C, respectively, and field trials (Livescu et al., 2014a).

Lubricant	Rotational Coefficients of Friction at 20 °C	Linear Coefficients of Friction at 98 °C	Field Coefficients of Friction
None	0.34	0.28	0.24–0.28
2% Lubricant 1	0.04	0.20	0.17–0.24
2% Lubricant 2	0.05	0.18	0.17–0.24
2% Lubricant 3	0.08	0.12	0.13–0.14

The rotational friction testing results presented in Table 1 were obtained at atmospheric temperature (i.e., 20 °C) and the contact surface roughness was not modified. The linear friction tests were performed at 98 °C with a CT coupon and a planar steel plate with average roughness of 5.68 and 0.67 μm, respectively. The differences between the rotational friction tests at 20 °C and linear friction tests at 98 °C are significant: 21% for the base case when no lubricant and no fluid friction reducer were mixed in brine (0.34 vs. 0.28); 80% for the first case with Lubricant 1 (0.04 vs. 0.20); 72% for the second case with Lubricant 2 (0.05 vs. 0.18); and 33% for the third case with Lubricant 3 (0.08 vs. 0.12). Also, the results shown in Table 1 indicate that the linear friction results, and not the rotational friction results, are very close to the field results. This conclusion is opposite to that presented by Barraez et al. (2014) who reported very close agreement between the coefficient of friction reductions obtained on a commercial laboratory rotational friction instrument and their field tests.

Based on the discussion above, the coefficients of friction vary with temperature. However, one coefficient of friction at a given temperature does not indicate the full friction-reducing potential of a lubricant. For instance, in Fig. 5, the temperature-dependent coefficients of friction are shown for 1% concentrations of the three lubricants from Table 1 (i.e., Lubricant 1, Lubricant 2, and Lubricant 3) mixed in sea water and 0.1% fluid friction reducer. At 98 °C the linear coefficient of friction of the Lubricant 3 solution (0.13) is 43% and 32% lower, respectively, than those of the Lubricant 1 (0.23) and the Lubricant 2 (0.19) solutions.

Although the linear friction tests those results are shown in Fig. 5 were performed at certain discrete temperatures (i.e., 20, 30, 40, 50, 60, 70, 80, 90, and 98 °C ± 2 °C), the coefficient of friction curves are smooth, indicating a continuous variance of the coefficients of friction with temperature. Note that the curves shown in Fig. 5 and the Stribeck curve appear more or less similar.

2.5. Field examples with CT lubricants

In addition to the scientific aspects related to understanding how lubricants work at downhole conditions, acquiring and interpreting CT friction data from field operations is another great challenge (Bhalla, 1995, Craig, 2003, Schneider et al., 2011, Hilling et al., 2012, Griffin and Nichols, 2012, Barraez et al., 2014, Livescu and Wang, 2014, Livescu and Craig, 2015, Castro et al., 2015, Kolle et al., 2016, Yeung et al., 2016, Yeung et al., 2017). In general, the lubricant benefit is evaluated by matching the pre-job predicted and post-job observed weights. This task is significantly easier and far more accurate when used in the planned phase for two reasons. First, steady-state weight data are available over a longer lubricated lateral section. Second, it is likely to use the situation when the well is unlubricated as a base case for matching the observed and predicted weights. Thus, the lubricant benefit could be quantified comparing the observed and predicted weights during RIH or POOH or both when the well is unlubricated versus when the well is lubricated.

A few case histories have been previously reported discussing different lubricant capabilities. Several papers detail various methods of analysing field data results to improve operational predictability (Hilling et al., 2012, Griffin and Nichols, 2012, Barraez et al., 2014). These methods are different from the authors’ practices published previously (Craig, 2003, Schneider et al., 2011, Livescu and Craig, 2015, Castro et al., 2015, Elrashidi et al., 2016). All these methods have merit in that data from multiple wells were sourced and a consistent technical process was followed in each case. More details about each method are given below.

Schneider et al. (2011) reviewed the results observed during milling operations from several shale formations across USA (Eagle Ford, Bakken, Haynesville, Barnett and the Marcellus) using both lubricants and fluid hammer vibratory tools. The lowest coefficient of friction observed was 0.16 (obtained in only one well), while the largest was 0.24 (obtained in several wells). From the weight matching analysis of all jobs, it was concluded that the lubricant benefits varied and were inconsistent.

Griffin and Nichols (2012) presented a case history with lubricants (called “metal-on-metal friction reducers” in their study) and a fluid hammer vibratory tool for a number of similar wells in the Bakken formation of the Williston Basin in the USA. They compared the differences between the predicted and actual (i.e., observed) weights and maximum depth before reaching lock-up. Their field results showed average extended-reach gains of 10% when both a lubricant and the fluid hammer tool were used. They also reported a 50% reduction in the circulating pressure and extending the CT fatigue life. Before the field trial, the lubricants were tested on a rotation friction instrument. The actual coefficients of friction were not reported, but friction reduction between 10 and 70% was obtained for field mixtures of lubricants and acrylamide copolymer (a fluid friction reducer). The lubricant samples were heated up at 50 °C, but no details were provided if the samples were heated up before or during testing. Field results for 17 wells were also reported. Using the same combinations of lubricants and fluid friction reducers from the laboratory tests, coefficients of friction between 0.18 and 0.29 were calculated from the actual weights during the 17 operations. However, the default coefficients of friction from the 17 operations were not provided. In addition, it is not clear in which wells the fluid hammer tool was used and how much it reduced the CT friction.

Barraez et al. (2014) continued the study by Griffin and Nichols (2012) and tested several lubricants (also called “metal-on-metal friction reducers” in their study), commonly used for well intervention operations in North America, using a rotational friction instrument. The actual coefficients of friction were not reported, but friction reduction between 44 and 78% was obtained in their rotational friction tests. The laboratory tests were performed at room temperature and at a unspecified “bottom hole temperature”. Like in the study by Griffin and Nichols (2012), no details were provided if the lubricant samples were heated up before or during testing. Among the conclusions of this study were that petroleum-based lubricants were consistently the lowest performants due to their poor miscibility with the brines and acids; sorbitan trioleate-based lubricants performed well with acids; and aluminium silicate lubricants performed poorly when mixed in brines. However, all these statements were vague and did not have any scientific proof.

Livescu and Craig (2015) presented five case histories with several lubricants: a 7-in monobore completion with a complex deviation survey in North Sea and four 5 ½-in ‘J shaped’ wells in Texas, USA. The lubricant from the North Sea operation reduced the coefficient of friction from 0.28 (when no lubricant was used) to 0.17. For the four operations in Texas, one lubricant reduced the coefficient of friction from 0.24 (when no lubricant was used) to 0.17; a second lubricant did not have any influence on the coefficient of friction; and a third lubricant reduced the coefficient of friction from 0.24 to 0.22 (when no lubricant was used) to 0.13 in two wells. Note that in the last well, the default coefficient of friction was 0.22 because a fluid hammer tool was used initially. Also, for all these five operations, the field coefficients of friction were calculated from both RIH and POOH data. These field coefficients of friction were validated by experimental results on a linear friction instrument for the given field fluid compositions and temperatures.

Castro et al. (2015) reported three extended-reach annular fracturing case histories from North America. In two wells, a lubricant reduced the coefficient of friction from 0.24 (when no lubricant was used) to 0.09 and 0.13, respectively. In the third well, the same lubricant reduced the coefficient of friction from 0.20 (when a fluid hammer tool, and no lubricant, was used) to 0.13. The field coefficients of friction were validated by experimental results on a linear friction instrument for the given field fluid composition and temperature.

Elrashidi et al. (2016) presented a case history in a 5 ½-in. ‘J shaped’ sand-screen-completed injector well from the North Sea. This is the first reported case in literature describing the lubricant use in a well severely filled with sand. A lubricant reduced the coefficient of friction from 0.36 (when no lubricant was used) to between 0.18 and 0.20. Note that the higher default coefficient of friction of 0.36 was due to the presence of sand. The field coefficients of friction were validated by experimental results on a linear friction instrument for the given field fluid composition, metal-on-sand surface roughness, and temperature.

Yeung et al. (2017) studied the friction reduction of several lubricants using a linear friction instrument similar to that developed by Livescu and Delorey (2016), without taking into account the effects of temperature, pressure, or contact surface roughness. They also presented the field data from eight wells. Their laboratory tests did not mimic the downhole conditions during CT operations and the laboratory results could not be extrapolated to real life applications. They speculatively claimed that “the field case studies have a similar conclusion where the applied lubricant combined with a friction-reducing tool could reduce the linear coefficients of friction by 40%.” At a careful reading, it is clear that this claimed friction reduction of 40% is from a much higher default coefficient of friction of 0.35 to 0.21. This means that, in order to claim a high friction reduction, the authors have artificially increased the Craig's default coefficient of friction from 0.24 to 0.35. This method of artificially increasing the default coefficient of friction to obtain higher friction reduction ranges was also used by Barraez et al. (2014).

Sherman et al. (2017) reported the details of a tribology and rheological analysis of three commercial lubricants in different oil field brines. They used a simplified version of the linear friction tester reported by Livescu et al. (2015)and Livescu and Delorey (2016), as all tests were performed at room temperature, and found that the tested lubricants reduced friction between 11 and 28%. Each test fluid sample consisted of 2% lubricant and 0.02% fluid friction reducer in brine water. The qualitative field results from five operations in North America were also reported. However, no field coefficients of friction were provided and no comparison between the laboratory and field coefficients of friction was shown. Without any supporting data, it was concluded that continuous pumping of lubricants may have yield more friction reduction that pumping lubricant slugs.

2.6. Field methods for CT extended-reach predictions: using a global coefficient of friction

As mentioned above, Craig (2003) obtained weight data with consistent well parameters (pressure, fluid composition, and presence or absence of lubricant) using the specific well profile and obtaining a global coefficient of friction by best curve fitting. In addition, the CT direction (i.e., RIH or POOH) did not affect the coefficient of friction.

The same field method was applied by Schneider et al., 2011, Livescu and Craig, 2015, Castro et al., 2015; and Elrashidi et al. (2016), independent of the well profile, lubricant and fluid chemistry, downhole temperature and pressure, well surface type (i.e., metal or sand), and how clean the well was (i.e., the wells were new, producing oil or gas, or had residual debris). A typical RIH and POOH weight matching for the same coefficient of friction is shown in Fig. 6.

2.7. Field methods for CT extended-reach predictions: using slopes and multiple discrete coefficients of friction

Hilling et al. (2012) reviewed RIH data from 70 wells with five groups of well deviation profiles while running three different vibratory tools to mill composite bridge plugs. They considered that post-job weight data matching while varying the coefficient of friction was not a suitable process to determine the overall vibratory tool effectiveness in terms of friction reduction. They used a commercial CT modelling computer program to determine the rate of weight loss per depth gain for various CT wall thicknesses in the lateral. The weight loss was normalized to 2-in. CT with an average wall thickness of 0.156-in. for each of the five well groups. The slopes were calculated from acquired weight versus depth data for each operation to determine the effectiveness of the vibratory tool in question. A graphical example for the slope method introduced by Hilling et al. (2012) is shown in Fig. 7.

As discussed previously, Griffin and Nichols (2012) reviewed RIH data from 40 Bakken extended-reach wells. They extended the Hilling et al. (2012) slope method by considering a coefficient of friction slope in the vertical section and another slope in the upper lateral. In addition, they added the third slope in the final section of the well when significant weight loss was observed as the CT became progressively more spiralled while approaching the lock-up. Griffin and Nichols created a force match for each well using over 20 discrete sections varying the coefficient of friction in each section to obtain a very smooth predicted to actual weight match (Fig. 8). The vibratory tool tensile benefit was modified in order to match the model with the actual field observed lock-up depth. Each of the three slopes had a single coefficient of friction determined by using a weighted average calculation.

It is worth mentioning that the slope method could be challenged if utilizing a tapered string (Newman, 2007, Galvan et al., 2017) and should be applied over a finite length of a lateral, depending on the CT tapering size. In addition, this method could be challenged if the well trajectory is complex with undulations due to drilling or continuous azimuth change and relatively minor wellboretortuosity.

The same methodology was used in the study by Barraez et al. (2014). Apparently, field coefficients of friction reductions in the range of 45–71% were obtained when comparing the base case (i.e., no lubricant) to the lubricated cases, but no absolute field coefficient of friction was reported. A different methodology for calculating these field coefficients of friction reductions is revealed in the paper. In order to obtain these significant friction reductions, Barraez et al. followed the method of Griffin and Nichols (2012) and divided the well in several sections with different coefficients of friction. Unspecified weight-averaged coefficients of friction were then used for the lubricated cases and were increased in a commercial CT modelling computer program until lock-up was obtained theoretically. Thus, the difference between the theoretical coefficients of friction for simulated lock-up and the weight-averaged coefficients of friction for the lubricated cases would represent the 45–71% friction reduction obtained for a lubricant.