Sealing Capacity of API ConnectionsTheoretical and Experimental Results

Catalin Teodoriu, Texas A&M, and Marius Badicioiu, UPG Ploiesti

Summary

Hydraulic fracturing has become a state-of-the-art stimulationtechnique. It has been proved over the years that significant pro-duction increase can be obtained by applying the right fracturingtechnique. Now, the most advanced techniques of geothermal-energy-recovery systems widely use hydraulic fracturing.

The following paper presents the experimental results of thetests carried out on four different compounds using the improvedgrooved-plate method. The tests have shown a large variation ofthe tested-thread-compounds sealing capacity. Starting from theexperimental results and the theoretical analysis of the AmericanPetroleum Institute (API) connection, a useful chart was built todetermine the real connection resistance, on the basis of its initialmakeup torque. The chart offers to engineers involved in thedesign of a fracturing process the possibility to estimate the max-imum pressure that may lead to a connection leak.

Introduction

Most of the published data show that a long fracture is the key tooptimum well stimulation. The desired length of the fracture canbe achieved by use of equipment capable to deliver the rightpressure and fluid volume. Because the hydraulic-fracturing tech-nique can also be applied to old wells, equipped with standardAPI connections, the high pressures that are achieved during thepumping phase require the understanding of leak resistance of APIconnections. It has also been proved that during the injection phase,the high pump rate may lead to additional pressure increase into thewell tubulars. The time and pressure values are two key parametersthat may affect the sealing capacity of the API connection.

Testing the sealing capacity of a casing connection is not aneasy task because it depends on many factors: thread type andform, thread compound, aging of the thread compound, andmakeup-induced stresses. Actually, there are no standards to eval-uate the seal capacity of a thread compound. To date, threeapproaches have been found in the literature:

The fixture designed during the project PRAC 88-51 thatconsists of two circular-steel plates having a spiral grove fromthe center to the exterior (Wood et al. 1990).

Full-scale testing of threaded assembly using a high loadpress (ISO 13678:2000 2000) in which not only the thread com-pound but the entire sealing capacity of the assembly is tested.

Small-sized connections as described by Hoenig and Obern-dorfer (2006).

There are many advantages and disadvantages for each one ofthe methods, but testing thread compounds separately requiresremoving all inconsistent parameters that may affect the evalua-tion process. The main parameters that may affect the thread-compound evaluation are the stress/strain state induced becauseof makeup and thread tolerances.

The fixture proposed by project PRAC 88-51 offers the advan-tage of comparing only the threaded compounds, by neglectingthe makeup- and tolerance-induced errors. This is why it has beenconsidered the use of the same experimental setup as the onedescribed by Wood et al. (1990). The experimental setup will bepresented in detail in this paper.

Thread Compounds for Oil-Country-TubularGoods (OCTG)

Typical thread compounds for OCTG are formed using basegrease in which solid particles are dispersed. The grease is stan-dard lubricating grease made of mineral oil, having a metal soapas thickener (i.e., aluminum stearate). In a very low amount,additives are added to the compound to improve the followingproperties: high-pressure resistance, wear protection, and corro-sion protection.

The role of solid particles is to provide antigalling resistanceand sealing properties of the compound. Powdered metals andnonmetallic particles such as graphite or ceramic spheres are usedas solid ingredients. Typical metals used for thread-compoundsmanufacturing are lead, copper, and zinc. The common nonmetal-lic solids used for compounds are graphite, polytetrafluoroethy-lene, and ceramics.

The so called green dope, or environmentally friendly com-pounds, have a totally metal-free composition. Fig. 1 shows aclassification scheme of thread compounds after Hoenig andOberndorfer (2006). Table 1 shows the composition of somecommon thread compounds used in the oil industry, including thetested thread compounds described in this paper.

According to API RP 5A3 (2003), the performance generalrequirements of thread compounds include consistent frictionalproperties; adequate lubrication properties; adequate sealing prop-erties; physical and chemical stability, both in service and instorage conditions; and properties that allow the efficient applica-tion of the compound on the connection surfaces. In addition, forrotary-shouldered-connection thread compounds, they shouldlubricate the connection during the makeup runs to achieve bear-ing stresses (buck-up force).

The sealing capacity, or according to some authors, leak tight-ness, is provided by the high viscosity of the thread compound andthe small free path inside of the threaded connection.

The API Threads

The API round thread (API 8ed) is one of the very first standar-dized-thread types used for casing and tubing. Being inexpensive,simple, and easy to manufacture, the API round threads have beenused extensively for low-cost wells. A proof of their importancefor the oil industry is given by the extensive tests carried out by atechnical advisory committee in the 1990s having Phil Pattillo asChairperson (API Report 88/89/91-50 1989; API Report 86-511987; API Report 84-51 1985). The tests were focused onimproved understanding of the mechanical behavior of the API-round-casing connection when subjected to service loads ofassembly interference, tension, and internal pressure (API Report86-53 1987).

The API 8rd is an open-type thread, which means that if noother material is applied on the thread before its makeup, no sealis provided. Between threads, a small gap remains that must befilled to provide a leak-tight connection. This space is filled by thethread compound that is applied on the pin and the box beforemakeup. The shape and dimensions of the API 8rd are given inFig. 2. As can be seen in Fig. 2, the leak path consists oftwo spiral spaces comprising the gaps between pin-thread crestand box-thread root, and pin-thread root and box-thread crest,respectively.

The buttress thread was introduced later to offer a connectionwith a very good tension resistance. It was patented 27 November

Copyright 2009 Society of Petroleum Engineers

This paper (SPE 106849) was accepted for presentation at the Production and OperationsSymposium, Oklahoma City, Oklahoma, 31 March3 April 2007, and revised for publication.Original manuscript received for review 31 January 2007. Revised manuscript received forreview 3 March 2008. Paper peer approved 17 April 2008.

96 March 2009 SPE Drilling & Completion

1956 by Samuel Webb and assigned to the US Steel Corporation.As stated in the patent, the high leak resistance should not beexpected unless the stab flank is closed. The shape of the buttresshas a different leak path than an API round thread, usually muchhigher. Early API tests (API Report 86-53 1987) showed no diffi-culties with the leak resistance, but it must be noted that at thattime, API compounds were used for tests only.

The leak-path size depends on the thread-manufacturing toler-ances. To calculate the gap volume, the minimum and maximumtolerances have been considered, as presented in Figs. 3 and 4 forAPI round thread, and API buttress, respectively.

The Experimental Setup

As explained in the Summary section, the grooved-plate setup hasbeen chosen for thread-compound analysis because it allows test-ing of thread compounds independent from thread tolerances and

stress/strain state. Later, full-scale specimens with controlled geo-metry have been used for reference.

The test was performed according to the following procedure:The grooved plate was filled completely with the dope to betested, and then the grooved plate was assembled over the sealingplate and placed into the hydraulic press. The center of thegrooved plate was connected to the high-pressure pump. Mineralhydraulic oil was used as the pressurizing medium. The pressurewas increased slowly, and the moment at which the dope wasexpelled was recorded.

To build the groove plates, we calculated the groove sizeaccording to the real dimensions of a 5-in. API short-roundthread, with a wall thickness of 7.72 mm with a J-55 grade. Thecalculated dimensions for the grooved plate are shown in Table 2.The grooved and seal plates are shown in Figs. 5a, 5b, and 6.

Finite-Element Analysis of the ThreadedConnection and Experimental Setup

To estimate the contact stress to be simulated between thegrooved plate and seal plate, a finite-element analysis was carriedout, using the ANSYS University program. The contact pressurebetween the thread turns plays an important role for the sealingcapacity of the connections because as long as the contact pres-sure is higher then the pressure to be sealed, the only leak path

Fig. 2API open-thread dimensions.

Fig. 1Classification scheme of thread compounds, afterHoenig and Oberndorfer (2006).

March 2009 SPE Drilling & Completion 97

oobHervorheben

remains the spiral path between thread turns, crest, and roots. Thesame conditions must be achieved between the two plates of theexperimental setup.

First, the 5-in. connection has been investigated by determin-ing the thread-turn contact pressure after makeup and under themakeup and axial-load case. The results were similar to thosereported (API Report 88/89/91-50 1989; API Report 86-51 1987;API Report 84-51 1985). According to the finite-element analysis,the contact pressure after the connection makeup with recom-mended torque is between 30 to 45 MPa and depends on how the

averaging is performed. Because of local contact problems withinthe incomplete thread-turns zone, some spots with high contactpressure have been found (see Fig. 7). These values will not beconsidered for the further analysis. It has also been observed thatthe pressure on the thread flank is not uniform (API Report 86-511987), but for the experimental setup an average value has beenconsidered.

The same analysis was carried out on several buttress-connec-tion sizes. Fig. 8 shows the flank-contact pressure for an 185/8-in.connection. The average contact pressure was 65 MPa.

Fig. 3Leak-path dimensions as a function of thread tolerances for API round thread ( = diameter).

98 March 2009 SPE Drilling & Completion

oobHervorheben

oobHervorheben

The third finite-element analysis was performed on thegrooved-plate model to investigate the groove deformation causedby axial load applied on the plate. The results have shown that thepressure distribution on the contact area is uniform, excepting thecontact zones at the end of the groove walls (see Fig. 9). Also, ithas been found that the shape of the groove is changing, aspresented in Fig. 10. The total amount of area shrinkage for thecontact pressures produced in a real connection is low. When thecontact pressure increases, the groove deformation becomesimportant, especially for the API-round-plate model. For a contactpressure of 50 MPa, the groove area changes for the buttress plateto 0.85% and for the API plate to 7.3%, which represents a changein area 9 times that of the buttress. These changes justify the

results and show that the API-plate leak resistance increasesalmost linearly compared to buttress-plate leak resistance(see Fig. 11).

Experimental Results

The first experiments were focused on testing four thread com-pounds using the grooved plate that mimics the buttress thread.The tested compounds are presented in Table 3. The first twothread compounds are proprietary types; therefore they will becalled T1 and T2; the third one is an API modified type, and thelast one is an API-type compound with polymers to increase theviscosity. The API modified type was used as reference for alltests.

Fig. 4Leak-path dimensions as a function of thread tolerances for API buttress thread.

March 2009 SPE Drilling & Completion 99

oobHervorheben

oobHervorheben

Fig. 10 shows the leak pressure as a function of contact pres-sure between plates. The leak pressure is the pressure at which thethread compound is expelled from the free end of the groove. Itcan be seen that all tested compounds show a higher leak pressurethan the API modified. Also, it has been observed that at highcontact pressures, the leak pressure tends to behave asymptoti-cally. The thread compounds T1 and T2, along with the API withpolymers, show very little differences at high contact pressuresbetween plates. This is explained by (1) the slight deformation of

the groove caused by the confining axial force applied to keep theplates together and (2) by the viscosity of the compounds.

The second set of tests was carried out on a grooved plate thatmimics the API short-round thread. For these tests only, the threadcompound T1 has been evaluated to compare the results with thefull-scale specimens that have been doped with this type of com-pound. Fig. 11 shows a comparison between the results obtainedusing the buttress-type groove and the API-round-type groove.One explanation for the asymptotic behavior of the API-buttressleak-resistance curve is the groove deformation caused by axialload of the plate. As presented in the preceding section, the groovedeformation can reduce its volume up to 7% for the API-stylegroove and up to 0.8% for the buttress-style groove. Because theAPI-groove size becomes smaller and smaller with the increasingcontact pressure, it is obviously why the API-plate leak resistanceis a direct function of contact pressure, and that of the buttressplate is not.

Discussions and Recommendationsfor Future Work

The experimental investigations presented in the precedingallow testing of thread compounds by qualitative comparison ofthe maximum expel pressure as a function of plate-contact pres-sure. The advantage of this approach is that it allows comparingthe leak capabilities of thread compounds independently of therhelogical properties. These can have a large impact on overallleakage when the dope behavior inside the groove has to bedescribed.

The critical parameter that dictates the ability of compoundsto withstand the applied forces is the dope-yield point. Unfortu-nately, API does not require such information to be provided by

Fig. 7Flank-contact pressure of a 51/2-in. API-round-threaded connection after optimum makeup torque.

Fig. 5Shape and dimensions of the grooved and seal plates used for the experiments.

Fig. 6Pictures of the grooved (top) and seal (bottom) plates.

100 March 2009 SPE Drilling & Completion

the dope manufacturers. For example, Fig. 12 shows an extract ofthe dope-properties specifications after API RP 5A3 (2003). Theonly remark in API RP 5A3 is that the rheological propertiesshould be tested according to ASTM D2196-05 (2005). Table 4

shows the tested-compound properties according to API that cangive the reader information about the rheological properties of thethread compound.

As shown in Fig. 9, groove deformation becomes significantfor API round threads, which directly affects the flow of a givencompound. In this case, the rheological properties will stronglyaffect the time after which the dope will be expelled, for a givenpressure difference, corresponding to a certain pressure gradient.This will influence the long term integrity of the well.

As a result of this study, we recommend that research beperformed that is focused on the relationships between doperheological properties, time, and flow behavior. To this aim, acomputational-fluid-dynamics (CFD) modeling study of the dopeflow is already under way at Texas A&M University.

Fig. 9Groove deformation at high contact pressure for APIround-type plate (bottom) and buttress-type plate (top).

Fig. 8Flank-contact pressure of a 185/8-in. API buttress threaded connection after optimum makeup torque and axial tension.

Fig. 10Leak-pressure curves for API buttress-type plate.

Fig. 11Comparison of API round and buttress leak-resis-tance-test results for Thread Compound T1.

March 2009 SPE Drilling & Completion 101

Practical Application of the Method

It is a common practice to record the applied torque while runningcasing. Knowing the minimum value for the applied torque of thethreaded connections, it will be easier to estimate the maximumpressure that may be applied without losing the connection tight-ness. In many fracturing operations, the part of the casing stringthat is subjected to internal pressure will have a low or zero axialtension. Therefore, the chart presented in Fig. 13 has been con-structed using the contact pressure calculated for the makeuptorque and internal-pressure case. Comparing the contact pressureinside the threaded connection to the minimum contact pressure atwhich the thread compound has been expulsed from the smallscale setup, it is possible to identify the actual leak resistance ofthe connection. The method does not consider the effect of toler-ances and thread-compound aging.

Conclusions

The tests performed on four different types of compounds haveshown that the API compound has the lowest leak resistance inconjunction with the API thread type.

The buttress leak resistance has an asymptotic behavior. Atcontact pressures higher than 100 MPa, the leak resistance isconstant.

The difference between API-round and buttress leak resistanceconsists in the contact-pressure dependency of the API-round leakresistance.

It is recommended that API-round connections be made upwith optimum makeup torque or higher.

As a result of this study, we recommend that research beperformed that is focused on the relationship between dope rheo-logical properties, time, and flow behavior. To this aim, a CFDmodeling study of the dope flow is already under way at TexasA&M University.

Nomenclature

A1 = thread gap area at flank root in box, mm2

A2 = thread gap area at flank root in pin, mm2

b = groove width, mmD = groove plate diameter, mmhn = height of thread box, mmhs = height of thread pin, mmH = theoretical thread height not truncated, mmMt = makeup torque, Nmp = thread lead or pitch, mmps = groove pitch, mm

scn = root truncation pin, mmscs = thread truncation pin, mmSrn = root truncation box, mmsrs = thread truncation box, mm

Acknowledgments

The authors would like to thank Vlad Ulmanu for his standingsupport and valuable advice.

References

API Report 84-51, API Research ReportProject #84-51, Investigation of

pipe thread compounds. 1985. Washington, DC: API.API Report 86-51, API Research ReportProject #86-51, Investigation of

pipe thread compounds. 1987. Washington, DC: API.

API Report 86-53, API Research ReportProject #86-53, Investigation of

leak resistance of API buttress connector. 1987. Washington, DC:API.

API Report 88/89/91-50, API Research ReportProject #88/89/91-51,

Investigation of pipe thread compounds. 1989. Washington, DC: API.

API RP 5A3, Recommended practice on thread compounds for casing,tubing and line pipe, third edition. 2003. Washington, DC: API.

ASTM D2196-05, Standard Test Methods for Rheological Properties of

Non-Newtonian Materials by Rotational (Brookfield type) Viscometer.

2005. Conshohocken, Pennsylvania: ASTM International.

Hoenig, S. and Oberndorfer, M. 2006. Tightness Testing of Environmen-

tally Friendly Thread Compounds. Paper SPE 100220 presented at the

SPE Europe/EAGE Conference, Vienna, Austria, 1215 June. DOI:

10.2118/100220-MS.

ISO 13678:2000, Petroleum and natural gas industriesEvaluation and

testing of thread compounds for use with casing, tubing and line pipe.

2000. Geneva, Switzerland: ISO.

Fig. 12API modified-thread-compound characteristics asreported in product specifications, after API RP 5A3 (2003).

102 March 2009 SPE Drilling & Completion

Wood, F., Dairymple, D., McKown, K., and Matthews, B. 1990. Thread-

compound test procedures being developed. Oil & Gas Journal 88(37): 7576.

Catalin Teodoriu is a research supervisor at the TechnicalUniversity of Clausthal and an adjunct assistant professor inthe Harold Vance Department of Petroleum Engineering atTexas A&M University. Teodoriu has an equivalent MS from theOil and Gas University of Ploiesti, Romania and two PhDs from

the Technical University of Clausthal and the Oil and Gas Uni-versity of Ploiesti. He has worked as a research and develop-ment engineer, as a researcher, and as a research supervisor.Teodoriu has been involved in research for casing resistanceunder extreme loads, swelling cements for gas wells, drillstringcomponents and makeup procedures, underbalanced dril-ling and formation damage, the evaluation of the casingfatigue and fatigue of casing connectors, and the develop-ment of laboratory testing devices and facilities. He has pub-lished more than 50 papers, from which more than 10 arepeer-reviewed. Marius Badicioiu is a lecturer at the Oil andGas University of Ploiesti and is involved in teaching graduatecourses for material technology. Badicioiu has an equivalentMS and a PhD from the Oil and Gas University of Ploiesti,Romania. Related research activities that Badicioiu has beeninvolved with are the testing of thread compounds, experi-mental stress analysis of tubular, welding processes and theiroptimization. He has published more than 20 scientific papers.

Fig. 13Chart for API-connection leak-resistance estimation knowing the thread-compound leak resistance.

SI Metric Conversion Factors

in. 2.54* E + 00 = cm*Conversion factor is exact.

March 2009 SPE Drilling & Completion 103

Sealing Capacity of API Connections-Theoretical and Experimental ResultsSummaryIntroductionThread Compounds for Oil-Country-Tubular Goods (OCTG)The API ThreadsThe Experimental SetupFinite-Element Analysis of the Threaded Connection and Experimental SetupExperimental ResultsDiscussions and Recommendations for Future WorkPractical Application of the MethodConclusionsNomenclature