Department of Energy, under the Energy Related Inventions Program.

This PCT Application is a C-I-P of Application 09/573,480 filed May 16, 2000 which is a C-I-P of Application 09/421,483 filed October 20,1999 upon both of which, priority is claimed.

TECHNICAL FIELD The use of screw threads to connect joints of pipe together so as to convey fluid, is a very old art that has progressed for hundreds of years in an effort to satisfy periodic needs for stronger and better sealing pipe connections.

Performance requirements for screwed pipe connections still vary widely today, such as for home piping with less than 80 psi fluid pressure with virtually no mechanical loads, to Oil Well Pipe that may be required to hold over 15,000 psi gas pressure and simultaneously, withstand extreme mechanical loadings and endure wide temperature fluctuations.

Due to the historical weakness of threaded pipe connections and their tendency to loosen, leak, and or break, their use in industrial plants and refineries has been limited by Industrial Codes to very small pipe sizes and low pressures.

However, because there is no reasonable alternative pipe connection for use within the very limited hole sizes drilled for Oil & Gas Wells, threaded pipe connections are still used today, so most research on and development of pipe connections has been directed toward such use.

In 1939 API adopted the 8Rd thread connection to connect joints of API tubing and casing which is still used to connect about 80% of well pipe today. My patent 5,427, 418 filed in 1992 explained why API 8Rd Connections loosen and leak and"API Item 2239 Work Group"discovered that fact in 1995 and recently adopted some principles of'418 in API"SR17 Supplemental Requirements for API LTC Connections with Specified Performance".

In an effort to provide pipe connections that sealed better than API 8Rd Connections, special"Premium"pipe connections were developed by numerous other parties who adopted thin annular sealing lips at the end of their pin threads as depicted in my patent 4,613, 717 which was itself, a variation of my patent

2,766, 998 that introduced proven reliable pressure-aided, high-pressure, high temperature metal seals to the nuclear and space industries as well as the oil and gas industry. However, such seals when used in oil-well pipe connections are fraught with several serious problems such as: leakage due to lip damage; excessive costs; and loss of connection efficiency when used in flush connections because the pipe wall thickness for supporting axial loads is reduced to form the lip and it's mating sealing surface. Today, most experts in the field believe that a lip seal is required"because pipe threads can't seal gas pressures above 10,000 psi".

There has been considerable confusion in the industry as to what constitutes a reliable qualification test for threaded pipe connections, which has resulted in too many sales claims reflecting hopes more than facts. New standard ISO-13679 gives promise to end that problem in that it allows one to choose the % efficiency ratings relative to the pipe ratings that a connection is to be tested and qualified for, under combinations of: internal pressure; external pressure; tension; compression; bending; temperature; and the choice of water or gas as the pressurizing fluid. It also specifies test procedures that accurately measure performance capability.

Therefore, it is expected that the number of new connections offered for sale will decline in face of such stringent standards, but that real progress should accelerate because users can for the first time, begin immediate use of a new ISO qualified connection with confidence. Application for the present invention is made with that realization in mind.

For purposes of the present invention, the following definitions will apply.

Flank angle = The acute angle in a plane coinciding with the pipe axis measured between a thread flank and a plane positioned 90 degrees to the axis, the angle being plus if the flank faces away from the axis, the angle being minus if the flank faces toward the axis.

Included Angle = The algebraic sum of the stab flank angle and load flank angle.

Thread Turn = A 360 degree portion of a screw thread.

Pin = A male threaded pipe end, the thread turn of smallest diameter being the first thread turn.

Box = A female threaded pipe end formed to mate with pin threads, the thread turn of largest diameter being the first thread turn.

Wedgethread = A screw thread form having a crest, root, stab flank and load flank, the load flank being formed on a greater helical angle than is the stab flank such that the axial length of the crest is least at the beginning of the first thread turn, the crest length gradually increasing to a maximum axial length at the end of the last thread turn, such that the box and pin may be screwed together to a desired position of full makeup at which, both stab flanks and load flanks contact and wedge against their respective mating flanks and thereby prevent further makeup of the connection.

Trapped Thread = A thread form having a negative included angle. Should the flanks have profiles other than straight lines, the effective flank angle shall be considered to be a straight line extending from the crest to the root of each respective flank.

Open thread = A thread form having an included angle not less than zero, as opposed to a trapped wedge thread that has a negative included angle which is less than zero.

Metal-to-metal seal = Continuos contact of a non-threaded surface formed completely around a portion of a box or pin for cooperation with a mating surface of the other, so as to effect a seal against fluid that is to be conveyed by the connection.

Pin wall thickness = A dimension measured radially at mid-length of the engaged threads, extending from the pin thread pitch diameter to the pin bore.

Box wall thickness= a dimension measured radially at mid-length of the engaged threads, extending from the box thread pitch diameter to the box outer diameter.

Stab pitch = axial length between stab flanks one thread turn apart.

Load pitch = axial length between load flanks one thread turn apart.

Full-strength connection= a pipe connection that will seal and not rupture under any combination of loads at which, the VME yield stress of the pipe body is not exceeded.

Bridge thickness dimension=the maximum gap width formed between assembled mating threads, that the thread dope used on the threads will seal.

Pin Critical Area = the cross-section area of the pin in a plane perpendicular to the axis, the plane being positioned at the large diameter end of thread engagement.

Box Critical Area = the cross-section area of the box in a plane perpendicular to the axis, the planed being positioned at the small diameter end of thread engagement.

BACKGROUND ART Moore 1,474, 375 discloses an early form of a trapped thread but not a wedgethread, that is sealed after assembly by compressing a malleable member between mating threads.

Stone 2,006, 520 discloses a square non-wedgethread form and seals elsewhere as at 39,20 and 28, for a flush joint connection efficiency that"may be 53%", per Col 8 line 10. He claims no thread seal as made evident in Col 3 In 5-9, Col 4 lines 3-9 and Col 6 lines 44-47. Rotation is stopped by contact of shoulders 20 and 34, not by the wedging of threads. Col 4 In 49-53 indicate that sealing "face 39 which is complimentary in angularity to and adapted to seat in fluid-tight relationship with face 18"which precludes its ability to form a reliable seal because rotation of the lower end of nose 39 will cause it to contact seat 18 only at the top of surface 39 whereupon, the lower end of surface face 39 will rotate and no longer be in contact with face 18. Had surface 39 been formed at an angle less than face 18 as taught by'998, then a surface seal may have been formed.

Blose Re. 30,647 discloses Trapped Wedgethreads that suggests a thread seal in Col 2 In 7-11 but does not teach how to accomplish a thread seal and in fact, cites clearances between roots and crests in Col 3 In 40-43. Conversely in Col 1 beginning at In 64, he cites, "Since back flank is intended to always be negative, thread strain reactions against this surface will cause the box member to be pulled <BR> <BR> radially inward and the pin member to be pulled radially outward. "The force of such pulling away of the box and pin threads from each other is transmitted through flanks having a small included which angle generates a very high axial force on the flanks which in turn, causes a premature resisting torque during makeup of the connection, which typically stops short of the desired position of full makeup.

Thus upon assembly, the box and pin are left pulling away from each other with their mating roots and crests apart, such that excessive clearances left by stopping short of the ideal position of makeup and by complex machining tolerances of the dovetailed trapped wedge threads, are sure to produce leak paths. Then it is no wonder that with such leak paths guaranteed, that he requires a"primary resistance to leakage--by a separately functional metal-to metal seal"per Col 2 In 12-15.

Such seals occupy radial space of the pipe wall which in turn, further reduce the connection strength against axial and fluid loads.

Blose 4,600, 224 discloses a trapped wedgethread and cites in Col 1 lines 56- 60 that he provides a"controlled clearance between mating roots and crests". No where does he claim a thread seal as he suggested in Re. 30,647, but provides metal-to-metal seals as at 12 of Fig 1, as at 24 & 26 of Fig 2, and as at 32 and 34 of Fig 3 for the three embodiments disclosed. Col 4 lines 52-54 state that crests and roots do not make contact upon full makeup but specifies no limiting gap width which precludes a reliable thread seal, and nowhere does he claim a thread seal which is understandable in light of experience with his prior invention. Col 5 lines 6-8 discuss the tendency of trapped threads to"hang up"during dis-assembly.

Hang-up and damage are a common deficiencies of sharp edge trapped wedge threads.

Blose 4,600, 225 adds additional clearance as at 43 between mating threads to further confirm that he was not able to form a wedge thread seal.

Ortloff 4,671, 544 further confirms lack of thread sealing attained by the inventions above having a common assignee, in that he provides a resilient seal (26) mid-point the mating threads and metal-to-metal seals as at 22 and 24 of Fig 1.

Col 2 In 18-20 he mentions the threads seal but does not teach how. If those threads did seal, then his resilient seal and metal-to-metal seals would not be needed. The embodiment shown in Fig 4 does not claim a thread seal, but a metal- to-metal seal as at 50.

Ortloff 4,703, 954 explains in Col 1 In 12-42 that trapped wedgethreads cause very high stress concentrations and the objects of his invention are to reduce those stresses. Although he claims a thread seal in Claim 1, he does not teach how to achieve it.

Reeves 4,703, 959 discloses a trapped wedgethread connection that seals on a soft seal such as polytetrafluoroethylene in Col 2 In 6-17. Again, he claims a thread seal but does not teach how to accomplish it. Col 2 In 50-59 explain that the seal material cannot escape from the groove which it completely fills, however, elevated temperatures that are encountered in a well will cause expansion of the seal which in turn will generate extreme pressures between box and pin sufficient to induce failure of the connection. If the threads sealed the contained fluid, the soft seal would not be needed.

Blose 4,822, 081 discloses a trapped wedgethread but nowhere does he claim a thread seal, having no doubt witnessed tests on several of his inventions listed above. Instead, he cites seals as at 51 and 54. In Col 8 In 36-37 he indicates that the thread must be trapped to be drivable, but he later witnessed driving tests on my 30"O. D. connection, and embodiment of Patent'418 having positive flanks, wherein his employer reported, "no energy loss", which confirmed that they were drivable indeed. Col 8 In 47-53 and Col 9 In 5-8 state, the critical area of the box will yield when being driven which precludes a high strength pipe connection particularly in tension, per Col 10 In 39-42. Such yielding causes an energy loss which is not desirable for a drivepipe connection because it kills some of the impact necessary to drive the pipe into the ground. Col 8 In 15 he states,"the thread flanks do not slide against one another, and in fact, the thread flanks do not make contact until final make-up of the joint"so therefore the large 40 foot long joint of heavy wall pipe must be tediously suspended and very carefully lowered while it is being rotated to carefully engage the sharp cornered mating threads before being tightened to the position of full make-up, explained in Col 4 In 22-24 and Col 6 In 44-52 and Col 7 In 47-57 Mott 5,454, 605 depicts a trapped wedgethread described in Col 2 In 48-61 and illustrated in Fig 3 & 4. The stab flank is formed in the same direction, but at a different angle than the load flank so as to trap the thread as opposed to the dovetail trapped wedge threads described in the patents above. He properly describes the assembly and disassembly problems and the damage susceptibility of dovetail wedge threads in Col 1 In 51-Col 2 In 2. His improvement is a connection particularly for use on drillpipe as detailed in Col 2 In 18-23 & In 48-55 which conveys drilling mud during drilling of the well only, which is typically less than a month. Mud is a much easier fluid to seal than oil or gas because solid particles in the mud collect to dam up whatever clearances may exist between the mating threads, whereas casing and tubing connections must seal against water, oil and/or gas for many years. In Col 2 In 60-65, Col 4 In 61-Col 5 In 14, he claims a thread seal but again, does not teach how to seal against even mud. He states,"when made-up, there is no clearance between the threads"so not even thread lubricant is entrapped therebetween"but unfortunately, such perfect confirmation is and will always be beyond machining capability, and particularly within cost limitations for

pipe connections. He depicts two embodiments of his invention, one in Fig 3 to reduce flank hangup during assembly but not during disassembly, and a second embodiment in Fig 4 to reduce flank hangup during disassembly but not during assembly. Drillpipe connections are typically made and unmade hundreds of times, so the reader is left to wonder which half of such damage is to be eliminated by use of his invention. Neither has he eliminated the sharp corners where the flanks and crest intersect or the damage he claims to be associated with such sharp edges.

Watts 2,766, 998 teaches how to form an elastic metallic lip-seal and mating seat so as to effect a high pressure seal against gas for many years while under conditions of extreme variations of pressure and temperature.

Watts 4,613, 717 depicts how to use the seal of'998 in a threaded tubular connection.

Watts 4,813, 717 teaches how to form a box in a given size plain-end pipe so as to maximize the strength of a threaded pipe connection, for connection with another pipe end having a pin thread per claims 18 & 19 or for connection by double-pin couplings per claim 2.

Watts 5,143, 411 teaches how to form both a box and pin on plain-end pipe so as to maximize the strength of an integral threaded flush pipe connection.

Watts 5,018, 771 teaches how to form a mating box and pin so as to facilitate assembly and prevent cross-threading, in order to gain a more reliable and cost- effective connection.

Watts 5,427, 418 teaches how to effect a thread seal which has been proven capable of sealing against high gas pressure, as high as the pipe rated pressure.

When formed on 30"x 1"wall drivepipe having 6 degree stab flanks that was later driven into the ground, a connection per'418 tested very successfully, with"no indication of energy loss".

Watts 5,516, 158 teaches a high-strength self swaging connection having advantages that increase with the size of the pipe diameter. However, because it is swaged after threads are formed, it is not particularly suited for pipe sizes smaller than 5 inches.

Tubular connections in accord with some of the Watts patents listed above have been used successfully under severe conditions that no other known threaded pipe connection was capable of operating under however, like most threaded pipe

connections, they require close torque control for proper make-up so as to withstand the loads, but not overstress the box and/or pin. The one proven good feature of wedgethread connections heretofore, is that they can withstand a wide range of makeup torque after repeated assemblies, however, experience has also proven them to leak, to be difficult to assemble, to be highly susceptible to damage, to be difficult to manufacture and inspect, and to lack strength against superimposed loadings.

Six of the wedgethread patents cited above have a common assignee and four name a common inventor which confirms that a long, careful and continuing improvement effort has been focused on wedgethreads for over 25 years. However, all efforts of record are devoted to trapped wedgethreads and none mention or suggest that an Open (non-trapped) wedgethread may be workable, or feasible to connect and/or seal a pipe connection.

Experience has taught a characteristic of all Trapped Wedgethreads listed above, that mating flanks typically engage before the ideal position of full makeup is reached which in turn, prematurely generates a very high torque sufficient to stop rotation short of full makeup. Such a condition allows the connection to leak and loosen, and such loosening reduces its ability to be driven or to serve as a mechanical support. Upon first rotation of such pins into such boxes, the crest- root gap and the load flank gap are very wide, so excess dope flows freely outwardly from the connection. The premature torque begins when the thread dope thickness trapped in the then widest gap between the mating thread surfaces is reduced toward the bridge thickness dimension. For included angles where that gap occurs between flanks, the applied torque generates very high fluid pressure on the dope which in turn, forces the stab flanks together which are already in intimate contact due to the weight of the pipe joint being stabbed, and the friction between stab flanks causes the high premature torque which stops rotation short of the full makeup position. For smaller included angles where that gap occurs between roots and crests, the pressurized dope prevents intimate root-crest contact because dope pressure acting between the crests and roots holds them apart while forcing all mating flanks into hard contact to thereby generate an extremely high torque which stops rotation short of the full makeup position. Hydraulic force of the dope is multiplied greatly by a factor equal to: cotangent (included angle/2 + angle of

friction). Should an included angle be chosen such that both the upper flank gap and the root-crest gap are reduced toward the bridge thickness simultaneously, then the premature torque will approximately the sum of the two torque's cited above, and the loosening tendency of the connection will be even greater.

SUMMARY OF THE INVENTION The present invention is for use in various configurations such as with: conventional double-box pipe couplings having a larger outer diameter than the pipe; double-pin couplings; integral joints; flush joints; upset pipe ends; for connecting pipe to heavy wall vessels, and for reversibly connecting tubular or bar posts to their mounting base.

Open wedgethreads in accord with the present invention may be formed on a taper or alternatively, may be formed cylindrically. The dependable wedging action of my Open wedge thread provides a firm, repeatable position and torque at full makeup because no negative flanks exist whereas heretofore, trapped wedgethreads often trap excess dope between negative flanks which prevents their makeup to their ideal position of full makeup, as I have explained above. Prior cylindrical threads required a shoulder or such as a lock-nut to provide a firm, repeatable position of full makeup. Use of the present invention for cylindrical wedgethreads requires the radial width of the load flank to equal the radial width of the stab flank, whereas radial width of a tapered wedgethread stab flank is greater than the load flank radial width.

The present invention does not suffer premature flank lock-up as described above for trapped wedgethreads, because no negative included angles are used.

When box and pin are formed with no difference in taper, my Open wedgethread allows the pin and box to momentarily deflect radially apart during makeup sufficiently for excess dope to escape, unlike trapped wedgethreads that cause lock- up when dope pressure prematurely engages the flanks. For very fast makeups, for very cold conditions, and for other conditions under which the dope may not fully escape from between my mating threads fast enough during makeup, the pin threads may be formed on a taper slightly less than the box taper, to insure that the first pin thread turn will contact the box threads before the other pin threads do, and then during makeup, the other pin threads successively contact the box threads which progressively extrudes excess dope towards the thread end of larger diameter as makeup continues, and thereby prevents entrapment of excess dope between the mating threads. When substantially cylindrical threads are to be used and this dope extrusion feature is desired, the pin threads may be formed cylindrical and the box threads formed with a very slight taper or, the box threads may be formed cylindrically and the pin threads may be formed with a very slight negative taper.

As an example of a workable taper difference, I have successfully used a difference in thread interference between the ends of thread engagement according to my formula below for the value"PTS". It is generally known that entrapment of excess dope between any type mating threads can cause major problems. Therefore, unless a method for extruding excess dope is provided, any sealing thread that uses thread dope is subject to dope entrapment problems.

Watts patents cited above teach certain features that are combined in embodiments of the present invention in various combinations with a new Open type wedge thread that will assemble easily with a selectively wide range of final make-up torque, stop rotation of the mating threads at the desired position of full makeup, effect a seal against high-pressure fluid, and effect a high-strength rigid connection unlike the trapped wedgethreads of record.

Industry experts have long held that to prevent thread jump-out, a negative thread flank angle was necessary and have therefore devised many complicated, expensive and hard to inspect thread designs, as the above cited trapped wedgethread patents disclose. However, tests run for me by an independent test lab disprove that widely-held belief and are explained as follows. When a conventional pin having a relatively thin wall is made up into a box having a much thicker wall, then a jump-out tendency may exist because upon application of high axial tension loads, the thin pin wall tends to contract radially according to Poisson's ratio more than the thick box wall contracts, because the box axial stresses are less than the pin axial stresses which in the absence of a radial restraining force acting on the pin, may cause separation of the mating threads, a leak, and then possibly jump-out as loads increase. If it's thread form has a high positive flank angle such as the 30 degree flank angle used on API 8Rd threads, jumpout tendency is increased by vector forces in proportion to tangent (flank angle-friction angle), a fact that is generally known. However not generally known until taught by my Patent'717, is the fact that when the box wall thickness is reasonably close to the pin wall thickness, that the box will contract radially with the pin when the connection is subjected to an axial tension load, and will expand radially with the pin when the connection is subjected to an axial compressive load, sufficiently to prevent separation of the mating threads if the load flank angles are small. The present invention defines workable values"M"for

diametrical thread interference to allow this feature to be safely used in combination with my Open type wedgethread to prevent thread separation.

For connections having box and pin walls of substantially different thicknesses, compressive axial loads imposed could cause a thinner box to expand radially away from the pin so as to lose the desired thread interference. Likewise, a tensile axial load imposed on a thin pin could cause it to contract radially away from the box so as to cause it to lose the desired thread interference. Therefore, axial loads must be limited independently to prevent such a loss, in accord with my formulas given below for values BCS and PTS.

The sharp corners of trapped wedgethreads of record are extremely subject to damage and hangup as Mott describes in'605 above and even more importantly, they are difficult and time consuming to stab and makeup in the field which wastes Rig-Time that can exceed a cost of $500, 000/day. They also increase the tendency for damage, galling and leaking which may require extremely expensive remedial work under dangerous conditions that often damage the environment. The present invention provides an Open wedgethread having generous radii that connect flanks and crests so as to prevent such difficulties.

Tests have historically confirmed that makeup torque of a non-wedgethread connection varies with many factors such as thread surface finish, box-pin radial thread interference, the dope used, the taper, the rotational speed of make-up, the temperature and the flank angles etc. so therefore, a makeup torque independent of such characteristics is desirable to reduce time, cost and effort required to assure proper make-up repeatability of the connection and most importantly, to prevent overstress of the box and pin. The present invention utilizes a new Open Wedgethread having a simpler and more reliable configuration than do trapped wedgethreads, so as to assure repeatable make-ups to the desired full makeup position with a given torque, and not be subject to so many torque variations.

The present invention provides a new Open wedgethread that is easier and less expensive to manufacture, inspect and assemble than trapped wedgethreads which have sharp corners and negative flank angles. Trapped wedgethreads have many dimensions that must be controlled such as the crest axial length, the root axial length, the load flank radial width, the stab flank radial width, the stab flank angle, the load flank angle, surface finish of the almost hidden stab flank, surface

finish of the almost hidden load flank, tiny almost hidden radii at the roots that induce severe stress concentration, tiny crest radii that induce damage and galling, the load flank lead, the stab flank lead, the diameters of the crests, the diameters of the roots, etc. and all dimensions must have close tolerance limits they cannot exceed in order to even screw the box and pin together. The difficulty of maintaining tolerances during manufacture increases with the Square of the number of related tolerances. Thus, it is evident why Blose and others provide multiple independent seals other than depend on trapped wedgethreads to seal. If the trapped wedgethreads of record did seal, a few thread turns would seal and all the other types of seals would not be necessary. The present invention provides a new Open wedgethread form having flanks positioned at or near 90 degrees to the pipe axis, which allow easy entry of cutting tools and easy chip removal during manufacture. Also, such threads prevent portions of the thread flanks, roots or radius from being hidden, and allow easy entry of gages and sight checks for simple thread inspection. Thus, the present invention allows the threads to be machined in a more conventional, simpler, more reliable and less expensive manner, such that the threads can effect a fluid seal. To measure the precise thread form during manufacture of pipe connections, it is a common inspection practice today to trace the threads as on a fifty-to-one scale, and/or to make a hard rubber mold of the thread form to check on an optical comparator. However, neither practice is suitable to check a trapped wedgethread form, which further impedes the ability to manufacture such pipe connections accurately, and explains additional reasons for their assembly and sealing problems so widely experienced.

Relatively wide stab flanks and generous radii of the preferred embodiment using tapered threads are made possible by my relatively steep taper.. When lowered into the box for assembly, the stab flanks rest upon a complimentary stab flank of the box to support the pipe joint being installed, which also positions the pin over halfway into the box as taught by my patent 5,018, 771, such that alignment is automatic, such that cross-threading cannot occur as Mott described in '605. I generally prefer such tapers to be 1/8 to 1/4 diameter change of any given axial thread length for instance, a 3/16 inch diameter change in an axial length of one inch however, other tapers are within the scope of the present invention.

Typically, the taper may increase in steps, with an increase of wall thickness as

necessary to prevent excessive thread lengths. To prevent taper lockup between pin crests and box crests when the pin is stabbed into the box, it is preferred that the crests and roots be positioned parallel to the pipe axis.

For tapered wedgethreads, the stab flank must have a greater radial width than the load flank, and the preferred values are defined in the formulas given below as"S"and"B". The difference in the load flank axial pitch and the stab flank axial pitch for a given diameter defines the wedging angle between adjacent thread turns and if it is too great the threads may not seal, but if it is too small the desired position of full makeup may not be reached. My preferred difference in axial pitch may be found by the formula given below for the value"J", and preferred values for the load flank axial pitch and the stab flank axial pitch may be found by the formulas for"LF"and"SF"respectively.

It is a further object of the present invention to provide an Open wedgethread that is highly resistant to handling damage during manufacture, shipment and installation, in that no sharp corners exist between the metal faces forming the corners, but have crests positioned substantially at 90 degrees to the flanks connected by generous radii so as to thereby survive reasonable handling without significant damage. To even further minimize effects of localized crest damage too small to notice visually during assembly, the present invention teaches that the thread flank widths be dimensioned and toleranced such that the pin roots contact the box crests upon assembly but hold the pin crests which are most subject to damage, apart from the box roots so as to prevent interference and galling therebetween. It is within the scope of the present invention for the box roots to contact the pin crests and for the pin roots be held apart from the box crests, but this damage prevention feature would be lost along with the best sealing configuration, explained as follows: At the small diameter end of thread engagement, the axial length of the pin root-box crest contact of a wedgethread is necessarily much longer than the axial length of the pin crest-box root contact, so my preferred embodiment provides that the longer length pin root-box crest lengths be in certain intimate contact to minimize leak tendency there where fluid pressure is typically the greatest, and that the shorter crest and root lengths define a gap not greater than the bridge dimension, to be sealed by thread dope.

So as to help maintain the previously discussed intimate pin root-box crest contact after make-up of the connection, a predetermined amount of radial interference should exist between the box and pin threads sufficient to generate an optimum tension hoop stress in the box and a compressive hoop stress in the pin.

Should box makeup hoop stress be excessive, then the connections rating against internal fluid pressure could be reduced, or should pin makeup hoop stress be excessive, then the connections rating against external fluid pressure could be reduced because such stresses are additive to hoop stresses generated by fluid pressures. Therefore, the present invention teaches that the preferred diametrical interference between the mating box and pin threads should approximate the quantity: 1/3 times the pipe O. D. times the pipe material yield strength, divided by the pipe material modulas of elasticity. For best performance, thread interference should extend all along the helical thread length to both ends of the engaged threads, which should be considered when diameter and taper tolerances are specified. Such a controlled interference should allow reliable service near maximum ratings. However, for service at lower load ratings, other magnitudes of radial interference may be used without departing from the teaching of the present invention.

For extreme service requirements or because a user specification requires it, metal-to-metal seals may be provided adjacent either or both ends of the engaged threads. Most often, such a sealing surface is positioned adjacent the pin end for cooperation with a mating sealing surface formed within the box adjacent the small diameter end of the box threads so as to reduce the sealing diameter against internal fluid pressure and thereby, reduce the fluid pressure load on the connection. Occasionally it may be required to provide a metal-to-metal seal adjacent the large diameter end of the mating threads by forming an inner diameter surface of the box adjacent it's end for cooperation with a mating sealing surface formed around the pin adjacent the large diameter end of the pin threads. Such an external seal will provide maximum resistance against external fluid pressures.

Such metal-to-metal seals may also be required to prevent corrosion of the threads in extreme cases.

When metal-to-metal seals are required in a pure flush-joint connection, then some of the wall thickness required for the seal lip thickness may be lost for

mechanical strength of the connection, however, by use of my In-Process-Swaging taught as follows, a full strength near-flush connection may be provided that has engaged mating threads selectively, that will traverse as much as the full wall thickness of the pipe. The box and/or pin pipe ends are formed plastically after preliminary machining but before finishing, so as to maintain the precise thread form required for the threads to engage precisely. One example may be to machine a suitable counterbore of predetermined configuration within the end of a plain-end pipe joint, insert a suitable mandrel into the counterbore and force it axially into the counterbore so as to enlarge the outer diameter of that pipe end to a diameter greater than the original pipe outer diameter while enlarging the inner surface of the counterbore to a desired configuration, then machine the desired box within that end of the pipe. To form the pin end, a portion of the pipe outer wall at that end may be removed so as to form a desired configuration around the pipe bore, place a tool having an inner surface of a desired configuration against the pipe end, force the tool axially around the pipe end so as to swage down the pipe bore at that end to a diameter smaller than the original pipe bore while forming the outer surface to approximate the configuration desired, and then machine the desired pin on that end of the pipe. It should be understood that partially forming such ends of thinner walls and for shorter lengths that the than the box or pin lengths, requires only a small fraction of the force or time that Upsetting or Swaging full pipe wall thickness over the full box or pin length would take so therefore, such forming may take place on the machine tool that threads the pipe ends, which greatly reduces time and cost for delivery of high strength threaded pipe joints. Depending on the pipe dimensions and pipe material, such forming may be done at room temperature, or after heating to the lower critical temperature of the material as with an induction coil.

Another feature of the present invention to enhance connection strength and sealability is to dimension the axial length of the gap to be sealed at a practical minimum, and to maximize the number of thread turns within the desired length of thread engagement for a given minimum starting crest length, which is accomplished by dimensioning the least axial crest length of the first pin thread turn to be substantially equal to the least axial crest length of the first box thread turn. My formulas given below for the mean axial pitch labeled"A"will allow

future designers to determine such dimensions for any size connection. Then, based on the value of"A", the stab flank lead and the load flank lead may be found by my formulas for"SF"and"LF"respectively shown below. Thus, other non- workable, wasteful and/or dangerous factors may be avoided during design and field testing of new size connections.

The radial gap width existing between crests and roots after mating flanks wedge in firm contact, should not exceed the bridge thickness dimension if a thread seal is desired. In the preferred embodiment of the present invention, that gap width is controlled by the radial width of the stab flanks and the radial width of the load flanks, the previously mentioned pin root-box crest radial interference being maintained. It is easy to say that the pin crest contacts the box root simultaneously as the box crest contacts the pin root, but tool wear and machining tolerances prevent such from happening, and to even approach that idealistic condition would be excessively expensive for the manufacture of pipe connections. Therefore, a maximum gap width must be defined and not exceeded in practice, for the threads to be able to seal. As I explained in'418, the maximum allowable gap when using API SA2 Modified Dope was measured as substantially 0.006". Dimensions and tolerances for flank widths of the present invention are best set to maintain, a gap width no greater than. 006" when the mating flanks are in firm wedging contact at the position of full make-up when API SA2 Modified thread dope is to be used for assembly of the connection. In keeping with standard production practice, the target gap width is preferably at midpoint between zero and the bridge thickness dimension.

Another feature of the present invention is explained as follows. At the position of full makeup, mating wedgethreads having a zero included angle are wedged tightly together and there is no tendency for threads to be forced out of their mating grooves by excessive torque. The same is true if the included angle is less than twice the resulting angle of friction for a given thread dope. If non- wedgethreads were formed with such a small included angle, flank lockup would occur when flanks first engage, long before the position of full makeup was reached to effect the desired radial interference between mating threads and therefore desired connection strength, would not be attained. With my Open wedgethreads, radial interference is effected by root-crest interference before and

independently of the final torque magnitude which can occur only when the flanks wedge at the full makeup position.

There are service applications for wedgethreads that do not require maximum torque resistance and/or a highly stressed box and pin that my previously described embodiments allow, and for them the following embodiment has certain advantages, such as to improve thread cutting tool geometry in certain instances.

The flanks may be formed with an included angle that is greater than twice the angle of friction. Axial wedging pressure on the flanks is many times greater than the pressure between roots and crests due to the very small, helically configured wedging angle between stab and load flanks. The radial force vector of the wedging force tending to separate the threads, equals the wedging force times tangent (included angle/2-angle of friction). It is therefore evident that no such force will exist if the positive included angle does not exceed twice the angle of friction between the flanks. Should it be desired that the included angle exceed twice the angle of friction, then engineering calculations must confirm that box and pin walls are thick enough to retain that force without being overstressed, before such an included angle is used. A zero degree included angle is preferred to eliminate any concern of such separating tendency, but wedgethreads having positive included angles greater than zero are clearly within the scope of the present invention.

Thus, it is now clear that the present invention teaches how to form a cost- effective high strength Open wedgethread pipe connection that may be easily and repeatedly assembled to a desired position of full make-up within a wide torque range, so as to seal high pressure gas or liquids, while simultaneously withstanding extreme mechanical loads.

BRIEF DESCRIPTION OF DRAWINGS Figure 1 depicts a fragmentary section of a box and pin of the connection at stab position.

Figure 2 depicts the connection of Figure 1 at the position of full make-up.

Figure 3 depicts an enlarged fragmentary thread section taken from Figure 1.

Figure 4 depicts a fragmentary section of another embodiment using a double-pin coupling.

Figure 5 depicts a fragmentary section showing a variation of the flank angles.

Figure 6 depicts a fragmentary section of another variation of the flank angles.

Figure 7 depicts a section of a swaged pin and box after assembly.

Figure 8 depicts a fragmentary section formed with cylindrical Open wedgethreads.

Figure 9 depicts a section of a box and pin after preliminary machining, before forming.

Figure 10. depicts the box and pin of Fig 9 after forming but before threading.

Figure 11. depicts the box and pin of Fig 9 after threading.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS Multiple-start wedgethreads are within the scope of the present invention but for brevity, only a one-start thread is described. Figure 1 depicts Pin 1 at stab position within Box 2 such that helically configured pin stab flank 3 and the weight of the pipe joint it is formed on, is supported by helically configured box stab flank 4.

Pin load flank 5 and box load flank 6 are not then in contact with each other. Stab position is attained by lowering the pin into the box without rotation, wherein thread turns of the pin pass downwardly through thread turns of the box until the pin stab flank of each turn contacts a box stab flank turn too small in diameter for it to pass through, such that the pin stab flank rests on the box stab flank whereupon, pin crest 7 is in horizontal alignment with box root 8. Both root and crest are preferably formed parallel to the connection axis so as to prevent taper lockup during stabbing. Axial length 9 and diameter 10 of the pin thread crest are least at their lower beginnings, gradually increasing to maximum axial length 11 and maximum diameter 12 at it's upper ending. At stab position, axial length 13 of box root 14 is a constant amount of length greater than then adjacent axial length 9 of pin crest 7 such that when relative rotation of the box and pin occurs to effect makeup of the connection, slippage occurs between stab flanks, and the pin thread crest moves helically downwardly and toward the box root and toward the full make-up position shown in Fig 2 at which, pin thread flanks wedge between the box thread flanks because then, adjacent lengths 9 of the pin thread crests and 13 of box roots become substantially equal, and pin load flank 5 abuts box load flank 6 to stop rotation at the desired position of full make-up. Shortly before such wedging occurs, radial interference between pin thread root 20 and box crest 21 occurs to insure that no gap exists between them upon make-up, but gap 22 of predetermined width is simultaneously formed between pin crest 7 and box root 8.

The radial width of stab and load flanks of both box and pin must be dimensioned and tolerance such that the width of gap 22 as shown in Fig 2 at the position of full make-up is no greater than the bridge thickness dimension. The pin root-box crest contact just described has advantages over pin crest-box root contact because: the pin crest is much more subject to damage than any other surface of the threads and therefore for the preferred embodiment, galling during make-up and leakage after installation is much less likely because the pin crest is held out of contact

with the box root; and because the certain intimate pin root-box crest contact axial length at the minimum diameter of thread engagement is greater than the axial length of the pin crest-box root contact, sealing ability is greatly enhanced.

As seen in Fig 2 at mid-length of thread engagement 23, the box thickness 24 is measured radially between the thread pitch diameter and box O. D. and the pin thickness 25 is measured radially between the thread pitch diameter and the pin I. D. Axial space may be formed where needed as at 26 and 27 to relax end-length tolerances while insuring that the shoulders will not abut before the threads become wedged to establish the position of full make-up. To maximize connection strength as is well known in the art, partial threads may extend beyond the ends of full depth thread engagement so as to increase the critical areas of the box and/or pin, however such partial threads will not seal.

Although my fully engaged Open wedgethread will seal against high pressure gas, when special user specifications require or when it is necessary to optimize the sealing area, mating metal-to-metal sealing surfaces taught by my'998 patent may be selectively provided as at 28 and 29 of Fig 2, or at 93 and 97 of Fig 7 or similarly, for the embodiment depicted in Fig 4. Such seals may comprise cylindrical or conical surfaces wherein pin sealing surface 93 of Figure 7 is of slightly greater diameter than is mating box sealing surface 95 so as to effect an interference fit upon assembly, and/or pin surface 97 is likewise of slightly greater diameter than is mating box surface 98 when radial interference is chosen to effect sealing pressure. With metal-to-metal seals, thread sealing may be optional.

The enlarged thread form detail of Fig 3 depicts generous concave radii 30 joining the pin root to flanks, and like radii 31 of the box which are typically but not necessarily about 15% of the load flank radial width, to reduce stress concentration. Also shown are even larger convex radii 32 joining pin crest to flanks and like radii 33 of the box which greatly reduce damage during handling, transport, storage and assembly, as compared to the damage susceptibility of sharp- edged trapped wedgethreads. Threads formed per Fig 3 also provide a stable support by contact of the stab flanks as at 34 for the pipe joint being stabbed, which avoids the tedious handling required during assembly of trapped wedgethread connections, and they also provide automatic axial and concentric alignment of the connection to automatically position pin crests adjacent box roots

so as to assist proper assembly required to assure a reliable connection. Included angle 35 measured in the gaps between thread flanks may equal zero degrees or it may be a positive angle less than twice the angle of friction, or it may be a positive angle greater than twice the angle of friction.

Load flanks are preferably formed parallel to the stab flanks such that included angle 35 in Figure 3 equals zero, primarily for the purpose of improving dimensional accuracy to assure a fluid seal and to reduce costs to manufacture and inspect the connection. The improved dimensional accuracy teaches a cost- effective wedgethread seal not hitherto attained. Both flanks are formed preferably at 90 degrees to the axis but not necessarily. For instance, when box wall thickness 24, and pin wall thickness 25 are reasonably close in dimension, then the wedging force and the diametrical interference will cause the box and pin to contract and expand together when under superimposed loads without need of a reserve radial restraining strength of the tubular member that the threads are formed on to prevent separation of the mating threads which in turn, prevents leaking and/or jumpout.

Should box and pin thickness difference be extreme and the axial load be so near the rated load that the radial thread interference in the thinner wall member is overcome by the radial force causing differential diametrical strain, then stab flanks 50,51 and load flanks 52,53 may slant upwardly away from the axis per Figure 5 if tension loads are of greatest importance for a given application, or stab flanks 60,61 and load flanks 62,63 may slant downwardly away from the axis per Fig 6 if compressive loads are of greatest importance for another application. In either case, they should slant at an angle sufficient to resist radial mechanical forces that would otherwise cause differential strain, to thereby prevent separation of the mating threads.

Although the preferred value of included angle 35 is zero for general use, a small positive included angle may be desired in certain cases, such as to improve thread cutting tool geometry for use on machines not having the full capabilities of modern threading machines. In such cases, the included angle may be as much as 2 degrees with little fear of thread separation when API Mod 2 thread dope is used because it is generally considered to have a 1.2 degree angle of friction between pipe threads. Should the required makeup torque and the load rating of a connection be low enough such that both box and pin have sufficient excess

strength to resist thread separation, then the included angle may be formed greater than 2 degrees providing that engineering calculations and/or proof tests confirm that fact before such a connection is placed in service. API considers the coefficient of friction for API Mod 2 thread dope to be 0.021 and the arctangent of . 021 is 1.2 degrees, which is the accepted angle of friction between pipe threads.

Thus, if included angles are no greater than twice 1.2 degrees, then the torque applied to the connection will not tend to cause the threads to be forced radially out of their mating grooves. When the included angle is more than twice the angle of friction, the tendency of the wedging force to force the threads out of their mating grooves increases directly as the tangent (included angle/2-angle of friction). For instance, if an included angle of 10 degrees is used, the force tending to separate the threads = tangent (10/2-1.2) x the wedging force = 1/16 of the wedging force. In such a case then, box and pin wall thicknesses must necessarily provide a reserve restraining force greater than 1/16th of the wedging force so as to prevent separation and withstand all other operating loads, while not overstressing the tubular walls on w, hich the threads are formed. Therefore, for such connections, engineering calculations and/or proof tests that consider box and pin wall strength, thread dimensions, the assembly torque and service loadings, must confirm the design to be reliable before it's use.

Figure 8 depicts an embodiment of the present invention wherein pin 101 and box 102 are formed with substantially cylindrical threads, box root 131 of like diameter as the adjacent box root 132, and box load flank width 141 being the same as box stab flank 142. However, should it be desired to positively extrude all excess dope during makeup, then the box threads may be formed with a very slight taper having greatest diameter at the box face, or the pin threads may be formed with a very slight negative taper of greatest diameter at the pin face, either case having a total radial thread interference well within the desired amount of radial thread interference for the connection defined herein.

Another object of the present invention is to provide wedgethread connections on any size plain-end pipe, having engaged threads that may traverse selectively as much as the full thickness of the pipe wall so as to provide very high strength connections, with or without metal-to-metal seals. Figure 9 depicts pipe ends machined prior to plastic forming wherein inner surface 70 of the counterbore,

and taper 76 which approximates the desired box thread taper between surface 70 and the pipe bore, are machined within the end of pipe 71 so as to form shaped length 72 having a desired configuration thinner than pipe wall 74 after which, mandrel 80 as shown in Figure 10 may be pushed axially therein, so as to form larger outer diameter 81 and frustro-conical portion 82 wherein surface 70 has been formed to substantially a continuation of taper 76. Another end may be machined to form surface 75 adjacent the pipe end and taper 73 may be machined intermediate 75 and the pipe outer diameter to produce a desired configuration and length that is thinner than pipe wall 74 after which, tool 83 shown in Fig 10 is forced axially therearound so as to plastically reduce bore 84 to be smaller than original pipe bore 85, and to reform surface 75 to be substantially a continuation of taper 73 which approximates the desired thread taper. It should be understood that means other than mandrel 80 and tool 83 may be used to form the ends, such as pinch-rolls which may be advantageous on large diameter pipe ends. Such plastic forming may be done at room temperature or after being heated to above the lower critical temperature of the pipe material. Figure 11 depicts the ends after threading wherein box thread 90 may extend from diameter 93 substantially as large as the pipe outer diameter, and may terminate at a diameter as small as the original pipe bore as at 95. Pin thread 91 may extend from diameter 94 equal to original pipe bore 85 and terminate at a diameter equal to the pipe outer diameter as at 92. So as to allow threads near the pipe ends to carry their proportionate share of the loads, the ratio of the wall strength at the beginning of expansion 96 as compared to the full pipe wall strength, should be at least as great as, the ratio of the engaged thread length outboard 96 to the total engaged thread length. Likewise, the ratio of the wall strength at the beginning of bore reduction 99 as compared to the full pipe wall strength, should be at least as great as the ratio of the engaged thread length outboard 99, to the total engaged thread length. Therefore it is now clear that the length of thread engagement may traverse as much as the full pipe wall thickness and thereby provide selectively, a connection as strong as the pipe. The length and configuration of surfaces 70 and 75 may be dimensioned to provide for thread length only, or it may be dimensioned longer to provide for metal-to-metal seals as at 95 and 97 in Figure 7.

Preferred dimensions for my wedge threads not taught heretofore by background art, may be determined per the following example however, the scope of the present invention is not intended to be limited thereby.

D=Pipe design O. D. =The mean diameter within the O. D. tolerances for pipe ends.

PPF = Nominal pounds per foot of length for the pipe to be connected. d = Pipe design I. D. = (D^2-PPF/2. 67) ^. 5 t = Pipe design wall thickness = (D-d)/2 W= Radial width of pin load flank =< t/6 = < Starting axial length of thread crest BL = Desired box thread P. D. at largest diameter of engagement.

BS = Desired box thread P. D. at the smallest diameter of engagement.

T = Conical thread taper = diameter change/axial length = 0.1875 preferred L = Length of engaged threads= (BL-BS) /T, (except when conical threads are chosen) J = Load flank axial pitch-Stab flank axial pitch=. 0025+. 00036 x D= (ANSI RC5 Fit@ 1 turn) A = Mean axial pitch = W + (W^2 + J x L) ^. 5 LF = Load Flank axial pitch = A + J/2 SF = Stab Flank axial pitch = A-J/2 N = Number of Thread turns = L/A S = Radial width of pin stab flank = W + A x T/2, (+/-. 001" tolerance) B = Radial width of box load flank = W +. 003", (+/-. 002" tolerance), with API SA2 Mod dope.

C = Radial width of box stab flank = S +. 003", (+/-. 002" tolerance), with API SA2 Mod dope.

Y = Pipe material unit yield strength E = Pipe material modulas of elasticity M= Preferred thread interference on diameter = 2/3 x D x Y/E PS = Pin thread P. D. at small end of thread engagement = BS + M PL = Pin thread P. D. at large end of thread engagement = BL + M PB = Pin bore PE = Pin % efficiency = 100 x [PL + W x (A-W)/A] ^2-PB^2/(D^2-d^2) BE = Box % efficiency = 100 x [D^2- (BS-W x (A-W) IA) ^2] 1 (D^2-d^2) PR = Poisson's ratio for pipe material

R = Ratio of, Box wall thickness to Pin wall thickness BCS = Box axial compressive stress limit = M x E/ [PR x D x (R+1)] PTS = Pin axial tension stress limit = R x BCS TID = Thread interference at small end minus interference at large end = 0. 002" x (BL + BS)/2