The invention relates to pipe swivel joints of the toroidal type which allow the passage of fluid from a stationary conduit to a conduit which is intermittently or continuously rotatable about the stationary conduit. Particularly but not exclusively, it is concerned with toroidal swivel joint units which may be stacked and thus allow a plurality of flow paths for the simultaneous transfer of a range of different fluids from a plurality of stationary conduits to a plurality of rotating conduits. The toroidal swivel unit(s) with which the invention is primarily concerned is (are) for the passage of contaminated hostile hydrocarbon fluids (liquids/gas) at high pressure and high temperature from an offshore well (or wells) to a floating vessel which is allowed to "weathervane" about the said well, i.e. rotate so as always to face into the wind.

In the past few years the search for offshore hydrocarbon energy sources has progressed into deeper waters. Furthermore the shallow water fields which are being discovered are becoming smaller thus making the conventional platform approach expensive in relation to the field's generated revenue. Due to both of the above, the industry is searching for a means where by the fuels can be recovered by means of a floating vessel rather than

from a fixed structure. Under the influence of weather the floating vessel will weathervane and a swivel mechanism is required to stop the weathervane motion from twisting the pipelines which connect the vessel with the sea bed. Since a multiple of pipe lines are involved, the swivel unit or stack must be able to accommodate a plurality of passages providing simultaneous and separate pathways for each fluid, i.e. various fractions of oil, as well as water and gas.

It is recognised that multi-line fluid swivels of the toroid type have been developed for offshore use. However most previous designs have had severe limitations when required to handle high pressure contaminated fluid. The high pressures cause the structural members to distort allowing extrusion gap increase at the seals which results in leakage and seal damage. The contaminants present in the fluid cause mechanical breakdown of the seals and/or the seal mating surfaces which causes leakage. This has a costly impact on field economics as it generally means that the well has to be shut-in, causing revenue loss, while the particular swivel joint is repaired/replaced. Furthermore, with some previous designs where two or more toroids rely on the one slew bearing, failure of one toroid and/or bearing can cause all the lines to be shut-in while the single toroid and/or bearing is repaired/replaced. In addition if

damage is caused to the sealing surface and the seal means is located in the main engineering elements then the swivel may have to be sent back to the factory for refurbishment thus causing prolonged well shut-down. In designs where the bearing is located adjacent to the main dynamic seals, seal leakage allows hydrocarbon fluid to enter the bearing and may cause irreparable damage to the bearing.

Due to the above described limitations, swivel units for high pressure contaminated fluid use have not been readily accepted by the industry. This invention has the object of overcoming problems of previously designed swivel joints.

The invention provides a pipe swivel joint unit for the passage of fluid from one conduit to an in-line adjoining conduit, where there is relative rotating movement between the conduits, comprising an inner annular member and an outer annular member concentrically arranged therearound, said members having mutually confronting cylindrical surfaces, a fluid-carrying passage defined by a portion of each of said confronting surfaces, bearing means adapted to permit relative rotational movement between the annular members, seal means adapted to seal said confronting surfaces with respect to each other at regions spaced from said passage, said seal means

including dynamic seal devices wherein said confronting surfaces include complementarily contoured stepped annular regions so as to define a gap therebetween, each step of each annular region comprising an annular land in a radial plane of the joint unit, lands of adjacent steps being of differing diametric dimensions.

Thus the sealing members interposed between the faces may act as face seals acting on flat surfaces lying in a radial plane rather than curved cylindrical faces and are located in cavities, each of which is displaced both radially and axially from its neighbour and are arranged so that the seals are subjected to a mainly compressive hoop stress. Conveniently, the bearing means may be ball, roller or other bearing devices and may be conventional slew bearings. Advantageously, the inner and outer annular members, more usually referred to as toroid members, may be composed of a plurality of unit portions secured together in an integral manner. Preferably, there is provided for each member a first unit portion providing the confronting surface defining the fluid passage and two flanking seal portions providing the stepped regions.

Preferably, silt seals may be included in addition to the dynamic seals. Advantageously, the silt seals are located in said first unit portion and the dynamic seals in said flanking seal portions.

Advantageously, each system of seal members, silt

seals and dynamic seals are arranged to comprise primary and secondary sealing members, and means are provided to monitor leakage past the primary seal member, the secondary seal member being capable of containing the pressure thereupon.

In examples to be described below, primary and secondary silt seals are provided in said gap inwardly of the dynamic seals. Thus the swivel unit according to the invention comprises an inner toroid body member and outer toroid body member so arranged that the inner face of the outer toroid member and the outer face of the inner toroid member provide an interface cylindrical gap above and below the main toroidal flow path, where the interface gap is provided with one or more contamination seals above the main toroidal flow path and one or more contamination seals below the main toroidal flow with the contamination seals nearest the toroidal flow path (known as primary silt seals) arranged to withstand the toroidal flow path pressure and thus stop silt in the toroidal flow path entering ⁵ the interface gap and the pressure behind the primary silt seal and before the secondary silt seal is held at the same pressure as that existing in the toroidal flow path, while the pressure behind the secondary silt seal is greater or smaller than that existing in the main toroidal flow path, the medium in this region being clean fluid.

Conveniently, the control over the pressures existing before and behind the secondary seals is exercised by using a fluid circuit including a pressure intensifier. Where pressure is to be maintained at two differing levels, high and low in the one system, the effect of the intensifier may be off-set as required by the use of a pressure attenuator. In an example of the device to be described below, the attenuator may be a device similar in construction to that of the intensifier and connected in a reverse manner in the circuit.

The toroid swivel unit may further comprise an inner toroid body member and an outer toroid body member so arranged that the inner face of the outer toroid body member and the outer face of the inner toroid body member provide an interface cylindrical gap, above and below the main toroidal flow path, where the interface gap provided with one or more contamination seals and the primary silt seal (the one nearest the toroidal flow path) is arranged to withstand the toroidal flow path pressure and thus stop silt entering the interface gap from the toroidal flow path, pressure behind the primary silt seal and before the secondary silt seal being a pressure greater than that in the toroidal flow path and the medium in this region (i.e. between the primary and secondary silt seal) being clean fluid.

The pressure behind the secondary silt seal may be at a pressure which can be higher or lower than that existing between the primary and secondary seal depending on the seal lubrication system (intensifier system).

In yet another aspect of the invention, the joint unit comprises an inner annular member and an outer annular member concentrically arranged therearound, wherein said bearing means comprise an outer and an inner race portion, one of the race portions being fixedly secured to one of the concentrically arranged members on a flat face extendin in a radial plane thereof and the other of the race portions being fixedly secured to the other concentrically arranged member on a radial face contiguous with the first mentioned radial face, each concentrically arranged member having an annular shoulder formed in said radially extending surface thereof, the two shoulders being arranged to face each other to define an annular groove adapted to receive and support the bearing means in axial and radial directions.

It will be understood that where a first unit portion j of the annular members is provided with two flanking seal portions, the abovementioned radially extending surface is provided upon one of said flanking seal portions.

It is proposed to discuss the mechanics/dynamics of toroidal swivel units as an assistance to an understanding of the invention. Reference is made to the accompanying drawings, in which certain Figures illustrate examples of the prior art. These are Figures 1, 2, 3, 5 and 6.

The remaining Figures relate to example of a device according to the present invention. It will be understood that so far as the following description relates to devices embodying the present invention, said description is given by way of example only, and not by way of limitation.

In the drawings:

Figs la-c show a sectional view through a device according to the art and two scrap sectional views illustrating the effect of low pressure and high pressure thereupon?

Figs. 2a-b show a second device according to the prior art in cross-section and scrap sectional view respectively;

Figs. 3a-b show similar views to those of Figs. 2a-b in relation, to a third device according to the prior art;

Figs. 4a-b show similar views to those of Figs _* 2a-b

I in relation to a first example of a device according to the invention;

Figs. 5a-b show a fourth example of a device according to the prior art;

Figs. 6a-b show a fifth example of a device according

to the prior art ;

Figs. 7a-b show a second, modified, example of a device according to the present invention;

Fig. 8 is a scrap sectional view of a third example of a device according to the present invention;

Fig. 9 is a similar view to that of Fig. 8 of a fourth example of a device according to the present invention;

Fig. 10 is a perspective view of a toroid swivel unit according to the invention and as shown in Fig. 7;

Fig. 11 is a sectional view on line XI-XI of Fig. 10;

Figs. 12 and 13 are a plan view and a side view respectively, partly in section, of the device of Fig. 11;

Figs. 14 and 15 are scrap sectional view of a first and a second modification to the device of Fig. 11, showing means for attaching bearing members;

Figs. 16 and 17 are identical to Fig. 14 emphasising aspects of the sealing means;

Fig. 18 is a scrap sectional view of a first alternative arrangement of lubrication passages to that shown in Fig. 8;

Fig. 19 is a scrap sectional view of a second alternative arrangement of lubrication passages;

Figs. 20 and 21 show circuit diagrams corresponding respectively to the arrangements of Figs. 18 and 19; and

Fig. 22 is a side view partly in section of a multi-path swivel unit stack having also an in-line axial swivel.

A toroidal swivel unit (Fig. 1) generally comprises an outer toroidal member 1 and an inner toroidal member 2 with a cylindrical interface gap 14 between the toroids sealed by one or more sealing elements 17, 18. When the seal is subjected to pressure in the toroidal flow path 13 the outer toroidal member 1 increases in diameter and decreases in length (depth) , while the inner member 2 decreases in diameter and grows in length. The cylindrical interface gap 14 between the inner 2 and outer 1 toroidal members grows in width as the pressure is raised. The sealing elements 17, 18, if located in this way are much distressed - not only must they have the inherent strength to withstand the pressure induced stress, they must also be very resilient to take account of increased clearance at the cylindrical interface gap 14. This combination of material characteristics is very hard to obtain and if attained is very expensive. Short seal life can be induced by the fluid borne contaminants. High density contaminant falls down while low density contaminant migrates upwards. The contaminant on entering the cylindrical interface gap 14 can be trapped by the seals 17, 18, and causes mechanical f ilure of the seals and the sealing surfaces . In an attempt to overcome the effect most previous designs employed a primary 17 and secondary 18 seal as shown. It should also be noted that seal damage necessitates the toroidal swivel being dismantled, (an onerous task

offshore) to allow seal replacement. If the damage is to the seal surfaces then the toroid must be sent back to the factory for refurbishment.

To overcome the difficulties discussed above some designs (Fig. 2) have included seal rings 3, 4, attached, say, by fasteners 9, 10, above and below the main toroid members 1, 2, to accommodate a primary dynamic seal 17 and a secondary dynamic seal 18. This arrangement overcomes the replacement problems associated with seal surface damage as it is relatively inexpensive to hold spare seal rings 3, 4, on board. Furthermore it reduces the "maintenance/replacement time-window" problem as the main engineering elements do not have to be stripped down when seal replacement is required. However this radially spaced seal configuration induces a cost and weight penalty in that the bending moment carried to the seal rings is high due to the large axial pressure region 82 being somewhat offset from the attachment means 9, 10. This requires the seal ring 3, 4, to have a deep section.

Another arrangement developed to accommodate the main dynamic seals away from the contaminant zone (Fig. 3) is to have axially displaced seal rings la, lb, 2a, 2b, attached, say by fasteners 10, 11, 12, above and below the main toroid members 1, 2, to include a primary dynamic seal 17 and a secondary dynamic seal 18 axially displaced

from the toroidal flow path 13. Due to the close proximity of the small axial pressure region 82 to the seal ring fasteners 11, 12, the bending moment is much ^• reduced, leading to thin seal rings. The limitation of this design is in the increase in the number of seal rings and static O-ring, (twice as many as that required for the radial arrangement) and the inordinately long fastener (bolt) length. In practice the axial arrangement is not significantly lighter than the radial arrangement, is more complex, more expensive and is prone to leakage by way of the static O-ring.

Fig. 4 illustrates the first of several examples of devices according to the invention. It will be understood that the description given of this and other examples is given by way of illustration only and not by way of limitation.

The device described in Fig. 4 overcomes the problems of the prior art by providing silt seals 15, 16, which are located in the cylindrical interface gap 14 and are not required to withstand full pressure. The primary silt seal 15 may have full pressure on both its free sides. The secondary silt seal 16 is only subjected to a small pressure difference caused by the seal lubrication (intensifier) system which is described below. Thus the silt seals 15, 16, are not distressed to the same extent as main dynamic seals located in this region (refer to Fig. 1). The main dynamic seals 17, 18, are in a combined

radial-axial, or stepped arrangement produced by the seal ring 3,4, geometry. This stepped seal arrangement over¬ comes the high degree of bending of the radial arrangement, reference Fig. 2, as the axial pressure region 82 is not as large and is in fairly close proximity to the attachment fasteners 11, 12. The stepped seal arrangement is superio to the axial arrangement, reference Fig. 3, as it requires fewer component parts, fewer O-rings, shorter fasteners and is less complex to manufacture. The steps include annular lands L each lying in a radial plane, lands of adjacent steps being of different diametric dimensions.

Another feature of the present example (Fig. 4) is its improved sealing efficiency due to the geometry and the dilation of the main toroid members 1, 2, discussed above. As the outer toroid member 1 is subjected to pressure it reduces in axial length which reduces the axial space 20, 24, which houses the seals. When the seals are subjected to pressure the outer seal ring 3 deflects away from the seals 17, 18, thus tending to increase the axial height of the space 20, 21, which houses the seals. The geometry permits these two effects to be cancelled out allowing the axial space 20, 21, to remain substantially at designed dimension. As the inner toroid 2 is subjected to pressure its axial length increases which causes the axial space 20, 21, which houses the seals to become smaller. These three effects work together to cause the axial space 20, 21, to reduce

in height with increasing pressure - i.e. increasing pressure increases seal squeeze, thus increasing sealing efficiency. Another element of some significance with respect to seal mechanics is that the pressurised fluid acts on the outside diameter of the seal. This causes the seal to operate in a co pressive stress field which tends to enhance seal life compared with arrangements which allow the seal to operate in a tensile stress field. The above discussions have been concerned with the "pressure vessel" aspects of toroidal swivel units. However for toroidal swivel units to function they must be capable of having relative rotation between the inner and outer elements.

Previous designs of toroidal swivels (Fig. 5) included the bearing arrangement 5 integrally with the main members. This was achieved by machining a semi-circular groove 7 in the external diameter of the inner toroid 2 to match with a machined semi-circular groove 6 in the inner diameter of the outer toroid 1. These semi-circular grooves 6, 7, constitute a track into which balls 8 are located, thus forming a bearing assembly. This arrangement is limited in that the accuracy and concentricity of the track is difficult to obtain. Also any leakage from the fluid path 13 passed the seals 17, 18, entered the bearing, causing early failure. Another difficulty in this arrangement is that when the balls 8

are removed the inner toroid 2 can move axially with respect to the outer toroid 1 which makes the total assembly unsafe from a pressure vessel aspect.

To overcome the difficulties associated with integral bearing arrangements more recent designs (Fig. 6) have concentrated on attaching standard bearings of the slew type to the main body members. In this approach a slew bearing 5 comprising an outer race 6, an inner race 7 and rolling elements 8 is utilised. The rolling elements may be of ball, cylindrical or conical type common to the bearing industry. The advantage in using a standard bearing is that the accuracy is significantly improved as is load capacity, while the cost is generally reduced compared to the integral bearing arrangement. In this approach the outer race 6 is attached to the outer toroid 1, the inner race 7 is attached to the inner toroid 2 and the rolling elements 8 are interspaced between the inner race 7 and the outer race 6, the interspace geometry generally being of a configuration typical of slew bearing applications. This arrangement simplifies bearing replacement as the main toroid members do not need to be disassembled. It has the same limitation as the integral bearing with respect to pressure vessel considerations when the bearing is removed. It should also be realised that unless the bearing 5 has a radial spigot/shoulder type location with respect to the inner and outer members there

is no guarantee that the relative rotation of the inner and outer elements will be concentric. Thus much of the inherent bearing accuracy may be lost in relation to the total assembly.

The bearing arrangement of the present example (Fig. 7) therefore utilises a standard slew bearing 5 with an inner race 7 and an outer race 6 with rolling elements 8 of the cylindrical or ball (or combination of both) types interspaced between the inner race 7 and the outer race 6. The interspace geometry by way of the rolling elements 8 is arranged to provide positive axial support and location and also radial support and location typical of the slew bearing technology common to the bearing industry. The outer race 6 is radially located and supported by a locating shoulder 45 in the outer seal plate 3 and axially located and supported on the outer seal plate 3 by means of fasteners 9 such as cap screws. The inner race 7 is radially located and supported by a locating shoulder 47 in the inner seal plate 4 and axially located and supported on the inner seal plate 4 by means of fasteners 10 such as cap screws. The outer seal plate 3 is radially located and supported by a locating shoulder 84 in the outer toroid member 1 and axially located and supported on the outer toroid member 1 by means of fasteners 9 such as cap screws. The inner seal plate 4

is radially located and supported by a locating shoulder 85 in the inner toroid member 2 and axially located and supported on the inner toroid member 2 by means of fasteners 10 such as cap screws. Thus the bearing arrangement is such that it supports and locates, in an axial and radial direction, the inner and outer elements of the swivel unit and allows relative rotation between them which occurs in a truly concentric manner. Hence the inherent accuracy of the slew bearing is conferred onto the total assembly. It should also be noted that when the bearing is removed the swivel unit remains captive due to the seal plate geometry and thus the assembly is safe from a pressure vessel point of view.

In an alternative arrangement, the fasteners attaching the inner and outer race of the bearing may be configured so as to act as a single functional element, that is, they are the only means of attaching the bearing assembly to the pressure vessel elements of the toroid. In this instance the fasteners pass through the slew bearing races and enter threaded holes in the seal ring whereby the foundation attachment between the seal rings and the bearings races is attained. However, the fasteners attaching the inner and outer race of the bearing may be configured to act as a double functional element, that is they not only attach the bearing assembly to the seal rings they also act (in combination with other

fasteners) to attach the seal rings to the main toroid members. In this instance the fasteners pass through the slew bearing races and also through holes in the seal ring and then enter threaded holes in the main toroid member. By this means the foundation attachment between the seal rings and the bearing races is attained and also the pressure resistant attachment between the seal rings and the toroid members is also attained.

The above arguments have been concerned with the pressure vessel aspects and the rotating mechanics of the swivel unit. However it is worthwhile to evaluate another aspect of the swivel unit in the area concerned with contamination and hostile fluids.

In early designs no attempt was made to protect the seals and other sensitive areas from hostile fluids and/or contamination except by overlaying the sealing surfaces with a noble or corrosion-resistant metal, such as stainless steel, to reduce the effects of corrosion by the hostile fluid. The next step in swivel unit evolution was to provide a back-up seal, i.e.a design configured with a primary and a secondary seal. The most recent previous designs attempted to enhance the life of the swivel unit by introducing a barrier fluid between the hostile contaminated fluid in the flow path and the main

dynamic seals, or by allowing the barrier fluid to be located between the primary and secondary dynamic seal. Generally this barrier fluid was at the same pressure, or in some instances, at a slightly higher pressure than that existing in the main fluid.

The present example (Fig. 8) provides double protection against contamination by means of a primary 15 and secondary 16 silt seal and double barrier fluid means by use of a primary barrier fluid cavity 86 and a secondary barrier fluid cavity 87 for acceptance of the pressure-intensified lubricating fluid. The double silt seals have been described above. A pressure tapping is led from the main toroidal flow path 13 and piped to an intensifier (not shown in Fig. 8). The intensifier amplifies the pressure from the flow path fluid to the lubricating fluid and pipes the lubricating fluid to the swivel unit. The lubricating fluid on reaching the swivel is diverted into two circuits. One circuit (primary barrier circuit) leads the lubricating fluid, via a stop off valve, to the cylindrical interface gap 86 formed by the secondary silt seal 16 and the barrier seal 19. The second circuit (secondary barrier circuit) leads the lubricating fluid, via a second stop off valve, to the cylindrical interface gap 87 formed by the barrier seal 19 and the primary dynamic seal 17. In this arrangement the primary silt seal 15 is in pressure balance; the

secondary silt seal 16 is subjected to a low pressure differential (i.e. the difference between toroidal flow path pressure and intensified lubricant pressure); the barrier seal 19 is operating in pressure balance and the primary dynamic seal 17 is subjected to full intensified lubricant pressure. If the silt seals 15, 16, become damaged or worn to such a degree that the output from the intensifier can not keep up with the flushing rate of lubricant into the main toroidal path 13 across the silt seals 15, 16, then the primary barrier circuit can be shut down by means of closing the first stop-off-valve which leaves the secondary barrier circuit in operation. In this arrangement the space between the primary and secondary silt seal is allowed to remain comparable to that which exists in the main toroidal flow path.

In another arrangement of the present example the fluid pressure in the region between the primary and secondary silt seal is fed from the intensifier and is at a pressure which is greater than that existing in the toroidal flow path. The pressure in the region between the secondary seal and barrier seal and the pressure in the region between the barrier seal and the primary dynamic seal is the same as that in the region between the primary and secondary seal. As the seals become damaged, the flow to the various regions can be closed off sequentially thus minimising the loss of lubricating fluid,

Furthermore the present arrangement can be constructed so that the region between primary and secondary silt seal is at high pressure while the pressure between the secondary silt seal and barrier seal and that between the barrier seal and the primary dynamic seal can be at low pressure by the use of a second intensifier cylinder connected to the toroidal flow path in an inverse manner. This can increase the life of the barrier seal and the primary dynamic seal as they operate, in this instance, at low pressure.

Another aspect of a design where the seals are radially displaced is that, if one of the primary seal leaks, the axial pressure above and below the toroidal flow path is different. This unbalanced axial pressure can induce heavy axial loads in the slew bearing which can lead to early bearing failure. This phenomenon is taken care of by means of interconnecting passages between the upper and lower seal rings (reference Fig. 9). These passageways allow the pressure to equalise. In the example, a cross drilling 100 in the outer seal ring connects the region between the primary dynamic seal 17 and the secondary dynamic seal 18 with a vertical passageway 101 in the outer toroid member 1. Thus leakage past the upper primary dynamic seal 17 is prevented from escaping to the environment by the upper secondary dynamic seal 18 and

flows along the upper ^* cross drilling 100 in the upper outer seal ring 3 until it drops down and through the vertical passageway 101 in the outer toroid member 1 when it connects with the lower cross drilling 100 in the lower outer seal ring 3 and flows on into the region between the lower primary dynamic seal 17 and lower secondary dynamic seal 18. Thus the axial pressure above and below the toroidal flow path 13 is in balance and the pressure is contained by the secondary dynamic seals 18 allowing the swivel to remain safe (from a pressure vessel standpoint) and operative as a mechanism.

The leakage past the primary seal can be monitored. This is achieved by providing a horizontal gallery 102 in the outer toroid body 1 so that it intercepts the abovementioned vertical passageway 101. The horizontal gallery 102 exits the outer toroid body 1 at a stop-off valve (not shown) . Opening of the stop-off valve .allows any pressure and/or fluid in the horizontal gallery 102 to be monitored.

The details of the various operating features of the preferred example will now be given with respect to Figs. 10-16.

Fig. 10 is a perspective view of the toroid swivel unit and Fig. 11 illustrates the toroidal flow path 13 formed by the internal diameter of the outer toroid member 1 and the external diameter of the inner toroid member

2. The outer seal ring 3 is attached to the outer toroid member 1 above and below the toroidal flow path 13. The inner seal ring 4 is attached to the inner toroid member 2 above and below the toroidal flow path. The interface gap formed by the inner stepped region of the outer seal ring 3 and the outer stepped region of the inner seal ring locates the primary dynamic seal 17, the secondary dynamic seal 18 and the environmental contamination excluder seal 22. The slew bearing 5 comprises an outer race 6 and an inner race 7 and a plurality of rolling elements 8 which run in the interface gap formed by the outer race 6 and the inner race 7. The toroid elements 1, 3, being attached to the outer bearing race 6 can rotate with respect to the toroid elements 2,4, which are attached to the bearing's inner race 7. The silt or contamination seals are located in the interface gap 14 above and below the toroidal flow path 13 formed by the inner surface of the outer toroid member 1 and the outer surface of the inner toroid member 2. Spaced from the toroidal flow path 13 these are the primary silt seal 15, secondary silt seal 16 and, if required, a barrier seal 19.

Figs. 12 and 13 show a single swivel unit in horizontal and vertical part section respectively and shows the outer seal ring 3 is attached to the outer toroid member 1 by means of a plurality of fasteners 9, the inner seal ring 4 is attached to the inner toroid member 2 by means of fasteners 10. An inlet port 31 on

the inner toroid member 2 is connected with the toroidal flow path 13 at an oblique angle by flow path 32. An outlet port 29 on the outer toroid member 1 is connected with the toroidal flow path 13 in a tangential manner by means of a straight flow path portion 30. An inspection port 33 is connected with the toroidal flow path 13 in a tangential manner by means of a straight flow path portion 34 and an inspection port 33 is blanked off by means of blind flange 35 which is plugged 36 and sealed 37. The plug 36 is so dimensioned that when removed it allows inspection instruments (e.g. an endoscope) to be inserted into the toroidal flow path. In compliance with general practice, the inlet port 31 and the outlet port 29 are of the same dimensions and are machined to provide standard flanges without any structural or engineering welding in the main body members 1, 2. The inspection port 33 may be machined identically with the other two ports 29, 31. The inspection port 33 and the outlet port 29 are supplied with vent plugs 41 and drain plugs 42. The inlet port 31 may also be provided with a vent plug 41. The bearing outer race 6 is shown attached by fasteners 11 and the inner race 7 is attached by fasteners 12.

Figs. 14 and 15 illustrate two of the preferred methods of attaching the slew bearing 5 to ensure axial and radial location and support while providing concentric rotation. The outer seal ring 3 is attached to the outer toroid member 1 by cap screw fasteners 9, and located on

the outer toroid member 1 by the deep shoulder location shown, thus ensuring that the two engineering elements are concentric. The inner seal ring 4 is attached to the inner toroid member 2 by fasteners 10, such as cap screws shown, and located on the inner toroid members 2 by the deep shoulder location shown, thus ensuring that the two engineering elements are concentric. It is now necessary to provide a system whereby the bearing 5 can be mounted and fixed so that it also provides axial and radial location and support while providing concentric rotation. This is achieved by machining a shoulder location for the bearing 5 on the inner and outer seal rings 4,3, to provide positive location to compensate for the fact that bolt holes in standard slew bearings are provided with clearance. The bearing outer race 6 is attached to the outer engineering elements 1,3, by means of cap screw fasteners 11 and located on the outer seal ring 3 by shoulder location shown, thus ensuring concentricity. The bearing inner race 7 is attached to the inner engineering elements 2,4, by means of fasteners 12 and located on the inner seal ring 4 by shoulder location shown, thus ensuring concentricity. Thus by means of employing a machined interface between all elements the inherent built-in accuracy of the bearing is conferred onto the elements ensuring that the swivel, when considered as a mechanism, operates in a true concentric manner. In Fig. 14 the bearing is shown as located,

clamped and fixed to the seal rings and the fasteners 11,12, act purely as a structural attachment. Fasteners 11 are interspaced between fasteners 9 and fasteners 12 are interspaced between fasteners 10. In an alternative arrangement shown in Fig. 15, the bearing is located and clamped to the seal rings but these are fixed to the main toroid members by the fasteners 11, 12, which pass through the seal rings. In this arrangement the fasteners act as a structural attachment (in holding the bearing in place) and also as a pressure vessel attachment in that they hold the seal rings to the main toroid members. Fasteners 11 replace the inner circle of fasteners 9 and fasteners 12 replace the outer circle of fasteners 10.

It should also be noted that the slew bearing 5 can be located above the toroidal flow path 13 as the upper and lower seal rings 3,4, are the same and are thus interchangeable. Also in extreme conditions two slew bearings 5 can be located one above and one below the main toroidal flow path.

Figs. 16 and 17 are identical with Fig. 14 and show a scrap section through the swivel unit illustrating the sealing arrangements. There is provided an interface toroidal gap 14 above and below the toroidal flow path 13 formed by the diametral dimensional clearance between the outer toroid member 1 and the inner toroid member 2. The primary silt seal 15, the secondary silt seal 16 and, if required, the barrier seal 19 are located in the interface

toroid gap 14 above and below the toroid flow path 13. The internal surface of the outer seal ring 3 has a stepped geometry such that the flat surfaces generated in a radial plane face towards the toroidal flow path 13. The external surface of the inner seal ring 4 has a stepped geometry such that the flat surfaces generated in a radial plane face away from the toroidal flow path 13. The stepped interface gap between the outer 3 and inner 4 seal rings is so arranged that at least two cavities 20,21, are formed. The cavities 20,21, have their surfaces remote from the flow path 13 formed by the axial face and cylindrical face of the stepped geometry of the outer seal ring 3 and have their opposite surfaces formed by the axial face and cylindrical faces of the stepped geometry of the inner seal ring 4. This method of forming a seal cavity or housing is advantageous as it avoids the need for machining re-entrant shapes which gives difficulty in achieving high accuracy and surface finish on large components especially when the re-entrant shape incorporates a hard metallic weld overlay - say such as stellite. The primary dynamic seal 17 is housed in the cavity 20 nearest to the toroidal flow path 13, the secondary dynamic seal 18 is housed in cavity 21. The interface stepped geometry of the seal ring 3,4, is so arranged along with dynamic seal 17,18, dimensions that the contact sealing surfaces are the flat faces rather than the cylindrical faces and the seals are supported by

the inner seal ring. . Thus the seals operate in a compressive hoop type stress field which tends to prolong seal life, the extrusion gap is at right angles to the face which is subjected to pressure thus minimising seal "tear out damage". Also the flat faces of the cavity tend to move towares each other as the pressure increases thus increasing the squeeze on the seal which increases sealing efficiency when compared to designs where the seal contacts the cylindrical surfaces which move away from each other when subjected to pressure.

The static interface between the seal rings 3,4, and the main toroid members 1,2, are sealed by static O-rings 25,26,27,28. Seals 22,25 and 28 stop environmental contamination entering the swivel unit. Static seals 26, 27, stop high pressure fluid in the swivel unit from escaping along a static interface to the environment.

Figs. 18 and 19 each show a scrap vertical section through the toroid swivel unit and illustrate alternative arrangements of cross drillings through the seal ring 3 and the outer toroid member 1. It should be noted that the arrangement on the weathervaning vessel dictates whether it is the outer element 1,3, or the inner elements 2,4, which have the cross drillings*, thus the description below may be modified for internal members having cross drillings although it only refers to external members. Fig. 18 illustrates the cross drillings which are primarily concerned with seal lubrication and the

prohibiting of silt and contaminant in the main toroidal flow path 13 reaching the dynamic seals 17,18. Fig. 19 illustrates the cross drillings which are primarily concerned with pressure equalisation and leakage monitoring.

In Fig. 18 the cross drilling 84 intercepts the main toroidal flow path 13 on a horizontal diameter and provides communication with the external surface of the outer toroid member 1, thus allowing fluid from the toroidal flow path 13 to leave the unit. Above and below the main toroidal flow path 13 cross drillings 85 continue from the interface gap 14, between silt seals 15, 16, to the external surface of the outer toroid member 1. Above and below the main toroidal flow path 13 a cross drilling 86 provides communication from the interface gap 14, between seals 16, 19, to the external surface of the outer toroid member 1. Above and below the toroidal flow path 13 the outer seal ring 3 is provided with a stepped cross drilling 87 which provides communications from the inner diameter of the outer seal ring 3 past the primary dynamic seal 17 (such that it intercepts the toroidal space between seals 17,19) to the static interface between the outer seal ring 3 and the outer toroid member 1 where it connects with stepped cross drilling 90 which exits at the external surface of the outer toroid member 1. Cross drillings 84, 85, 86, 90, are arranged to meet up with cross drillings in a valve manifold block which is

attached to the outer toroid member 1 by means such as fasteners, not shown.

In Fig. 19 the outer seal ring 3 above and below the main toroidal flow path 13 is provided with stepped cross drilling 100 which commences at the inner surface of the outer seal ring 3, past seal 18 (such that it intercepts the toroidal space between seals 17,18) and continues radially outwards until it changes direction and emerges from the outer seal ring 3 by way of the static interface at the outer toroid member 1 when it communicates with vertical through drilling 101 in the outer toroid member 1. A horizontal cross drilling 102 communicates with the vertical cross drilling 101 and exits the outer toroid member 1 on its outside surface. Leakage past the primary dynamic seal 17 migrates along cross drilling 100 to the opposite cross drilling 100 via vertical cross drilling 101 and thus equalises pressure above and below the toroidal flow path 13 and stops axial load build-up on the bearing. Leakage past primary seal 17 can be monitored via cross-drilling 102 by stop-off valve (not shown) at the end of cross-drilling 102.

Figs. 20 and 21 show two alternative means whereby pressurised lubricant may be fed to the seals. Fluid from the toroidal flow path 13 leaves the toroid via cross-drilling 84 and passes on through the valve manifold to exit the manifold at stop-off valve 150.

In Fig. 20 it passes to lubrication unit 128 where it

enters the intensifier 125. The intensifier amplifies the pressure in the lubricant which exits the bottom of the intensifier and enters the valve manifold block at valve 151. The intensified lubricant can then move to the space between the primary 15 and secondary 16 silt seals via cross-drilling 85 if valves 157, 158 and 159 are open. Similarly intensified fluid can move along cross-drilling 86 to the space between the secondary silt seal 16 and seal 19 if valves 153 and 156 are open. Likewise intensified fluid can move along cross-drilling 90 to the space between seal 19 and 17 if valves 153, 154 and 155 are open. Thus if the silt or contaminant in the main toroidal flow ^' path 13 causes wear to the primary silt seals 15 and the intensified lubricant leaks at too fast a rate into the toroidal flow, the lubricant flow to the space between seals 15 and 16 can be shut off (by closing valves 157, 158 and 159) and diverted to the space between seals 16 and 19 (by opening valves 153 and 156); thus prolonging the operating life of the swivel unit. This process can be repeated when seal 16 becomes worn by diverting lubricant flow to the space between seals 19 and 17.

In Fig. 21 the lubrication unit 127 includes two intensifier units. The intensifier 125 amplifies the pressure and intensifier 126 is arranged to act to attenuate the pressure. Thus lubricant at high pressure (from intensifier 125) and at low pressure (from

attenuator 126), relative to the pressure in the toroidal flow path 13, can be fed to the swivel unit. High pressure lubricant enters the valve manifold via valve 151 and low pressure lubricant enters the valve manifold via valve 155. High pressure fluid reaches the spaces between the seals by setting the valves in open and closed condition as outlined above. Low pressure lubricant can reach the space between seal 19 and 17 via cross-drilling 90 by having valves 152,155, open. Low pressure lubricant can reach the space between seals 19 and 16 via cross drilling 86 by having valves 152, 154 and 156 open. Thus lubricator 127 allows lubricant at high and/or at low pressure with reference to fluid path pressure to be fed to the swivel unit and also, by means of the valve manifold, for the different pressure levels to be fed to different areas within the swivel.

Fig. 22 shows a multipath swivel unit stack comprising three toroid swivel units 200, 201, 202, and an inline axial swivel unit 203 mounted and contained within a foundation frame 204 which houses a torque drive mechanism 205. Although the swivel units are illustrated as being of the same diameter it will be understood that they may be of different sizes, if desired. The lowest toroid swivel unit 200 is supported on and located by the foundation frame 204. The middle toroid swivel 201 is supported on the lowest toroid swivel 200 by an interposed sandwich ring 206. The topmost toroid swivel 202 is

supported on the middle toroid swivel 201 by a further, interposed, sandwich ring 207. The inline axial swivel 203 is supported on the topmost toroid swivel unit 202 by a sandwich ring 208. Each sandwich ring 206, 207, 208, is provided with several apertures or openings which allow access to the flanges F in the pipes to the inlet ports to the toroid swivels and in the axial swivel. Sandwich rings 206, 207, 208, are manufactured such that they have the necessary rigidity and strength in all directions (axial, radial and tangential) to give a stiff foundation path. Torque or slew transmission between the lowest toroid swivel 200 and the middle toroid swivel 201 is by means of inner sandwich ring 209. Torque or slew transmission between the middle toroid swivel 201 and the topmost toroid swivel 202 is by means of inner sandwich ring 210. Torque or slew transmission between the topmost toroid swivel 202 and the inline axial swivel 203 is by means of inner sandwich ring 211. Each sandwich ring 209, 210, 211, is provided with a plurality of apertures or openings which allows access to the flanges in the pipes leading to the inlet ports in the toroidal swivel units and in the axial swivel unit. The sandwich rings 209, 210, 211, are manufactured such that they are compliant in the axial and radial direction but are rigid to resist twisting (torque) in a tangential direction, thus ensuring that proper slew or azimuth alignment is

maintained between all swivel units. The compliant properties of sandwich rings 209, 210, 211, in the radial and axial direction accommodates the structural dilations of the swivel units when they are under pressure.

A number of safety features appropriate to use in a hazardous area are conveniently incorporated in devices according to the invention. For example, all parts of the operating system are actuated pneumatically, for example, all switches are pneumatically operated microswitches. Moreover, damage to, and the resulting mal-function of, valves controlling the lubrication fluid supply are contained within a manifold connection rather than being mounted in relatively exposed positions upon pipes and conduits.