WO2010094936A1

WO2010094936A1 - Compression method and means

Info

Publication number: WO2010094936A1
Application number: PCT/GB2010/000309
Authority: WO
Inventors: John Philip Roger Hammerbeck
Original assignee: John Philip Roger Hammerbeck
Priority date: 2009-02-20
Filing date: 2010-02-22
Publication date: 2010-08-26
Also published as: US20120070326A1; GB0902792D0

Abstract

The invention is a compression method having characteristics of smooth compression and internal cooling of the gas. Embodiments of the invention employ a cylindrical chamber and an orbiting rotor to create a moving duct or chamber whose walls converge, relative to a static gas packet drawn into the moving duct, at a "pinch point". Preferably the closing speed of the walls is subsonic and the speed of the pinch point is supersonic. This enables high pressure to co-exist, at the narrowing end of the duct, with low pressure elsewhere in the duct, because of the pressure information barrier produced by the supersonic advance of the pinch point. The invention also discloses means for providing gas inlet and outlet functionality, and means for providing a variable flow compressor.

Description

COMPRESSION METHOD AND MEANS

This application relates to the field of gas pumping and compression.

Background of the invention

Gas compression devices used in refrigeration, air conditioning and industry consume a large portion of electrical power generated. An increase in gas pumping efficiency will result in reduction of carbon dioxide emissions. Proposals to sequester carbon dioxide at pressure underground or in the ocean depths are dependent on using compression methods that are efficient and can also overcome problems such as phase change and the material erosion of compressor parts when compressing impure gas mixture. Small changes in compressor efficiency may determine commercial viability.

In an existing compressor design according to international publication WO 2008/122781 there is provided a compressor comprising a cylinder and a rotor, whereby the rotor traverses the internal circumference of the cylinder and a pinch point is formed at the closest point of the rotor periphery to the internal wall of the cylinder. The rotor traverses the internal circumference of the cylinder such that the pinch point moves at high, preferably supersonic speed. In an embodiment, the rotor rolls around the internal circumference of the cylinder such that the speed of the rotor surface, relative to the cylinder wall, is low or zero, thus reducing wear and frictional heating of the components and of the gas to be compressed, termed herein "rolling", thus aiding compressor efficiency. Optionally, a strip valve arrangement on the rotor surface allows entry of gas into the chamber formed between rotor and cylinder. Optionally, a strip valve arrangement on the cylinder wall allows exit of gas from the chamber and optionally incorporates actuation means to control its opening position.

In another existing compressor design there is provided a compressor comprising a cylinder and a rotor, whereby the rotor traverses the internal circumference of the cylinder and a pinch point is formed at the closest point of the rotor periphery to the internal wall of the cylinder. The rotor moves such that the pinch point moves at high, preferably supersonic speed. The rotor rotates around the internal circumference of the cylinder such that a fixed point on the rotor periphery is maintained adjacent to the pinch point - termed herein "rotating". Optionally, ports in the rotor allow entry and exit of gas via passages communicating with the axial ends of the cylinder.

Known types of compressors typically suffer from problems which tend to reduce efficiency, including but not limited to those described herein, namely:

• low inter-stage compression rise

• large physical size relative to gas processing rate

• Limited ability to offer variable flow rate • Complicated inlet/outlet porting and/or valving

• Losses associated with gas leaks

• Lowered efficiency related to irregular gas flow

• Difficulties in achieving balance at high rotational speeds

In particular, efficiency losses can result when a variable flow compressor is desired. This is at least partly because typical existing variable flow compressors vary their flow by changing the operating speed of the compressor. However, since such compressors are usually optimised for a particular operating speed, changing the operating speed can result in reduced efficiency.

Variable speed compressors also require expensive control electronics to provide variable speed drive.

Some existing compressors also suffer from stalling if the inlet charge density undergoes a step, or rapid, change.

Other existing compressors achieve variable flow by recirculating or exhausting a portion of the compressed gas, or by throttling the input of the compressor, however this reduces efficiency.

Many of the above problems also apply to expanders and companders.

Brief description of invention

The invention is set out in the claims.

In a first embodiment of the invention there is provided a compressor as defined in claim 1 of the appended claims.

Embodiments of the invention will now be described, by way of example, with reference to the figures which are as follows:

Fig. 1 - Schematic view of compressor housing and rotor Fig. 2a to 2g - 'Rolling' rotor operation

Fig. 3a to 3g - 'Orbiting' rotor operation

Fig. 4a to 4g - 'Rotating' rotor operation

Fig. 5 - Strip valve arrangement Fig. 6 - Rotor port arrangement

Fig. 7 - Rotor surface features

Fig. 8 - Spiral duct embodiments

Fig. 9 - Rotor balancing/drive arrangement

Fig. 10 - Strip valve and balancing arrangement Fig. 1 1 - Strip valve actuation arrangement

Fig. 12a - Dual Lobed Rotor

Fig. 12b - Dual Lobed Rotor, alternative embodiment

Fig. 13 - Disc valve arrangement

Fig. 14a - Flow control flanges, perspective view Fig.14b - Flow control flanges, axial end view

Fig. 14c - Angled rotor surface

Fig. 15 - Pressure balancing flange

Fig. 16a - Sliding flange

Fig. 16b - Sliding flange, alternative embodiment Fig. 17 - Input scoop

Fig 18a - Sliding flange, alternative embodiment

Fig. 17 - Input scoop

Fig. 18a - Open channel pump with variable capacity

Fig 18b - Closed channel pump with variable capacity Fig. 19 - Outlet redirection feature

Fig. 20 - Movable compressor floor, perspective view

Fig. 21 - Movable compressor floor, cross sectional view

Fig. 22 - Movable compressor floor, cam actuation Detailed description

Fig. 1 shows an existing compressor that has the desired characteristics of smooth compression and internal cooling of the gas. This compressor employs a cylindrical chamber (10) and rotor or orbiter (20) to create a moving duct or chamber (40) of unchanging geometry and size, whose walls converge relative to a static gas packet drawn into the moving duct (40). In use, (40) walls converge at a lower speed than the point of closest approach of the walls [hereinafter called the pinch point (50)] moves along the duct (40). In preferred operation the closing speed of the walls is subsonic and the speed of the pinch point (50) is supersonic. As the pinch point (50) advances, the volume in which gas is at highest pressure/temperature also advances to areas of the walls that have been cooled since last being adjacent to the high temperature gas. When such a compressor is operating with the pinch point (50) moving at supersonic speeds, information about the pressure rise caused by narrowing of the duct (40) cannot propagate forward and push the gas forward. This enables high pressure to co-exist, at the narrowing end of the duct (40), with low pressure elsewhere in the duct (40) because the volumes are physically separated by the pinch point (50) and the pressure information barrier (40) produced by the supersonic advance of the pinch point (50). This provides a compressor that has the high pressure ratio capability of positive displacement compressors combined with the smooth pulse-less outflow of centrifugal and axial machines.

Various embodiments employ a duct (40) created between an inner circumference of a cylinder (10) and a shaped wall (20) moving within the cylinder (10) so as to form a narrowing of the duct (40) at the point of closest approach of the two members (50). Three embodiments demonstrating variations on the movement of the rotor (20) within the cylinder (10) will now be described.

As shown in Figs. 2a to 2g, as can be used in a first class of embodiments described below, a 'rolling' rotor (20) rolls around the inner circumference of the cylinder (10) as the rotor (20) traverses the inner circumference of the cylinder (10). The orientation of the rotor (20) is shown by respective arrows A, B, C in Fig. 2a. The sequence of six illustrations shown consecutively in Figs 2b to 2g illustrates (see arrow A in each) how the orientation of the rotor (20) changes with respect to the cylinder (10) as the rotor (20) rolls around the inner circumference of the cylinder (10). The rotor changes orientation as it rolls such that the speed of the rotor (20) surface, relative to the surface of the inner circumference of the cylinder (10) is substantially low or zero. The rotor (20) can be arranged to substantially contact the inner surface of the cylinder (10) or the two surfaces can be spaced slightly apart. The rotor (20) can be arranged to roll by means of contacting the inner surface of the cylinder (10) or can be rotated by other means such as gears or by entrainment by the gas being compressed. This feature results in a substantially low or zero rubbing speed between the surface of the rotor (20) and the inner surface of the cylinder (10), which in turn results in improved wear performance of those surfaces. Other results of this feature are lower frictional losses, lower kinetic energy imparted to the gas being compressed (lower entrainment) and lower frictional heat imparted to the gas being compressed. These results all contribute to greater efficiency of the compressor.

As shown in Figs. 3a to 3g, as can be used in the first class of embodiments described below, an orbiting rotor (20) does not change orientation with respect to the cylinder (10) as the rotor (20) traverses the internal circumference of the cylinder ( 10). Fig 3a shows sequential position 20a, 20b, 20c and corresponding orientations with arrows A, B, C. Figs 3b to 3g show the sequential rotor positions and corresponding orientation A. An orbiting rotor (20) results in a greater relative speed between the surface of the rotor (20) and the inner surface of the cylinder (10) than with the rolling rotor (20) of Fig. 2, but a lower relative speed than with a rotating rotor (20) as will be described in the following paragraph. Efficiency losses when an orbiting rotor (20) is employed tend therefore to be in a range between those of the rolling rotor (20) and those of the rotating rotor (20).

As shown in Figs. 4a to 4g, as can be used in the first class of embodiments or a second class of embodiments described below, the rotating rotor (20) changes orientation as the rotor (20) traverses the internal circumference of the cylinder (10), in such a way that a fixed point on the rotor (20) surface A, B, C in the sequential positions 20a, 20b, 20c in Fig 4a is adjacent to the pinch point (50). The movement of point A can be seen in the sequential position shown in Figs 4b to 4g. A rotating rotor (20) results in a greater relative speed between the surface of the rotor (20) and the inner surface of the cylinder (10) than either the rolling rotor (20) of Fig. 2 or the orbiting rotor (20) of Fig. 3. Efficiency losses when a rotating rotor (20) is employed tend therefore to be in a range which is higher than those of the rolling rotor (20) of Fig. 2 or the orbiting rotor (20) of Fig. 3. An advantage of the rotating rotor (20) of Fig. 4 is that a greater range of valve arrangements can be practically used than with the other two rotor (20) types. A compressor incorporating the rotating rotor (20) can be made with fewer moving parts than a compressor incorporating the other two types of rotor.

As shown in Fig. 5, in a first class of embodiments, the duct (40) is a chamber formed between two cylinders, one relatively static (10) and acting as a stator and the other (20) acting as a rotor - rolling, orbiting or rotating it within it. Using a valving mechanism described below, gas is drawn into the duct (40) by a rarefaction caused by the widening of one end the duct (40) (i.e. when the rotor is adjacent an opposing side of the stator). It passes through inlets in the walls of either of the cylinders (10, 20) or of the end walls and is expelled at higher pressure at the other end of the duct (40) after it has been compressed by a relative narrowing of the duct (40) caused by the orbiting component (20) approaching the stator wall. By mounting blades inside the rotor (20), a degree of pre-compression can be achieved. In such an embodiment the rotor (20) may have a rolling or rotating surface or may orbit without rotation.

In a device built according to such a first class of embodiments, as shown in Figs 5 and 10, there is provided a cylindrical rotor (20), within a cylinder (10). The rotor (20) is provided with a surface channel (210), of depth equal to the thickness of strip (220) that fits within the channel (210). The strip (220) is of larger circumference than the rotor (20) circumference, so that when the strip (220) is pressed onto the rotor (20) it forms a gas tight seal. However because the strip (220) is of larger circumference than the rotor portion (20), the strip (220) will always protrude above the rotor (20) surface circumference, away from the final point by virtue of the squeezing force exerted on it there, allowing gas to flow through openings (230) in the base of the channel (210) to outside the rotor (20).

The cylinder ( 10) is similarly provided with a channel (240) and strip (250) on the outside, allowing gas to pass from the inside of the cylinder (10) to ducting means on the outside. This outlet strip (250) may be provided with reinforcement across its width to support it against high gas pressures.

In operation the rotor (20) orbits, preferably at a speed that results in the pinch point (50) between the rotor (20) and cylinder (10) rotating at supersonic speed. As the pinch point (50) rotates, low pressure follows behind (in terms of the direction of rotation) the pinch point (50), pulling the strip valve (220) away from the rotor (20) and continuously inducing gas into the chamber (40). At the other end of the chamber (40) the converging surfaces of rotor (20) and cylinder (10) compress previously inducted gas and force it out of the chamber (40) through the exit strip (250), which is forced to and held in an open position by the pressure of gas in front of the pinch point (50). To prevent the exit strip (250) overlapping the pinch point (50) and allowing gas to escape from the high pressure volume into the low pressure volume, the exit strip (250) may be actuated by mechanical, electrical or magnetic means to control the distance of its opening (270) from the pinch point (50). As shown in Fig. 11, a cam (261) on the drive shaft (660) operates a pushrod (260) which operates to lift the exit strip (250). This actuation is also helpful for controlling start and shutdown conditions and to give a degree of capacity control. In rolling or rotating rotor embodiments (see below) where the high pressure exit side of the pinch point (50) can be separated by some distance from the low pressure inlet side, actuation of the strip (250) may be used to restrict the area of the outlet by moving the opening (270) partially past the pinch point (50) and so controlling the pressure ratio of the device. Hence, in an embodiment, a strip is deformable by mechanical actuation, in particular by an actuator such as a cam and pushrod coupled to the rotor, for example the rotor drive shaft.

As shown in Fig. 6, a blind passage (275) or passages are provided within the rotor, open on the axial face and terminating adjacent the inside of the rotor surface. This passage (275) communicates with the axial face of the rotor (20), so that cooling fluid may be circulated behind the rotor (20) circumferential surface. The walls and end plates of the chamber (10) are additionally provided with passages for the circulation of cooling fluid. Finned means (276) may be provided to increase the heat flow from the chamber (10) walls to be cooled into the cooling fluid.

In operation the rotor (20) is rotated with the inlet conduit (330) leading, so that the duct (40) rotates with the rotor (20) at a speed preferably in excess of the local speed of sound. Appropriate curvature of the inlet conduit (330) passage way causes gas to be drawn from an axial face of the rotor (20) into the conduit (330) in a substantially radial direction. As the duct (40) - i.e. the space between rotor and stator - rotates around the stator, the gas is confined to a converging duct (40) formed between the surface of the rotor (20) and the cylinder (10) wall. The supersonic speed of the approaching pinch point (50) does not give time for information about increasing pressure to propagate upstream. The gas is steadily compressed until, as the pinch point (50) reaches the gas, it is permitted to escape at high pressure through the outlet (340) and through passages within the rotor (20) to a radial end of the rotor (20) from whence it is ducted out of the device. During compression in the duct (40) the gas temperature increases. The heat of compression is transferred continuously, both through the wall of the rotor (20) into the cooling fluid circulating behind the wall and through the chamber walls (10) into the cooling fluid (277) circulating there.

As shown in Fig. 7 the surface of the rotor (20) may be provided with spiral grooves (400) and/or passages (410) to conduct high pressure gas that passes the pinch point (50) or along the axial ends of the chamber back to a selected or controlled point (420) in the duct (40). This gas is cooled on its passage back to the chamber and this is more advantageous for the efficiency of the device than allowing it to re-emerge at the inlet end of the chamber (330). In complex devices it would be possible to bleed this gas through micro pores (430) in the rotor surface to promote laminar flow. In this second class of embodiments the device may include a rotor (20) with the converging duct (40) formed between the cylinder (10) wall and the surface of the rotor (20) (as shown in Fig. 6) or the converging duct (40) may be formed between the cylinder wall (10) and a channel of reducing cross-section (330) on the rotor (20) where the rotor is concentric with the stator (as shown in Fig. 8).

Referring, for example, to Figs 7 and 8, in order to avoid accelerating gas, it is important that for a given cross-section of duct (40) the ratio of stationary to moving duct (40) surfaces should be as high as possible. In embodiments using a channel (330) within the rotor (20), the duct (40) may be formed by a groove (580) which winds spirally down the circumferential face of the rotor (20) so that all parts of the duct (40) including the high pressure/high temperature end of the duct (40) are continuously exposed to fresh surface areas to conduct heat from the duct (40). In such an embodiment heat transfer up the cylinder stationary wall ( 10) may be reduced by flanges (350) behind the surface (see Fig 6). The rotor (20) may be further cooled by internal fluid flow along the sides of the duct (40) and side of rotor (20).

As shown in Fig. 8, a second class of embodiments employ a rotating rotor (20). In such a device there is provided a rotor (20) within a cylinder (10), the rotor (20) being profiled so that a substantial part of the rotor (20) circumferential surface remains in rotatably close proximity to the inner wall of the cylinder ( 10) as it traverses the inner wall. The remaining circumferential surface of the rotor (20) is shaped or cut out so as to create a duct or groove (40) with a narrowing end (530) between it and the cylinder (10) wall. A wider end of the duct (540) is provided with an inlet conduit (520) communicating with the central part of an axial face of the rotor (20). Spaced from the duct (40) the circumferential surface of the rotor (20) is provided with an exit conduit (550) communicating with another portion of the axial face of the rotor (20). In large devices there may be provided more than one shaped duct (40).

In a spiral duct (500) embodiment the output pressure ratio may be controlled by providing a moveable sleeve (510) between the rotor (20) and cylinder (10). Tn operation, gas inlet (520) is through one axial end of the chamber (40) and outlet (550) through the other. Moving the sleeve (510) axially with respect to the rotor (20) changes the outlet area and so changes and controls the pressure ratio of the device.

Any of the above embodiments may be provided with means to adjust the offset of the rotor (20) from the central axis of the containing cylinder (10) and so adjust the clearance between the rotor (20) and cylinder (10) at the pinch point (50). This is advantageous for wear compensation, adjusting for different rates of thermal expansion, reducing leakage and to control capacity.

As shown in Fig. 9, an arrangement for driving the assembly, and additionally adjusting the offset of the rotor (20) from the central axis of the containing cylinder ( 10), and thereby adjusting the clearance between the rotor (20) and the cylinder ( 10) at the pinch point (50), is described herein. In overview, the rotor (20) has a rotor axis (670) each end of which is coupled to a drive rotor support and an idler rotor support (680, 690) respectively, each of the drive rotor support and the idler rotor support (680,690) in turn are coupled to a drive shaft and an idler shaft (660, 650) respectively, which are arranged such that they are on the central axis of the cylinder (10) and are each supported by a bearing support (630). In more detail, an end of the rotor axis (670) is joined by a coupling (600) to a drive rotor support (680), and an other end of the rotor axis (670) is joined by a coupling (600) to an idler rotor support (690). The idler rotor support (690) is joined by a coupling (600) to a fixed shaft (650). The drive rotor support (680) is joined by a coupling (600) to a drive shaft (660). Both drive shaft (660) and idler shaft (650) are arranged to be parallel to the rotor axis (670) and to lie on the central axis of the cylinder (10). Each rotor support (680, 690) is arranged to support the rotor axis (670) such that the rotor (20) surface is substantially positioned close to the inner circumference of the cylinder (10). Both idler shaft (650) and drive shaft (660) are supported by a bearing support (630) and are rotatable within, and axially constrained relative to said bearing support (630). Each bearing support (630) is arranged such that its axial distance from the centre of the rotor axis (670) is equal to that of the other bearing support (630) and is controllable. By controlling the distance of the bearing supports from the centre of the rotor axis (670) it is possible to vary the position and angle of each rotor support (680, 690) and resultantly it is possible to vary the running clearance between the rotor (20) and the housing (10).

Three classes of coupling (600) can be advantageously employed in the preceding arrangement. A first class of coupling (600) includes couplings which are suitable for forming a joint which is articulated in two axes between two shafts, but not capable of transmitting any axial torque. An example of a commonly known coupling (600) falling into the first class is a ball joint. A second class of coupling (600) includes couplings which are suitable for forming a joint which is articulated in two axes between two shafts, and capable of transmitting axial torque. An example of a commonly known coupling (600) falling into the second class is a constant velocity joint, a Hardy-Spicer universal joint, certain types of rubber couplings or compliant rubber tubing. A third class of coupling includes couplings witch are suitable for forming a joint which is capable of articulating in one axis and capable of transmitting axial torque. An example of such a joint is a hinged joint.

The drive shaft (660) can transmit rotational torque via a drive coupling (640). The drive shaft (660) is coupled to the drive rotor support (680) by a coupling (600) of the third class. The end of the drive rotor support (680) which is coupled to the rotor axis (670) is thereby constrained to orbit in a circular motion around the draft shaft (660) axis.

In embodiments employing a rolling rotor, the drive rotor support (680) is coupled to the rotor axis (670) by a coupling (600) of the first class. The rotor axis (670) is coupled to the idler rotor support (690) by a coupling of the first or second class. In such embodiments, either at least one of the coupling (600) which couples the rotor axis (670) to the idler rotor support (690) and the coupling (600) which couples the idler rotor support (690) to the idler shaft (650) are of the first class, and/or the idler shaft (650) is free to rotate. The rotor (20) is thereby free to roll independently of the drive shaft (640) orientation and the idler shaft (650) orientation, but the rotor (20) is compelled to traverse the inner circumference of the cylinder (10) by the drive transmitted from the drive shaft (640) to the drive rotor support (680).

In embodiments employing an orbiting rotor, the drive rotor support (680) is coupled to the rotor axis (670) by a coupling (600) of the first class. In such embodiments, both of the coupling (600) which couples the rotor axis (670) to the idler rotor support (690) and the coupling (600) which couples the idler rotor support (690) to the idler shaft (650) are of the second class, and the idler shaft (650) is fixed so that it cannot rotate. The rotor (20) is thereby constrained so as to maintain its orientation with respect to the cylinder (10) by virtue of its connection to the fixed idler shaft (650). The rotor (20) is compelled to traverse the inner circumference of the cylinder (10) by the drive transmitted from the drive shaft (640) to the drive rotor support (680).

In embodiments employing a rotating rotor, the drive rotor support (680) is coupled to the rotor axis (670) by a coupling (600) of the second or third class. The rotor axis (670) is coupled to the idler rotor support (690) by a coupling of the first or second class. In such embodiments, either at least one of the coupling (600) which couples the rotor axis (670) to the idler rotor support (690) and the coupling (600) which couples the idler rotor support (690) to the idler shaft (650) are of the first class, and/or the idler shaft (650) is free to rotate. The rotor (20) is thereby constrained to maintain its orientation with respect to the drive rotor support (680), and is unconstrained relative to the idler rotor support (690) orientation, and as a result, a fixed point on the rotor (20) surface is maintained adjacent to the pinch point (50). The rotor (20) is compelled to traverse the inner circumference of the cylinder (10) by the drive transmitted from the drive shaft (640) to the drive rotor support (680).

Although the rolling, orbiting and rotating rotor constraint arrangements have been herein described with reference to the use of specific combinations of the aforementioned classes of coupling, it will be appreciated that the rotor characteristics described herein can be accomplished by other combinations not described. Accordingly, the descriptions of the orbiting, fixed, and rotating rotor constraint arrangements described herein are not intended to be limiting to the scope of the invention, the invention being set out in the claims.

As shown in Fig. 9, a means for counterbalancing the rotor (20) is provided. The drive rotor support (680) is extended past the coupling (600) which couples the drive rotor support (680) to the drive shaft (660), in a direction away from the rotor (20). A counterbalance weight (620) is provided either separately, or integrally with the drive rotor support (680) extension. Similarly, the idler rotor support (690) is extended past the coupling (600) which couples the idler rotor support (690) to the idler shaft (650), in a direction away from the rotor (20). A counterbalance weight (620) is provided either separately, or integrally with the idler rotor support (690) extension. Each counterbalance weight (620) is arranged to have a weight and a distance from the central axis of the cylinder (10) such that the weight of the rotating components on the opposite side of the central axis of the cylinder (10) is balanced. The mass or position of the counterbalance weights (620) can be adjusted during operation of the compressor, to compensate for thermal expansion or other effects which would otherwise upset the balance of the rotating components of the compressor. This can be achieved by the use of actuators to adjust the position of the counterbalance weights (620) on the rotor supports (680, 690), Alternatively, the mass of the counterbalance weights (620) can be altered, for example by pumping fluid or gas in or out of the counterbalance weights (620) which can incorporate a fluid or gas reservoir.

Fig. 10 shows an alternative arrangement for counterbalancing the rotor (20) where the drive shaft (660), rotor axis (670) and counterbalance weights (620) are housed within the cylinder (10), this being advantageous in that sealing of the chamber is facilitated.

Although the manner in which the various chambers are sealed and ducted are not described in all cases in detail it will be appreciated that in embodiments of this invention the usual sliding seal means of the compressor art are provided to prevent leakage of gas from high pressure volumes to low pressure volumes. Ducting means to direct low pressure gas into devices and high pressure gas away from the device are also provided. In any of the above embodiments conventional control means of the art, such as valves, may be used in combination to control and regulate flow.

Although embodiments have been described with a static cylinder (10) and a movable rotor (20), other embodiments may employ a moving cylinder (10) and static rotor (20) or both moving rotor (20) and cylinder (10).

A compressor may be reversed, with appropriate valving, to operate as an expander.

Advantages of such a compressor as described above, and other embodiments, are that high efficiency of compression and high stage pressure rise are achieved by compressing gas while imparting as little kinetic and friction energy to the gas. Embodiments also allow cooling of the gas while it is being compressed.

In axial and centrifugal compressors the necessity of multiple stages, caused by the low pressure rise per stage can be exploited to provide inter-cooling between stages. For high efficiency of compression all surfaces enclosing the gas may be cooled and the gas and/or surfaces continuously changed so that the gas is brought into contact with freshly cooled surfaces during compression. Preferably the gas should not flow relative to the walls as this causes frictional heating.

Further advantages of embodiments follow:

By employing supersonic rotation of the pinch point, the simple mechanical layout of the invention is made possible, since high pressure cannot propagate to the low pressure areas of the chamber and therefore no mechanical separation between low and high pressure regions of the chamber is required. The continuous rotational compression means of the invention allows for smooth continuous compression. By employing smooth and continuous compression means, the invention advantageously reduces the energy imparted to the gas being compressed.

By employing adjustable running clearance means, and/or a rotor which rolls as it traverses the internal circumference of the cylinder, frictional losses are reduced, which reduces heating of the gas to be compressed and thereby increases efficiency.

The fixed chamber volume of some embodiments allows for enhanced heat transfer properties because the maximum chamber surface area is always in contact with the gas being compressed. This allows the gas being compressed to be more effectively cooled, which in turn aids compressor efficiency.

The amount of gas processed in each revolution is greater than the volume of the interior volume of the cylindrical chamber and the volume of the rotor. The swept volume is the cylinder volume less the volume of a rotor having a radius equal to: {radius of the rotor minus the radial offset of the rotor axis from the cylinder axis J . In other words, the sweep path of the rotor surface diametrically opposite the pinch point defines the swept volume.

A further advantage of a compressor according to embodiments is that it exhibits high flow properties compared to, for example, axial or centrifugal compressors of a similar physical size. As a result, a compressor can be made physically smaller than some other known compressors. Further developments in compressor technology will now be described in detail below.

A dual-lobed rotor embodiment of the compressor described above is shown in Figure 12a. An elliptical rotor 20 within a circular housing 10, or inner cylinder, has an axis of rotation. The clearance between the elliptical rotor and the housing forms a pair of ducts 40, each bounded by a pair of pinch points 50 at the parts of the rotor 20 having the furthest radius from the axis. The rotor 20, in embodiments, incorporates a pair of inlet flow passages 1200 arranged to communicate axially with the housing and with the duct 40 via the surface of the rotor 20. In use, flow passage 1200 carries inlet fluid (e.g. gas) from an axial end of the compressor and discharges the fluid into the inlet end of the duct 40. The gas is then compressed in the way previously described for the single duct embodiments above. An advantage of the dual lobed embodiment is enhanced balancing properties at high rotational speeds.

Alternatively, the rotor 20 may be shaped as shown in Figure 12b, so as to provide a different profile of duct cross-sectional area variation with respect to angular position of the rotor, the rotor having a comparatively narrow region near the axis and lobes at each radial end. In use, the inlet flow passages 1200 (not shown in Figure 12b but equivalent to those of Figure 12a) advantageously impart a degree of centrifugal compression to the inlet gas and aid inlet charge filling of the compressor duct(s) 40, thus allowing a reduction in rotational speed for a given required compression capability. The ends of the rotor lobes are optionally flattened so as to enhance separation of the high pressure gas close to the pinch point 50 in front of the advancing pinch point 50 from the low pressure gas behind the moving pinch point 50. This helps to reduce leakage past the pinch point 50. In a further development, as shown in Figure 13, high pressure gas in front of the advancing pinch point 50 is allowed to leave the compressor by the provision of resilient deformable discs 1300 in the axial end of the housing 10, so as in use to seal against the axial ends of the housing 10 under low or negative internal pressure, but arranged to deform outwardly and allow high pressure gas to exit the housing 10. This provides a simplified output valving arrangement.

The above embodiments comprise a compressor having a rotor 20 within a stator chamber (or housing) 10, where the rotor 20 is shaped so as closely approach or touch the chamber wall at one or more pinch points 50, the rotor 20 having no abrupt changes of surface direction between pinch points 50. When the rotor 20 operates at supersonic speed high compression ratios can be achieved. Variations include operating the compressor with both a rotor surface that rotates relative to the stator chamber 10 and a circular surface that rolls relative to the stator chamber 10 wall. The latter is the most efficient mode of operation but more complex to manufacture.

An improvement, according to the present invention, which simplifies the port arrangement in these devices and to reduces the reduction of leakage from high pressure volumes to low pressure volumes will now be described.

Embodiments also relate to providing an unimpeded gas path through the device and to reduction of leakage.

Reduction of leakage has previously involved very tight tolerances between high speed surfaces. If they touch, severe damage may result.

In a simple compressor shown in Figure 14a, the rotor 20 has one or more pinch points 50 with no abrupt changes of surface between pinch points 50 and rotates within a stator chamber 10. The portion of the rotor 20 which extends relatively close to the stator 10 wall (thereby forming the pinch point 50) may be wide, i.e. the pinch point 50 need not be a narrow region of close approach of the rotor 20 to the stator wall 10, but can optionally be a relatively wide area on the rotor 20 circumference. This pinch point portion 50 may have axially orientated grooves 1400 to create a labyrinth seal against fluid flow in a circumferential direction. The rotor is provided with rotatable side flanges

1410, 1420 (inlet = 1410, outlet side = 1420), on both axial sides of the rotor 20, that extend radially to closely fit within the chamber 10, and between the rotor portion 20 is eccentrically mounted. The periphery of the flanges 1410, 1420 are provided with a series of circumferential grooves 1430 to provide labyrinth seals against fluid flow in an axial direction. The stator housing 10 is also optionally shaped to improve sealing.

A small section 1440 of the flange 1410, 1420 on one axial end is cut away near the leading side in the direction of rotation of a pinch point 50. This provides an outlet (or exhaust) port 1440. A larger section 1450 of the flange is cut out or reduced in diameter on the other axial end of the rotor 20, so as to form a gap, near the trailing side in the direction of rotation of a pinch point 50, to provide an inlet port 1450. This may extend to where the rotor 20 surface is furthest from the stator chamber wall 10 (point 1480 in Figure 14b), i.e. almost 180 degrees around the rotor 20 circumference (for a single pinch point 50 embodiment). The cut out outlet 1440 and inlet 1450 should be designed to be aerodynamically efficient and may also be shaped to encourage flow onto and off the rotor 20 surface. The outlet cutout 1440 is ideally situated at a high pressure region close to the pinch point 50. The inlet cutout 1450 is ideally situated away from the high pressure region so as to reduce shock waves reflected into the inlet port of the compressor (not shown). The rotor 20 surface may also be angled on either side of the pinch point 50 to reduce dead flow volume (i.e. portions of the volume of the duct 40 which are aerodynamically relatively inaccessible to fluid due to their location in sharply defined recesses) as shown in Figure 14c.

In order to prevent flow in the circumferential labyrinth grooves on the flanges providing a leak path into the cut away section, the labyrinth grooves are sealed at or near the edge of the cut away sections (as shown in Figure 14a, item 1405). To prevent damage from touch down between the tips of the seals on the flanges and the chamber wall and to allow the compressor to run-in to a very gas tight condition, abradable material can be used in sections of the chamber wall 10 adjacent the rotating flange labyrinths 1430 of the rotor 20.

When run at high pressure ratios, considerable thrust is produced at the high pressure end of the compressor. This may be counteracted by providing a second sealed flange wall 1540 in the outlet chamber 1530, as shown in Figure 15.

In operation the rotor 20 turns at high speed, gas flow is induced through the larger inlet cut away 1450 by a rarefaction following the pinch point 50. Induction ceases where the inlet 1450 ends and the rotor/stator gap is largest ( 1480 in Figure 14b). Inducted gas is now in a converging duct 1490 and is compressed and forced out of the compressor in front of the approaching pinch point 50, via the outlet 1440. Because a high relative speed of gas and rotor 20 is desirable, the gas should enter the compressor on a path that is static relative to the stator 10 or angled to oppose the rotor 20. Gas flows in an almost straight path through the compressor, relative to the stator 10, and moves relative to the rotor 20.

A further improvement will now be described which provides for changing the flow of compressors and pumps without altering the rotor speed. This is advantageous as it removes the need for complicated and expensive variable speed drive control apparatus. In existing positive displacement compressors this is commonly achieved by throttling the inflow or re-circulating flow. These methods reduce the efficiency of such a device, and in order to provide an improved device the present invention alters the displacement by changing the cross-sectional area of the flow channel or duct 40 while the compressor is operating. This also changes the swept volume per rotation. Thus the flow per revolution of the rotor 20 can be adjusted to anywhere between the maximum design flow and zero flow.

Adjustment of displacement in devices with open gas ports disclosed above may be achieved by axially moving one or both of the flanges 1410, 1420 so that the cross sectional area of the flow channel (or duct) 40 is changed.

In a simple embodiment of this invention as shown in Figure 16a there is provided a flange 1410 at the inlet axial end of the rotor 20. This flange may be an integral part of the rotor 20 or fixed to the rotor 20. A second flange 1620 is fixedly mounted on a first end of a hollow member 1610 that fits closely round the rotor 20 so that the hollow member 1610 can slide inwards and outwards on the rotor 20 to vary the separation of the flanges 1410, 1620. In operation the rotor 20 turns, gas is inducted into one end of the chamber 1640 and forced out at the other as described above. To vary the displacement and so change the volume of fluid that is pumped per rotation, the hollow member 1610 and the flange 1620 are moved towards or away from the inlet flange 1410 by moving the bearing housing 1660 at the output end of the device and causing the hollow member 1610 to slide along the drive shaft 1280 to which the rotor 20 is fixed. A splined connection between the hollow member 1610 and the shaft 1280 prevents torque from the drive being transmitted entirely through the rotor 20 to the hollow member 1610 and distorting it and unbalancing the assembly. As will be apparent to the skilled reader, movement of the bearing 1660 can be achieved by any appropriate mechanical, pneumatic, hydraulic or magnetic means. Of course, the inlet flange 1410 could be moved instead of or as well as the outlet flange 1620.

An alternative arrangement is shown in Figure 16b, in which the movable flange 1620 is supported on the rotor by means of opposed engaging ends 1630, 1640 axial Iy movable in cooperating axial slots 1650, 1660 in the rotor 20, thereby obviating the need for a hollow member 1610 on which the flange 1620 is supported in the embodiment shown in Figure 16a. The flange 1620 is shaped to fit closely between the rotor 20 and the housing 10, with an outside shape similar to that of the inside shape of the housing 10, and an inside shape similar to that of the surface of the rotor 20, thereby to minimise leakage past the flange 1620.

As shown in Figure 17, flow into the device may be encouraged by angling the leading edge of the inlet opening 1450 of the flange 1410 on the inlet side, away from the rotor 20 in an axial direction, so as to form a scoop for encouraging fluid into the chamber (or duct) 1640. This angle may be varied during operation by appropriate mechanical or electromechanical actuation means to suit the desired flow.

Clearly a device of this construction can also be used at moderate speeds for pumping liquids. Variable displacement at a fixed speed is useful in applications where variable proportions of fluids have to be mixed in industrial processes. Fluid flow through a device as described can also be reversed so that the device functions as a variable expander or turbine to extract work from fluid flowing from high pressure to low pressure.

Furthermore connecting a variable displacement compressor according to this invention with a variable displacement expander according to this invention enables work to be recovered from the expansion process (for example, a refrigerant expansion process) rather than energy being wasted in throttled expansion (i.e. operation in expansion mode with a throttled, or partially closed off, inlet port). This recovered work can employed to reduce the work required for compression by a direct drive connection to the compressor. This is not normally practicable in compression systems because of the varying volumes that have to be processed by compressor and expander. However, employing independently variable compressor and expander displacements overcomes this problem. Thus higher efficiency can be achieved in refrigeration processes with unprecedented control over conditions in the condenser and evaporator. Such control is also required in Fuel Cell systems where the volumetric flow in and out of the compressor varies so significantly that it has not hitherto proved practical to directly feed the work recovered in the expander to the compressor.

Further, in existing pumps and fans it has been known to change flow by dynamically changing the angle of attack of rotating blades and by changing the angle of non rotating vanes at the inlet to the rotor, so as to change the angle at which flow impinges on the turbine blades. These methods have the problem that they reduce efficiency by imposing an unnecessary acceleration on the fluid. The present method of altering the channel hydraulic cross- sectional area described above can also be applied to turbines both with enclosed or shrouded flow channels and open or unshrouded flow channels. In another embodiment of a device where displacement is adjusted by varying the hydraulic cross-section of a flow channel as shown in Figure 18a, as applied to unshrouded and shrouded turbine flow channels of a type commonly used in liquid pumps, there is provided a shaft 1280 on which is mounted a rotor member 1810 with channels 1825 formed between strakes 1820.

Slideably mounted on the shaft is a second rotor member 1830 with projecting members 1840 that fit closely within the channels of rotor member 1810. In operation, axially moving the second rotor member 1830 that has projecting members 1 840 moves the projecting members 1840 in and out of the channels between the strakes 1820 to change the hydraulic area of the channels encountered by flow G. In open turbines the projections form the floor 1850 of the open channel and are moved axially in and out to change the effective hydraulic cross-sectional area formed by the rotating channel and the stator chamber wall. The projection forming the floor may have a curved surface to more approximately conform to a circular cross-section and so maximise the flow area to wall area and so reduce friction losses. An alternative embodiment having closed channels is shown in Figure 18b.

In devices built according to this invention it is not always necessary to actuate the sliding member in the direction that widens the channel as there is sufficient pressure in the channel to provide a returning force to widen the channel again. Actuation is only required to narrow the channel and this will normally be achieved by mechanically pushing on the moveable member 1830 from an axial direction (for example by direct, hydraulic or pneumatic action) or by advancing a magnetic field to repel the moveable member 1830). Of course, movement in either direction can be augmented or opposed by appropriate biasing means such as a spring. Compressors and pumps can generally operate in reverse to act as engines by converting pressure and kinetic energy into work. The above description of varying flow by changing the width of the channel on the rotor 20 is also industrially useful in such engines.

Moving the bounding surfaces of a rectangular cross sectional duct 40 in a compressor or pump changes the cross-sectional area and so can be employed to change the fluid flow without changing the rotational speed of the device. This offers advantages over changing fluid flow by changing rotational speed, or by throttling the inlet, which are the established methods of providing variable flow. This method of varying the volume displaced per revolution can be applied to converging or diverging ducts within a compressor or pump, as well a ducts of unvarying cross-sectional area.

In a compressor intended to operate at supersonic speed of rotation, the high speed stresses in the material are challenging. This is particularly the case when moving the radial plane boundaries (i.e. flanges 1620 defining a duct 40 boundary) in an axial direction. A problem with moving flanges 1620 axially is that axial movement also alters the axial balance of the rotor and induces vibration. A further embodiment has been developed in which the inner boundary of the duct 40, comprising the rotor 20 surface, hereinafter termed the floor 2000, is moved to adjust the flow through a duct 40. This overcomes the stress challenges of moving the radial plane boundaries and also enables streamlining of the inlet to reduce entry turbulence losses.

In an embodiment of such a device (Figure 20 & Figure 21 , showing one duct only) adapted for compressing gas, there is provided a rotor 20 with two half crescent ducts 40. The or each duct 40 is bounded at the outer surface by the circumferential wall of the cylindrical chamber 10, in which it rotates. The rotor 20 is provided with two side walls (or flanges 1410, 1420) extending in the radial plane which define the axial width of the duct 40. Either of these walls (or flanges 1410, 1420) may be tapered inwards towards the inlet 1450 and outlet gas ports 1440 which are provided at each end of the duct 40. This improves gas flow. An inlet port 1450 is provided through a first side wall

1410 at the larger cross-sectional end 2010 of each duct 40 and communicates with a source of gas to be compressed. An outlet port 1440 is provided in the second side wall 1420 at the smaller cross-sectional end 2020 of each duct. The outlet port 1440 communicates with a collection and diffusion chamber 1530 which is bounded by the second side wall 1420 and a balancing side wall 1540 which acts to cancel the axial thrust otherwise created by high pressure gas. High pressure gas exits radially through outlet passages 2150 arranged in the stator housing 10 adjacent to the diffusion chamber 1530.

The floor 2050 of the duct 40 is shaped to provide a converging duct 40 towards the outlet 1440, in co-operation with the chamber wall 10. The floor 2050 is supported and restrained at the outlet end of the duct 2020 by means such as a hinge or pivot axis 2060 which defines an axis which is near the outer circumference of the rotor 20 and about which axis the floor 2050 can rotate towards or away from the rotor centre through a small arc. The inlet end 1450 of the duct 40 is bounded by an angled or tapered surface 2070, that is arranged so as to smooth the path of inlet fluid into the duct 40, for example, a duct which gradually widens in a rotor axial direction towards the area near the inlet port 1450.

The side wall 1410 on the inlet side is provided with an opening 2080 to create an inlet port 1450. In operation the floor 2050 is moved around its pivot axis 2060 to change flow by changing the cross-sectional area of the duct 40, while leaving the overall geometry of the chamber 40 substantially unchanged. The opening 2080 must be large enough to act as an inlet port 1450 of suitable size for the designed flow. To smooth flow in the entry chamber of the compressor/pump and into the port, a plate 2090 of the same width as the opening 2080 may be attached to the axial end of the floor 2050 or be integral to it, so that, as the floor 2050 moves to and fro, the plate 2090 closes the unopened area level with the outer surface of the side wall 1410 and so reduces turbulence in the entry chamber adjacent the inlet side wall 1410.

The entry chamber (not shown) may also be provided, as is known in the art of pumps and compressors, with angled inlet vanes or ducts either on the stator chamber 10 or rotating as part of the rotor 20, or both, to turn the inflow in a direction that increases efficiency (i.e. contrary to the direction of motion of the pinch point 50). In the device the intention is to impart velocity to the inlet gas in the opposite direction to the rotor direction, so as to lower the speed that the rotor 20 has to turn to create super sonic speed relative to the gas. Such pre- rotation devices may also include angling the trailing edge of the inlet 1450 outwards to induce gas into the duct as shown in Figure 17.

Similarly the outlet port 1440 may be provided with angled vanes to compound some of the kinetic energy of the fluid leaving the chamber 40, as shown in Figure 19. In the embodiment described here the fluid (G) will leave the duct (chamber 40) with high kinetic energy in an axial direction and will be slowed in the diffusion chamber 1530. In some designs it may be advantageous to turn this flow using a vane 1910 (or ramp) or vanes against the direction of rotation to provide a thrust (F) on the rotor 20 in the direction of rotation, as shown in Figure 19. This acts both to slow the fluid (G) and increase its pressure and to recover energy from the fluid. Advantageously, in a supersonic compressor, wherein the pinch point 50 is moving at or in excess of supersonic velocity, such energy recovery is possible without imparting back pressure to the inlet. In the embodiments of Figures 20 and 21, the outer edges of the duct 40 walls may be angled inwards to prevent the floor 2050 moving too far outwards and touching the chamber wall 10.

In operation it is preferable to have low pressure beneath (i.e. radially inwards of) the floor 2050 because this enables higher pressure in the outlet end of the duct 2020 to act on the floor 2050, pushing it inwards to counteract centrifugal forces. To achieve this, the floor 2050 may be sealed against passage of gas from the duct 40 to the inner part of the rotor 20. The seals may take the form of lip seals (not shown) mounted on sides of the floor 2050. The lips are pressed against the rotor 20 side walls (4, 4a) by pressure in the duct 40. To adjust the level of the floor 2050 without requiring high force to overcome the friction of the seals, the lip seals may be released by admitting high pressure gas into the interior of the rotor 20.

The duct floor 2050 may also be adjusted by means of a hydraulic (or pneumatic) actuator 2095 comprised in the rotor 20. Hydraulic fluid can be passed into the hydraulic actuator of the rotor 20 via the shaft 1280 and a rotating type annular seal (not shown). In practice, the maximum rotational speed of a shaft relative to a rotational annular seal may be limited by a maximum rotational speed capability of the seal. Since, in embodiments, the rotor 20 is required to operate at a rotational speed such that the pinch point 50 is moving at a super sonic velocity, the rotational speed of the shaft 1280 may be in excess of the rotational speed capability of the rotational seal. In such cases, an intermediate rotating member 2025 can optionally be placed axially adjacent to the rotor 20 assembly (comprising the rotor 20 and at least the output flange 1420), and operated at a rotational velocity which is below that of the rotor 20. Hydraulic fluid can then be passed into the intermediate member 2025 by means of a rotational seal, and then from the intermediate member 2025 to the rotor 20 by means of a further rotational seal. Thus, by way of example when the intermediate member 2025 is operated at a rotational speed equal to half that of the rotor 20, each rotational seal need only be operated at a rotational speed of half that which would otherwise be required if the intermediate member 2025 was not employed. Of course, more than one intermediate member 2025 can be used if necessary.

In a further embodiment, the duct floor 2050 can be actuated mechanically, as shown in Figure 22. For example, a wedge shaped cam 2210 attached to a actuation shaft 2200 is arranged so that it can be moved axially by axial movement of the actuation shaft. A cam follower 2220 is attached to the duct floor 2050. Thereby, in operation, by moving the actuation shaft 2200 axially, the wedge shaped cam 2210 is moved, and the sloping surface of said wedge shaped cam 2210 acts on the cam follower 2220, thereby urging it in a direction perpendicular to the cam 2210. The duct floor 2050 is thereby moved outwardly. A spring or other means of urging the duct floor 2050 towards the inward position may be employed, so as to ensure that the cam follower 2220 remains in contact with the cam 2210. Centrifugal force acting on the duct floor 2050 and cam follower 2220 may in some embodiments and operating conditions overcome gas pressure in the duct 40, and thereby tend to force the duct floor 2050 in a radially outward direction. The spring or equivalent urging member is arranged to resist this action. An alternative arrangement is that the cam follower 2220 is hook shaped 2225, or similar, such that it engages with a sloped surface 2215 of the cam 2210 which is on the opposite side of the axis from the floor 2050. Thereby, the floor can be forced inwards by cam action. A further alternative arrangement is that the actuation shaft is rotated relative to the rotor shaft 2015 to which the rotor 20 is attached, and the actuation shaft 2200 has an offset lobe shaped element (not shown) attached to its end inside the rotor 20, the lobe being similar to that employed on a cam shaft of an internal combustion engine. Thus, in use, by rotation of the actuation shaft 2200 relative to the shaft 1280, in this embodiment the lobe is caused to rotate relative to the rotor 20, and the cam follower 2220 or hook 2225 bears upon the lobe, thus forcing the floor 2050 in a radially outward or inward direction.

Thus, an alternative arrangement for adjusting the level of the duct floor 2050 is effected. Of course, other means of actuation such as electro-mechanical means can be employed, using control coupling means equivalent to the rotating hydraulic seals of the previously described hydraulic embodiment above.

It will be appreciated that any appropriate fluid, not limited to gas or liquid, can be operated on according to the invention, for purposes of compression, expansion or other operation on a property of the fluid.

Although the invention has been explained in relation to its preferred embodiments, these are not intended to limit the invention. It will be understood by those skilled in the art that many other modifications and variations are possible without departing from the scope of the invention as claimed. Embodiments and features of embodiments may be juxtaposed or interchanged as appropriate.

Claims

1. A compressor or expander having a rotor and a stator forming a closed volume therebetween and arranged for relative traversal, so as to form a pinch point at the point of nearest proximity, wherein at least one of the rotor and the stator has a movable portion such that the closed volume is controllably variable.

2. An apparatus according to claim 1 wherein the pinch point is arranged to move at substantially supersonic velocity.

3. An apparatus according to any preceding claim wherein the rotor has a variable geometry formed by said movable portion.

4. An apparatus according to claim 3 wherein said movable portion comprises a rotor surface portion which is movable in a direction having a radial component relative to an axis of rotation of the rotor.

5. An apparatus according to claim 4 wherein the movable rotor surface portion is arranged to be actuated by fluid pressure.

6. An apparatus according to claim 5 wherein the fluid pressure is hydraulic pressure.

7. An apparatus according to claim 5 wherein the fluid pressure is pneumatic pressure.

8. An apparatus according to claim 3 comprising a first flange arranged for rotation in the stator and for defining a boundary of the closed volume.

9. An apparatus according to claim 8 comprising a second flange arranged for rotation in the stator and for defining another boundary of the closed volume.

10. An apparatus according to claim 8 or 9 wherein one of the flanges comprises said movable portion, movable in an axial direction of the rotor.

1 1. An apparatus according to claims 8 to 10 wherein the movable flange is supported on an axially movable member.

12. An apparatus according to claims 8 to 10 wherein the movable flange is supported in an axial groove in the rotor.

13. An apparatus according to claims 8 to 12 wherein the first flange has an inlet opening for allowing fluid to enter the compressor, the inlet opening being arranged such that in use it trails the pinch point in the direction of rotation.

14. An apparatus according to claim 13 wherein a portion of the leading edge of the inlet opening extends at least partly away from the rotor in an axial direction so as to form a scoop.

15. An apparatus according to claims 6 to 14 wherein the second flange has an outlet opening for allowing fluid to leave the compressor, the outlet opening being arranged such that in use it leads the pinch point in the direction of rotation.

16. An apparatus according to any preceding claim wherein the rotor includes a labyrinth seal in the region of the pinch point.

17. An apparatus according to claim 15 comprising a third flange arranged for rotation in the stator and axially spaced from the second flange, so as to define an output diffusion chamber.

18. An apparatus according to claim 17 wherein the stator comprises an outlet passage, and the output diffusion chamber and the outlet passage are arranged for radial communication therebetween.

19. An apparatus according to claim 17 wherein the output diffusion chamber comprises a ramp adjacent the outlet opening, the ramp oriented such that gas exiting the compressor axially through the outlet opening is redirected contrary to the direction of rotation.

20. An apparatus according to claims 5 to 19 wherein each flange incorporates one or more labyrinth seals.

21. An apparatus according to claim 20 wherein the labyrinth seal comprises circumferential grooves extending from a point adjacent the pinch point to another point adjacent the pinch point and on the other side of the pinch point, such that the grooves are sealed in the region of the pinch point.

22. An apparatus according to claims 1 to 16 further comprising outlet valving means comprising a resilient deformable disc arranged to seal against an axial end of the stator.

23. An apparatus according to any preceding claim wherein the rotor is shaped so as to form a plurality of pinch points at respective points of proximity with the stator.

24. An apparatus according to claim 23 wherein the rotor is generally elliptical in shape.

25. An apparatus according to claim 23 wherein the generally elliptical rotor includes a relatively narrow central portion and lobes either side thereof.

26. An apparatus according to any preceding claim in which the rotor incorporates a passage arranged to communicate axially with the housing, and with the duct.

27. An apparatus according to any preceding claim wherein the rotor surface includes a portion adjacent to the pinch point in which proximity of the portion to the stator varies as the rotor is traversed in an axial direction.

28. An apparatus according to claim 27 in which the proximity of the portion decreases with distance along the axis from an inlet side.

29. An apparatus according to claim 27 in which the proximity of the portion decreases with distance along the axis from an outlet side.

30. A compressor or expander having a rotor and a stator forming a closed volume therebetween and arranged for relative traversal, at least one of the rotor and stator being shaped so as to form two or more pinch points at respective points of closest proximity with the other of the rotor and stator.

31. An apparatus as claimed in claim 30 in which one of the rotor and stator is generally circular in shape and the other of the rotor and stator is generally elliptical in shape in a radial plane.

32. An apparatus as claimed in claim 30 or 31 in which the generally elliptical one of the rotor and stator includes a relatively narrow central portion and lobes either side thereof in a radial plane.

33. A compressor or expander having a rotor and a stator forming a closed volume therebetween and arranged for relative traversal, so as to form a pinch point at the point of nearest proximity, further comprising a flange arranged for rotation in the stator and for defining an axial boundary of the closed volume.

34. The apparatus of claim 33 wherein the flange incorporates an inlet or outlet gap for admitting or exhausting fluid.

35. The apparatus of claims 30 to 34 wherein the pinch point is arranged to move at substantially supersonic velocity.

36. A turbine or pump comprising a compressor or expander according to any preceding claim.

37. A method of compressing or expanding a fluid comprising the steps of: arranging a rotor and a stator to form a closed volume therebetween, and for relative traversal, so as to form a pinch point at the point of nearest proximity, and arranging at least one of the rotor and stator to have a movable portion, such that the closed volume is controllably variable.

38. A compressor or expander having a first rotor having first formations defining flow channels therebetween, and a second rotor, at least one of the first and second rotors being axially moveable relative to the other, the second rotor having second formations cooperable with the first formations to vary a dimension of the flow channels upon relative movement.

39. A device or method substantially as described herein or shown in the drawings.