GB2585093A

GB2585093A - Rotating machine interface arrangement

Info

Publication number: GB2585093A
Application number: GB1909389.7A
Authority: GB
Inventors: Tarald Kibsgaard Svend; Brenne Lars; Mazzali Carlo
Original assignee: Equinor Energy AS
Current assignee: Equinor Energy AS
Priority date: 2019-06-28
Filing date: 2019-06-28
Publication date: 2020-12-30
Also published as: GB201909389D0

Abstract

A rotating machine, such as a multi-stage centrifugal compressor, comprises a housing (8, fig 1) and a rotor 4 configured for rotation relative to the housing. The rotor has a first interface 24 and a second interface 18a positioned adjacent the rotor to engage the first interface. A piezoelectric member 20 coupled to the second interface moves the second interface relative to the housing when a voltage is applied across the piezoelectric member. A controller 26 controls the voltage applied by a voltage source across the piezoelectric member. A method of operating the rotating machine comprises rotating the rotor and controlling a position of the second interface by applying a voltage across the piezoelectric member. A method of removing fouling on the first and/or second interface comprises causing the second interface to vibrate by supplying an oscillating voltage to the piezoelectric member.

Description

ROTATING MACHINE INTERFACE ARRANGEMENT

TECHNICAL FIELD

The present disclosure relates to an interface between a rotating component and a stationary component within a rotating machine, such as a compressor, and particularly to a sealing or bearing arrangement within the rotating machine that permits a clearance gap to be dynamically adjusted during operation of the rotating machine.

BACKGROUND

A rotating machine, such as a compressor, will typically include multiple seals. For example, in a typical, multi-stage, centrifugal compressor, seals are provided at each axial end of a rotor shaft of the compressor, as well as between each compressor stage.

In order to avoid damage during operation of the compressor, it is necessary to provide a clearance gap between the stationary interface of the seal and the rotating interface of the seal. This clearance gap prevents a stationary portion of the seal from contacting a rotating portion of the seal, which could damage the seal or indeed the compressor. However, in order to account for transient situations, this clearance gap is often larger than required during steady state operation of the compressor.

One common transient situation is startup of the compressor. A typical compressor configuration is designed such that normal operation of the compressor does not excite any resonant modes of compressor. However, during startup, the rotational speed of the compressor increases from stationary to its operational speed, and therefore may pass through one or more rotational speeds that do excite resonant modes of the compressor. When the rotational speed of the compressor coincides with one of the resonant frequencies of the compressor, increased deflection of the rotor occurs at one or more anti-nodes.

In order to permit safe operation of the compressor, it is necessary to size the clearance gap of the seals to be sufficient to avoid contact, even when operating during such transient situations. However, this results in compressor inefficiency due to fluid escaping through the clearance gap of the seal during steady state operation. -2 -

SUMMARY

The present invention provides a rotating machine comprising a housing; a rotor configured for rotation relative to the housing, the rotor having a first interface; a second interface positioned adjacent the rotor so as to engage the first interface; a piezoelectric member coupled to the second interface so as to move the second interface relative to the housing when a voltage is applied across the piezoelectric member; a voltage source connected to the piezoelectric member so as to apply the voltage across the piezoelectric member; and a controller configured to control the voltage applied by the voltage source across the piezoelectric member.

With this configuration, the distance between the first interface and the second interface may be adjusted. For example, the distance may be increased in order to provide increased safety and avoid damage during transient operating conditions, and may be decreased in order to provide increased efficiency by reducing leakage during steady state operating conditions. In some embodiments, the position of the second interface may be adjusted relative to the housing, for example to realign a central axis of the second interface.

The voltage source may be configured to apply the voltage across the piezoelectric member in a substantially radial direction, i.e. relative to the axis of rotation of the rotor. For example, the voltage may be applied in a direction within 20° of the radial direction, preferably within 10° of the radial direction, and more preferably within 5° of the radial direction. The application of a voltage in a radial direction may cause the piezoelectric member to expand in the radial direction, and consequently contract in the directions perpendicular to the radial direction.

The voltage source may be connected to the piezoelectric member via first and second electrodes. The first electrode may be formed on a radially inner side of the piezoelectric member and the second electrode may be formed on a radially-outer side of the piezoelectric member. One or each of the first and second electrodes may comprise a conductive film, which is preferably a metallic film (e.g. aluminium, Duplex stainless steel).

The voltage source may be capable of applying a voltage across the piezoelectric member greater than 500 volts, preferably greater than 1000 volts.

The piezoelectric member may exhibit a relative contraction, in response to an applied voltage, of greater than 0.01%, preferably greater than 0.05%. -3 -

The second interface may be coupled to the piezoelectric member such that it deforms when a voltage is applied across the piezoelectric member, wherein the deformation may be plastic or elastic. An elastic deformation is preferred, as the second interface will then return to its neutral position when the voltage is removed.

The controller may be configured to control the piezoelectric member based on one or more of: a position of the rotor, and/or a deflection of the rotor and/or a rotational speed of the rotor.

The second interface may be a radially-inner surface of the piezoelectric member such that deformation of the second interface occurs through deformation of the piezoelectric member.

The rotating machine may comprise a position sensor for detecting a position of the rotor, preferably for detecting a transverse deflection of a rotation axis of the rotor from a reference axis. Controlling the position of the second interface based on the position of the rotor can prevent damage to the machine by ensuring that the second interface does not contact any of the rotating components of the rotating machine and/or increase efficiency of operation by minimising a clearance gap between the first interface and the second interface.

The rotating machine may comprise a speed sensor for detecting a rotational speed of the rotor. The controller may be configured to provide predictive control of the voltage based on the rotational speed of the rotor. For example, upon start-up, the rotational speed of the rotor may excite a resonant mode of the rotor. Commonly this may be at a frequency between 3000rpm and 5000rpm. Thus, the controller may detect that the speed of the rotor is approaching a predetermined threshold, and adjust the voltage accordingly. For example, the distance between the first and second interface can be increased to improve machine dynamics and safety.

The piezoelectric member may have a tubular or partially-tubular shape. The second interface may comprise an annular or partially-annular body, which may be within the tubular shape of the piezoelectric member. The second interface may be a radially-inner surface of the tubular shaped piezoelectric member.

The second interface may comprise a plurality of second interface portions, and/or the piezoelectric member may be at least one of a plurality of piezoelectric members arranged so that each piezoelectric member may respectively move a second interface portion relative to the housing. -4 -

The plurality of second interface portions may be axially offset from one another. That is to say, axially offset with respect to the rotational axis of the rotor. The axial offset may, for example, permit a different clearance gap at one axial end compared to the other axial end. This arrangement may provide improved sealing as well as rotordynamic performances (i.e. stability) in some situations.

The plurality of second interface portions may additionally, or alternatively, be circumferentially offset from one another. This allows for control of the centre of the second interface. For example, in the case of a seal, this may allow for adjustment where the rotor is not correctly centred with the second interface. In the case of a bearing, this may permit re-centring of the rotor by adjusting the centre of the bearing.

The plurality of second interface portions may have different diameters and/or may have different distances between the first interface and second interface when the voltage across the plurality of piezoelectric members is zero.

Each of the plurality of piezoelectric members may be connected to the voltage source, preferably such that a voltage applied across each of the plurality of piezoelectric members may be controlled independently.

The voltages applied across each of the plurality of piezoelectric members may be of different magnitudes and polarities.

The voltages applied across each of the plurality of piezoelectric members may be chosen in order to control any one or more of the axial displacement of the second interface, the radial displacement of the second interface and the lateral displacement of the second interface.

The first interface may be a first seal surface, and the second interface may be a second seal surface. Together, the first seal surface and the second seal surface engage to form a seal. The seal may be one of a labyrinth seal, wear ring or a pocket damper seal (i.e. honeycomb, hole pattern, etc.).

The second seal surface may have a first configuration in which it provides a relatively wide clearance gap with the first seal surface, and a second configuration in which it provides a relatively low clearance gap with the first seal surface. The at least one piezoelectric member may be configured to move the second seal surface from the first configuration to the second configuration, and/or from the second configuration to the first configuration. The advantage of being able to move a seal to a wide-clearance configuration would be to prevent contact of the first and second seal interfaces, thus reducing the likelihood of damage to the seal or the -5 -machine and increasing safety. The advantage of being able to move a seal to a low-clearance configuration would be to limit gas loss and thus increase efficiency. Typically, a wide-clearance configuration would be used during transient situations, such as start-up and shut-down of the rotating machine, whilst the low-clearance configuration would be used during steady state operations.

The seal may be a seal at a balance piston of a compressor. The piezoelectric member may be configured to cause the second seal surface to take a convergent configuration in which it provides a relatively wide clearance gap with the first seal surface at a high pressure side of the seal and a relatively low clearance gap with the first seal surface at a low-pressure side of the seal. The piezoelectric member may be configured to cause the second seal surface to take a divergent configuration in which it provides a relatively low clearance gap with the first seal surface at a low pressure side of the seal and a relatively wide clearance gap with the first seal surface at a high pressure side of the seal.

When the seal at the balance piston is in a convergent configuration, the seal coefficient may influence the resonant properties (stiffness) of the rotor and the critical speed of the rotor of the compressor may be increased.

When the seal at the balance piston is in a divergent configuration, the seal coefficient may influence the resonant properties (stiffness) of the rotor and the critical speed of the rotor of the compressor may be decreased.

The controller may be configured to control a degree of convergence and/or divergence of the first seal surface and the second seal surface. The controller may be configured to control the degree of convergence and/or divergence to minimise vibration of the rotating machine, or to optimise one or both of a stiffness of the rotor and a critical speed of the rotating machine.

In another embodiment, the first interface may be a first bearing interface, and the second interface may be a second bearing interface. Together, the first bearing interface and the second bearing interface may form a bearing.

The bearing may be a hydrodynamic bearing, such as a foil bearing, a tilting-pad bearing, or may be a rolling-element bearing. These bearings currently have no way of aligning or moving a rotational axis of the rotor. However, application of the above described arrangement to these types of bearing permits slight adjustments to be made during operation.

The controller may be configured to control the piezoelectric member based on a load profile around the circumference of the bearing. Even distribution of the -6 -load around the circumference of the bearing has the advantage of increasing bearing lifetime.

The rotating machine may comprise at least one strain and/or temperature detector to detect the load profile around the circumference of the bearing.

The bearing may be a magnetic bearing, preferably wherein the magnetic bearing comprises a plurality of electromagnets.

The controller may be configured to control the plurality of electromagnets of the magnetic bearing in addition to the piezoelectric member.

The controller may be configured to control the voltage applied across the piezoelectric member based on detection or prediction of a transient state of operation of the rotating machine.

The controller may be configured to control the voltage applied across the piezoelectric member based on detection or prediction of a failure condition of operation of the rotating machine, e.g. slugging. Detection and/or prediction of the above states prevent the rotating machine from being damaged and increase lifetime.

The piezoelectric member may comprise a monolithic ceramic material. The piezoelectric member may comprise a perovskite powder, wherein the perovskite may include PZT (Lead Zirconate Titanate).

The rotating machine may comprise a compressor. The compressor may comprise a multi-stage centrifugal compressor. The compressor may be suitable for compressing natural gas, preferably unrefined natural gas. The rotating machine may be capable of continuous operation for at least six months, preferably at least twelve months, more preferably at least twenty-four months. Commonly, such machine are capable of running continuously for many years, except in the case of breakdown, such as caused by fouling or machine dynamics problem.

The rotating machine may comprise a pump, a turbine, a turboexpander and/or any other type of rotating machine.

The present invention further provides a method of operating a rotating machine, comprising rotating a rotor of the rotating machine, the rotor having a first interface; and controlling a position of a second interface that is engaged with the first interface by applying a voltage across a piezoelectric member coupled to the second interface.

The rotating machine may be a rotating machine as described above, optionally including any one or more or all of the optional features thereof. -7 -

The rotating may comprise rotating the rotor from stationary to an operational rotation, wherein the speed of the rotor at operational rotation may be greater than 4000 rpm.

The controlling may comprise moving the second interface between a first configuration and a second configuration. The first configuration may be where the second interface is away from the first interface. The second configuration may be where the second interface is close to the first interface.

The controlling may comprise moving the second interface before the rotating begins from stationary. By moving the second interface away from the first interface before start-up, damage from transient behaviour during start-up can be avoided.

The controlling may comprise moving the second interface before the rotor has reached operational rotation. During start-up, moving the second interface in response to vibration prevents damage, but the second interface can be moved closer if vibration is minimal to improve machine dynamics.

The controlling may comprise moving the second interface after the rotor has reached operational rotation. Once the rotor has reached steady-state operational rotation, the second interface may be moved close to the rotor to maximise the operational efficiency of the machine.

The controlling may comprise receiving rotor data from a sensor, and moving the second interface in response to the data.

The rotor data may be related to the position and/or deflection and/or rotational speed of the rotor.

The rotating machine may comprise a compressor, which may be a multi-stage centrifugal compressor. The rotating may cause the compressor to compress natural gas.

The controlling may comprise predicting or detecting a transient state of operation of the rotating machine, and moving the second interface in response to the transient state.

The controlling may comprise predicting or detecting a failure condition of operation of the rotating machine, and moving the second interface in response to the failure condition.

The controlling may comprise moving at least one of a plurality of second interface portions, wherein the plurality of second interface portions may be moved -8 -by applying a voltage across each one of a plurality of piezoelectric members to which the plurality of second interface portions are coupled.

The controlling may comprise applying a different voltage across each one of the plurality of piezoelectric members so that the plurality of second interface portions may be moved independently.

The first interface may be a first bearing interface, and the second interface may be a second bearing interface, wherein the first bearing interface and the second bearing interface comprise a bearing.

The controlling may comprise aligning the axis of the rotor and the axis of bearing.

The present invention further provides a method of removing fouling on at least one of a first interface and a second interface of a rotating machine, the first interface being coupled to a rotor of the rotating machine and the second interface being positioned to engage the first interface, the method comprising causing the second interface to vibrate by supplying an oscillating voltage to a piezoelectric member coupled to the second interface. Being able to run the piezoelectric member at high (ultrasonic) frequencies, and clean parts of the rotating machine without the intervention of an operator, not only reduces costs and increases efficiency but increases the time that the machine is operational.

The oscillating voltage may be applied at a frequency of at least 1Hz, preferably at least 20kHz, and more preferably at least 40kHz.

The fouling may comprise salts, such as glycol salts.

The first interface may be a first seal surface, and the second interface may be a second seal surface, wherein the first seal surface and the second seal surface engage to form a seal. The seal may be one of a labyrinth seal and a pocket damper seal. There is currently no economic method available for cleaning these seals.

The bearing may be a hydrodynamic bearing (e.g. a foil bearing, a tilting-pad bearing) or a rolling-element bearing. There is currently no economic method available for cleaning these bearings.

BRIEF DESCRIPTION OF THE FIGURES -9 -

Certain preferred embodiments of the present disclosure will now be described in greater detail, by way of example only and with reference to the accompanying drawings, in which: Figure 1 shows a cross-section through a multi-stage compressor; Figure 2 shows a detail of the multi-stage compressor of Figure 1; Figure 3 shows a perspective, cut-away view of a shaft labyrinth seal for use in the multi-stage compressor of Figure 1; Figure 4 shows a cross-sectional view of the shaft labyrinth seal; Figure 5 shows a cross-sectional view of a magnetic thrust bearing for use in the multi-stage compressor of Figure 1; and Figure 6 shows a piezoelectric member for use with a seal or a bearing of the compressor of Figure 1.

DETAILED DESCRIPTION

Figure 1 shows a schematic cross-section of multi-stage compressor 100 for compressing natural gas. The multi-stage compressor forms one example of a rotating machine, but it will be appreciated that the techniques described herein are applicable to many other types of rotating machine. Such rotating machines may include pumps, turbines, turboexpanders and the like.

The multi-stage compressor 100 comprises a housing 8 which supports a rotor assembly and a system of stationary internal components.

The rotor assembly comprises a rotor shaft 4 and a plurality of impellers 1 mounted on the rotor shaft 4. Additional components that may form part of the rotor assembly include impeller spacers, seal sleeves, and a thrust balance drum (balance piston) 14.

The system of stationary internal components comprises an inlet nozzle 10, a diffuser 2 and a return channel 3 for each impeller 1, and a discharge nozzle 11. The system of internal stationary components guides the natural gas along a flow path from the inlet nozzle 10, sequentially through each of the compression stages comprising an impeller 1, diffuser 2 and return channel 3, and finally to the discharge nozzle 11. In the embodiment shown in Fig. 1 there are five compression stages, however any number of compression stages may be used.

When gas enters through the inlet nozzle 10, it is directed to an intake of a first compressor stage impeller la in a way that uniformly distributes the flow to the impeller la. Stationary inlet guide vanes can be positioned adjacent to the intake of the impeller la. Variation of the inlet guide vane angles can be employed to adjust the flow capacity of the compressor 100. The gas then exits the first compressor stage impeller la at a high velocity and enters an intake of a first compressor stage diffuser 2a.

In the diffuser 2a the gas is slowed down and the kinetic energy (high velocity) is converted into static pressure. The diffuser may be a vaned diffuser 5, a vaneless diffuser 6, or a hybrid version, such as a wedge or channel diffuser. In the embodiment shown in Fig. 1, a combination of vaned diffusers 5 and vaneless diffusers 6 is shown, but any combination of diffuser types may be used.

After exiting the first compressor stage diffuser 2a, the flow encounters a return bend, which creates a 180-degree turn in the direction of flow (i.e., from radially outward to radially inward). Following the return bend, the flow enters a first compressor stage return channel 3a that directs the flow inward to the intake of the second compressor stage impeller 1 b. The return channel 3a delivers flow uniformly to the next impeller stage lb with minimal pressure losses. Inlet guide vanes can be positioned at the exit of the return channel 3a, similar to those in the inlet nozzle 10.

The flow path the gas takes through each impeller 1, diffuser 2 and return channel 3 of each compressor stage is substantially the same as for the first stage.

The gas exits the last compressor stage impeller le at a very high velocity, which may be near the speed of sound. The last compressor stage diffuser 2e converts the high dynamic pressure (high kinetic energy) of the gas to static pressure, and discharges the gas to the discharge nozzle 11, where it is collected to be delivered to a downstream pipe. The discharge nozzle 11 may be any one of a discharge plenum, volute or a scroll.

The multi-stage compressor 100 uses a configuration of bearings and seals to operate effectively and efficiently. The seals act to prevent flow of air between stages or out of the compressor 100, and the bearings act to support the rotor assembly.

The rotor assembly is supported by radial bearings 9 and thrust bearings (not pictured) which position the rotor shaft 4 within the housing 8 and the stationary internal components. The thrust bearing is used to ensure the compressor rotor is maintained in its desired axial position.

The radial bearings 9 may be magnetic bearings or may be hydrodynamic bearings, which can be lubricated with oil, air or water. The radial bearings 9 may be sized to be large enough to support the rotor shaft 4 weight, yet small enough to operate at sufficiently low peripheral speeds required to limit operating temperature to acceptable levels.

With reference now to Figure 2, internal seals 12, 15, 16 of the compressor 100 are provided to minimise compressor leakage and internal recirculation losses.

In Fig. 2 the leakage flow paths, represented by arrows, show how these losses occur.

Some losses occur when gas escapes from the compressor 100 and can occur where there is not a gas-tight seal between the rotor assembly and the housing 8. This primarily occurs due to leakage at the balance piston 14, where one side of the balance piston 14 is at suction pressure and the other side of the balance piston 14 is at discharge pressure.

Internal recirculation losses occur when there is not a gas-tight seal between the inlet side and the outlet side of an impeller 1, as well as when there is not a gas-tight seal between the outlet of one impeller e.g. the first stage impeller la and the inlet of a subsequent impeller e.g. the second stage impeller lb. The pressure at the outlet of an impeller 1 is greater than that at the inlet of the impeller 1, such that the pressure difference causes a recirculation of gas, which reduces the flow rate at the outlet of the impeller 1. Similarly, the pressure at the outlet of a first impeller e.g. la is less than the pressure at the inlet of a subsequent impeller e.g. lb due to the conversion of kinetic energy into static pressure, such that the pressure difference causes some gas to flow back along the rotor shaft to the first impeller la. The recirculating flow and leakage flow results in recompression of the recirculated gas, reducing the efficiency of the compressor 100.

The shaft seals 12 prevent leakage between the plurality of impellers 1, the impeller eye seals 16 prevent leakage between the inlet and outlet of a single impeller 1, and the balance piston seal 15 prevents leakage across the balance piston.

These seals 12, 15, 16 are shown in Figure 2 as labyrinth seals, but may be any other type of suitable seal e.g. a pocket damper seal.

It is desired that seal clearances are as tight as practically possible to minimize leakage, but they must be set to avoid heavy rubbing (and therefore damage) with the rotor assembly. However, upon start up and shutting down of the compressor 100, the rotation of the rotor shaft 4 passes through one or more critical speeds. A critical speed is the rotational speed at which acting dynamic forces cause the rotor shaft 4 to vibrate at its natural frequency. This results in larger vibrations than usual, and this radial deviation could result in damage to seals if they have been set to a very tight clearance. The compressor 100 is typically started in a manner such that critical speeds are passed through as quickly as possible, but it is still necessary to have a greater clearance for the seals during these times. In one non-limiting example, the first critical speed of an unloaded rotor shaft 4, supported by bearings 9 at each end, may be around 4500 rpm. Large radial movements of the rotor 4 can also be caused by unforeseen transient issues and/or faults (such as slugging).

Figures 3 and 4 shows a labyrinth seal 12 suitable for use as the shaft seal 12. Whilst the following description is directed towards the shaft seal 12, it will be appreciated that a similar construction may also be applied to the balance piston seal 15 or to the impeller eye seal 16, for example. Furthermore, the following modifications could be made to any other type of seal suitable for use in a multi-stage compressor 100, such as a pocket damper seal.

The shaft seal 12 is coupled to the housing 8 of the compressor 100 and is positioned adjacent to the spinning rotor 4. The rotor 4 has a first interface 24 that is configured to engage with a second interface 18a of the labyrinth seal. The first interface 24 and second interface 18a define a tortuous clearance gap between them. The width of the clearance gap is preferably optimised for operational efficiency, but taking into account damage avoidance. In the illustrated example, both interfaces 18a, 24 comprise a series of grooves configured to interlock with one another to define the clearance gap. However, in some embodiments, one of the interfaces 18a, 24 may comprise a smooth surface.

As shown in Figures 3 and 4, a piezoelectric member 20 is positioned between the second interface 18a and the housing 8. The piezoelectric member 20 is directly coupled to the second interface 18a. The piezoelectric member 20 is positioned on the opposite side of the second interface 18 in a radial direction from the rotor shaft 4. In the illustrated embodiment, the second interface 18a is formed as a component separate from the piezoelectric member 20, for example from a non-piezoelectric material. However, in an alternative arrangement, the second interface 18a may comprise a radially-inner surface of the piezoelectric member 20.

The piezoelectric member 20 has a thin-walled tubular shape. The piezoelectric member 20 is metalized on the inner and outer surfaces. The inner and outer surfaces are connected via electrodes to a controller 26, which is configured to apply a voltage radially across the piezoelectric member 20. When a voltage is applied radially across the walls of the piezoelectric member 20, the piezoelectric member 20 will locally expand in the radial direction and locally contract in the circumferential and axial directions. This causes in an increase in the wall thickness and a decrease in circumferential length of the piezoelectric member 20, resulting in a contraction of the internal diameter of the piezoelectric member 20. The size of the contraction is proportional to the magnitude of the applied voltage.

Some piezoelectric materials may be sensitive to the polarity of the applied voltage, such that a piezoelectric member 20 formed from such a material may permit the internal diameter to also expand in response to certain applied voltages. However, the preferred embodiment uses a piezoelectric material that responds independently of the polarity of the applied voltage.

In one example, the tubular piezoelectric member 20 has a diameter 10mm, a wall thickness 1mm and a length of 20mm. A maximum operating voltage of the piezoelectric member 20 is 1000V, i.e. such that the applied field strength is 1kV/mm. The converse piezoelectric effect shows a relative contraction of approximately 0.05%, giving a circumferential contraction of approx. 15pm. This results in a radial contraction of 4.7pm. The controller 26 can therefore tightly control the diameter of the piezoelectric member 20.

The piezoelectric member 20 is coupled to the second interface 18a such that when it contracts or expands, the second interface 18a is deformed. This deformation is relatively small and typically occurs within the elastic deformation region, and may decrease the distance (clearance gap) between the first interface 24 and the second interface 18a.

The advantage of the present embodiment is that the controller 26 may therefore change the clearance of the shaft seal 12, decreasing the distance in order to increase efficiency or increasing the distance in order to prevent contact and damage.

In Figure 5, a second embodiment is shown where a piezoelectric member 20 is used to position a bearing 9.

The bearing 9 supports the rotor 4, and in this example is a magnetic bearing 9 that supports the rotor 4 by generation of a magnetic field by a plurality of electromagnets 30. In Figure 5 two pairs of electromagnets are shown but any number of electromagnet pairs may be used. The plurality of electromagnets 30 are positioned concentrically adjacent to the rotor 4 such that the electromagnets form a second interface 18b of the bearing opposing a first interface 24 of the bearing provided by the rotor 4.

A controller 26 independently controls the strength of the magnetic field generated by each of the plurality of electromagnets 30. Thus the position of the rotor 4 within the magnetic bearing 9 can be adjusted in order to align their respective axes.

A piezoelectric member 20 is directly coupled to a radially outer surface of each of the plurality of electromagnets 30. The piezoelectric member 20 comprises a plurality of piezoelectric elements 20a-d, each one being directly coupled to an outside surface of at least one of the plurality of electromagnets 30.

The controller 26 applies a voltage across the piezoelectric member 20, and can independently control the voltage applied across each of the plurality of piezoelectric elements 20a-d. Thus by changing the relative position of each of the plurality of electromagnets 30, the relative position and alignment of the rotor 4 can be adjusted.

The controller 26 is connected to a sensor 28, from which the controller 26 receives operational information. The sensor 28 is a position sensor. There may also be more than one sensor 28. The sensor 28 records an axial position of the rotor shaft 4 and sends it to the controller 26. The controller 26 can then control the plurality of electromagnets 30 and/or the piezoelectric member 20 in order to adjust the position of the rotor 4 in response to the received data.

It should be noted that while Figure 5 shows the specific embodiment of a magnetic bearing 9, any such bearing (e.g. hydrostatic, or hydrodynamic) that is suitable to support a rotating shaft may be controlled using a similar piezoelectric member 20. The control over the bearing 9 that the piezoelectric member 20 provides, by adjusting alignment and concentricity during operation, is especially useful for those bearings that currently have fixed centres.

Furthermore, a piezoelectric member 20 composed of a plurality of separately controllable piezoelectric elements 20a-d may also be used in combination with a seal, such as the labyrinth seal discussed above. This permits not only control of the clearance gap, but also minor adjustments to the central axis of the second interface 18a of the seal.

Figure 6 shows a further embodiment of a piezoelectric member 20. The piezoelectric member 20 is divided into a plurality of piezoelectric elements 20a-h. In the embodiment shown in Figure 6 there are eight piezoelectric elements, however any number of piezoelectric elements may be used. Piezoelectric elements 20a-d are arranged in two axially-offset rings, each including four of the piezoelectric elements. Within each of the rings, the piezoelectric elements are spaced circumferentially from each other. The plurality of piezoelectric members 20a-h may each be connected to separate outputs from the controller 26, such that the controller 26 may control the piezoelectric elements 20a-h independently of one another. This arrangement of piezoelectric members can be used in any of the seals 12, 15, 16 or bearings 9 as in previously discussed embodiments, to provide more complex and/or compound adjustments.

For example, if piezoelectric elements 20c and 20f are charged and piezoelectric elements 20a and 20h are not charged, then the tube will be laterally displaced from the resulting contraction and expansion on opposing sides of the piezoelectric member 20, which allows control of the central axis of the bearing or seal.

In another example, the piezoelectric elements 20e-h, which are axially spaced from piezoelectric elements 20a-d, could be on the outer side of a seal 12, where a smaller clearance gap could be used if there is less deviation from the rotor shaft 4. Thus, a higher voltage could be applied to piezoelectric elements 20e-h than piezoelectric members 20a-d such that the clearance of the seal 12 is smaller at the end close to the piezoelectric elements 20e-h. The clearance gap can be made to vary along the entire length of the seal 12 through varying the voltage ratios applied to the plurality of piezoelectric elements 20a-h.

This arrangement of piezoelectric members 20a-h is particularly advantageous when used in the balance piston seal 15 in order to control the divergence and convergence of the balance piston seal 15.

If the piezoelectric elements 20a-d are in a wide-clearance position and piezoelectric elements 20e-h are in a low-clearance position, then the balance piston 15 will provide a convergent configuration. That is to say, a configuration in which there is a relatively wide clearance gap between the second seal surface 18a and the first seal surface 24 at the high-pressure side of the seal 15, and a relatively low clearance gap between the second seal surface 18a and the first seal surface 24 at the low-pressure side of the seal 15.

In the converse configuration, where the piezoelectric elements 20a-d are in a low-clearance position and piezoelectric elements 20e-h are in a wide-clearance position, the balance piston 15 will provide a divergent configuration. That is to say, a configuration, in which there is a relatively low clearance gap between the second seal surface 18a and the first seal surface 24 at the high-pressure of the seal 15, and a relatively wide-clearance gap between the second seal surface 18a and the first seal surface 24 at the low-pressure of the seal.

The divergent or convergent configuration of the second seal surface changes the stiffening effect applied by the balance pistol seal 15 to the rotor shaft 4 at the point of the seal, which affects the vibration modes and the critical speed of the rotor shaft 4.

In one exemplary situation, in the convergent configuration, the balance piston seal 15 may provide a stiffening effect on the rotor shaft 4 such that it acts like one of the bearings 9. This has the effect of increasing the first critical speed of the rotor shaft 4, in this example to around 7000rpm.

When in the divergent configuration, the balance piston seal 15 provides a negative stiffening effect that has the effect of lowering the critical speed of the rotor shaft 4, in this example to around 2000rpm.

By precisely controlling the stiffness of the rotor shaft 4 using the balance pistol 15, it is possible to optimise the critical speed of the rotor shaft 4 in order to minimise vibration of the rotor shaft 4. Typically, this is achieved by maximising the critical speed, i.e. in a maximally convergent configuration.

Whilst this technique is particularly prevalent at the balance pistol seal 15 because of the high pressure differential arising across the balance piston seal 15, it will be appreciated that it may be applied to any of the seals 12, 15, 16 within the compressor 100.

The piezoelectric elements 20a-h may be formed as discrete piezoelectric components, which are then connected together. Alternatively, they may be formed as single piezoelectric components having separate electrodes such that different portions of the component can be actuated independently, as the piezoelectric elements 20a-h.The arrangements of the piezoelectric member 20 of any of the previous embodiments may also provide a method of cleaning at least the second interface 18a,b by removing fouling. The nature of the fouling that occurs will depend on the particular gas being compressed the rotating machine by the compressor, but in the case of a compressor 100 compressing natural gas, it will commonly include the accumulation of salts and/or glycol.

In order to clean the second interface 18a,b, the controller 26 may be configured to apply an AC voltage across the piezoelectric member, inducing high speed relaxation/contraction of the piezoelectric member 20. The resulting vibration has a maximum frequency that is limited by the response time of the piezoelectric member 20, but this is of the order of nanoseconds. Therefore the controller 26 may control the piezoelectric member 20 to vibrate at ultrasonic (for example) frequencies in order to shake free fouling and clean the second interface 18a,b.

Claims

CLAIMS: 1. A rotating machine comprising: a housing; a rotor configured for rotation relative to the housing, the rotor having a first interface; a second interface positioned adjacent the rotor so as to engage the first interface; a piezoelectric member coupled to the second interface so as to move the second interface relative to the housing when a voltage is applied across the piezoelectric member; a voltage source connected to the piezoelectric member so as to apply the voltage across the piezoelectric member; and a controller configured to control the voltage applied by the voltage source across the piezoelectric member.
2. A rotating machine according to claim 1, wherein the controller is configured to control the piezoelectric member based on at least one of: a position of the rotor, a deflection of the rotor, and a rotational speed of the rotor.
3. A rotating machine according to claim 1 or 2, wherein the controller is configured to control the piezoelectric member based on detection or prediction of a transient state of operation of the compressor.
4. A rotating machine according to any preceding claim, wherein the piezoelectric member has a tubular or partially-tubular shape, and wherein the second interface comprises an annular or partially-annular body within the tubular shape of the piezoelectric member.
5. A rotating machine according to any preceding claim, wherein the second interface comprises a plurality of second interface portions, and wherein the piezoelectric member is one of a plurality of piezoelectric members arranged so that each piezoelectric member will move a second interface portion relative to the housing.
6. A rotating machine according to claim 5, wherein the plurality of second interface portions are axially offset from one another.
7. A rotating machine according to claim 5, wherein the plurality of second interface portions are circumferentially offset from one another.
8. A rotating machine according to any preceding claim, wherein the first interface is a first seal surface, and the second interface is a second seal surface.
9. A rotating machine according to claim 8, wherein the second seal surface has at least a first configuration in which it provides a wide-clearance seal with the first seal surface, and a second configuration in which it provides a low-clearance seal with the first seal surface, and wherein the at least one piezoelectric member is configured to move the second seal surface between the first configuration and the second configuration.
10. A rotating machine according to claim 8 or 9, wherein the controller is configured to control a degree of convergence and/or divergence of the first seal surface and the second seal surface in order to minimise vibration of the rotating machine.
11. A rotating machine according to any of claims 1 to 7, wherein the first interface is a first bearing interface, and the second interface is a second bearing interface.
12. A rotating machine according to any preceding claim, wherein the rotating machine comprises a multi-stage centrifugal compressor for compressing natural gas.
13. A rotating machine according to any preceding claim, wherein the piezoelectric member comprises a monolithic ceramic material.
14. A method of operating a rotating machine, comprising: rotating a rotor of the rotating machine, the rotor having a first interface; and controlling a position of a second interface that is engaged with the first interface by applying a voltage across a piezoelectric member coupled to the second interface.
15. A method of removing fouling on at least one of a first interface and a second interface of a rotating machine, the first interface being coupled to a rotor of the rotating machine and the second interface being positioned to engage the first interface, the method comprising: causing the second interface to vibrate by supplying an oscillating voltage to a piezoelectric member coupled to the second interface.