Condition monitoring arrangement

Abstract

A method of monitoring the condition of a fluid-film bearing arranged to support the generator of a direct-drive wind turbine is provided. The method allows providing a hydrophone configured to convert acoustic noise to an output signal; immersing the hydrophone in the lubricating fluid of the fluid-film bearing; providing access to the hydrophone output signal at the exterior of the fluid-film bearing; and evaluating the hydrophone output signal to determine the condition of the fluid-film bearing. Disclosed embodiments further include a condition monitoring arrangement of a fluid-film bearing; and a direct-drive wind turbine comprising a fluid-film bearing and the condition monitoring arrangement.

Claims

1. A method of monitoring a condition of a fluid-film bearing arranged to support a generator of a direct-drive wind turbine, which method comprises: providing a hydrophone configured to convert acoustic noise to an output signal; immersing the hydrophone in a lubricating fluid of the fluid-film bearing; providing access to the hydrophone output signal at the exterior of the fluid-film bearing; evaluating the hydrophone output signal to determine the condition of the fluid-film bearing.

2. The method according to claim 1, wherein the step of evaluating a hydrophone output signal comprises computation of an acoustic frequency spectrum and comparison with an expected acoustic frequency spectrum.

3. The method according to claim 1, wherein an expected acoustic frequency spectrum is established for the fluid-film bearing in a pristine condition.

4. The method according to claim 2, wherein a respective expected acoustic frequency spectrum is established for each operating mode of the wind turbine.

5. The method according to claim 1, wherein the hydrophone output signal is evaluated on the basis of hydrophone data obtained from multiple equivalent wind turbine installations.

6. The method according to claim 1, wherein the hydrophone output signal is evaluated during one or more operation modes of the wind turbine.

7. The method according to claim 1, wherein the hydrophone output signal is evaluated on the basis of ancillary data collected for the wind turbine.

8. A condition monitoring arrangement of a fluid-film bearing arranged to support a generator of a direct-drive wind turbine, comprising: at least one hydrophone immersed in a lubricating fluid of the bearing and configured to convert acoustic noise to an output signal; and a data evaluation arrangement configured to receive the acoustic output signal during operation of the wind turbine and to determine the condition of the fluid-film bearing on the basis of the hydrophone output signal.

9. The condition monitoring arrangement according to claim 8, comprising a plurality of hydrophones immersed in the lubricating fluid of the bearing.

10. The condition monitoring arrangement according to claim 8, wherein the data evaluation arrangement is configured to perform trend analysis on the acoustic output signal.

11. The condition monitoring arrangement according to claim 8, comprising a number of ancillary sensors arranged to collect ancillary wind turbine data, wherein ancillary data comprises any of: momentary power output; temperature data; an oil particle count; wind speed; wind direction; rotor speed.

12. A direct-drive wind turbine comprising: a fluid-film bearing arranged between the rotor and the stator of its generator, a condition monitoring arrangement comprising: at least one hydrophone immersed in a lubricating fluid of the bearing and configured to convert acoustic noise to an output signal; and a data evaluation arrangement configured to receive the acoustic output signal during operation of the wind turbine and to determine the condition of the fluid-film bearing on the basis of the hydrophone output signal, wherein the condition monitoring arrangement is adapted to determine the condition of the fluid-film bearing using the method of claim 1.

13. The direct-drive wind turbine according to claim 12, wherein the fluid-film bearing is a hydrodynamic bearing.

14. The direct-drive wind turbine according to claim 12, wherein the fluid-film bearing is adapted for a sliding speed of at most 5 m/s.

15. A computer program product comprising: a computer program that is directly loadable into a memory module of a condition monitoring arrangement comprising: at least one hydrophone immersed in a lubricating fluid of the bearing and configured to convert acoustic noise to an output signal; and a data evaluation arrangement configured to receive the acoustic output signal during operation of the wind turbine and to determine the condition of the fluid-film bearing on the basis of the hydrophone output signal, wherein the computer program comprises program elements for performing the method according to claim 1 when executed by a processor of the condition monitoring arrangement.

16. The direct-drive wind turbine according to claim 12, wherein the fluid-film bearing is adapted for a sliding speed of at most 5 m/s.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0030] FIG. 1 shows a fluid-film bearing of the type that may be used to support the generator of a direct-drive wind turbine;

[0031] FIG. 2 shows the fluid-film bearing of FIG. 1 adapted for use with the inventive condition monitoring arrangement;

[0032] FIG. 3 shows an exemplary embodiment of the inventive condition monitoring arrangement;

[0033] FIG. 4 shows an exemplary acoustic noise signature;

[0034] FIG. 5 shows an acoustic noise signature and an acoustic noise spectrum obtained during operation;

[0035] FIG. 6 shows an exemplary set of acoustic noise signatures for multiple operating modes of a wind turbine implementing the inventive condition monitoring arrangement;

[0036] FIG. 7 shows a further exemplary embodiment of the inventive condition monitoring arrangement;

[0037] FIG. 8 shows a prior art condition monitoring arrangement.

[0038] In the diagrams, like numbers refer to like objects throughout. Objects in the diagrams are not necessarily drawn to scale.

DETAILED DESCRIPTION

[0039] FIG. 1 shows a fluid-film bearing 20 of the type that may be used to support the generator of a direct-drive wind turbine. The bearing 20 can be assumed to have a diameter of several metres. The fluid-film bearing 20 comprises an outer ring 201, an inner ring 202 and other housing parts (not shown), shaped to define an oil-filled cavity 203. The outer ring 201 can be a surface of an outer rotating ring, while the inner ring 202 can be a surface of a stationary support structure, as will be clear to the skilled person. Self-aligning axial pads 204A and radial pads 204R are arranged in the cavity 203. The drawing shows an exemplary axial pad 204A arranged between a bearing housing (not shown) and the outer ring 201 to make contact with a side face of the outer ring 201; and an exemplary radial pad 204R arranged between the inner and outer rings 201, 202 to make contact with the inward-facing surface of the outer ring 201. As the skilled person will be aware, the bearing can be quipped with any number of pads 204A, 204R, and these can be arranged in any suitable configuration.

[0040] Over the service lifetime of the bearing 20, wear and tear on the sliding surfaces (particularly the surfaces of the rotating outer ring 201) and the alignment pad surfaces can develop. Particle contamination in the lubricating oil 205 can encourage surface deterioration. It is generally not possible to detect such developments in this type of bearing using conventional condition monitoring techniques such as orbit analysis.

[0041] FIG. 2 shows the fluid-film bearing 20 of FIG. 1, adapted for the inventive condition monitoring arrangement 1. The diagram shows several hydrophones 10 arranged in the oil-filled cavity 203. In this exemplary embodiment, three hydrophones 10 are arranged equidistantly about the bearing cavity 203. Each hydrophone 10 is placed so that it does not affect the bearing's function. A hydrophone 10 detects pressure alterations in the fluid in which it is immersed, and generates a corresponding electrical signal. As illustrated in the diagram, the output signal 10D of each hydrophone 10 is brought to the exterior of the bearing 20, so that the output signal 10D can be read in a subsequent signal processing stage 11 of the condition monitoring arrangement 1.

[0042] FIG. 3 shows a simplified block diagram of the inventive condition monitoring arrangement 1 as deployed in a direct-drive wind turbine 2. The output 10D from a hydrophone is analog (in the form of a voltage or an electrical current), which is converted into a digital signal 11 out in an ADC stage 11 comprising an amplification module 113, a low pass filtering module 115 (to avoid aliasing in the signal caused by unexpected higher frequency components) and a digital sampling module 117. After analogue-to-digital conversion in this way, the ADC stage 11 presents the hydrophone information as a digital signal 11 out for a subsequent analysis stage 12.

[0043] In a fault-free bearing 20, the acoustic signal as detected by the hydrophone arrangement will have essentially the same characteristics as the acoustic fingerprint of that bearing. FIG. 4 shows an exemplary acoustic fingerprint in the form of a signature spectrum S. The signature spectrum S comprises various frequencies at certain amplitudes, and can be computed in an initial stage 121 by processing the sampled data using a suitable transformation algorithm as will be known to the skilled person.

[0044] However, if some kind of physical damage has occurred somewhere inside the bearing 20, the acoustic signal will no longer be described by the signature spectrum. FIG. 5 shows a signature spectrum S for a pristine bearing 20 and an observed acoustic noise spectrum 122 obtained for that bearing 20 at some point during its service life. The actual spectrum departs significantly from the signature spectrum S, indicating that a problem is developing in the bearing 20.

[0045] The nature of the acoustic noise in the fluid-film bearing 20 of a wind turbine 2 will change depending on the wind turbine's operating mode. The level of noise and the frequencies present in the noise can depend on whether the wind turbine is being operated in a low wind-speed mode, a maximum wind-speed mode, etc. A wind turbine can have several distinct operating modes. FIG. 6 illustrates a set of exemplary acoustic noise signatures S.sub.M1, S.sub.M2, . . . S.sub.Mn for n operating modes. These can be established in a preliminary stage, for example using a pristine bearing 20 of a new wind turbine installation, and stored in a memory 124.

[0046] Therefore, the analysis stage 12 is preferably configured to choose the most appropriate acoustic signature S from such a set S.sub.M1, S.sub.M2, . . . S.sub.Mn on the basis of operational conditions 250 of the wind turbine. Relevant parameters can be rotor speed, output power, rotor azimuth position, etc. as described above, and this ancillary information can be obtained from various sensors deployed in the wind turbine 2 as will be known to the skilled person.

[0047] The observed spectrum 122 can be analysed in a subsequent stage 123 to identify any frequency component whose amplitude is significantly different than its signature amplitude, i.e. the amplitude of that frequency component in the relevant signature noise spectrum. To this end, each frequency component of the signature spectrum S can be assigned a predetermined threshold. If one or more frequency components exceed the predetermined threshold(s), this can be indicative of deterioration of the fluid-film bearing 20. For example, in the observed spectrum 122 of FIG. 5, the amplitudes of the frequency components up to 2 Hz are significantly greater than in the signature spectrum S, and this difference can be indicative of insufficient lubrication which, if not remedied, can lead to severe damage and costly repairs.

[0048] The spectrum analysis stage 123 can then report the results of the analysis. A report 12out can contain relevant information such as which frequency component(s) exceeded a threshold. Of course, this report can be in a form that can readily be converted to a graphic representation of the momentary noise spectrum 122 of the fluid-film bearing compared to the signature noise spectrum S.

[0049] The report 12out may also include a health indicator such as a condition number. A condition number can be forwarded to the respective wind turbine controller, which can take it into consideration with other parameters (e.g. wind speed) when adjusting its power references. A condition number can also be used by the wind turbine controller to track the performance of the fluid-film bearing over time.

[0050] The signal pre-processing and analysis stages 11, 12 can be implemented locally in the wind turbine 2. Alternatively, signal pre-processing can be implemented in the wind turbine 2, and the output of the final stage can be sent to a remote application (park controller, back office etc.) configured to implement the analysis stage 12 as shown in FIG. 7. The diagram shows a hydrodynamic bearing 20 arranged between the rotor and stator at the drive end of the wind turbine generator (not shown). The condition monitoring arrangement 1 comprises a number of hydrophones 10 arranged to record acoustic signals during operation of the wind turbine 2. Acoustic noise propagating through the oil film undergoes reflections at various surfaces as explained above and is picked up by the hydrophones.

[0051] The remote application 12 computes the frequency spectrum from the digital hydrophone data 10D collected during operation of the wind turbine 2, and compares the observed spectrum 122 to a signature spectrum S. The report 12out can be used as appropriate by a control module. For example, a local or remote wind turbine controller may conclude that the wind turbine 2 should operate at reduced power output until the hydrodynamic bearing 20 can be serviced, or that the bearing 20 has a serious fault and the wind turbine 2 should be stopped until the fault can be repaired. The various components of the wind turbine 2 are controlled using references 25ref generated by the wind turbine controller 25.

[0052] As indicated above, a wind turbine is generally equipped with ancillary sensors arranged to measure various parameters during operation of the wind turbine. The diagram indicates ancillary data 250 from such sensors being input to the wind turbine controller 25 for use in the decision-making process.

[0053] FIG. 8 shows a roller bearing 4 which could be deployed between the rotor and stator of a direct-drive wind turbine with a generator that has dimensions similar to those of the embodiment described in FIG. 1. Condition monitoring of this type of bearing 4 is well established, for example by deploying sensors 45 such as case-mounted accelerometers to monitor vibration in the bearing 4. Various kinds of damage to the bearing 4, for example deterioration of the rollers 40, can be deduced (in a local or remote module 43) from easily-detected changes in the data delivered by the sensors 45, for example by changes in kinematic frequencies that are clearly defined by the bearing geometry. Damage to this type of bearing 4 is therefore easily detected and generally does not occur suddenly, but instead develops over time, giving the operator sufficient opportunity to undertake preventive measures. This prior art approach to condition monitoring is suitable for roller bearings, but is unsuitable for fluid-film bearings. This is because, unlike in a roller bearing, the force transmission path to a case-mounted sensor is highly non-linear.

[0054] Although the present invention has been disclosed in the form of preferred embodiments and variations thereon, it will be understood that numerous additional modifications and variations could be made thereto without departing from the scope of the invention. For example, while the inventive approach to condition monitoring is particularly suited to a fluid-film bearing with a large diameter and low sliding speed, it can of course be applied to a fluid-film bearing with any diameter and any sliding speed.

[0055] For the sake of clarity, it is to be understood that the use of a or an throughout this application does not exclude a plurality, and comprising does not exclude other steps or elements.