Patent Application
20180291878
The invention relates to a system for early error detection in a drive system of a wind generator. The invention also relates to a wind generator comprising this system and to the use of this system. Furthermore, the invention relates to a method for early error detection in a drive system of a wind generator.
One of the fundamental operating components in a wind generator is a pitch drive system for adjusting the pitch of rotor blades during operation. In the pitch drive system, a pitch bearing represents one of the critical components. When the pitch drive system is subject to an error during use, this error can relate to various mechanical faults, for example bearing breakage, gear tooth damage, unbalance, misalignment or even shaft breaks. Mechanical damages can result from corrosion, friction, lack of grease, fatigue, etc. A malfunction in the pitch drive system increases the risk of blade pitch blocking, which can be critical for the wind generator. Moreover, service of the pitch drive system often results in extensive and undesired downtimes of the wind turbine.
It is therefore desirable to survey the operation of the pitch drive system. Document EP 2 329 141 B1 discloses a pitch control system, which incorporates a condition monitoring system for the pitch drives. The system applies a test signal to the blade pitch system, representing a desired pitch action. The actual pitch response, which is performed in reaction to the test signal, is subsequently detected. When the actual pitch angle differs from the command value by a certain threshold, this indicates a potential malfunction in the pitch drive. This, however, requires that there is already an initial damage of the pitch drive system. Furthermore, it may be undesirable to perform artificial test pitch actions, which are initiated by the condition monitoring system.
Detection of errors in the pitch drive system at a very early state, in particular prior to any serious damage and malfunction of the pitch drives, is therefore desirable.
It is an object of the invention to provide a system for early error detection in a drive system of a wind generator and a wind generator comprising this system, which are enhanced with respect to the technical deficiencies in the prior art. Furthermore, it is an object of the invention to provide an enhanced method for early error detection in a drive system of a wind generator and an advantageous use of the system for early error detection.
In one aspect of the invention, a system for early error detection in a drive system of a wind generator is provided. The system comprises a first rotation sensor, which is coupled to an input shaft of a drive gear or transmission stage. This first rotation sensor is configured to sense an input angle of rotation of the input shaft. Furthermore, the system comprises a second rotation sensor, which is coupled to an output shaft of the drive gear or transmission stage. This second sensor senses an output angle of the output shaft. The output angle, which is captured at the output shaft, is the response of the drive gear to the input angle.
The system according to aspects of the invention further comprises a control unit, which is configured to simultaneously capture a first and a second time dependent signal. The first time dependent signal is indicative of the input angle and is acquired at the first rotation sensor. The second time dependent signal is indicative of the output angle of the output shaft. This signal is captured at the second rotation sensor. The control unit is further configured to determine a time dependent angle difference from the first and the second time dependent signals under consideration of the transmission ratio of the drive gear or transmission stage. This time dependent angle difference Δφ(t) is then:
Δφ(t)=φIN(t)−(DR*φOUT(t),
wherein iDR is the inverted transmission ratio of the drive gear or transmission stage. In the ideal case (without any error), Δφ(t) is then supposed to be zero. Deviations from zero can be small or larger which already indicates a behavior of the transmission stage. The time dependent angle difference Δφ(t) is analyzed by the control unit, so as to perform an early error detection of the drive system.
Advantageously, the system for early error detection is capable of detecting errors in the drive gear or transmission stage at an early state, in particular prior to occurrence of serious damage. The first and the second rotation sensors are, in particular, high resolution angular sensors. These are capable of detecting the input angle and the output angle of the drive gear or transmission stage very precisely. The sensors are high precision sensors in terms of time and angle resolution. Consequently, small variations between the input and the output angle can be detected.
In contrast to conventional condition monitoring systems, the system for early error detection according to aspects of the invention dispenses with a numeric model for establishing a theoretical output signal, which—in accordance with a specific application of the system—can be matched with an actual angle, a pitch drive angle, for example. Condition monitoring systems, which apply a numerical model, always include considerable uncertainties with respect to the interpretation of the results. Advantageously, the system according to aspects of the invention operates at a significantly higher level of reliability.
Pitch drive monitoring systems, which are known from the prior art, are frequently based on a comparison between a command value and an actual value of the pitch angle. For detection of an error, these systems, however, require an initial damage in the pitch drive system. Advantageously, the system according to aspects of the invention detects errors at a significantly earlier stage. Secondary damage in the drive system, which results from initially small failures, can be avoided. Maintenance work can be optimized because the operator of the wind power plant is timely alerted.
In addition to this, the system for early error detection according to aspects of the invention can monitor the condition of the drives during normal operation of the wind generator. The system advantageously dispenses with artificial test operations, which are required in some conventional systems. These forced operations more or less disturb the production of electricity in the wind generator and interfere with the main control system. Advantageously, this can be avoided. The system for early error detection according to aspects of the invention works more or less autonomously. Its operation is independent from the main control of the drives. This simplifies the upgrade of existing wind generators.
The system according to the invention is basically applicable to any drive system of a wind generator, in particular to a pitch drive system (for adjusting the pitch angle of a rotor blade), a nacelle drive system (yaw or azimuth drive) and/or a main drive system (driving the main shaft of the wind generator). Depending on the actual drive system, the rotation sensors are coupled to input and output shafts of the drive system.
According to an advantageous embodiment of the invention, the first sensor is coupled to a driving shaft of a drive motor, for example a pitch drive motor. This drives the input shaft of the drive gear, for example a pitch drive gear. The second sensor is an angle sensor for determining an angle of a component, for example a rotor blade, which is driven by the drive system. Many wind generators include a rotation sensor for detecting the rotor angle of the drive motor. Furthermore, angle sensors are typically available to detect the actual angles of a component, for example the nacelle or the rotor blades (pitch). In other words, sensors, which are already present in many wind generators, can be used as the first and second rotation sensors. Advantageously, this simplifies the upgrade of existing wind generators still further.
In an aspect of the invention, the control unit is further configured to perform a spectral analysis, in particular an order tracking analysis (order analysis), on the amplitude of the time dependent angle difference. The input angle may serve as a basis for the order tracking analysis. This order tracking analysis results in an amplitude spectrum and in an envelope curve spectrum.
In various experiments, particular features in the spectra have been identified with certain errors in the drive system. A collection of predetermined features was established. This can be stored in a database in the control unit. The plurality of features can include for example: certain patterns of peaks, certain peak shapes or shapes of the spectrum, peaks exceeding certain thresholds or groups of peaks at certain intervals, etc. One or more features or a combination of predetermined features can be identified with a certain error, for example with a bearing failure or a shaft breakage, gearing damage, unbalance and/or a misalignment of gears in the drive gear.
Furthermore, the experiments revealed that the envelope curve spectrum in particular indicates a bearing failure or a shaft breakage. Certain features in the amplitude spectrum can be identified with a gearing damage, unbalance and/or a misalignment of gears.
In an embodiment of the invention, the control unit is configured to match at least one predetermined feature with features in the order tracking analysis spectra. In particular, an error relative to a bearing failure or a shaft breakage can be identified with a feature in the envelope curve spectrum. For determination of a gearing damage, unbalance and/or a misalignment of gears, the control unit matches at least one predetermined feature, which is assigned to one of these errors, with a feature in the amplitude spectrum.
Upon successful matching of at least one of the predetermined features with a feature in a spectrum, an error message relative to the error, which is assigned to the matched feature, is output by the control unit.
Advantageously, the system for early error detection is not only capable of detecting errors at a very early stage but is also capable of identifying the type of error. In particular for offshore wind generators, this represents valuable information. Upon entry of an error message, the service technicians can collect the potentially required spare parts to be taken to the offshore site. Since transport to offshore sites is always costly, it is advantageous if no superfluous but all necessary spare parts are transported. Savings with respect to service expenses can be expected.
In another aspect of the invention, a wind generator, in particular an offshore wind generator having a drive system, which is equipped with the system for early error detection according to aspects of the invention, is provided.
Similar advantages, which have been mentioned with respect to the system for early error detection, also apply to the wind generator. In still another aspect of the invention, a method for early error detection in a drive system of a wind generator is provided. A first time dependent signal, which is indicative of an input angle of rotation of an input shaft of a drive gear or transmission stage, is captured. Simultaneously, a second time dependent signal, which is indicative of an output angle of rotation of an output shaft of the drive gear or transmission stage, is captured. A time dependent angle difference is determined from the first and the second signals. This time dependent angle difference is analyzed to perform early error detection in the drive system.
According to an advantageous embodiment of the invention, the input angle of rotation is an angle of rotation of a rotor of a drive motor. The drive motor drives the input shaft of a drive gear. The output angle of rotation is an angle of a driven component, as for example the nacelle or a rotor blade or any component, which is driven by the drive system.
In still another embodiment of the invention, the step of analyzing the time dependent angle difference includes performing of a spectral analysis, in particular an order tracking analysis. This is performed on the amplitude of the time dependent angle difference. The order tracking analysis takes the input angle as a basis.
Predetermined features, which are indicative of a certain error, are matched with features, which reside in the spectrum of the order tracking analysis. In particular, the order tracking analysis reveals an envelope curve spectrum. At least one predetermined feature, which is assigned to an error relative to a bearing failure or a shaft breakage, is matched with a feature in the envelope curve spectrum. Furthermore, the order tracking analysis can reveal an amplitude spectrum. At least one predetermined feature, which is assigned to an error relative to a gearing damage, unbalance and/or a misalignment of gears, is matched with a feature in the amplitude spectrum.
According to an embodiment of the invention, upon successful match of at least one of the predetermined features with a feature in a spectrum, an error message relative to the error, which is assigned to the matched feature, is output.
The same or similar advantages, which have been already mentioned with respect to the system for early error detection also apply to the method for early error detection in the same or similar way and will therefore not be repeated.
Further aspects and features of the invention ensue from the following description of preferred embodiments of the invention with reference to the accompanying drawings, wherein
In
In case of a pitch drive system, the pitch drive gear 20 comprises an output shaft 22 carrying a drive bevel 24. The drive bevel 24 is configured to engage an internal gear at a blade root of the rotor blade 6 of the wind generator 2. This is to rotate the rotor blade 6 by a pitch angle φ (
The system 12 for early error detection in the drive system comprises a first rotation sensor 26 for sensing an input angle of rotation φIN(t). The input angle φIN(t) represents an input to the drive gear 20 (or transmission stage 20). By way of an example only, the first rotation sensor 26 is coupled to the input shaft 18 of the drive gear 20. In an alternative embodiment, the first rotation sensor 26*, which is shown in dashed lines, is coupled to the driving shaft 16 of the drive motor 14. The first rotation sensor 26, 26* communicates a signal, which is indicative of the input angle φIN(t) via a first data link 28 to a control unit 30. Furthermore, the system 12 for early error detection comprises a second rotation sensor 32, which is coupled to the output shaft 22 of drive gear 20. The second rotation sensor 32 acquires an output angle φOUT(t), which is communicated to the control unit 30 via the second data link 34. The data links 28, 34 can be configured according to frequently known data communication technology. It can be a wired or a wireless link.
The system for early error detection 12 is configured in that the signals for the input angle φIN(t) and the output angle φOUT(t) are acquired simultaneously. In other words, the values refer to a common time scale. Furthermore, the output angle φOUT(t) represents an output of the drive gear 20, which is performed in response to the input angle φIN(t).
In particular, the first rotation sensor 26 and the second rotation sensor 32 are high precision sensors. This applies to the resolution on a time scale and to the angular resolution.
Both input and output signals φIN(t), φOUT(t), which are captured by the control unit 30, are time dependent values. The control unit 30 is configured to calculate a time dependent angle difference Δφ(t) from the two time dependent signals. By way of an example only, the angle difference Δφ(t) can be calculated by subtracting the output angle from the input angle φIN(t). However, the output angle φOUT(t) should be corrected by the inverted transmission ratio iDR of the transmission stage 20. This generally provides that the so obtained difference Δφ(t) is:
Δφ(t)=φIN(t)−iDR*φOUT(t)
In the ideal case, Δφ(t) is then supposed to be zero. Deviations from zero can be small or larger which already indicates a behavior of the transmission stage. Any other mathematical operation, which reflects the difference between the input and the output angles can, however, also be applied.
The control unit 30 is configured to perform a spectral analysis of the amplitude of the time dependent angle difference Δφ(t). In particular, an order tracking analysis can be performed on the amplitude of the time dependent angle difference Δφ(t). This order tracking analysis applies the input angle φIN as a basis. An example for a spectrum of an order tracking analysis is shown in the simplified diagram of
The order tracking analysis comprises an envelope curve spectrum 42, which is drawn in a dashed line. Furthermore, the plot includes an amplitude spectrum 44, which is drawn as a plurality of single points.
Various experiments revealed that the amplitude spectrum 44 can be analyzed for characteristic features, for example peaks exceeding certain thresholds, regular distanced peaks, certain patterns, etc. These features are indicative of certain gear errors. The allocation between the features and the errors can be stored in a database 31 in the control unit 31.
In addition to this, the experiments revealed that features, which can be identified in the amplitude spectrum 44, mainly indicate a gearing damage, unbalance and/or misalignment of gears in the drive gear 20.
Also the envelope curve 42 was analyzed with respect to particular features. When certain features can be identified in the amplitude spectrum 42, the various practical experiments revealed that these mainly indicate a bearing failure or a shaft breakage in the drive gear 20.
Within the context of this analysis, the high resolution of the sensors 26, 32 is rather crucial. This applies to the angular resolution as well as to the time resolution. Consequently, small variations between the input angle φIN(t) and the output angle φOUT(t) can be detected. Advantageously, this enables the system 12 to perform early error detection prior to serious secondary damages, which typically result from small initial damages.
In
The condition monitoring starts (step S1) with the acquisition of a first time dependent signal, which is indicative of the input angle of rotation φIN(t) of the input shaft 18 of the drive gear 20. Simultaneously, a second time dependent signal is captured (step S2). This is indicative of an output angle of rotation φOUT(t) of the output shaft 22 of the drive gear 20. A time dependent angle difference Δφ(t) is subsequently determined, for example by subtracting the output angle φOUT(t) from the input angle φIN(t), while the output angle φOUT(t) is multiplied with the transmission ratio iDR in accordance with the above equation (step S3). A spectral analysis is performed. For example, an order spectrum is generated. The order spectrum can be determined with respect to the number of rotations of a motor, i.e. the spectrum is indicated as a function of multiples of the rotation frequency. The order spectrum can also be determined as function of multiples of a fixed rotation angle. In the present example, the amplitude of the oscillations in the time dependent angle difference Amp(Δφ(t)) is plotted as a function of the input angle αIN (step S4).
An example for an order spectrum showing the amplitude of the angle difference Amp(Δφ) as a function of the input angle φIN, is shown in
By way of an example only, there is a bearing failure at the input shaft 18 of the drive gear 20. This can be derived from the envelope curve spectrum 42, which shows maxima at regular intervals. The distance between these maxima on the abscissa corresponds with the input angle φIN of the input shaft 18. For example, an input angle φIN having the arbitrary number 1 indicates a single revolution of the input shaft 18. The peaks in the envelope spectrum 44 occur for every revolution of the input shaft 18, i.e. at approximately φIN=1.5-2.5-3.5 etc.
The amplitude spectrum 44 and the envelope curve spectrum 42 are analyzed with respect to characteristic features therein. These features can, for example, include peaks exceeding certain thresholds, peaks occurring at regular intervals, characteristic curve spectra, characteristic curve shapes, etc. The control unit 30 matches the predetermined features in the database 31 with the spectra. Various experiments revealed that the envelope curve spectrum 42, if it includes certain characteristic features, gives a strong hint towards a bearing failure or shaft breakage. In other words, if certain characteristic features can be identified in the envelope curve spectrum 42, this indicates that there is a bearing failure or a shaft breakage in the drive gear 20. Similarly, the amplitude spectrum 44 is analyzed with respect to certain features. These can be identified with, for example, gearing damage, unbalance and/or misalignment of gears in the drive gear 20, as they occur.
In the flow chart of
When no characteristic features can be identified, neither in the amplitude spectrum nor in the envelope curve spectrum (steps S61 and S62), the question is whether the condition monitoring should be continued (step S8). When the condition monitoring continues, the method follows a branch “YES” and returns to step S2. If the condition monitoring should be terminated, the method of early error detection ends in step S9.
Advantageously, the condition monitoring of drives according to aspects of the invention can be performed during normal operation of the wind generator 2. No artificial test action is required. Consequently, the system 12 does advantageously not interfere with standard operation. Furthermore, errors in the drive system can be detected at a very early stage, i.e. prior to occurrence of serious secondary damages in the system. This will likely shorten the service downtimes of the wind generator 2. The system 12 for early error detection is particularly advantageous for offshore wind generators.
In an advantageous embodiment, the drive system is a pitch drive system. The motor is a pitch drive motor and the drive gear is pitch drive gear. The driven component is then a rotor blade.
Although the invention has been described hereinabove with reference to specific embodiments, it is not limited to these embodiments and no doubt further alternatives will occur to the skilled person that lie within the scope of the invention as claimed.
| Number | Date | Country | Kind |
|---|---|---|---|
| 14181970.6 | Aug 2014 | EP | regional |
This application is the U.S. national phase of PCT/EP2015/069255, filed 21 Aug. 2015, which claims priority to EP 14 181 970.6, filed 22 Aug. 2014.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2015/069255 | 8/21/2015 | WO | 00 |
