The invention relates to condition monitoring of wind turbine pitch bearings by measuring variations in distance between inner and outer rings of a pitch bearing during an angular rotation.
Wind turbine pitch bearings undergo high loadings during operation. Predicting failure of pitch bearings is problematic due to the inherent variability of operation under varying conditions. Typical usage of a pitch bearing will involve a maximum range of rotation of up to 90 degrees but, for a majority of the time a pitch bearing is operational, the amount of rotation may be much smaller, for example only a few degrees. Small repeated and unpredictable variations, for example in response to wind speed to optimise loading of the wind turbine, typically results in heavy wear that may eventually result in cracking and, in extreme cases, catastrophic failure. It is therefore important to be able to regularly monitor the condition of pitch bearings over the operational lifetime of a wind turbine. This may typically be done by regular visual inspections, checking for any signs of excessive wear or cracking and periodically refreshing lubricant, for example as disclosed in EP 2937564 B1. Excessive loading that may eventually lead to cracking can in some cases be prevented or reduced by installing a bearing compression strap, for example as disclosed in EP 3344884 B1. This will not, however, prevent or reduce wear of the internal parts of a pitch bearing, i.e. the bearings and races of the pitch bearing structure.
Measuring an amount of vibration of a pitch bearing during a pitching movement may be used to determine a condition of the pitch bearing, as for example disclosed in EP 3511562 A1. Measuring vibration or acoustic emission can, however, be complicated by sources of vibration other than those resulting from the bearings and rings themselves. It would therefore be advantageous to provide a way of monitoring the condition of a wind turbine pitch bearing that can avoid or possibly augment existing vibration measurement techniques.
In accordance with a first aspect of the invention there is provided a method of monitoring a condition of a pitch bearing of a wind turbine, the pitch bearing comprising a first ring attached to a blade of the wind turbine and a second ring attached to a hub of the wind turbine, the method comprising:
In a typical wind turbine pitch bearing, the first ring is an outer ring and the second ring is an inner ring. In alternative examples, the inner ring may be connected to the blade and the outer ring to the hub.
The displacement sensor may be mounted to measure a distance parallel to a rotational axis of the pitch bearing. The actual distance measured does not need to be exactly parallel, provided that a measurable component of the distance measured is parallel to the rotational axis. The distance measured enables a measure of how uniformly the rings rotate relative to each other, since any non-uniformity will tend to result in an axial displacement. In some cases, it may be advantageous or preferable to mount the displacement sensor to measure displacement orthogonal to the rotational axis, for example where a uniform planar surface of one of the rings is not accessible or available and a curved surface of the ring may instead only be available to measure against.
The step of recording may be carried out while a main rotor of the wind turbine is stationary, i.e. while the wind turbine is not operational. During such recording, one of the (typically) three blades of the wind turbine may have its longitudinal axis aligned vertically. In some cases, the step of recording may be carried out while the main rotor is rotating, i.e. while the wind turbine is operational. Carrying out the step of recording while the wind turbine rotor is stationary avoids any load variations resulting from rotation of the wind turbine rotor from interfering with the displacement measurements. Carrying out the step of recording while the main rotor is rotating may in some cases be useful, for example to provide more continuous monitoring of one or more pitch bearings of the wind turbine during service. An orientation of the wind turbine rotor may also be recorded during such recordings, so that the position of the blade over time can be known to allow for fluctuating loads to be taken into account. An orientation of the main rotor may be available from a controller of the wind turbine or may be measured, for example by use of an optical encoder or orientation sensor on the main rotor.
The recorded angular position may be derived from a measurement of:
Measuring the angular position from a gravity vector may for example be done by mounting an accelerometer or orientation sensor on the ring that is rotating, i.e. the ring attached to the blade. The orientation as measured may then be determined based on a known orientation of the blade relative to the horizontal. The rotational axis of the pitch bearing being measured may for example be oriented horizontally and the angular range measured by a change in gravity vector between vertical and horizontal, i.e. over a 90 degree range, representing a typical full normal operating range of a wind turbine pitch bearing.
Measuring an elapsed time and a rotation rate may alternatively be used to determine the angular position, since the starting and ending position will be known and the rate of rotation may be uniform.
An alternative measure of angular position may be provided by detecting a position of the second ring relative to the first ring, for example from an encoder on the pitch bearing. A position may in some cases be determined by detecting the passage of bolts on the pitch bearing, which will tend to be placed at regularly spaced intervals.
The method may comprise determining a variation in the measured distance over the angular range and estimating a condition of the pitch bearing based on the variation.
In accordance with a second aspect of the invention there is provided a method of determining a condition of a pitch bearing of a wind turbine, the pitch bearing comprising a first ring attached to a hub of the wind turbine, a second ring attached to a blade of the wind turbine and a rotational axis, the method comprising:
In accordance with either the first or second aspect, estimating a condition of the pitch bearing may comprise comparing the variation in the measured distance with one or more of:
A previously stored variation in measured distance may for example be one or more previously measurements carried out on the same bearing in situ, i.e. on the wind turbine, or may be a measurement carried out prior to installation. The previously stored variation may thereby provide a baseline to compare the measurement with, enabling a change over time to be detected.
A modelled variation in distance of the pitch bearing may be used for comparison instead or as well as previously stored measurements. A modelled variation may be determined from various parameters relating to the pitch bearing, for example the size and stiffness of the components and clearances between the inner and outer rings, together with known loadings resulting from rotation of the pitch bearing in situ.
A recorded variation for other pitch bearings of the same type may for example relate to other pitch bearings on the same wind turbine and/or on another wind turbine. When, for example, determining the condition of multiple pitch bearings over a number of wind turbines on a common wind turbine installation, if all are of the same type a useful measure of variation, particularly if no prior recordings are available, is to identify any particular outliers, i.e. pitch bearings having variations that are at the extreme end of variation in measured distance so that these can be investigated in more detail.
Determining a variation in the measured distance over the angular range may comprise determining a quality value from one or more of:
Each of these quality values will provide a measure of how much the displacement varies over the angular range. In general, a higher peak to peak or RMS value will tend to indicate a poorer condition. These measures may, however, in some cases not be able to identify an increase in extreme variations across the angular range, in which case a measure of deviation from a mean value may be useful. This measure may for example be kurtosis, i.e. a measure of the shape of the distribution of displacements over the angular range. Kurtosis can thereby determine whether a distribution has outlier values, which may be evident in displacements having large isolated peaks.
The quality value may be compared to a predetermined threshold value, with an increased quality value indicating a poorer condition. If the quality value exceeds the threshold, a notification output may be provided. This can, for example, be useful when analysing multiple measurements taken across a number of pitch bearings, enabling particular pitch bearings to be identified for further analysis, investigation, repair or replacement.
The method may be performed upon being triggered by an event. The event may be time-based or detection of a pitching operation of the pitch bearing.
Pitching operation of the pitch bearing may be detected by a rotation sensor configured to detect rotation of the first ring relative to the second ring.
In accordance with a third aspect, there is provided a method of determining a condition of a pitch bearing of a wind turbine, the pitch bearing comprising a first ring attached to a blade of the wind turbine and a second ring attached to a hub of the wind turbine, the method comprising:
The triggering event may be one or more of: a time-based event; a measure of rotation of the first ring; and a measure of displacement by the displacement sensor outside a preset threshold.
The set period for recording may be defined by a set time period, a number of samples or a measure of rotation of the first ring.
Other features of the first and second aspects may also be applied to the third aspect.
In accordance with a fourth aspect of the invention there is provided a computer program comprising instructions for causing a computer to perform the method according to the second or third aspects. The computer program may be stored on a non-volatile storage medium.
The invention is described in further detail below by way of example and with reference to the accompanying drawings, in which:
The recorder 101 may for example be a general-purpose computer comprising an interface arranged to receive various sensor measurements or may be a dedicated recorder arranged to receive and store measurement data and provide this data periodically to an external computer for analysis. Recording of the measurement data and analysis of the data may be carried out by the same recorder 101 or may be carried out separately, for example by transmitting or transferring recorded data to a remote computer.
The recorder 101 is connected to a displacement sensor 105, which is mounted to the outer ring 102 and positioned to detect a distance to a planar surface of the inner ring 103. The displacement sensor 105 shown in
To record an angular position of the inner ring 103 the recorder 101 may be connected to an angular position sensor 106 and/or an accelerometer or orientation sensor 107. The angular position sensor 106 may for example be a digital proximity sensor fixed to the outer ring 102 and positioned to detect each of the bolts 104 as they pass so that the angular position of the inner ring can be determined. The accelerometer or orientation sensor 107 may alternatively or additionally be used by detecting an orientation of the inner ring 103, which can also be used to determine an angular position. The accelerometer 107, or another accelerometer or acoustic sensor, may alternatively be used to measure vibration during rotation of the pitch bearing 100. The recorder 101 may alternatively be connected to receive an encoder signal indicating an angular position of the inner ring 103, which may be provided as a signal from a controller of the wind turbine. By recording signals from the displacement sensor 105 as a function of time and one or more other signals that provide an indication of an angular position of the pitch bearing as a function of time, recorded measurements can be used to determine a measure of displacement as a function of relative angular position between the inner ring 103 and outer ring 102. A vibration or acoustic emission signal may also be recorded as a function of time.
In an example measurement, each of three pitch bearings of a wind turbine were fitted with displacement sensors to measure a distance to a planar surface of the inner ring of each pitch bearing, in a similar arrangement to that illustrated in
In this case, the peak-to-peak displacement 207a is around 1.5 mm and the overall trace shows multiple peaks over the angular range. The vibration trace 206 also shows multiple peaks, notably around 50 to 70 degrees on the forward pitch angle movement and towards the end of the reverse movement, showing a series of sharp impacts spaced at around 0.3 Hz. A comparison of the two bearing vibration traces with a further pitch drive vibration trace revealed this to be indicative of a fault in the bearing rather than in the pitch drive because peaks were observed from the pitch bearing around 20 ms before that in the pitch drive.
In each of the displacement traces, a general feature is that the traces are symmetrical, i.e. show similar shape traces in each pitch direction. This may be used as a check to determine whether the displacement measurements have been performed correctly. If the displacement measurement taken over the angular range in a first direction is sufficiently close to that taken over the angular direction in a second opposite direction then the measurement may be determined to be correctly taken. An error measure may for example be determined from data series of displacement and pitch in the first and second directions to provide a displacement measurement quality value. This may for example be in the form of an R2 value, which will be closer to 1 if the displacement measurements are closely matched. A lower R2 value, for example below around 0.9, will tend to indicate a measurement error. In a general aspect therefore, a measure of symmetry of displacement as a function of angular position may be determined by comparing displacement as a function of angular position in first and second opposing directions. The measure of symmetry may for example be a measure of fit between displacement in the first and second opposing directions. A notification may be output if the symmetry, for example as measured by the measure of fit, is below a threshold value. The threshold may for example be around 0.9 of an R2 measure of fit.
A further feature to note from the displacement traces is that significant deterioration in a pitch bearing may not be evident from vibration analysis alone. In
In the analysis stage, the relevant data may first be trimmed (step 504), for example to remove excess portions of recorded data prior to and after movement of the pitch bearing. A trimming operation may be necessary when handling longer term recordings, in which pitch angle changes may be infrequent. The data may be analysed to determine at what points the pitch angle changes by more than a threshold value, for example by more than 10 degrees and this portion used for further analysis. Alternatively, some or all of the recorded data may be analysed using displacement as a function of pitch angle. If the measurements are taken while the wind turbine is operational, i.e. when the main hub is rotating, account may be taken of the varying load expected as functions of angular position and rotational speed of each blade. The angular position of the wind turbine hub may for example be determined and recorded by an encoder mounted on the hub.
The data is then analysed to calculate one or more metrics (step 505), as described above. Possible metrics may include one or more of the following:
In addition to calculating metrics based on displacement, calculations may also be taken based on a speed of rotation during data acquisition, such as measuring a variation in pitch speed (e.g. by calculating a standard deviation about a nominal speed) and comparing the pitch speed with a motor current and/or hydraulic pressure. Significant variation in a nominal pitch speed may indicate a problem with the pitch motor.
After calculation of metrics, a comparison may then be made (step 506) of the metrics with a numerical model, previous measurements on the same bearing or on another bearing of the same type. If damage is indicated (step 507), for example by the quality value indicated by the metric(s), a notification may be provided (step 508), otherwise a measurement schedule may be continued with (step 509).
An additional check may be included in step 507 if further measurement data is acquired such as vibration information. If, for example, a quality value based on the displacement measurements is over a threshold that indicates the pitch bearing is damaged, an additional check on the vibration data may be done to determine whether the vibration is also above a threshold value. A notification may be provided if both of these criteria are met. As discussed above, however, a vibration measurement may not necessarily show that a pitch bearing is damaged when the displacement measurement does show excessive displacements.
After notification, further actions may be indicated such as inspection, derating the wind turbine (to prevent further damage), stopping operation or changing the frequency of data collection if damage is accumulating but is not yet critical.
As described above in relation to the method illustrated in
In alternative implementations, a rotation sensor may be absent or not used. Instead, displacement data may be recorded upon being triggered by a time-based or action-based event as above, with a measurement recorded only in the form of displacement as a function of time. An action-based event could be triggered by a measure of displacement being detected outside of a predetermined range, which indicates either that the pitch bearing has suffered damage or that an error has occurred with the displacement sensor mounting.
In any of the implementations described herein, the output of the displacement sensor may be continuously monitored to check the status of the sensor. If, for example, the mean, median or standard deviation of a measured displacement changes by more than a predetermined amount over a predetermined time period, this may indicate a problem with the sensor. An alarm output may then be triggered, allowing the sensor to be checked and action to be taken.
In a general aspect therefore, a method of determining a condition of a pitch bearing of a wind turbine, the pitch bearing comprising a first ring attached to a blade of the wind turbine and a second ring attached to a hub of the wind turbine, may comprise:
The triggering event may be time-based, for example at regular time intervals, or may be a measure of rotation of the first ring or may be a measure of displacement by the displacement sensor outside of a preset threshold.
The set period for recording may be defined by a set time period, a number of samples or a measure of angular rotation of the first ring, for example by detecting passage of a number of features passing the rotation sensor.
Other embodiments are intentionally within the scope of the invention as defined by the appended claims.
Number | Date | Country | Kind |
---|---|---|---|
2103319.6 | Mar 2021 | GB | national |
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/GB2022/050629 | 3/10/2022 | WO |