DK201970767A1

DK201970767A1 - Masking tonal noise from a wind turbine

Info

Publication number: DK201970767A1
Application number: DKPA201970767A
Authority: DK
Inventors: Dam Madsen Kaj; Gupta Mranal; Daniel Costa Goncalves Silva Cruz Luis
Original assignee: Vestas Wind Sys As
Priority date: 2019-12-12
Filing date: 2019-12-12
Publication date: 2020-12-11

Abstract

A method of masking tonal noise from a wind turbine. A parameter is measured which is indicative of a frequency of the tonal noise. A masking noise bandwidth is obtained on the basis of the parameter, and masking noise is generated which masks the tonal noise. A bandwidth of the masking noise is based on the masking noise bandwidth. The parameter may be an operating condition of the wind turbine. Alternatively the parameter may be measured by picking up sound with a microphone, and analysing the sound to identify a tone and determine its frequency.

Description

DK 2019 70767 A1 1

MASKING TONAL NOISE FROM A WIND TURBINE

FIELD OF THE INVENTION The present invention relates to a method, and associated apparatus, for masking tonal noise from a wind turbine.

BACKGROUND OF THE INVENTION A known method of masking pure tones within sound from a noise generating source is described in US 2010/0272285. A sound exhibits a pure tone when the sound pressure in a given one-third octave band is at least five decibels above the sound level in each of two adjacent one-third octave bands One or more masking sounds are generated which are capable of masking only the pure tones.

SUMMARY OF THE INVENTION A first aspect of the invention provides a method of masking tonal noise from a wind turbine, the method comprising: measuring a parameter which is indicative of a frequency of the tonal noise, obtaining a masking noise bandwidth on the basis of the parameter; and generating masking noise which masks the tonal noise, wherein a bandwidth of the masking noise is based on the masking noise bandwidth. The invention may prevent or mitigate the introduction of unnatural noise characteristics by the masking noise. The parameter may be an operating condition of the wind turbine. For example the parameter may be a rotation speed of a rotor of the wind turbine, a power produced by a generator of the wind turbine, a speed of a cooling fan, a pitch or yaw angle of a blade of the wind turbine, or a temperature. The masking noise bandwidth may be obtained on the basis of the operating condition without determining the frequency of the tonal noise. Alternatively, the operating condition may be used to estimate the frequency of the tonal noise, and masking noise bandwidth then obtained on the basis of the estimated frequency.

DK 2019 70767 A1 2 Only a single operating condition (such as a speed of a cooling fan) may be measured and used to set the masking noise bandwidth, or multiple operating conditions may be measured and used. Alternatively the parameter may be measured by picking up sound with a microphone, and analysing the sound to identify a tone and determine its frequency. Typically the higher the frequency of the tonal noise, the higher the masking noise bandwidth.

The method may be used to mask only a single tone, or plural tones.

In the case of plural tones, the masking noise for each tone may be assigned a respective different bandwidth. For example the method may comprise: identifying plural tones; obtaining a different masking noise bandwidth for each tone (optionally on the basis of a frequency of the tone); and generating masking noise which masks the tones, wherein a bandwidth of the masking noise for each tone is based on the masking noise bandwidth obtained for that tone. Typically the higher the frequency of the tone, the higher the masking noise bandwidth.

The masking noise bandwidth may be obtained by calculation based on the parameter - for example by an equation. Alternatively the masking noise bandwidth may be obtained in other ways, for instance by a look-up table.

Optionally the method further comprises: obtaining a background masking noise level which is indicative of a level of background masking noise which partially masks the tonal noise; and calculating an additional masking noise level on the basis of the background masking noise level and a target total masking noise level, wherein a level of the masking noise is based on the calculated additional masking noise level. Thus the level of the masking noise is controlled in accordance with a predefined total masking noise level. As a result the total masking noise level may be prevented from becoming undesirably high. The additional masking noise level may be calculated on the basis of a difference between the target total masking noise level and the background masking noise level.

DK 2019 70767 A1 3 The use of a target total masking noise level may also enable the total masking noise level to be accurately controlled, reducing uncertainty during the operation of the wind turbine. This reduced uncertainty enables components of the wind turbine to be designed on the basis of the target total masking noise level.

The method may further comprise determining an operating condition of the wind turbine (for instance power, rotor speed or turbine speed) and setting the target total masking noise level on the basis of the operating condition of the wind turbine. Alternatively the target total masking noise level may be fixed, or it may vary on the basis of some other parameter such as wind speed.

The masking noise may be generated by a loudspeaker, and/or by one or more further sources.

A further aspect of the invention provides apparatus for masking tonal noise from a wind turbine, the apparatus comprising: a masking noise source; and a masking noise control system configured to: measure a parameter which is indicative of a frequency of the tonal noise, obtain a masking noise bandwidth on the basis of the parameter, and cause the masking noise source to generate masking noise which masks the tonal noise, wherein a bandwidth of the masking noise is based on the masking noise bandwidth.

The apparatus of the further aspect may be configured to mask tonal noise from the wind turbine by a method according to the first aspect of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS Embodiments of the invention will now be described with reference to the accompanying drawings, in which: Figure 1 shows a wind turbine system with a loudspeaker for generating additional masking noise; Figure 2 shows a sound pressure time series recorded by the microphone; Figure 3 shows a sound pressure level spectrum for a 10s time period, including tonal noise;

DK 2019 70767 A1 4 Figure 4 shows tonal noise in the spectrum of Figure 3, and background masking noise in a pair of critical bands; Figure 5 shows the spectrum of Figure 4 with additional masking noise added; Figure 6 shows background masking noise energy levels and a pair of target masking level curves; Figure 7 shows the spread of the background masking energy; Figure 8 is a pair of graphs contrasting high masking and low masking; Figure 9 shows the graphs of Figure 8 with tonal noise added; Figure 10 shows a low background masking noise with a high additional masking noise; — Figure 11 shows total masking based on Figure 10; Figure 12 shows a high background masking noise with a low additional masking noise; Figure 13 shows total masking based on Figure 12; and Figure 14 shows an additional masking noise with a selected bandwidth. DETAILED DESCRIPTION OF EMBODIMENT(S) Figure 1 shows a wind turbine system including a horizontal axis wind turbine 1. The wind turbine 1 comprises a tower 2 supporting a nacelle 4 to which a rotor is mounted. The rotor comprises a plurality of wind turbine blades 3 that extend radially from a central hub. In this example, the rotor comprises three blades 3. The system includes an International Electrotechnical Commission (IEC) microphone 5 which records a time history of sound pressure as shown in Figure 2. A ten second period 10 of this time history is shown in the lower part of Figure 2.

Figure 3 is a spectrum of sound pressure versus frequency, presented as a bar graph with each bar representing sound pressure level (in dB(A)) in a respective frequency range. The range of each bar may be the same for each bar, or each bar may represent a one-third octave band. Typically each bar covers a range of 1-2 Hz, although for ease of illustration Figure 3 gives a lower resolution example in which each bar represents a range of the order of 10 Hz. The spectrum of Figure 3 is based on the ten second time period 10 from Figure 2, and includes a tonal noise 20 - i.e. a pure tone which occupies a narrow frequency range. If each bar represents a 10 Hz range, then the tonal noise 20 in this example may be in the 300-310Hz frequency range. Figure 4 shows the tonal noise 20, and background

DK 2019 70767 A1 masking noise 21, 22 in a pair of critical bands 21a, 22a on either side of the frequency of the tonal noise 20. In this example the critical bands 21a, 21b each have a range of about 30 Hz, but in other embodiments the widths of the critical bands 21a, 21b may be higher or lower.

5 The background masking noise 21, 22 only partially masks the tonal noise 20, so the tonal noise 20 may be annoying.

The background masking noise 21, 22 may originate from a number of sources, including aero-acoustic noise from the rotor, ambient noise (for example seasonal noise from flora and fauna), or noise from wind turbine auxiliaries such as cooling fans. The level of the background masking noise 21, 22 in the critical bands 21a, 21b can thus vary depending on a number of factors, including wind speed, turbine operating conditions, time of year and so on. Figure 6 is a graph illustrating how the level of the background masking noise 21, 22 in the critical bands 21a, 21b can vary with respect to turbine speed (in rpm).

Each data point on Figure 6 represents an energy level (in dB(A)) obtained by summing the six bars in Figure 4 associated with the background masking noise 21, 22. This gives an indication of the energy or power level of the background masking noise in the critical bands 21a, 21b. Note that the energy level in the 10Hz range containing the tonal noise 20 may be excluded from the sum.

The masking energy levels shown in Figure 6 may be determined according to the International Standard IEC 61400-11 Ed.3, section 9.5.6, pg 37. Alternatively the masking energy levels may be determined in any other suitable way.

Two exemplary data points 30, 31 are labelled in Figures 6 and 7, each associated with the same turbine speed S. The data point 30 has a higher energy level than the data point 31. As explained above, the difference in these energy levels may be due to a number of factors including operating conditions of the wind turbine (for instance yaw angle, blade pitch angle, time of year and so on) and wind conditions such as wind speed.

DK 2019 70767 A1 6 As shown in Figure 7, there is a large spread in the energy level of the background masking noise, generally varying between an upper boundary 34 and a lower boundary

35. This spread has a significant effect on the masking effect as shown in Figures 8 and 9.

Figure 8 contrasts a spectrum of a low background masking noise 40 (which peaks at a relatively low sound pressure level 42) with a spectrum of a high background masking noise 41 (which peaks at a relatively high sound pressure level 43). Figure 9 adds a tonal noise 50 which peaks at a level 51. As can be seen in Figure 9, the tonal noise 50 is masked more by the high background masking noise 41 than by the low background masking noise 40. A method of masking tonal noise will now be described with reference to Figures 4 and

5.

First, a frequency of the tonal noise 20 is identified. The tone may be a stationary tone, or its frequency may vary depending on turbine speed. Next a background masking noise level is obtained, which is indicative of an energy level of the background masking noise 21, 22 summed over the critical bands 21a, 22a adjacent to the tonal noise 20 on either side. The background masking noise level may be obtained by estimation on the basis of a time of year and on the basis of various operating conditions of the wind turbine. For example the data point 30 may be obtained on the basis of a known rotor speed, pitch angle, yaw angle and time of year. This may be achieved by inputting the time of year and operating conditions into a look-up-table, or by inputting them into a theoretical model which obtains the background masking noise level by simulation. Alternatively the background masking noise level may be obtained by measuring sound in a vicinity of the wind turbine with a sensor, for instance the microphone 5. Next, an additional masking noise level is obtained on the basis of the background masking noise level and a target total masking noise level. Figures 6 and 7 show a target curve 32 indicating how the target total masking level may increase with turbine speed. Optionally the target total masking noise level may plateau at high turbine

DK 2019 70767 A1 7 speeds, as indicated by an alternative target curve 33 in Figure 6. The target curve 32, 33 is predefined, and data recording the target curve (i.e. the relationship between turbine speed and target total masking level) is stored in a memory - for instance in the form of a lookup table.

The target curve 32, 33 could be generated based on simulations and measured experience. In this example, the target total masking level is defined by only a single parameter: the turbine speed. Other parameters, such as tone line level, could govern the masking energy target curve (i.e. the other parameters could be part of the lookup table). The target total masking level is set by determining the turbine speed, and setting the target total masking noise level accordingly. So for the turbine speed S associated — with data points 30 and 31, a target total masking noise level Lpn (target) is read from the lookup table. The additional masking noise level is then calculated on the basis of a difference between the target total masking noise level and the background masking noise level, by the formula: Lpn (additional) = Lpn (target) - Lpn (ambient + WTG) where Lpn (additional) is the additional masking noise level, Lpn (target) is the target — total masking noise level, and Lpn (ambient + WTG) is the background masking noise level. The background masking noise includes ambient noise, together with masking noise coming from the wind turbine itself. Once Lpn (additional) has been calculated, then additional masking noise is generated accordingly which further masks the tonal noise 20. An example is shown in Figure 5. Figure 4 shows a spectrum without additional masking noise, and Figure 5 shows a spectrum with the additional masking noise added. It can be seen that the masking noise 21b, 22b in Figure 5 further masks the tonal noise 20 so that the tonal noise 20 no longer stands out.

DK 2019 70767 A1 8 An energy level of the additional masking noise, summed over the pair of critical bands, is based on the calculated additional masking noise level Lpn (additional), so that the total energy level of the masking noise 21b, 22b in the critical bands 21a, 22a is equal to Lpn (target).

As shown in Figure 7 for the two data points 30, 31, the level 30a of the additional masking noise associated with the data point 30 is less than the level 31a of the additional masking noise associated with the data point 31.

— The additional masking noise may be generated by a loudspeaker 101 shown in Figure 1, and/or by one or more further sources. Such further sources may include the wind turbine rotor (the noise of which can be controlled by controlling the rotor speed and blade pitch); static aero-dynamic devices on the blades 3 such as vortex generators, or dynamic aerodynamic devices on the blades 3 such as flaps.

Figure 10 shows an example of low background noise 40, tonal noise 50 and high additional masking noise 61. Figure 11 shows the total masking noise 62 generated by the sum of the low background noise 40 and the high additional masking noise 61.

Figure 12 shows an example of high background noise 41, tonal noise 50 and low additional masking noise 63. Figure 13 shows the total masking noise 64 generated by the sum of the high background noise 41 and the low additional masking noise 63. Since the background noise 41 of Figure 12 has a higher energy level than the background noise 40 of Figure 10, the additional masking noise 63 has a lower energy level than the additional masking noise 61.

A Gaussian masking spectrum profile is scaled as required, so that the energy level of the additional masking noise in the critical bands is approximately Lpn (additional). The additional masking noise 61, 63 is then generated in accordance with the scaled masking spectrum profile.

The bandwidth of the additional masking noise 61, 63 is also based on the frequency of the tonal noise 20, 50. By way of example, the bandwidth of the additional masking noise 61, 63 may be based on equation [30] from the International Standard IEC 61400-11, Edition 3, which is:

DK 2019 70767 A1 9 i Tor HE. Critical bandwidth = 2581758. 1+1,4./— 4 | 1000 | | For fe > 70Hz, where f. is the frequency of the tonal noise.

So for all tone frequencies above 70Hz, the bandwidth of the additional masking noise 61, 63 may be set to be equal to the critical bandwidth in the equation above. Thus the higher the frequency of the tonal noise f., the higher the bandwidth of the additional masking noise.

For all tone frequencies less than 70Hz, the critical bandwidth may be fixed at 100 Hz (the additional masking noise extending from 20Hz to 120Hz).

The equation above defines a critical bandwidth, and the bandwidth of the additional masking noise may be based on this critical bandwidth in a number of ways. For example, Figure 14 shows Gaussian additional masking noise 70 with a maximum at the frequency 71 of the tonal noise, a full-width half maximum (FWHM) 72 and a full bandwidth 73. The sum of the background noise and the additional masking noise 70 has a peak 74 with a peak bandwidth 75. The additional noise 70 may be generated so that either the FWHM 72 of the additional noise, the full bandwidth 73 of the additional noise, or the peak bandwidth 75, is equal to the critical bandwidth defined > above. In another example, the additional masking noise may have a square profile in the frequency domain, in contrast to the Gaussian profile of the additional masking noise 70 in Figure 14. In this example, the bandwidth of the square profile may be set to be equal to the critical bandwidth defined above.

Thus the bandwidth of the additional masking noise 61, 63, 70 is obtained by first obtaining the frequency of the tonal noise 20, 50, then determining the bandwidth of the additional masking noise 61, 63, 70 based on the frequency.

The bandwidth of the tonal noise is a crucial parameter in associating annoyance with the tone. The additional masking noise has the potential of producing unnatural noise characteristics, and defining the bandwidth of the additional masking noise (based on

DK 2019 70767 A1 10 the need) may prevent or mitigate such unnatural noise characteristics. Limiting the additional masking noise within a defined bandwidth may also limit its effect on the overall noise.

The frequency of the tonal noise may be obtained in a number of ways.

In one example, sound may be picked up by a microphone near the wind turbine (such as the microphone 5 of Figure 1) then analysed as defined by section 9.5.4 of the standard IEC 61400-11 Ed.3 to identify one or more tones and determine their frequency (or frequencies). In this example the parameter obtained by the standard is the frequency of the tonal noise - in other words it is a parameter which is directly indicative of the frequency of the tonal noise.

In other examples, rather than directly obtaining the frequency of the tonal noise 20, 50, one or more other parameters may be measured which are indirectly indicative of the frequency of the tonal noise, and these parameters used to determine the bandwidth.

By way of example, the generator speed and power production may be used as an input to get the slot passage frequency of the generator or the tooth mesh frequency of the gearbox. The cooling fan speed could also be determined on the basis of the turbine operation or on the basis of a temperature measurement. Based on the different speeds and prior knowledge about the components, the potential frequency (or frequencies) of the tonal noise can be easily determined.

As an example, for a fan with n blades and rotating with a speed of k rpm, the frequency of the tonal noise can be obtained by calculation as n x k /60.

The system of Figure 1 includes various apparatus for masking tonal noise as described above. The additional masking noise source in this example is a loudspeaker 101. A masking noise control system 100 is configured to obtain a background masking noise level which is indicative of a level of background masking noise which partially masks the tonal noise. This may be obtained from the microphone 5, another microphone on the tower or nacelle, or by estimation or simulation as described above.

DK 2019 70767 A1 11 The control system 100 is configured to calculate an additional masking noise level on the basis of the background masking noise level and a target total masking noise level, and instruct the masking noise source 101 to generate additional masking noise which further masks the tonal noise. As described above, a level of the additional masking noise may be based on the calculated additional masking noise level, and a bandwidth of the additional masking noise may be based on the frequency of the tonal noise. The control system 100 is also configured to assign a bandwidth for the additional masking noise by the methods described above. That is, the control system 100 is configured to: measure a parameter which is indicative of a frequency of the tonal noise, obtain a masking noise bandwidth on the basis of the parameter, and cause the masking noise source 101 to generate masking noise which masks the tonal noise, wherein a bandwidth of the masking noise is based on the masking noise bandwidth.

The examples given above show only a single tonal noise 20, 50 but there may be plural tones which are simultaneously masked by respective multiple additional masking noises.

In the case of plural tones, each tone may be assigned a respective different bandwidth. If plural tones are identified, then a different masking noise bandwidth may be obtained for each tone on the basis of a frequency of the tone (for instance using equation [30] from the International Standard IEC 61400-11, Edition 3). Masking noise is then generated which masks the tones. A bandwidth of the masking noise for each tone is based on the masking noise bandwidth obtained for that tone.

Although the invention has been described above with reference to one or more preferred embodiments, it will be appreciated that various changes or modifications may be made without departing from the scope of the invention as defined in the appended claims.

Claims

DK 2019 70767 A1 12 CLAIMS

1. A method of masking tonal noise from a wind turbine, the method comprising: measuring a parameter which is indicative of a frequency of the tonal noise, obtaining a masking noise bandwidth on the basis of the parameter; and generating masking noise which masks the tonal noise, wherein a bandwidth of the masking noise is based on the masking noise bandwidth.

2. The method of claim 1, wherein the parameter is an operating condition of the wind — turbine.

3. The method of claim 1, wherein the parameter is measured by picking up sound with a microphone, and analysing the sound to identify a tone and determine its frequency.

4. The method of any preceding claim, wherein the higher the frequency of the tonal noise, the higher the masking noise bandwidth.

5. The method of any preceding claim, wherein the masking noise bandwidth is obtained by calculation based on the parameter.

6. The method of any preceding claim, further comprising: obtaining a background masking noise level which is indicative of a level of background masking noise which partially masks the tonal noise; and calculating an additional masking noise level on the basis of the background masking noise level and a target total masking noise level; wherein a level of the masking noise is based on the calculated additional masking noise level.

7. The method of any preceding claim, wherein the masking noise is generated by a loudspeaker.

8. The method of any preceding claim, wherein the masking noise is generated by a loudspeaker and one or more further sources.

9. The method of any preceding claim, comprising:

DK 2019 70767 A1 13 identifying plural tones; obtaining a different masking noise bandwidth for each tone; and generating masking noise which masks the tones, wherein a bandwidth of the masking noise for each tone is based on the masking noise bandwidth obtained for that tone.

10. The method of claim 9, wherein the higher the frequency of the tone, the higher the masking noise bandwidth.

11. Apparatus for masking tonal noise from a wind turbine, the apparatus comprising: a masking noise source; and a masking noise control system configured to: measure a parameter which is indicative of a frequency of the tonal noise, obtain a masking noise bandwidth on the basis of the parameter, and cause the masking noise source to generate masking noise — which masks the tonal noise, wherein a bandwidth of the masking noise is based on the masking noise bandwidth.

12. Apparatus according to claim 11, wherein the apparatus is configured to mask tonal noise from the wind turbine by a method according to any of claims 1 to 11.

13. Apparatus according to claim 11 or 12, wherein the masking noise source comprises a loudspeaker.