JP5207074B2 - Wind turbine blade abnormality determination method, abnormality determination device, and abnormality determination program - Google Patents

Description

  The present invention relates to an abnormality determination method, an abnormality determination device, and an abnormality determination program for a blade that rotates by receiving wind in a wind turbine.

  A wind turbine for wind power generation is constructed at a height of about 60 m to 100 m in a location with good visibility, and therefore there is a high risk of blade damage due to lightning strikes. As a method for inducing lightning strikes, a method is used in which a small metal plate called a receptor is attached to the tip of a blade of a windmill so that a lightning current can flow safely to the ground.

  However, this receptor can only handle lightning strikes to a certain extent, and infrequent occurrences of large or small lightning strikes may cause blade damage or damage to the receptor itself. Yes. When the receptor itself melts, it may not be able to cope with lightning that occurs thereafter, and lightning strikes the blade directly, leading to damage to the blade.

  Generally, damage to blades due to lightning strikes is rarely a fatal damage, but small damage (openings) due to lightning strikes can cause fatal damage to blades due to breakage due to the force of air resistance and other factors caused by blade rotation. It will be damaged. The fatal blade damage causes a great amount of damage due to the replacement of the blade and the loss of power sales opportunities during the replacement period.

  Although lightning strikes on windmills are inevitable, the discovery of blade damage caused by lightning strikes at an early and small stage is highly requested by businesses as a risk avoidance in wind power generation. At present, measures are taken to detect damage by periodic visual inspection using binoculars, but this is a work that requires skill and is not a method that anyone can easily do.

  In Patent Document 1, in a wind power generator composed of a nacelle section that houses a blade rotation mechanism and a generator, and a tower section that indicates the nacelle section, a vibration sensor that detects vibration is attached to each windmill blade of the windmill. In addition, a wind turbine blade breakage detection device having a determination means for comparing and determining vibration patterns obtained from these vibration sensors is disclosed. The windmill blade breakage detection device detects breakage of a windmill blade (blade) at an initial stage. It has been proposed to prevent breakage and scattering.

JP 2001-349775 A

  However, in the wind turbine blade breakage detection device disclosed in Patent Document 1, it is necessary to provide a vibration sensor for each wind turbine blade (blade), and the introduction cost is high. Moreover, there is a limit to the detection of breakage by detecting minute vibrations of the wind turbine blade, and an abnormality in the very initial stage cannot be detected.

  In the field where wind turbines for wind power generation are operated and operated in this way, a method that can detect the abnormality of the blades at low cost and early and reliably without requiring the skill of the operator is desired. Yes.

  For this reason, the abnormality determination method for a wind turbine blade of the present invention collects an acoustic signal below the blade, analyzes the collected acoustic signal over the frequency axis and the time axis, It is characterized in that the blade abnormality is determined by detecting the presence of a Doppler shift component in which a frequency component having a high sharpness changes with the change.

  Furthermore, in the wind power generation wind turbine blade abnormality determination method of the present invention, a case where a time transition of a frequency component having a high sharpness has a predetermined inclination is determined as a blade abnormality.

  Furthermore, in the wind power generation wind turbine blade abnormality determination method of the present invention, the predetermined inclination for detecting the presence of the Doppler shift component is changed by the rotational speed of the wind turbine.

  In addition, the abnormality determination device for a wind turbine blade according to the present invention includes a frequency-time analysis unit that analyzes an acoustic signal collected below the blade over a frequency axis and a time axis, and an analysis by the frequency-time analysis unit. It is characterized in that it includes a determination unit for determining blade abnormality by detecting the presence of a Doppler shift component in which a frequency component having a high sharpness changes with time.

  The abnormality determination program for wind turbine blades according to the present invention includes a frequency-time analysis process for analyzing an acoustic signal collected below the blade over a frequency axis and a time axis, and a frequency-time analysis process. It is characterized in that a determination process for determining blade damage is performed by detecting the presence of a Doppler shift component in which a frequency component having a high sharpness shifts with time transition in the analysis result.

The figure which shows the wind power generator windmill made into the measuring object in embodiment of this invention. The figure which shows the mode of the wind noise generation | occurrence | production from a braid | blade. The figure which shows the structure of the measuring apparatus utilized for the abnormality determination method which concerns on embodiment of this invention. The figure which shows the sound collection range in the abnormality determination method which concerns on embodiment of this invention. The figure which shows the directional characteristic of the microphone with respect to the blade in the abnormality determination method which concerns on embodiment of this invention. The figure which shows the frequency-time analysis result which concerns on embodiment of this invention. The figure which shows the frequency-time analysis result which concerns on embodiment of this invention. The figure for demonstrating the detection of the Doppler component which concerns on embodiment of this invention. The figure which shows the abnormality determination apparatus of the wind power generator windmill blade which concerns on embodiment of this invention. The figure which shows the abnormality determination apparatus of the wind power generation windmill blade which concerns on other embodiment of this invention.

  FIG. 1 is a diagram showing a side view and a front view of a wind turbine that is a measurement target in the embodiment of the present invention. The wind turbine 10 is of a propeller type including a tower 11, a nacelle 12, three blades 13a to 13c, a head 14, and receptors 15a to 15c attached to the blades. 68 [m], and the diameter of the circle drawn by the rotation of the blade 14 is 61.4 [m], which belongs to a large class in Japan.

GFRP (Glass Fiber Reinforced P) is light and strong.
Lastic: glass fiber reinforced plastic) is used, but CFRP (Carbon Fiber Reinforced Plastic) can also be used. These three blades 13 a to 13 c are connected to the head 14.

  A nacelle 12 that can be rotated in accordance with the wind direction is installed on the top of the tower 11. Inside the nacelle 12, a generator for generating electric power by the rotation of the head 14 is installed. Further, a speed increaser (speed increase gear) that increases the rotational speed of the head 14 may be provided between the head 14 and the generator. The wind power generation wind turbine 10 as the measurement target of the present embodiment employs an upwind format in which the blades 13a to 13c are arranged on the windward side and the nacelle 12 is arranged on the leeward side, but has an arrangement relationship opposite to this. It may be a downwind format. In addition, a lightning rod for protecting the anemometer from lightning strikes is installed at the top of the nacelle 12, and at the time of a lightning strike, a current is passed to the ground via a lightning conductor connected to the lightning.

  Wind power generation wind turbines are classified into two types, a constant rotation type and a variable speed rotation type, depending on the power generation method. The constant rotation type is a method of generating electricity by rotating the windmill at a constant speed regardless of the wind speed. On the other hand, the variable speed rotation type is a system that rotates at an arbitrary number of rotations at which the maximum output is obtained according to the wind speed. The same type is also adopted in this embodiment.

  This constant rotation type wind power generation wind turbine is designed to generate power at about 3 [m / s] to 25 [m / s], although the numerical value varies slightly depending on the manufacturer and model. The minimum wind speed at which power generation is possible is called the cut-in wind speed, and the maximum wind speed is called the cut-out wind speed. In the constant rotation type, the rotation speed between the cut-in wind speed and the cut-out wind speed is constant, and is approximately 20 [rotations / minute]. Some models allow the number of rotations to be varied in multiple stages.

  In actual operation, when this cut-in wind speed is reached, the windmill does not start rotating suddenly, but if there is a certain amount of air even below the cut-in wind speed, the windmill is rotated by receiving power supply and generating power. For waiting. On the other hand, when the wind speed exceeds the cut-out wind speed, the wind turbine rotation is stopped because it exceeds the allowable capacity of the generator or the like.

  Receptors 15a to 15c for preventing damage caused by lightning strikes are provided near the tips of the blades 13a to 13c. The receptor 15 is a metal plate having a diameter of about 20 [mm]. When lightning is arrested at the receptor 15, the current is sent to the ground via the lightning conductor connected to the receptor 15, the fixing part of the head 14, the metal part of the tower 11, and the grounding conductor from the bottom of the tower 11, and the blade 13 caused by lightning strike Prevent damage.

  FIG. 2 is a diagram showing how wind noise generated by the blade 13 is generated. The frequency characteristics of wind noise are observed over a wide band.

  In addition, when measurement is performed with a microphone installed at a fixed position such as below the wind turbine, the blade 13 that generates the wind noise approaches the measurement position by its rotation, and then moves away. Repeatedly, with the so-called Doppler effect, the frequency rises in the process of increasing the velocity component relative to the measurement position of the velocity vector accompanying the movement of the sound source, and decreases in the process of decreasing the speed component. Is considered to be measured.

  On the other hand, the present inventors perform repeated measurements and associate the measurement results with the state of the blade 13 to break the blade 13 itself, break the receptor 14 attached to the blade 13, or mounting failure. When an abnormality occurred in the blade 13, it was confirmed that a specific wind noise different from the normal wind noise was generated from the abnormal portion.

The wind noise from this abnormal part, like the wind noise from the entire blade 13, may have its directional characteristics behind the moving direction of the blade 13, and compared with the wind noise emitted from the entire blade 13. In this case, the level is higher and the frequency characteristic is measured as having high sharpness. And the wind noise from this abnormal part is measured with the Doppler effect by the movement of the sound source position accompanying the rotation of the blade 13 when measured at the lower position of the windmill.

  The present invention focuses on the characteristics of wind noise emitted from this abnormal part, analyzes the acoustic signal collected below the windmill over the frequency axis and the time axis, and detects wind noise from this abnormal part. After processing into an easy-to-determine product, visual inspection on the display device, or automatic diagnosis by the judgment unit and judgment processing eliminates the need for expert work, and is simple and reliable at low cost. In addition, the abnormality of the blade 13 at an early stage can be detected.

  FIG. 3 is a diagram illustrating a configuration of a measurement apparatus used in the abnormality determination method according to the embodiment of the present invention. The measuring device is composed of a microphone and an information processing device 20. As the microphone, an appropriate type such as non-directional or unidirectional can be used. A dedicated measuring instrument may be used for the information processing apparatus 20 or may be realized by executing a program on a commercially available personal computer. The measuring device can be realized in a size that can be easily carried, such as a notebook computer being used as the information processing device 20.

  The information processing apparatus 20 includes a frequency-time analysis unit 21 and a display unit 22 as main components. The frequency-time analysis unit 21 performs frequency axis conversion for each sound signal picked up by the microphone, and outputs an analysis result indicating a level distribution on the frequency axis and the time axis. Although frequency processing can be simplified by using Fourier transform, the phenomenon can be grasped more clearly by using wavelet transform with high frequency-time resolution.

  The display unit 22 is a means for displaying this analysis result to the measurer. For example, the display unit 22 displays the level distribution by the difference in color and brightness, with the horizontal axis representing the time axis and the vertical axis representing the frequency axis. Various other display forms such as a three-dimensional graph can be adopted. The measurer visually determines the analysis result displayed on the display unit 22 and determines the abnormality of the blade 13 by determining the characteristics of wind noise from the abnormal part.

  FIG. 4 is a diagram showing a preferable sound collection range, that is, a position where a microphone can be installed in the abnormality determination method of the embodiment of the present invention. Sound collection by the microphone is performed below the blade 13, but in the present invention, the sound collection range is shown in the figure in order to make it possible to detect a change in frequency due to the Doppler effect of the acoustic signal. .

  Specifically, the angle α with respect to the vertically lower side of the head 14 positioned at the center of the plurality of blades 13 is within a range of 30 [°], and is a range excluding a predetermined range behind the tower 11. By setting it as such a sound collection range, the position change of the blade 13 with respect to the sound collection position can be increased, and the Doppler shift component of the wind noise can be clearly captured. Further, by removing the predetermined range behind the tower 11 from the sound collection range, there is no concern that the wind noise is blocked by the tower 11. Since the wind power generation wind turbine 10 of the present embodiment is an upwind type, the predetermined range behind the tower 11 is set, but in the case of the downwind type, the predetermined range in front of the tower 11 is excluded from the sound collection range.

  As shown in the drawing, a preferable sound collection position (sound collection optimum position) for efficiently collecting the Doppler shift component is vertically below the rotation center of the blades 13a to 13c. In addition, considering the risk of contact with the blade 13 and the fact that sound can be easily collected, the height of the sound collection position is preferably about 1 to 2 m from the ground surface.

FIG. 5 is a diagram showing the directivity characteristics of a preferred microphone in the abnormality determination method of the present invention. As known as the Doppler effect of a sounding body that rotates, a sound source is generated at points A and C where a straight line connecting the observation position where the microphone is installed and the rotation center of the blade 13 intersect with the arc drawn by the blade 13. The original frequency f0 emitted is observed. Further, the highest frequency due to the Doppler effect is observed when there is a sound source at point D, and the lowest frequency due to the Doppler effect is observed when located at point B.

  Further, when the contact points of the arc drawn by the blade 13 with the tangent drawn from the observation point are B and D, the frequency observed by the Doppler effect is as long as the blade 13 passes from D to A through B. It goes down. On the other hand, while the blade 13 passes from B through C through D, the observed frequency increases. However, in actual measurement, since the distance from the observation position is far, the frequency rise component (generated between points B to C and D) is buried in other sounds, and the frequency fall component (down Doppler shift). Component) will be measured remarkably.

  From such a relationship, it is desirable that the directivity characteristics of the microphone include the generation range of the descending Doppler shift component. Further, as shown in FIG. 4, the optimum sound collection position is set to be vertically below the rotation center of the blade 13 because of such a relationship that the frequency transition due to the Doppler effect is the largest. Is.

  A non-directional microphone can be used. However, by adopting such a positional relationship between the directional microphone and the blade 13, noise generated in the surroundings can be suppressed, and the blade 13 can be moved away from the microphone. It is possible to reliably collect the generated Doppler shift component.

  FIG. 6 and FIG. 7 show the frequency-time analysis results of the acoustic signal actually collected at the measurement apparatus described with reference to FIG. 3 and the measurement position described with reference to FIG. 4 and FIG. In this frequency-time analysis result, time is plotted on the horizontal axis and frequency is plotted on the vertical axis, and the magnitude of the level is indicated by the difference in luminance. In the figure, a portion with a white color (high luminance) indicates a high level portion, and a portion with a black color (low luminance) indicates a low level portion.

From the overview of FIG. 6, it can be seen that three large stripes repeat approximately every 4.5 seconds. These three large stripes are due to the acoustic signals generated by the three blades 13a to 13c, and indicate that the wind turbine 10 is in a standby state of 13 [rotation / minute] at the time of this measurement. The wind power generation wind turbine measured in this embodiment is a constant rotation type, and rotates at a constant speed of about 20 [rotations / minute] during power generation.

FIG. 7 is a diagram showing the same frequency-time analysis results as in FIG. 6, but with additional explanation, a partly enlarged diagram is attached. The large stripes described in FIG. 6 can be roughly divided into two components. The first component is a component A having a generation time of about 0.5 seconds from 12 to 17 [kHz] surrounded by a frame A in the figure. This is when the blade 13 passes near the tower 11. , The sound component generated by the interaction. In the drawing, the subscripts A1 to A3 indicate that the three different blades 13a to 13c are used.

The second component is a component having a generation time of about 1.5 seconds from 9 to 13 [kHz] surrounded by a frame B in the figure. This component is a wind noise component accompanied by the Doppler effect generated from the entire blade 13, and is a component that changes as the time goes down, that is, while decreasing the frequency. Further, it is a component that occurs after a predetermined time with respect to the aforementioned component A. The number indicated by the subscript indicates that the component is generated from the same blade 13. The component of wind noise generated from the entire blade 13 appears as a component that slowly changes at a relatively low level over a relatively long time in a wide frequency range.

  Thus, the blade 13 having no abnormality is composed of the component A due to the interaction between the blade 13 and the tower 11 and the component B due to wind noise generated from the entire blade 13. On the other hand, in the blade 13 in which an abnormality has occurred, different components can be confirmed in addition to these two components. In the figure, the component that appears as a bright line that stands out in the lower left of the B2 frame is a component that occurs due to the abnormality of the blade 13.

  This bright line, that is, a component generated by a phenomenon in which a frequency component having a high sharpness shifts with time transition (a “Doppler shift component” in the present invention) is a component generated by abnormality of the blade 13 itself or the receptor 15. Normally, it is difficult to confirm with human hearing, but by performing frequency-time analysis in this way, it can be manifested as a phenomenon that is visually easy to visually recognize. In this measurement, the attachment failure of the receptor 15 was confirmed in the blade 13 generating the component. As described above, the abnormality determination method in the present embodiment analyzes the acoustic signal collected under the blade over the frequency axis and the time axis, and detects the presence of time transition of the frequency component with high sharpness, It is possible to easily determine the abnormality of the blade. As in the present embodiment, in the analysis result in which the signal level is expressed by the luminance of the image, the Doppler shift component can be confirmed as a bright line. However, when the signal level is expressed by a three-dimensional graph, it is visually observed due to the steep undulation. The existence can be confirmed.

  The Doppler shift component due to the abnormality of the blade 13 appears in a state where the frequency band differs depending on the type and state of the abnormality. By using the frequency change (slope) with respect to, it becomes possible to more accurately determine the presence or absence of a Doppler shift component.

  Since the inclination of the Doppler shift component depends on the moving speed of the blade 13 and the sound pickup position of the acoustic signal, the Doppler shift component is predetermined in the case of a constant-rotation type wind power generation wind turbine in which power generation continues at a constant speed. It has the inclination of.

  FIG. 8 is a schematic diagram for explaining the detection of the Doppler shift component, and shows a state in which frequency components having high sharpness are arranged on the time axis. As shown in this figure, for example, the inclination of the projection of the peak value of the frequency component having a high sharpness, that is, the frequency change B (inclination: B / A) with respect to the time change A depends on the speed of the blade 13. By determining whether or not it is within the predetermined range, it is possible to easily and accurately determine whether or not the time transition of the frequency component having a high sharpness is a Doppler shift component generated due to an abnormality of the blade 13. It becomes possible.

  On the other hand, when the rotational speed of the windmill is variable, such as in a standby state in a variable speed rotating wind turbine or a constant rotational wind turbine, the Doppler shift component of the wind turbine varies depending on the rotational speed. The slope will change. Therefore, it is possible to detect the presence / absence of the Doppler shift component more accurately by varying the inclination to be detected according to the rotational speed of the windmill. The rotational speed of the windmill may be detected from the main shaft connected to the head 14 of the windmill, but is detected by determining the periodicity of the component A shown in FIG. 7 from the result of the time-frequency analysis, for example. You can also.

The inclination of the Doppler shift component is also related to the sound collection position, that is, the microphone installation position. Therefore, when changing the sound collection position, the inclination to be detected is changed in consideration of the sound collection position. It is good as well. The Doppler shift component can be detected only by the number of revolutions of the windmill by setting the microphone installation position in advance or providing a margin for the detection range of the inclination to be detected. However, more accurate detection can be performed by considering the rotational speed of the windmill and the sound collection position.

  Moreover, it is good also as using whether the repetition has arisen on the same braid | blade 13 as a reference | standard which judges whether the frequency component with high sharpness changes with time changes is a Doppler shift component. It is considered that there is a low probability that abnormality will occur simultaneously on the plurality of blades 13. Therefore, it is possible to easily increase the accuracy of the determination by adding to the determination condition that the time transition of the frequency component having a high sharpness is repeatedly generated with the same blade 13. Also for the Doppler shift component of FIG. 7, it can be confirmed that the time transition of the frequency component having the same sharpness is repeatedly generated in the plurality of regions B2.

  FIG. 9 is a diagram illustrating an abnormality determination device according to an embodiment of the present invention. This abnormality determination apparatus implements the abnormality determination method described so far as an apparatus, and is configured by the information processing apparatus 20 in the present embodiment. More specifically, it includes a frequency-time analysis unit 21, a determination unit 23, a display unit 22, a communication unit 24, and the like. Further, when the information processing apparatus 20 is a general-purpose personal computer or the like, it can be realized as a program executed by the information processing apparatus 20.

  In the measurement apparatus of FIG. 3, the frequency-time analysis result is displayed on the display unit 22 and the above determination of the blade 13 is made by visual observation of the measurer, whereas in this abnormality determination apparatus, the determination unit 23 performs automatic determination. It is different in point. The determination in the determination unit 23 is executed by various methods similar to the abnormality determination method. As described above, according to the abnormality determination device that performs automatic diagnosis by the determination unit 23, it is possible to determine the abnormality of the blade 13 without human intervention.

  Although the determination result in the determination unit 23 may be displayed on the display unit 22 of the information processing device 20 and notified, the administrator possesses the communication unit 20 provided in the information processing device 20 via the communication network. The information processing terminal device may be notified by e-mail or voice message. Thus, remote management can be performed by combining communication means with automatic diagnosis, and the maintenance management cost of the wind turbine can be reduced.

  FIG. 10 is a diagram illustrating an abnormality determination device for a wind turbine blade according to another embodiment. The abnormality determination device described in FIG. 9 is different in that a plurality of wind turbines 10a to 10c are managed by one information processing management device 20 and an acoustic signal is transmitted through a communication network. It has become. The acoustic signals A1 to A3 collected by the microphones M1 to M3 are transmitted to the information processing apparatus 20 via the communication network. In the information processing apparatus 20, the frequency-time analysis unit 21 performs analysis for each acoustic signal, and the determination unit 23 determines abnormality for each of the wind turbines 10 a to 10 c. In addition, it is good also as providing the imaging | photography part which image | photographs each wind power generation windmill 10 or several wind power generation windmills 10, and transmitting simultaneously not only acoustic signal A1-A3 but image information, such as a moving image or a still image.

  When managing a plurality of wind power generation wind turbines 10a to 10c in this way, an identifier indicating the wind power generation wind turbine 10 is attached to each of the acoustic signals A1 to A3. It is possible to confirm whether an abnormality has occurred in the blades 13 of the wind turbines 10a to 10c. The identifier indicating the wind turbine 10 can be identified by mixing a predetermined frequency component in the acoustic signal in addition to information such as an identification number. According to such a configuration, when the frequency of the transmitted acoustic signal is analyzed, it is possible to confirm which wind turbine 10a to 10c has an abnormality at the same time.

Furthermore, it is good also as employ | adopting the structure which can confirm which blade 13 has abnormality among several blades 13. FIG. This configuration can be easily realized by attaching a synchronization signal indicating that the blade 13 has passed a predetermined position to an acoustic signal picked up by the microphone. Further, at least one of the plurality of blades 13 of the wind power generation wind turbine 10 is provided with a whistle-like sounding portion so that a sound having a predetermined frequency component is emitted by the rotation of the blade 13. It is possible to easily confirm which blade 13 is abnormal from a sound signal that has been sounded even in a remote place. Moreover, the wind power generation windmills 10a to 10c can be discriminated by making the predetermined frequency of the sound blowing unit different for each wind power generation windmill 10.

  As described above, according to the present invention, it is possible to detect an abnormality of a blade in a wind power generation wind turbine early and reliably at low cost without requiring skill of an operator.

  Note that the present invention is not limited to these embodiments, and embodiments configured by appropriately combining the configurations of the respective embodiments also fall within the scope of the present invention.

DESCRIPTION OF SYMBOLS 10 ... Wind-powered windmill, 11 ... Tower, 12 ... Nasser, 13 ... Blade, 14 ... Head, 15 ... Receptor, 20 ... Information processing device, 21 ... Frequency-time analysis part, 22 ... Display part, 23 ... Determination part, 24 ... Communication Department

Claims (5)

The acoustic signal is collected under the blade,
Analyzing the collected sound signal over the frequency and time axes,
An abnormality determination method for wind turbine blades characterized by detecting the presence of a Doppler shift component in which the frequency component having a high sharpness changes with time in the analysis result of the acoustic signal. . The method for determining an abnormality of a wind turbine blade according to claim 1, wherein a transition of a frequency component having a high sharpness has a predetermined inclination is determined as a blade abnormality. The wind power generation wind turbine blade abnormality determination method according to claim 2, wherein the predetermined inclination for detecting the presence of the Doppler shift component is changed according to the number of rotations of the wind turbine. A frequency-time analysis unit for analyzing an acoustic signal collected below the blade over the frequency axis and the time axis;
In the analysis result in the frequency-time analysis unit, a determination unit for determining blade abnormality by detecting the presence of a Doppler shift component in which a frequency component having high sharpness changes with time transition is provided. A wind turbine blade abnormality determination device. A frequency-time analysis process for analyzing an acoustic signal collected below the blade over the frequency axis and the time axis;
In the analysis result in the frequency-time analysis process, the determination process for determining damage to the blade is performed by detecting the presence of a Doppler shift component in which a frequency component having a high sharpness changes with time. Yes Wind power generation wind turbine blade abnormality judgment program.