A fan blade fault diagnosis system and method based on a helical waveguide tube

Abstract

The application belongs to the technical field of wind power generation, and particularly relates to a fan blade fault diagnosis system and method based on a spiral waveguide tube, which comprises the following steps: S1: obtaining sound pressure of a received sound wave at a measuring point and a receiving time, wherein the sound wave is propagated to the measuring point by different propagation paths; S2: comparing the sound pressure value with a preset threshold value; if the sound pressure value is lower than the threshold value, returning to S1; if the sound pressure value is higher than the threshold value, entering S3; S3: calculating a time difference of the sound waves propagated by different propagation paths and arriving at the measuring point; and S4: calculating three-dimensional coordinates of a sound source according to a sound wave transmission model, geometric constraints of an internal structure of the fan blade and the time difference. Compared with the prior art, the application solves the problems of the prior art, such as being easily disturbed, limited detection distance, complex structure and high cost based on acoustic signal detection. The scheme realizes low-cost detection and positioning of a fan blade fault position through a simple acoustic sensing structure.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of wind power generation, and particularly relates to a fan blade fault diagnosis system and method based on a spiral waveguide tube. BACKGROUND

[0002] As a kind of renewable clean energy, the stable operation of the equipment of wind power generation is of great significance to guarantee power output. The health status of fan blades, as an important component of wind turbines, directly affects the power generation efficiency and operation safety. Especially in large-scale wind turbines, blade structure fatigue fracture accidents often cause significant losses. Most large and medium-sized fan blades are composed of a hollow shell structure made of multiple layers of composite materials. The blade fault detection methods mainly include visual detection, vibration analysis and acoustic emission detection. Visual detection can be used to explore the cracks and surface damage on the inside and outside surfaces of the shell; vibration analysis mainly assesses the health status through blade structural dynamics characteristics and modal analysis, which has limited detection effect before the significant change of structural dynamics parameters, and is difficult to be used for early fault detection.

[0003] Acoustic emission detection can be used for early fault detection or monitoring of multi-layer structure wall damage by sensing local delamination or cracking on the blade structure. Some improvements have been made to the acoustic detection method for locating the damage of wind turbine blades in the prior art, such as:

[0004] CN116429902A discloses a fan blade multi-crack acoustic emission monitoring method and system, which determines the emission source positioning sensor array layout method according to the sound velocity distribution properties of the fan blade material; uses the sensor array with the layout method to obtain the acoustic emission signals of the fan blade; performs noise reduction processing on the acoustic emission signals obtained by each sensor, wavelet change and decomposition in the time domain, and extracts the characteristic parameters of multi-crack damage; based on the characteristic parameters, establishes an acoustic signal diagram, determines the time difference of acoustic wave arrival between each sensor, and determines the fatigue damage position of the fan blade multi-crack acoustic emission source according to the time difference.

[0005] CN113406201A discloses a wind turbine blade damage positioning detection method, which comprises the following steps: arranging multiple groups of triangularly arranged sensors on the fan blade; receiving sound signals through the multiple groups of sensors; obtaining the positioning of the blade damage or defect through the sound signals and the position coordinates of the sensors.

[0006] However, in the prior art, the fan blade crack positioning monitoring is often performed by installing multiple sets of acoustic sensors inside or on the surface of the blade, and the arrival time difference of the sound waves received by different sensors is used for triangulation positioning. For example, CN116429902A discloses a method of arranging a sensor array, in which three or four acoustic emission sensors are arranged on the target blade according to the distribution characteristics of the material sound velocity (in L or Z type arrangement), and the crack position is calculated by using the time difference of the received acoustic emission signals; CN113406201A proposes to uniformly arrange multiple sets of triangular sensor arrays on the suction surface and pressure surface of the blade, each set of three sensors being an equilateral or isosceles triangle, and the blade defect position is calculated by using the sound signal and sensor coordinates. These methods use multiple probes for time difference positioning, which seems to be able to locate the crack, but in actual fan operating conditions, the positioning effect is often difficult to achieve the expected result:

[0007] Firstly, the dispersion of the sensors brings complexity in structure and arrangement. In order to cover the entire fan blade, structural supports need to be installed at multiple places on the fan blade, which increases the complexity and difficulty of weight arrangement of the internal structure of the fan blade, and the long wiring of multiple sensors and their leads leads to an increase in system cost. Moreover, as studies have shown that the accuracy of the sensor (probe) position will seriously affect the positioning accuracy, it is necessary to accurately arrange the probes into an isosceles or equilateral triangle to obtain accurate results, but in the actual installation process, it is difficult to provide and guarantee such ideal internal space on the fan blade, and small positional deviations in the installation process will directly affect the positioning calculation accuracy. Furthermore, the fan blade is monitored in a rotating state, and the sensors arranged at different positions on the fan blade also need to consider the different additional disturbances and calibration synchronization problems caused by the rotating motion of different sensors, which may introduce timing and positioning errors.

[0008] Secondly, the non-uniformity of the fan blade material and internal structure makes the sound wave propagation path extremely complex. The fan blade is usually composed of multiple layers of composite materials, and sound wave reflection and refraction occur at the interface between each layer. At the same time, the anisotropy of the material itself also causes the sound speed to change in different directions, combined with the change in the cross-sectional thickness and chord direction of the fan blade, making the sound wave propagation speed in the fan blade not constant. In existing research, the sound speed is usually assumed to be constant when calculating the positioning, but the real wave speed field may be complex and variable. When there is a deviation between the input wave speed assumption and the actual wave speed in the material, the position error caused by the time difference positioning calculation will be amplified with the length of the propagation path, especially when the sound source is close to the probe, the error is more significant. In addition, sound waves undergo complex and uncontrollable phenomena such as attenuation, waveform change, refraction, dispersion, etc. during propagation, which will introduce unknown time delay deviations. Since the fan blade geometry is a three-dimensional curved surface, image distortion will occur when sound waves propagate along the path, further causing positioning errors. These natural propagation effects cannot be accurately compensated for by simply using time difference algorithms, so the positioning results often appear as a cluster of positioning points around the real sound source rather than an accurate point, and the error is difficult to converge. The dispersion arrangement of sensors and the corresponding mounting brackets used in existing technologies will further increase the complexity of the sound wave propagation path, thereby further exacerbating the introduction and accumulation of errors.

[0009] Thirdly, the fan operating environment noise has a significant impact on positioning accuracy. Blade vibration in the wind field and noise from the transmission mechanism, generator, and surrounding wind will produce a large amount of background sound signals, which will directly superimpose on the target sound emission signal. Literature indicates that in actual applications, cabin noise and other environmental noise have a coupling effect on sensor signals, which may cause false alarms and signal pollution. Even with signal processing methods such as filtering and gain (such as the filtering of signals below 100 Hz in CN113406201A), high-frequency mechanical vibration noise and airflow noise are still difficult to completely remove. Since the amplitude of the sound emission signal is usually low, it is easy to be mixed with interference in a noisy environment; when the signal is overwhelmed by noise or distorted, the time difference extraction error increases, ultimately leading to a shift in the positioning result. If the sound source signal itself is weak, it may not be detected by at least three probes simultaneously, and the positioning algorithm may not even converge. In addition, small differences in sensor amplifier gain, trigger threshold, and other parameters will also introduce time difference measurement errors. Generally, these microsecond-level time difference errors will bring millimeter or even centimeter-level positioning deviations when converted into distances.

[0010] The time-difference positioning-based triangulation positioning algorithm has inherent limitations. In theory, solving the time-difference equation with three probes will obtain the intersection of two hyperbolas, one of which is the real sound source and the other is a pseudo intersection, which needs a fourth probe to distinguish the true and false. Even if the number of probes is increased, it can only reduce the positioning area and is difficult to eliminate the influence of a single error source. The existing patents mostly use simple geometric formulas to solve the sound source coordinates, but this calculation is highly sensitive to the input time difference, probe coordinates and assumed sound speed. Once any of the above conditions deviates, the output position will deviate from the true value. In addition, when multiple sources are coupled, the time-difference method is prone to confusion, and the positioning accuracy depends on the geometric conditions such as probe spacing and array shape.

[0011] Finally, due to the large damping of the composite structure blade, in order to improve the fatigue strength and reduce the blade flutter, further measures to improve the structural loss factor are usually taken during the production and assembly process of the blade, resulting in a large attenuation of the propagation of structural elastic sound waves on the shell; but early cracks are often brittle medium-high frequency sound, which can only transmit a very short distance on the blade structure, and large and medium-sized generator blades are tens of meters or even hundreds of meters long. If the traditional acoustic emission technology is used, dozens or even hundreds of acoustic emission detection nodes need to be arranged on the blade, resulting in poor practicability of the traditional method.

[0012] In summary, the existing blade crack or delamination positioning technology using distributed acoustic emission sensors is limited by multiple factors such as sensor arrangement method, blade material and structural complexity, environmental noise interference, and defects in the algorithm itself, but the node detection distance is short, the cost is high, and the accuracy is low. It cannot meet the requirements of large-scale wind power plants for low-cost wide-range detection and positioning of blades. SUMMARY

[0013] The purpose of the present application is to solve at least one of the above problems by providing a fan blade fault diagnosis system and method based on a spiral waveguide tube to solve the problems of limited detection distance, poor noise interference resistance, complex sensor arrangement, high system cost, low integration, and dependence on external power supply in the existing fan blade crack acoustic detection technology. The present application realizes low-cost detection and positioning of fan blade fault positions by combining the spiral waveguide tube of the spiral structure with the acoustic sensing structure of the double acoustic sensors.

[0014] The purpose of the present application is achieved by the following technical solutions:

[0015] The first aspect of the present application discloses a fan blade fault diagnosis system based on a spiral waveguide tube, which is installed inside the blade shell. The blade shell is divided into a leading edge chamber, an intercostal chamber and a trailing edge chamber by a web. The system comprises:

[0016] at least one spiral waveguide arranged along the length of the blade shell, the spiral waveguide having a tube wall with at least eight sound wave transmission holes spaced along the length of the tube wall, the sound wave transmission holes being in communication with the leading edge cavity and / or the web cavity and / or the trailing edge cavity;

[0017] a first acoustic sensor having a receiving end arranged at one end of the spiral waveguide, for receiving sound wave signals propagating through the spiral waveguide;

[0018] a second acoustic sensor arranged within the blade shell and adjacent to the first acoustic sensor, for receiving sound wave signals propagating through the internal cavity of the blade shell;

[0019] a control component electrically connected to the first acoustic sensor and the second acoustic sensor, for obtaining the sound wave signals received by the two sensors and determining the fault location on the fan blade based on the time difference of arrival of the two signals.

[0020] Preferably, the spiral waveguide is provided with a plurality of spiral waveguides, each spiral waveguide having a different lead (axial distance) and / or spiral outer diameter (radial dimension), and each spiral waveguide is not in communication with each other and arranged in a spatially staggered manner.

[0021] A separate first acoustic sensor is arranged at one end of each spiral waveguide, and a second acoustic sensor is arranged adjacent to the first acoustic sensor; and a sound insulation structure is arranged at the intersection area of each spiral waveguide.

[0022] When a plurality of spiral waveguides are used, the sensors arranged correspondingly can also filter out interference through matching algorithms. Specifically, the matching algorithm is used to establish a corresponding relationship between candidate events in multiple acoustic channels and to eliminate interference. The matching algorithm includes but is not limited to: generalized cross-correlation time difference estimation (GCC, GCC-PHAT), template-based matched filtering, dynamic time warping-based waveform similarity matching (DTW), feature matching based on mutual power spectrum coherence or phase consistency (MSC, PLV), matching based on beamforming and spatial spectrum peak consistency (SRP-PHAT, MVDR), sparse representation-based dictionary matching (OMP / LASSO), and multi-channel consistency matching under geometric constraints (RANSAC or M-estimation) robust outlier rejection. Preferably, the matching process includes: arrival time detection and threshold pre-screening, propagation constraint gating of candidate pairs, coherence / phase consistency verification, and multi-channel robust fitting. Finally, the time difference set verified for consistency and the positioning result are output.

[0023] More preferably, when a plurality of spiral waveguides are provided, the directions of rotation of adjacent spiral waveguides are opposite.

[0024] Preferably, the shape of the inner cross-section of the spiral waveguide includes one of a circle, an ellipse and a polygon, wherein the inner diameter of the circular cross-section or the equivalent diameter of the polygonal cross-section is 2-20 mm, and the minor axis of the elliptical cross-section is at least 1 mm and the major axis is at least 2 mm;

[0025] The spiral waveguide has a spiral structure, and the spiral outer diameter (from one side of the spiral outer edge to the opposite side of the spiral outer edge) of the spiral waveguide is 3-200 mm, and the pitch is 1-12 m; and / or, the diameter of the sound wave transmission hole is 1-10 mm, and the spacing is 1-10 m.

[0026] More preferably, the major axis of the spiral waveguide with an elliptical cross-section is at least 5 mm; and the spiral outer diameter of the spiral waveguide is 20 mm.

[0027] More preferably, the cross-section of the spiral waveguide is circular, the inner diameter is 2-20 mm, and the wall thickness is 0.5-10 mm; and / or,

[0028] The spiral waveguide can be divided into dense type (pitch 1-4 m and not including 4 m, corresponding number of turns 100-25 turns and not including 25 turns), moderate type (pitch 4-8 m, corresponding number of turns 25-13 turns) and sparse type (pitch greater than 8 m, corresponding number of turns less than 13 turns); or, the spiral waveguide has a segmented structure, each segment has the same or different structure (pitch), and each segment is connected by a connecting piece.

[0029] Preferably, the inner wall surface of the spiral waveguide is provided with an acoustic reflection coating, and the reflection coefficient of the acoustic reflection coating is not less than 0.85; and the spiral waveguide is provided with sound-absorbing material at the end away from the first acoustic sensor, and the sound-absorbing coefficient of the sound-absorbing material is not less than 0.9.

[0030] Preferably, the first acoustic sensor and the second acoustic sensor are each a piezoelectric microphone or an electret microphone.

[0031] Preferably, the control assembly includes a signal processing unit, a communication module and an energy supply unit;

[0032] The signal processing unit is electrically connected with the first acoustic sensor and the second acoustic sensor, and is used to acquire the sound wave signals received by the first acoustic sensor and the second acoustic sensor and locate the fault position on the fan blade through time difference;

[0033] The communication module is electrically connected with the signal processing unit, and is used to transmit the detection result to the outside of the system;

[0034] The energy supply unit is electrically connected with the first acoustic sensor, the second acoustic sensor, the signal processing unit and the communication module, and is configured to supply power to the power-consuming units in the system; wherein the energy supply unit comprises a guide rail arranged along the radial direction of the fan blade, a mass block reciprocatingly sliding along the guide rail, a reset elastic member connected between one side surface of the mass block and one end of the guide rail, a one-way transmission mechanism, a micro generator and an energy storage module; the mass block is mechanically connected with the input part of the one-way transmission mechanism (such as a crankshaft connecting rod mechanism or other mechanism capable of converting reciprocating motion into one-way circumferential / rotary motion), and the output part of the one-way transmission mechanism is coaxially connected with the rotor shaft of the micro generator. The centrifugal force generated when the fan blade rotates drives the mass block to move along the guide rail away from the blade root, and when the rotation weakens, the mass block is returned by the reset elastic member (such as a spring), the one-way transmission mechanism rectifies the reciprocating displacement of the mass block into one-way torque to drive the micro generator to generate electricity, and the output of the micro generator is charged to the energy storage module through the power management circuit and supplies power to the first acoustic sensor, the second acoustic sensor, the signal processing unit and the communication module.

[0035] More preferably, the micro generator is one or a combination of an electromagnetic energy harvester, a piezoelectric energy harvester or a capacitive energy harvester.

[0036] The spiral transmission structure design of the spiral waveguide can significantly lengthen the path of sound wave propagation, thereby significantly improving the time difference resolution of the sound wave signals received by the two acoustic sensors. This structure design makes it easier to distinguish the sound wave signals emitted by weak defects in the time domain by increasing the difference in propagation path.

[0037] The sound waves emitted by the fault are directly propagated by the air in the fan blade and also enter the spiral waveguide inside the fan blade to propagate through the air and are transmitted to the two acoustic sensors respectively. The two sensing channels have significantly different propagation paths, making the signal paths complementary and redundant; and the two acoustic sensors built into the fan blade monitor the sound waves propagated by the two paths respectively, forming cross verification, and the overall anti-interference ability is enhanced.

[0038] Based on this, the system improves the detection sensitivity and positioning accuracy through the combined design and structural means of the spiral waveguide that lengthens the sound wave path, the dual acoustic sensors and the like.

[0039] The system is also configured to adopt a self-powered design (energy supply unit), and the energy supply unit is integrated inside the fan blade, uses the potential energy generated by the rotation of the fan blade to drive the micro generator to realize power supply, and does not need to be additionally connected to an external power source. The overall structure is simple, the wiring is less, and it is suitable for large-scale deployment.

[0040] The second aspect of this invention discloses a method for diagnosing wind turbine blade faults based on a helical waveguide, which is performed using any of the systems described above;

[0041] The method includes the following steps:

[0042] S1: Obtain the sound pressure of the sound wave received at the measurement point inside the wind turbine blade and the time when the sound wave is received. The sound wave propagates to the measurement point through different propagation paths in the wind turbine blade.

[0043] S2: Compare the relationship between sound pressure level and preset threshold:

[0044] If the sound pressure is less than the preset threshold, the fan blades are not faulty, and the process returns to step S1.

[0045] If the sound pressure is greater than the preset threshold, the fan blades are faulty, and proceed to step S3;

[0046] S3: Calculate the time difference between the arrival of sound waves propagating along different paths at the measurement point;

[0047] S4: Calculate the three-dimensional coordinates of the sound source based on the sound wave transmission model, the geometric constraints of the internal structure of the wind turbine blades, and the time difference obtained in step S3.

[0048] The preset threshold is the sound pressure of the background noise at the measurement point when the wind turbine blades are operating normally;

[0049] The different propagation paths include: a first propagation path that propagates in a spiral shape through a spiral waveguide and a second propagation path that propagates along the length direction inside the wind turbine blade, with different propagation distances between the first propagation path and the second propagation path.

[0050] Preferably, the preset threshold is an adaptive threshold dynamically set based on the background noise, and the adaptive threshold is calculated using the following formula:

[0051] Threshold = μ noise + k · σ noise ;

[0052] In the formula, μ noise To determine the average sound pressure level of the background noise at the measurement point, σ noise To determine the standard deviation of the background noise sound pressure level measured at the measurement point, k is a coefficient.

[0053] The fixed threshold (background / environmental noise) in the traditional detection method is susceptible to external environment changes, resulting in missed weak signals and false signals. The adaptive threshold setting method can dynamically adjust the judgment criteria according to environmental changes, effectively suppressing false alarms and reliably improving weak signal recognition rate.

[0054] Preferably, the sound wave transmission model is:

[0055] d 1= v 1· ( t 1- t 0);

[0056] d 2= v 2· ( t 2- t 0);

[0057] wherein, d 1 is the distance of the sound wave propagating from the sound source to the measurement point through the first propagation path, d 2 is the distance of the sound wave propagating from the sound source to the measurement point through the second propagation path; v 1 is the sound speed of the sound wave in the first propagation path, v 2 is the sound speed of the sound wave in the second propagation path; t 0 is the emission time of the sound source, t 1 is the time when the sound wave propagates from the sound source to the measurement point through the first propagation path, t 2 is the time when the sound wave propagates from the sound source to the measurement point through the second propagation path.

[0058] Preferably, the medium in the propagation path is air;

[0059] The sound speed is the sound speed after temperature and humidity correction:

[0060] v ( T , H ) = 331.3 + 0.606 · T + 0.0124 · H ;

[0061] wherein, T medium temperature, ℃; H is the relative humidity, %RH;

[0062] When the first propagation path and the second propagation path are in different air environments, v 1= v ( T 1, H 1), v2= v ( T 2, H 2)。

[0063] Preferably, the three-dimensional coordinates of the sound source are calculated by a time difference positioning algorithm and based on a least square method iteration. Specifically, according to the time difference Δt of the sound waves collected by the two acoustic sensors via different propagation paths, and by using the geometric constraints inside the fan blade and the sound speed model (sound wave transmission model), the distance difference of the sound source to each sensor is solved by least square method iteration, so as to calculate the three-dimensional coordinates of the sound source.

[0064] The working principle of the present application is as follows:

[0065] When the internal structure of the fan blade is damaged, such as micro-cracks, delamination, material debonding, etc., the acoustic emission signals generated will propagate along different paths to the two acoustic sensors, and the control assembly will synchronously collect and analyze the time difference of the two signals, so as to calculate the three-dimensional coordinates of the sound source and realize early fault positioning.

[0066] Compared with the prior art, the present application has the following beneficial effects:

[0067] 1. Low-cost long-distance detection capability: the present application realizes real-time collection of fault sound and fault positioning of large fan blade structure by using simple acoustic structure design and few sensors.

[0068] 2. Strong anti-interference capability: the structure of the double acoustic sensors is combined with the double propagation paths with significant distance difference, which effectively enhances the recognition ability of weak signals; specifically:

[0069] 1) The sound waves generated by the blade structure failure propagate in a spiral shape through the first propagation path of the spiral waveguide tube, and the sound wave attenuation is caused by the slight propagation loss of air, and the geometric dispersion is very small, thus giving the weak signal a very long propagation distance; while the continuous cavity structure along the length direction inside the blade approximates to a waveguide structure, forming a second propagation path for sound wave propagation, and the propagation attenuation is still much lower than that of the structural elastic sound waves detected by the conventional acoustic emission method.

[0070] 2) The second acoustic sensor receiving the direct conduction of the fan blade and the first acoustic sensor receiving the conduction of the spiral waveguide tube record signals respectively, and the two-way collection can be verified with each other, effectively reducing the influence of background noise; at the same time, if the number of sensors and spiral structures is appropriately increased, the multiple sensor data can also be filtered by a matching algorithm to improve the weak signal recognition rate and positioning accuracy.

[0071] 3) The fixed threshold used in the prior art is prone to "trigger delay or miss low amplitude signals"; the present scheme also uses an environment-adaptive threshold for judgment, which can dynamically adjust the judgment reference according to the noise level, while suppressing false alarms, it can improve the detection probability of weak signals; for example, weak acoustic emission events may be ignored under the traditional fixed 45dB threshold, while the adaptive threshold can capture these low amplitude events.

[0072] Therefore, the present scheme effectively amplifies the measurable time difference of weak sound events by extending the air propagation path and the structure of double sensor monitoring, and improves the signal recognition degree by using redundant sensors and dynamic adaptive threshold technology, thereby enhancing the detection capability of weak defect acoustic emission.

[0073] 3. Simple structure, easy to maintain: the overall integrated design of the system is arranged in the fan blade, reducing external wiring, suitable for large-scale deployment.

[0074] 4. Self-powered feature: due to the few sensors, significantly less than the scheme used in the prior art, the system does not require additional external power supply, and can use the rotation of the wind turbine driven fan blade to realize system power supply, effectively reducing maintenance costs.

[0075] 5. Strong applicability: can detect various structural defects such as micro-cracks, delamination, and debonding.

[0076] 6. Remote monitoring capability: supports multiple wireless communication methods to realize real-time data upload and early warning. BRIEF DESCRIPTION OF DRAWINGS

[0077] Figure 1 It is a structural schematic diagram of the whole wind turbine fan blade;

[0078] Figure 2 It is a structural schematic diagram of the fan blade fault diagnosis system based on the spiral waveguide tube;

[0079] Figure 3 It is a cross-sectional schematic diagram of the spiral waveguide tube at the sound wave inlet hole;

[0080] Figure 4 It is a structural schematic diagram of the cross section of the wind turbine fan blade;

[0081] In the figure: 100 - support beam; 101 - fan blade; 102 - hub; 103 - motor cabin; 200 - blade shell; 201 - second acoustic sensor; 202 - first acoustic sensor; 203 - spiral waveguide tube; 204 - sound wave inlet hole; 205 - sound-absorbing material; 206 - acoustic reflective coating; 207 - leading edge cavity; 208 - web; 209 - inter-web cavity; 210 - trailing edge cavity. DETAILED DESCRIPTION

[0082] In order to better understand the purpose and technical solutions of the present application, the present application is further described in detail below in combination with the drawings and examples. It should be noted that the following examples are only used to explain the present application and do not constitute a limitation on the present application.

[0083] The present application aims to provide a fan blade 101 fault diagnosis system and method based on an internal spiral waveguide 203 and double acoustic sensors, which can realize accurate collection and positioning of internal fault signals of the fan blade 101, improve detection accuracy and system stability, and is suitable for online monitoring and fault warning of the fan blade 101. It should be additionally pointed out that, as shown in the figure, the internal cavity of the fan blade 101 is sequentially divided into a leading edge chamber 207, a trailing edge chamber 210 and an inter-web chamber 209 along the length direction by the web 208. Figure 4

[0084] Research shows that damage in composite materials usually occurs in the form of "matrix cracks, delamination, interface debonding, and fiber breakage". Many existing detection technologies are only optimized for specific defect types: for example, ultrasonic detection is commonly used for adhesive layer debonding or surface cracks, while vibration / image detection may only target macroscopic cracks; they often have difficulty in covering all defect modes at once. The present scheme relies on acoustic emission (AE) principle to detect internal damage of the fan blade; acoustic emission signals are naturally sensitive to multiple damage modes and can capture elastic wave signals of micro-cracks, delamination, and interface debonding processes at the same time.

[0085] Example 1

[0086] A fan blade 101 fault diagnosis method based on a spiral waveguide 203, comprising the following steps:

[0087] S1: Obtain the sound pressure of the sound wave received at the measurement point in the fan blade 101 and the time when the sound wave is received, wherein the sound wave propagates to the measurement point by different propagation paths in the fan blade 101;

[0088] S2: Compare the relationship between the sound pressure and the preset threshold value:

[0089] If the sound pressure is less than the preset threshold value, the fan blade 101 is fault-free, and returns to step S1;

[0090] If the sound pressure is greater than the preset threshold value, the fan blade 101 has a fault, and enters step S3;

[0091] S3: Calculate the time difference of the sound waves propagating by different propagation paths reaching the measurement point;

[0092] S4: According to the sound wave transmission model, the geometric constraints of the internal structure of the fan blade 101 and the time difference obtained in step S3, calculate the three-dimensional coordinates of the sound source; ​

[0093] The different propagation paths include a first propagation path propagating in a spiral shape through the spiral waveguide 203 and a second propagation path propagating along a length direction inside the fan blade 101, the first propagation path and the second propagation path having different propagation distances therebetween.

[0094] The preset threshold is a sound pressure of background noise of the fan blade 101 at a measuring point in a normal operation.

[0095] A fan blade 101 fault diagnosis system based on a spiral waveguide 203, as shown in Figures 1-3 The spiral waveguide 203 is installed inside the blade shell 200, the blade shell 200 is divided into a leading edge chamber 207, an inter-web chamber 209 and a trailing edge chamber 210 by a web 208, and comprises:

[0096] At least one spiral waveguide 203 is arranged along a length direction of the blade shell 200, a tube wall of the spiral waveguide 203 is provided with not less than eight sound wave transmission holes 204 spaced along a length direction of the spiral waveguide 203, the sound wave transmission holes 204 are in communication with the leading edge chamber 207 and / or the inter-web chamber 209 and / or the trailing edge chamber 210;

[0097] A first acoustic sensor 202 is arranged at one end of the spiral waveguide 203, and is used to receive a sound wave signal propagating in the spiral waveguide 203;

[0098] A second acoustic sensor 201 is arranged in the blade shell 200 and adjacent to the first acoustic sensor 202, and is used to receive a sound wave signal propagating in a cavity of the blade shell 200;

[0099] A control assembly is electrically connected with the first acoustic sensor 202 and the second acoustic sensor 201 respectively, and is used to acquire sound wave signals received by the two acoustic sensors and determine a fault position on the fan blade 101 based on a time difference of arrival of the two signals.

[0100] More specifically, in the embodiment:

[0101] Figure 1A typical wind turbine blade 101 part structure is shown, which is mainly composed of a support beam 100, a wind turbine blade 101, a hub 102 and a motor cabin 103, based on the cooperation of each component to achieve efficient conversion of wind energy to electrical energy. The support beam 100, as a key load-bearing structure, connects the blade and the hub 102, has good rigidity and fatigue resistance, and can effectively transmit the dynamic load generated by the blade under wind load. At the same time, its design takes into account the aerodynamic characteristics to reduce the interference with the flow field. The wind turbine blade 101 is the core component of the wind turbine to capture wind energy, which adopts an aerodynamic optimized airfoil section design, gradually thins along the length direction and has a twist angle to adapt to the incoming flow conditions at different radii and improve the aerodynamic efficiency; the blade surface is smooth and has a low friction coefficient, and the internal structure can be further integrated with reinforcing ribs and other structures to enhance the overall strength. The hub 102, as a connecting component between the wind turbine blade 101 and the main shaft, plays a key role in transmitting the energy captured by the wind turbine blade 101 to the main shaft, and is usually made of high-strength cast iron or composite materials, which has excellent load-bearing capacity and fatigue resistance; In addition, the hub 102 can also integrate bearings and variable pitch mechanisms inside as needed to achieve precise control of the blade angle. The motor cabin 103 is the energy conversion center of the wind turbine, which is installed with a generator, a gear box, a control subsystem and a yaw driving device, responsible for converting mechanical energy from the main shaft into electrical energy, and adjusting the wind turbine operating state through the external control subsystem. Figure 4 The internal structure of the typical wind turbine blade 101 is shown by the cross-sectional structure, specifically, the internal structure of the wind turbine blade 101 is sequentially separated into a leading edge chamber 207, a trailing edge chamber 210 and an inter-web chamber 209 along the length direction by a web 208.

[0102] The above description of the wind turbine structure is only a simple introduction to the part structure related to the present scheme, and is not an improvement of the wind turbine itself, and all can use existing technology; In addition, some structures in the conventional wind turbine are not described in the above, only the correlation with the present scheme is weak, and no additional description is made in the embodiment of the present scheme, which can use the conventional existing technology available to those skilled in the art.

[0103] As shown in Figure 2 , an embodiment of the system proposed by the present scheme is installed in the internal structure of the wind turbine blade 101, specifically, the cross-web installation is installed in the internal structure of the leading edge chamber 207, the inter-web chamber 209 and the trailing edge chamber 210. Specifically, as shown in Figure 2As shown, the fan blade 101 is internally provided with at least one spiral waveguide 203, which is fixedly arranged in the fan blade 101 and used to guide the sound wave signals to propagate along a specified path, prolong the propagation time, and enhance the time difference resolution; the spiral waveguide 203 extends along the length direction of the fan blade 101, and can adopt any form of arrangement in a spiral structure and a segmented structure according to the structural design, wherein the segments (leadings) in the segmented structure can be the same or different, and the segments can be connected through flexible joints (preferably made of the same material as the spiral waveguide 203). The spiral waveguide 203 is a flexible pipe body, which has bendability, sufficient hoop stiffness, and acoustic guiding performance; the material thereof is selected from polyurethane reinforced glass fiber composite material, epoxy resin-based glass fiber composite material, or polyimide-based fiber composite material; alternatively, a flexible polymer or fiber reinforced composite material with equivalent mechanical and acoustic performance can be used as an equivalent material. A first acoustic sensor 202 is also installed at the root of the wind blade 101, away from the first end of the sound-absorbing material 205, at the bottom of the spiral waveguide 203; the receiving end of the first acoustic sensor 202 is connected to one end port of the spiral waveguide 203, and used to receive the sound waves propagated through the spiral waveguide 203; in addition, a second acoustic sensor 201 is also installed at the root of the wind blade 101 and adjacent to the position of the first acoustic sensor 202, and used to receive the sound waves directly propagated through the inner cavity of the wind blade 101 (the receiving end of the second acoustic sensor 201 is not connected to the spiral waveguide 203). The positions of the two acoustic sensors can be considered as measurement points, and both are used to synchronously collect sound wave signals for subsequent time difference analysis and sound source positioning. On the wall surface of the spiral waveguide 203, sound wave transmission holes 204 are arranged at intervals, for sound waves to enter the inside of the pipe; the sound wave transmission holes 204 are not less than eight on the spiral waveguide 203, and the spiral waveguide 203 is communicated with the leading edge cavity 207, the trailing edge cavity 210, and the inter-web cavity 209 of the wind blade 101 through the sound wave transmission holes 204; sound-absorbing material 205 is arranged at the end of the spiral waveguide 203, with a sound absorption coefficient not less than 0.9, for preventing interference caused by sound wave reflection at the end of the pipe; the sound-absorbing material 205 can be polyurethane foam or polyester fiber sound-absorbing felt, etc. An acoustic reflection coating 206 is also arranged on the inner wall of the spiral waveguide 203, with a reflection coefficient not less than 0.85; the material of the acoustic reflection coating 206 can be selected from aluminum powder coating, silver plating coating, or polytetrafluoroethylene coating, etc., to enhance the conduction efficiency of sound waves in the pipe.

[0104] When micro-cracks, delamination, material debonding and other structural damages occur inside the wind turbine blade 101, the acoustic emission signals generated by the damages will propagate along different paths (air medium inside the spiral waveguide 203 and air medium inside the blade shell 200) to the two acoustic sensors, and then the signal processing unit will synchronously collect and analyze the time difference of the two acoustic signals, so as to calculate the position of the sound source and realize early fault positioning.

[0105] In this embodiment, the key structural parameters of the spiral waveguide 203 include: the spiral waveguide 203 in a spiral structure, with an inner diameter of 10 ± 5 mm, a wall thickness of 2 ± 0.5 mm, an outer diameter of 30 mm, a lead of 2 m, and the number of spiral turns determined according to the length of the blade, usually 100-10 turns. The purpose of setting the spiral waveguide 203 with a spiral structure is to lengthen the propagation path of the acoustic wave in the air. Compared with direct propagation to the first acoustic sensor 202, the signal arrival time will be significantly delayed (the delay time is about 10-30 ms) after passing through the extended path of the spiral waveguide 203, so as to form a sufficient time difference with the structural propagation path, which is beneficial to improve the resolution accuracy of the sound source positioning. The significant time delay brought by the long path also enhances the distinguishability between the micro-structural damage signal and the environmental noise. Especially in the case of large external noise interference during the operation of the wind turbine, it can more accurately distinguish and judge whether there is a structural damage signal. In other embodiments, a segmented spiral waveguide 203 with different lead lengths for each segment can also be selected.

[0106] As Figure 3As shown, the cross section of the spiral waveguide 203 is circular, and a small hole with a diameter of 5 ± 1 mm is opened on the wall of the spiral waveguide 203 every 5000 ± 500 mm along the length direction as a sound wave transmission hole 204 for the lateral propagation of the sound wave and entering the inside of the spiral waveguide 203. In this embodiment, the sound wave transmission hole 204 is set as follows: hole diameter: 5 ± 1 mm, to ensure that the sound wave can effectively enter the pipeline; hole spacing: 5000 ± 500 mm, to match the wavelength of the main detected sound wave; hole opening angle: 45° ± 5° with the tangent of the spiral waveguide 203, to improve the sound wave collection efficiency; hole inner wall treatment: the hole opening adopts a chamfered (horn mouth or inverted triangular shape) design, to reduce the sound wave reflection loss. Thus, the spiral waveguide 203 provided with the sound wave transmission hole 204 can: the hole spacing matches the wavelength of the sound wave, to reduce the signal attenuation and improve the transmission efficiency; the hole is laterally opened, to enhance the sound response capture of the small cracks in the inside of the fan blade 101 and cover a larger monitoring range; the acoustically treated inner wall of the spiral waveguide 203 can reduce the sound wave energy loss and improve the detection sensitivity; the sound absorption structure at the end of the spiral waveguide 203 can eliminate the echo interference and improve the signal quality. More preferably, the sound wave transmission hole 204 can be set to face the blade shell 200 (sound source), so that when the blade shell 200 or the internal beam and other structures have cracks, delamination and other damages and emit sound waves, the sound wave transmission hole 204 facing these parts can more accurately capture the sound signals, with high signal-to-noise ratio and detection sensitivity.

[0107] Therefore, based on the spiral structure and the multi-hole guide, the directionality of the sound wave propagation can be effectively improved, and the recognition ability of the sensor to abnormal signals in different directions can be enhanced.

[0108] In other embodiments, the key structural parameters of the spiral waveguide 203 can also be set as follows: the inner diameter is 2-20 mm, the wall thickness is 0.5-10 mm, the spiral outer diameter is 3-200 mm, and the lead is 1-12 m; the cross section of the spiral waveguide 203 can also be an ellipse or a polygon, wherein the short diameter of the elliptical cross section is at least 1 mm, and the long diameter is at least 2 mm, and the equivalent diameter of the polygonal cross section is 2-20 mm. The structural parameters of the sound wave transmission hole 204 can also be set as follows: the diameter is 1-10 mm, the spacing is 4000-6000 mm, and the angle between the hole opening of the sound wave transmission hole 204 and the tangent of the spiral waveguide 203 is 40-50°.

[0109] The system effectively improves the detection distance, and is especially suitable for blades with a length of tens of meters to hundreds of meters. According to experimental verification, the system is significantly better than the traditional acoustic emission monitoring system.

[0110] In addition to the physical structure described above, the system further comprises a control component part, which mainly consists of a signal processing unit, a communication module and an energy supply unit.

[0111] The signal processing unit includes a preamplifier, a band-pass filter, an analog-to-digital converter (ADC) and a digital processing unit (DSP / FPGA), which can be connected in a conventional manner. The signal processing unit is electrically connected to the first acoustic sensor 202 and the second acoustic sensor 201 respectively to collect and process the acoustic signals recorded by the two acoustic sensors.

[0112] The communication module can use any one or more of LoRa, NB-IoT, BLE, Wi-Fi or ZigBee, and in this embodiment it is a low-power LoRa module. It is only activated when an abnormal signal is detected, in cooperation with the edge computing event triggering mechanism, which can effectively reduce energy consumption. The communication module is provided, and the system can transmit the calculation results (three-dimensional positioning coordinates) and monitoring data to an external host computer or a remote monitoring platform.

[0113] The energy supply unit uses a centrifugal energy recovery device to drive the micro-generator to work and supply power to the system through the potential energy generated by the rotation of the blade, achieving complete self-power supply. Specifically, the energy supply unit includes a mass block installed inside the fan blade 101, a micro-generator, a voltage stabilizing module and an energy storage module.

[0114] More specifically, first, a guide rail is provided in the radial direction of the fan blade 101, and the mass block can move reciprocally in the guide rail; further, a reset elastic member (such as a spring) is connected between the inner surface of the guide rail and one end of the guide rail on the side of the mass block, which can reset the mass block; secondly, a one-way transmission mechanism (such as a crankshaft connecting rod mechanism) is connected between the mass block and the micro-generator to convert the reciprocating linear motion of the mass block in the guide rail into one-way rotary motion, and the output end of the one-way transmission mechanism is specifically connected to the rotor shaft of the micro-generator, thereby driving the rotor of the micro-generator to rotate; finally, the output end of the micro-generator is electrically connected to the energy storage module (such as a super capacitor or a lithium battery) through a power management circuit to stabilize the output voltage and store electrical energy, and the low-voltage direct-current line of the energy storage module output end supplies power to each power unit in the system, such as the first acoustic sensor 202, the second acoustic sensor 201, the signal processing unit and the communication module, etc., to realize continuous power supply and self-power supply to the entire detection system, ensuring autonomous operation without external power supply. The micro-generator is a combination of one or more of a piezoelectric energy harvester, an electromagnetic energy harvester and a capacitive energy harvester.

[0115] Embodiment 2

[0116] The embodiment is based on the system of embodiment 1, and a set of fault diagnosis methods are designed.

[0117] The first acoustic sensor 202 receives the air-borne acoustic wave through the first propagation path (air filled in the spiral waveguide 203), and the second acoustic sensor 201 receives the acoustic wave from the second propagation path (air filled in the fan blade 101). The two acoustic sensors are arranged adjacent to each other and are located at the measuring point inside the fan blade 101. In this embodiment, the first acoustic sensor 202 and the second acoustic sensor 201: since the frequency component of the acoustic wave in the air is low, the signal amplitude is small and is easily affected by the environmental noise, a piezoelectric microphone is selected to capture the low-frequency to medium-frequency signal, the frequency response range is 20 Hz - 20 kHz, the sensitivity is -45 ± 3 dB, and the signal-to-noise ratio is not less than 65 dB. In other embodiments, the first acoustic sensor 202 and the second acoustic sensor 201 can also select an electret microphone with appropriate parameters.

[0118] Since the propagation distances of the acoustic wave in the two paths are different (the propagation distance in the spiral waveguide 203 will be significantly longer than the propagation distance in the fan blade 101), the time when the two acoustic sensors receive the signal is obviously different. By recording the signal arrival time (time) t 1and t 2, the spatial position of the sound source can be inversely deduced according to the acoustic wave transmission model combined with the propagation path:

[0119] d 1= v 1· ( t 1- t 0);

[0120] d 2= v 2· ( t 2- t 0);

[0121] In the formula, d 1is the distance of the acoustic wave propagated from the sound source to the measuring point through the first propagation path, d 2is the distance of the acoustic wave propagated from the sound source to the measuring point through the second propagation path; v 1is the sound speed of the acoustic wave in the first propagation path, v 2is the sound speed of the acoustic wave in the second propagation path; t 0is the emission time of the sound source, t 1is the time when the acoustic wave is propagated from the sound source to the measuring point through the first propagation path, t 2is the time when the acoustic wave is propagated from the sound source to the measuring point through the second propagation path.

[0122] To further improve the positioning accuracy of the sound source, the sound speed correction model is introduced in the embodiment, further considering the influence of temperature and humidity on sound speed:

[0123] v T H T H

[0124] In the formula, T medium temperature, ℃; H relative humidity, %RH;

[0125] When the first propagation path and the second propagation path are in different air environments, v 1= v T 1, H 1), v 2= v T 2, H 2)。

[0126] Further, the three-dimensional coordinates of the sound source can be calculated by solving the simultaneous equations by the least squares method, combined with the geometric constraints of the internal structure of the fan blade 101. x y z

[0127] Test verification:

[0128] To verify the accuracy of the above three-dimensional positioning method based on the time difference (TDOA) of the double sensors and the sound speed correction model, the embodiment further carries out finite element / wave simulation and physical test under laboratory conditions. The test and simulation steps, parameters and results are as follows.

[0129] 1. Equipment and materials

[0130] 1). Test scale model: fan blade 101 (length L = 6.0 m, composite laminated plate, material properties are set according to common parameters of laboratory engineering composite materials);

[0131] 2). Spiral waveguide tube 203: flexible tube body made of flexible material (tube inner diameter 3 mm, spiral outer diameter 40 mm, lead 500 mm), the tube wall is treated according to embodiment 1 and provided with an opening as a sound wave transmission hole 204 (set according to the scale ratio of the test scale model);

[0132] 3). Acoustic sensor: piezoelectric microphone (20 Hz - 20 kHz, SNR ≥ 65 dB);​​​​​​​​​​

[0133] 4) Data acquisition: Multi-channel ADC (sampling rate ≥ 500 kS / s / channel, synchronous sampling);

[0134] 5) Environmental control: Temperature cabinet or air heater (adjustable T: -10 ℃ to +50 ℃), humidity control device (adjustable H: 10%RH - 90%RH);

[0135] 6) Excitation source: Micro-impactor (for simulating crack / debonding impact) and controllable point sound source (horn or piezoelectric power supply);

[0136] 7) Simulation method: Acoustic finite element.

[0137] 2, Simulation (finite element simulation)

[0138] 1) Modeling: Establish the acoustic cavity model of the fan blade 101; The actual size of the spiral waveguide 203 and the orifice of the sound wave inlet hole 204 is modeled according to the setting of embodiment 1;

[0139] 2) Boundary conditions: External air sound field coupling, spiral waveguide 203 end absorbing boundary (absorption coefficient ≥ 0.9) of sound absorbing material 205;

[0140] 3) Excitation: A number of artificial "fault points" (N = 20) are pre-set inside the fan blade 101, which are excited by short-time impact or narrow-band pulse, and the time sequence signals reached by the air path in the spiral waveguide 203 and the air path in the fan blade 101 are recorded;

[0141] 4) Parameter scanning: The environmental temperature T (0 ℃, 15 ℃, 30 ℃) and humidity H (20%, 50%, 80%) are scanned in combination, and the sound speed correction is applied in the simulation according to v ( T , H ) = 331.3 + 0.606 · T + 0.0124 · H , v 1= v ( T 1, H 1), v 2= v ( T 2, H 2);

[0142] 5). Solution and post-processing: Extract the signal arrival time recorded by each acoustic sensor (using cross-correlation / envelope peak method), and solve the TDOA equation set under the constraint of the internal structure of the fan blade 101, to obtain the estimated position.

[0143] 3. Physical test

[0144] 1). Two acoustic sensors are arranged according to the structure design of embodiment 1, and 20 reference fault points are artificially set on the surface / inside of the fan blade 101 (the positioning coordinates are measured and recorded in advance);

[0145] 2). Trigger the excitation under static and simulated wind noise conditions (simulate different noise levels with a wind noise generator) (repeat 10 times for each reference fault point, a total of 200 samples), and change the environmental temperature and humidity (T = 5 / 20 / 35℃, RH = 20% / 50% / 80%); T H = 5 / 20 / 35℃,

[0146] 3). After data collection, extract the arrival time and calculate the positioning result according to the same signal processing procedure as simulation.

[0147] 4. Data processing and statistical indicators

[0148] 1). Define the positioning error e i = || p est,i - p true,i || (Euclidean distance, unit: cm);

[0149] 2). Statistics: average error μ e , standard deviation σ e , maximum error e max and 95% upper confidence limit;

[0150] 3). Judgment criteria: if μ e + 2 · σ e ≤ 5 cm, it is considered that the positioning accuracy can be stably achieved within ± 5 cm (empirical criterion, or confidence interval method can also be used).

[0151] 5. Sample results

[0152]

[0153] Note: The maximum error in the above table is less than 5.0 cm, and the​μ e + 2 · σ e = 3.0 + 2 × 1.1 = 5.2 cm (slightly greater than 5 cm). Using the more stringent 95% one-sided confidence limit rule, the 95% site error can be obtained as approximately 4.9 cm (the specific value is calculated based on the data distribution). Overall, it is proven that under most operating conditions, the positioning error can be stably controlled within the range of ± 5 cm.

[0154] In summary, this scheme relies on the principle of acoustic emission (AE) to detect damage inside the blade. Dual-channel acoustic sensing and a wide frequency response allow for the acquisition of signals in different modes. For example, fiber fracture generates short pulses and high-frequency signals, while delamination generates low-frequency continuous waves; both acoustic sensors can detect these different characteristics. Furthermore, the adaptive thresholding and multi-channel collection employed in this scheme give the system and method greater versatility and coverage, enabling simultaneous monitoring of various structural defects.

[0155] Example 3

[0156] Compared to the system given in Embodiment 1, this embodiment sets the spiral waveguides 203 into a pair to form a double spiral pipe structure in order to further improve the fault location accuracy.

[0157] In this embodiment, the helical waveguide 203 is divided into a main sound transmission duct and a secondary sound transmission duct. Both are arranged in a helical structure along the length of the fan blade 101, and are independent of each other but spatially staggered, with different leads (forming a sound path difference). Corresponding to the main sound transmission duct, a first main acoustic sensor and a second main acoustic sensor are provided; corresponding to the secondary sound transmission duct, a first secondary acoustic sensor and a second secondary acoustic sensor are provided. The arrangement of the acoustic sensors and the helical waveguide 203 is still implemented according to the structural layout given in Embodiment 1.

[0158] The main parameters of the two helical waveguides 203 are as follows:

[0159] 1) Main sound transmission duct: main spiral, inner diameter of tube 4 mm, outer diameter of spiral 40 mm, lead 5 m, counterclockwise direction;

[0160] 2) Secondary sound transmission duct: secondary spiral, inner diameter of tube 3 mm, outer diameter of spiral 30 mm, lead 10 m, clockwise direction.

[0161] Several (no fewer than eight) sound wave receiving points (sound wave inlet holes 204) are evenly distributed on each spiral waveguide 203, and a sound insulation structure is set in the intersection area of ​​the main sound transmission pipe and the auxiliary sound transmission pipe to prevent sound wave crosstalk.

[0162] Subsequently, the signal processing unit will analyze the four signals collected according to the method in Embodiment 2 above, construct a four-element equation set, and solve it through iterative calculation by the least square method to realize higher-dimensional accurate positioning. Compared with the single spiral structure, the double spiral pipe structure can eliminate the redundant and ambiguous area of the path, and significantly improve the analysis ability and spatial resolution.

[0163] Embodiment 4

[0164] This embodiment is based on Embodiment 2, and further introduces the collection of background noise and the establishment of adaptive threshold to improve the accuracy of fault judgment.

[0165] Specifically,

[0166] Before the system is put into operation or during operation, the first acoustic sensor 202 and the second acoustic sensor 201 collect the background noise signals inside the fan blade 101 in a period without obvious structural impact and abnormal events. The signal processing unit analyzes the collected background noise signals in the frequency domain and the time domain, extracts characteristic parameters such as the root mean square (RMS) and the power spectral density (PSD) of the noise, and calculates the statistical average and the standard deviation within a set collection period.

[0167] Based on the above statistical characteristics, the system dynamically calculates the adaptive threshold, and the calculation formula is as follows:

[0168] Threshold = μ noise + k · σ noise ;

[0169] In the formula, μ noise is the average sound pressure of the background noise measured at the measurement point, σ noise is the standard deviation of the sound pressure of the background noise measured at the measurement point, k is a coefficient; wherein k The general value range is 2-3, which is determined by experiment according to the actual environmental noise level.

[0170] When the amplitude or energy index of the acoustic signal detected by the first acoustic sensor 202 and the second acoustic sensor 201 in real time exceeds the adaptive threshold, the system determines that there is an abnormal sound source, i.e. a fault sound source, and automatically triggers the early warning and sound source positioning process. Through this method, the noise level can be dynamically adapted to changes under different operating conditions and external environmental conditions, effectively reducing the false positive rate and improving the accuracy of diagnosis.

[0171] Embodiment 5

[0172] The difference between this embodiment and embodiment 1 is only that the spiral waveguide adopts a moderate spiral structure (pitch 4-8 m, corresponding to about 25-13 turns), which can achieve a good balance between detection coverage and resolution, and can be applied to most engineering deployments: it can significantly extend the sound path and improve the resolution, without causing excessive pipe length and heavy layout.

[0173] Embodiment 6

[0174] The difference between this embodiment and embodiment 1 is only that the spiral waveguide adopts a dense spiral structure (pitch 1-4 m and not 4 m, corresponding to about 100-25 turns and not 25 turns), which is usually used for: the highest positioning resolution is required, and the detailed detection of critical sections (such as the middle and rear edges of the blade or the vulnerable zone). The spiral waveguide with the dense spiral structure has the characteristics of longer total pipe length and longer air path extension, so the time delay difference is more obvious, which will further help to distinguish weak fault signals; but correspondingly, the longer spiral waveguide will also bring a relatively more complex structure and sound energy attenuation, so it can only be partially arranged (such as combined into a segmented structure) or independently arranged for specific purposes.

[0175] Embodiment 7

[0176] The difference between this embodiment and embodiment 1 is only that the spiral waveguide adopts a sparse spiral structure (pitch > 8 m, corresponding to less than 13 turns), which is usually used for: the internal space of the blade is limited, and the weight needs to be reduced as much as possible, or only rough positioning / monitoring of the approximate fault area of the whole blade is required; among them, the spiral waveguide with a spiral structure with a pitch of 8-12 m (corresponding to about 12.5-8.3 turns) is more preferred. The spiral waveguide with the sparse spiral structure has the characteristics of fewer pipe turns, simple structure, low cost, small impact on blade counterweight, and small sound energy attenuation; but its path extension is limited, the time difference resolution is low, and it is only suitable for preliminary screening or low-resolution applications.

[0177] Embodiment 8

[0178] The difference between this embodiment and embodiment 1 is only that the spiral waveguide adopts a segmented structure, each segment of which can be the same or different, such as it can adopt any combination of the spiral (different pitches) structures given in embodiments 1, 5, 6, and 7. The connection between the segments is achieved through a connecting piece.

[0179] If the connecting piece is used for connection, a flange structure can be designed at the end of the pipe section, reliable connection between sections is realized through the flange, and a cylindrical guide sleeve (inner lining sleeve) is coaxially sleeved at the connection between the sections, so that the inner surface of the spiral waveguide pipe forms a continuous inner surface, which is beneficial to reduce edge scattering and sound wave reflection. The flanges between the sections are connected through a threaded part and sealed between the flanges through a sealing ring, and the guide sleeve and the spiral waveguide pipe can be fixed by adhesive, so as to ensure the air tightness in the pipe, maintain the continuity of the air medium in the pipe, and ensure the reliable propagation of the sound wave. Moreover, the connection mode of the flange is beneficial to on-site disassembly and repair.

[0180] In addition, the connecting piece can be connected in the following forms: 1) threaded sleeve connection + inner lining sleeve + structural adhesive, 2) clamp / quick-mount clamping piece + elastic sealing ring (convenient for quick repair), and 3) flexible bellows joint + flange (allowing for slight misalignment / thermal expansion and contraction).

[0181] The structure of the segmented structure at the connection between the sections is known and fixed, so the influence of the segmented structure on sound propagation is controllable and calculable, and therefore, the form of the segmented structure basically does not affect the fault positioning of the fan blade.

[0182] In summary, the present scheme has the following advantages compared to the defects of the prior art:

[0183] Existing defect one: multiple points are dispersedly arranged, and the structure is complex and may affect the counterweight;

[0184] The prior art needs to arrange multiple sensors and mounting structures inside or on the surface of the fan blade, which not only complicates the wiring, but also may change the weight distribution and internal design of the fan blade.

[0185] Advantages of the present scheme: The present scheme discards the traditional multi-point distributed sensor layout, and can realize fault positioning through the combination of the spiral waveguide pipe and a pair of acoustic sensors arranged at a single measuring point, greatly simplifying the wiring and installation complexity, without the need to distribute sensors in a large range inside the fan blade, significantly reducing the number of arrangements and installation complexity, and avoiding adverse effects on the structure and counterweight of the fan blade.

[0186] Existing defect two: the time difference accuracy is limited due to uneven sound speed and complex propagation path;

[0187] The sound speed in the composite material fan blade has anisotropy and interlayer reflection and refraction phenomenon, and the prior art often uses the constant sound speed assumption, which causes the time difference calculation error to be amplified.

[0188] The scheme advantage: through the spiral waveguide tube prolongs the propagation path in the air medium, makes the same fault signal form significant propagation time delay difference in two propagation paths, improves the time difference resolution, thereby improves the positioning accuracy, further combines the least square iteration algorithm of temperature and humidity compensation, controls the positioning error within 5 cm, and significantly improves the positioning accuracy.

[0189] The existing defect three is that noise interference is serious, and weak signals are easy to be covered.

[0190] Mechanical vibration and pneumatic noise in the operation of the fan can significantly interfere with the acoustic emission signal, and the noise suppression effect in the prior art is limited, and weak defect signals are often covered.

[0191] The scheme advantage: through the cross verification mechanism of the double-channel sensing, the robustness and anti-interference ability of the signal are improved; meanwhile, an adaptive threshold method based on the statistical characteristics of background noise is introduced, the environmental noise level is dynamically adapted, the positioning mechanism is started only when the acoustic signal is significantly higher than the dynamic threshold, the wind noise and vibration interference are effectively suppressed, the anti-interference ability is enhanced, the recognition rate of weak signals is effectively improved, the false alarm is reduced, and the false alarm rate is reduced.

[0192] The existing defect four is that the triangular positioning algorithm is highly sensitive to sensor layout and parameters, and the positioning accuracy can only reach centimeter level.

[0193] The triangular positioning and multi-point time difference algorithm are highly dependent on the sensor position, synchronization accuracy and sound speed assumption, and any slight deviation will cause centimeter error.

[0194] The scheme advantage: the present application only relies on a pair of acoustic sensors, combines the geometric constraint condition of the fan blade and the improved time difference positioning algorithm, uses the least square iteration to solve the sound source coordinates, effectively reduces the dependence on the sensor layout accuracy, and the positioning result is more stable and reliable, and the accuracy is much higher than that of the traditional method.

[0195] At the same time, the simplified hardware structure and built-in design also significantly reduce the system cost, and the sensor, sound collecting structure, control module and energy supply unit are all integrated in the blade, so that the system integration is greatly improved. In particular, the present application introduces an energy supply unit (centrifugal energy recovery device), uses the rotation of the fan blade to drive the internal ball to drive a micro generator to realize energy collection, so that the system can run independently for a long time without external power supply, and solves the limitation of the prior art which depends on external power supply.

[0196] In summary, this invention combines the structural innovation of "dual sensors + (multiple) helical waveguides" with an adaptive threshold signal processing algorithm to achieve integrated acoustic monitoring that is anti-interference, low-complexity, low-cost, highly integrated, and fully self-powered. This significantly improves the reliability and practicality of wind turbine blade structural health monitoring, demonstrating outstanding technological progress and promising engineering application prospects.

[0197] The above description of the embodiments is provided to enable those skilled in the art to understand and use the invention. It will be apparent to those skilled in the art that various modifications can be made to these embodiments, and the general principles described herein can be applied to other embodiments without inventive effort. Therefore, the present invention is not limited to the above embodiments, and any improvements and modifications made by those skilled in the art based on the disclosure of the present invention without departing from the scope of the invention should be within the protection scope of the present invention.

Claims

1. A wind turbine blade fault diagnosis system based on a helical waveguide, installed inside a blade shell (200), wherein the blade shell (200) is divided by a web (208) to form a leading edge chamber (207), an inter-web chamber (209), and a trailing edge chamber (210), characterized in that, The system includes: At least one helical waveguide (203) is arranged along the length of the blade shell (200). The wall of the helical waveguide (203) is provided with at least eight acoustic wave inlet holes (204) at intervals along its length. The acoustic wave inlet holes (204) are connected to the leading edge chamber (207) and / or the inter-spindle chamber (209) and / or the trailing edge chamber (210). When there are multiple helical waveguides (203), each helical waveguide (203) has a different lead and / or helical outer diameter. Moreover, each helical waveguide (203) is not connected to each other and is arranged in a spatially staggered manner. An independent first acoustic sensor (202) is arranged at one end of each helical waveguide (203), and a second acoustic sensor (201) is arranged adjacent to the first acoustic sensor (202). Furthermore, a sound insulation structure is provided in the intersection area of ​​each helical waveguide (203). The first acoustic sensor (202) has its receiving end disposed at one end of the spiral waveguide (203) and is used to receive the acoustic wave signal propagating inside the spiral waveguide (203); The second acoustic sensor (201) is disposed inside the blade housing (200) and adjacent to the first acoustic sensor (202), and is used to receive the sound wave signal propagating through the cavity of the blade housing (200); The control component is electrically connected to the first acoustic sensor (202) and the second acoustic sensor (201) respectively, and is used to acquire the acoustic wave signals received by the two and determine the fault location on the wind turbine blade (101) based on the arrival time difference of the two signals.

2. The wind turbine blade fault diagnosis system based on a helical waveguide according to claim 1, characterized in that, The inner cross-section of the spiral waveguide (203) includes one of the following shapes: circular, elliptical, and polygonal. The inner diameter of the circular cross-section or the equivalent diameter of the polygonal cross-section is 2-20 mm, and the minor axis of the elliptical cross-section is at least 1 mm and the major axis is at least 2 mm. The spiral waveguide (203) has a spiral structure, with a spiral outer diameter of 3-200 mm and a lead of 1-12 m; and / or, the diameter of the acoustic wave inlet (204) is 1-10 mm and the spacing is 1 m-10 m.

3. The wind turbine blade fault diagnosis system based on a helical waveguide according to claim 1, characterized in that, The inner wall of the spiral waveguide (203) is provided with an acoustic reflective coating (206), the reflective coefficient of which is not less than 0.85; the spiral waveguide (203) is provided with a sound-absorbing material (205) at the end away from the first acoustic sensor (202), the sound-absorbing material (205) having a sound absorption coefficient of not less than 0.

9.

4. The wind turbine blade fault diagnosis system based on a helical waveguide according to claim 1, characterized in that, The first acoustic sensor (202) and the second acoustic sensor (201) are each a piezoelectric microphone or an electret microphone.

5. The wind turbine blade fault diagnosis system based on a helical waveguide according to claim 1, characterized in that, The control components include a signal processing unit, a communication module, and a power supply unit; The signal processing unit is electrically connected to the first acoustic sensor (202) and the second acoustic sensor (201) to acquire the acoustic wave signals received by the first acoustic sensor (202) and the second acoustic sensor (201) and locate the fault location on the fan blade (101) by time difference. The communication module is electrically connected to the signal processing unit and is used to transmit detection results to the outside of the system; The energy supply unit is electrically connected to the first acoustic sensor (202), the second acoustic sensor (201), the signal processing unit, and the communication module, and is used to supply power to the power-consuming units in the system. The energy supply unit includes a guide rail arranged along the radial direction of the wind turbine blade (101), a mass block that reciprocates along the guide rail, a reset elastic element connected between one side of the mass block and one end of the guide rail, a one-way transmission mechanism, a micro generator, and an energy storage module. The mass block is mechanically connected to the input part of the one-way transmission mechanism, and the output part of the one-way transmission mechanism is coaxially connected to the rotor shaft of the micro generator.

6. A method for fault diagnosis of wind turbine blades based on helical waveguides, characterized in that, The system described in any one of claims 1-5 is used; The method includes the following steps: S1: Obtain the sound pressure of the sound wave received at the measurement point inside the wind turbine blade and the time when the sound wave is received. The sound wave propagates to the measurement point through different propagation paths in the wind turbine blade. S2: Compare the relationship between sound pressure level and preset threshold: If the sound pressure is less than the preset threshold, the fan blades are not faulty, and the process returns to step S1. If the sound pressure is greater than the preset threshold, the fan blades are faulty, and proceed to step S3; S3: Calculate the time difference between the arrival of sound waves propagating along different paths at the measurement point; S4: Calculate the three-dimensional coordinates of the sound source based on the sound wave transmission model, the geometric constraints of the internal structure of the wind turbine blades, and the time difference obtained in step S3. The preset threshold is the sound pressure of the background noise at the measurement point when the wind turbine blades are operating normally; The different propagation paths include: a first propagation path that propagates in a spiral shape through a spiral waveguide and a second propagation path that propagates along the length direction inside the wind turbine blade, with different propagation distances between the first propagation path and the second propagation path.

7. The method for fault diagnosis of wind turbine blades based on helical waveguides according to claim 6, characterized in that, The preset threshold is an adaptive threshold dynamically set based on the background noise, and the adaptive threshold is calculated using the following formula: Threshold = μ noise + k · σ noise ； In the formula, μ noise To determine the average sound pressure level of the background noise at the measurement point, σ noise To determine the standard deviation of the background noise sound pressure level measured at the measurement point, k is a coefficient.

8. The method for fault diagnosis of wind turbine blades based on helical waveguides according to claim 6, characterized in that, The sound wave transmission model is as follows: d 1 = v 1· ( t 1 - t 0)； d 2 = v 2· ( t 2 - t 0)； In the formula, d 1 represents the distance a sound wave travels from the sound source through the first propagation path to the measurement point. d 2 represents the distance the sound wave travels from the sound source through the second propagation path to the measurement point; v 1 represents the speed of sound in the first propagation path. v 2 represents the speed of sound in the second propagation path; t 0 represents the moment the sound source was emitted. t 1 represents the time it takes for the sound wave to travel from the sound source through the first propagation path to the measurement point. t 2 represents the time when the sound wave travels from the sound source to the measurement point via the second propagation path.

9. The method for fault diagnosis of wind turbine blades based on helical waveguides according to claim 8, characterized in that, The medium in the propagation path is air; The speed of sound is the speed of sound after temperature and humidity correction. v ( T , H ) = 331.3 + 0.606 · T + 0.0124 · H ； In the formula, T Medium temperature, °C; H This represents relative humidity, %RH. Specifically, when the first propagation path and the second propagation path are in different air environments, v 1 = v ( T 1, H 1), v 2 = v ( T 2, H 2).

Images