A method for identifying the icing state of a fan blade by using acoustic and stress wave multi-modal identification

Abstract

The application discloses a kind of acoustic fingerprint and stress wave multi-modal identification method of fan blade icing state, including, quantum mel filter group is constructed, and the acoustic fingerprint signal of fan blade icing state is converted into quantum state coding;Quantum convolution feature extraction operation is carried out to the quantum state coding, and according to the result of quantum convolution feature extraction is converted into classical feature vector;Icing state classification model is constructed, and the classical feature vector is input into the icing state classification model and is handled;The application can consider the acoustic fingerprint and stress wave under the icing state of fan by constructing icing state classification model, and provides a kind of efficient, accurate method for wind turbine blade icing detection.

Description

Technical Field

[0001] This invention relates to the field of wind turbine blade monitoring technology, and in particular to a method for multimodal identification of acoustic signature and stress wave in the icing state of wind turbine blades. Background Technology

[0002] As wind turbines expand into higher latitudes and more complex climate regions, blade icing monitoring has become a core technological challenge for ensuring the safe operation of wind turbines. Current mainstream monitoring methods primarily rely on acoustic vibration analysis and guided wave detection technology. This involves acquiring dynamic response signals from the blades using acoustic emission sensor arrays and extracting modal parameter shifts during ice formation using time-frequency analysis. In recent years, researchers have proposed a voiceprint recognition method based on Mel-frequency cepstral coefficients (MFCC). Utilizing a filter bank design based on a biomimetic auditory system, this method enhances the extraction of micro-vibration characteristics in the early stages of icing within a specific frequency band (80Hz-8kHz) of the wind turbine. Simultaneously, Lamb wave detection technology excited by piezoelectric ceramic arrays can invert the ice thickness distribution by analyzing the dispersion distortion along the stress wave propagation path. Furthermore, at the data processing level, the fusion architecture of deep convolutional networks and support vector machines has proven effective in improving the classification accuracy of multi-source heterogeneous data.

[0003] However, existing technologies still have significant limitations. First, traditional Mel filter banks face a trade-off between frequency domain resolution and computational complexity during quantum reconstruction. Their linear triangular window function struggles to adapt to the nonlinear frequency offset effect caused by the dynamic growth of ice layers, leading to an increased feature loss rate of high-frequency harmonic components. Furthermore, classical convolutional networks suffer from Hilbert space mismatch when extracting quantum state features. Conventional tensor dimensionality reduction operations can destroy the phase-sensitive information contained in quantum entangled states, increasing recognition errors. Summary of the Invention

[0004] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0005] In view of the aforementioned existing problems, this invention is proposed. Therefore, this invention provides a method for multimodal identification of acoustic signatures and stress waves in the icing state of wind turbine blades, to solve the problems mentioned in the background art.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for multimodal identification of acoustic signature and stress wave in the icing state of wind turbine blades, comprising:

[0007] A quantum Mel filter bank was constructed to convert the acoustic signature signal of the icing state of the wind turbine blades into quantum state encoding;

[0008] The quantum state encoding is subjected to quantum convolution feature extraction, and the result of quantum convolution feature extraction is converted into a classical feature vector.

[0009] An icing state classification model is constructed, and the classic feature vector is input into the icing state classification model for processing, thereby realizing the identification results of the original wind turbine blade acoustic signature signal and stress wave in the icing state.

[0010] As a preferred embodiment of the acoustic signature and stress wave multimodal identification method for the icing state of wind turbine blades described in this invention, wherein: the construction of the quantum Mel filter bank includes:

[0011] The frequency range of the original wind turbine blade acoustic signature signal is obtained, and the frequency range is regarded as the frequency space.

[0012] The frequency space is mapped to the quantum state space using the Mel scale, thus obtaining the Mel frequency range corresponding to each quantum state in the quantum state space.

[0013] As a preferred embodiment of the acoustic signature and stress wave multimodal identification method for the icing state of wind turbine blades described in this invention, wherein: the frequency range of the original wind turbine blade acoustic signature signal includes:

[0014] Vibration signals of wind turbine blades were acquired using a metasurface acoustic array, and acoustic signature signals ranging from 0.1 to 20 kHz were extracted using a lock-in amplifier circuit.

[0015] As a preferred embodiment of the acoustic signature and stress wave multimodal identification method for the icing state of wind turbine blades according to the present invention, the method includes: converting the acoustic signature signal of the icing state of wind turbine blades into quantum state encoding, comprising:

[0016] Based on the icing thickness of the wind turbine blades, the high-frequency and low-frequency changes caused by icing are non-uniformly sampled and assigned to the Mel frequency range corresponding to each quantum state for encoding.

[0017] As a preferred embodiment of the acoustic signature and stress wave multimodal recognition method for the icing state of wind turbine blades described in this invention, the extraction operation of quantum convolution features from the quantum state encoding includes:

[0018] A quantum convolution kernel sensitive to icing characteristics is constructed. Based on the high-frequency and low-frequency changes caused by icing, a controlled rotation gate is applied to the quantum convolution kernel to obtain the quantum convolution output state.

[0019] As a preferred embodiment of the multimodal recognition method for acoustic signature and stress wave of wind turbine blade icing state described in this invention, the method involves converting the results of quantum convolution feature extraction into classical feature vectors, including:

[0020] The quantum convolution output state is converted into a set of classical probability values ​​through quantum measurement, and the probability values ​​are mapped to obtain a classical feature vector.

[0021] As a preferred embodiment of the acoustic signature and stress wave multimodal recognition method for the icing state of wind turbine blades described in this invention, the method includes: constructing an icing state classification model and inputting the classical feature vector into the icing state classification model for processing, including:

[0022] Classical feature vectors and classical stress wave features are used as inputs to the icing state classification model, and are fused in the regenerative kernel Hilbert space.

[0023] As a preferred embodiment of the acoustic signature and stress wave multimodal identification method for the icing state of wind turbine blades described in this invention, it further includes:

[0024] Support vector machines are used in the icing state classification model, and aerodynamic constraints are embedded in the support vector machines.

[0025] As a preferred embodiment of the acoustic signature and stress wave multimodal identification method for wind turbine blade icing state described in this invention, the aerodynamic constraints include:

[0026] Pressure gradient value on the surface of the wind turbine blades.

[0027] Compared with existing technologies, the beneficial effects of the invention are as follows:

[0028] 1. This invention extends the traditional Mel scale to the quantum state space through the nonlinear mapping mechanism of the quantum Mel filter bank, breaking through the limitations of the Shannon sampling theorem, and can effectively identify the acoustic signature signals of ice layers at high and low frequencies;

[0029] 2. By applying a controlled rotation gate to the quantum convolution kernel, it is possible to achieve the equivalent extraction of multidimensional classical eigenvectors with a small number of qubits, preserving most of the quantum phase-sensitive information and reducing computational complexity. Attached Figure Description

[0030] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0031] Figure 1 This is a flowchart illustrating the overall process of a method for identifying acoustic signatures and stress waves in the icing state of wind turbine blades according to an embodiment of the present invention. Detailed Implementation

[0032] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0033] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0034] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0035] This invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not adhering to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.

[0036] Furthermore, in the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used solely for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0037] Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" in this invention should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; similarly, they can refer to mechanical connections, electrical connections, or direct connections, or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0038] Example 1

[0039] Reference Figure 1 This is the first embodiment of the present invention, which provides a method for multimodal identification of acoustic signature and stress wave in the icing state of wind turbine blades, including:

[0040] S1. Construct a quantum Mel filter bank to convert the acoustic signature signal of the wind turbine blade icing state into quantum state encoding;

[0041] Specifically, a metasurface acoustic array is used to collect vibration signals of wind turbine blades, and a lock-in amplifier circuit is used to extract acoustic fingerprint signals from 0.1 to 20 kHz.

[0042] Specifically, this metasurface acoustic array is composed of a topological insulator material, and its acoustic wave focusing efficiency η satisfies the formula:

[0043]

[0044] Among them, P f P0 represents the focused sound pressure level, while P0 represents the ambient noise sound pressure level.

[0045] It should be noted that the harmonic distortion caused by icing of wind turbine blades is non-uniform. For example, the frequency shift has a non-linear relationship with time. This change is closely related to the ice growth rate, and traditional methods are unable to capture this dynamic characteristic.

[0046] Furthermore, the frequency range of the original wind turbine blade acoustic signature signal is obtained, and the frequency range is regarded as the frequency space.

[0047] It needs to be explained that frequency space refers to the frequency domain representation of a signal, usually derived from the time domain signal through Fourier transform, and it encompasses the frequency range of the signal; for example, from 0Hz to a certain maximum frequency f. max Hz;

[0048] Furthermore, the frequency space is mapped to the quantum state space using the Mel scale to obtain the Mel frequency range corresponding to each quantum state in the quantum state space;

[0049] It should be explained that the Mel scale is a scale based on the non-linear perception of sound frequency by the human ear. It converts the actual frequency into Mel frequency (unit: Mel) to better match the perceptual characteristics of different frequencies by the human ear.

[0050] It needs to be explained that the quantum state space is a Hilbert space spanned by qubits. For n qubits, the space dimension can be represented as 2. n That is, there are 2 n There are three ground states, denoted as |k>, where k = 0, 1, 2, ..., 2 n -1;

[0051] Specifically, by using the Mel scale formula to convert the actual frequency range of the frequency space into the Mel frequency range, we can obtain: Mel(0) = 0, Mel(f max Its corresponding value is also the maximum Mel frequency value;

[0052] Furthermore, the Mel frequency range is divided into 2 equal parts. n Given a set of intervals, each corresponding to a quantum ground state, the width of each interval is expressed as:

[0053]

[0054] Therefore, the range of the k-th Mel frequency interval can be expressed as:

[0055] (k·ΔMel,(k+1)·ΔMel]

[0056] Specifically, each quantum ground state is assigned to the k-th Mel frequency interval, then the ground state |k> is equivalent to the characteristics of the original wind turbine blade acoustic signature signal within that Mel frequency interval;

[0057] Furthermore, based on the icing thickness of the wind turbine blades, the high-frequency and low-frequency changes caused by icing are non-uniformly sampled and allocated to the Mel frequency range corresponding to each quantum state for encoding.

[0058] It should be noted that since icing on wind turbine blades alters their surface acoustic properties, causing changes in the spectrum of acoustic signals, this needs to be taken into account. For example, under high-frequency changes (under a thinner ice layer), it indicates that icing can cause distortion or attenuation of high-frequency harmonics; under low-frequency changes (under a thicker ice layer), it indicates that icing can enhance low-frequency resonance or change the low-frequency response.

[0059] Specifically, as can be seen from the above, when the ice layer is thin, it mainly affects the high-frequency part of the acoustic signal, which corresponds to a larger k value in the Mel frequency range, such as |m+1>~|2. n-1>;When the ice layer is thicker, it mainly affects the low-frequency part of the voiceprint signal, which corresponds to a smaller k value in the Mel frequency range, such as |0>~|m>;

[0060] It should be noted that setting m and representing it as the dividing point is mainly for determining the relationship between ice thickness and frequency variation;

[0061] Specifically, considering the icing thickness on the wind turbine blades, the encoded quantum state |ψ> can be represented as:

[0062]

[0063] Wherein, amplitude a k It is related to the signal characteristics of the k-th Mel frequency range;

[0064] It should be noted that, depending on the thickness of the ice layer, the amplitude 'a' needs to be dynamically adjusted during encoding. k Therefore, it can be adjusted using non-uniform sampling as follows:

[0065] a = 0.03h 2 +1.7h+2.4

[0066] Where h is the ice thickness, i.e., the ice layer thickness;

[0067] S2. Perform quantum convolution feature extraction on the quantum state code, and convert the result of quantum convolution feature extraction into a classical feature vector;

[0068] Furthermore, a quantum convolution kernel sensitive to icing characteristics is constructed, and a controlled rotation gate is applied to the quantum convolution kernel according to the high-frequency and low-frequency changes caused by icing, respectively, to obtain the quantum convolution output state;

[0069] It should be explained that the quantum convolution kernel is a specially designed quantum operator, which is mainly used in this invention to capture the acoustic signature signal of the wind turbine blades caused by icing, so as to provide a basis for the subsequent application of the controlled revolving door.

[0070] It should be explained that a controlled rotation gate is a type of quantum gate that can apply a rotation operation to a target qubit according to control conditions (quantum convolution kernel);

[0071] Specifically, the quantum convolution kernel contains the encoded quantum state |ψ> and the weight value corresponding to the quantum state |ψ>;

[0072] Furthermore, the quantum convolution kernel is used as a control condition to adjust the rotation angle of the revolving door based on the high-frequency and low-frequency changes caused by icing.

[0073] Specifically, in high-frequency changes, the rotation angle corresponding to the Mel frequency range where the k value is located is increased; in low-frequency changes, the rotation angle corresponding to the Mel frequency range where the k value is located is decreased.

[0074] It should be noted that by increasing and decreasing the rotation angle corresponding to the Mel frequency range where the k value is located, the controlled rotation gate can increase the sensitivity of the quantum convolution kernel to icing features;

[0075] Specifically, a quantum convolution operation is constructed by applying a controlled rotation gate to the quantum convolution kernel, the quantum convolution operation being performed through a unitary operator U. conv To achieve this, a quantum convolution kernel is applied to the encoded input quantum state |ψ> to obtain |ψ>. out >=U conv |ψ>, where |ψ out > represents the quantum convolution output state, which contains extracted icing feature information;

[0076] Furthermore, through quantum measurement, the quantum convolution output state is converted into a set of classical probability values, and the probability values ​​are mapped to obtain a classical feature vector;

[0077] Specifically, the quantum convolution output state is measured by selecting a computational basis, where the computational basis refers to a basis composed of all possible qubit states, to obtain the probability of each ground state |k>:

[0078] P k =| <k|ψ out >| 2

[0079] Among them, P k This represents the probability that the measurement result is the ground state |k>.

[0080] Specifically, by performing quantum measurements on the output state of quantum convolution, a set of classical probability values ​​can be obtained, which can be expressed as:

[0081]

[0082] It should be noted that due to the randomness of quantum measurements, multiple measurements are required in practice; for example, if several quantum measurements are performed, the frequency of each |k> occurrence is statistically analyzed to obtain an estimated value. Determine the Is it close to the real P? k ;

[0083] Furthermore, by mapping the obtained classical probability values, we obtain a 2 n The classic eigenvector f of dimension 1 is as follows:

[0084]

[0085] It should be noted that the probability values ​​can be transformed depending on the specific task; for example, calculating the expected value or weighted sum of certain ground states.

[0086] S3. Construct an icing state classification model, input the classic feature vector into the icing state classification model for processing, thereby realizing the identification results of the original wind turbine blade acoustic signature signal and stress wave in the icing state;

[0087] Furthermore, classical feature vectors and classical stress wave features are used as inputs to the icing state classification model, and fused in the regenerative kernel Hilbert space.

[0088] Specifically, the classical stress wave characteristics are features extracted from stress wave signals by traditional signal processing methods (such as time-frequency analysis), which reflect the physical properties of the wind turbine blade surface (stress distribution or changes in vibration modes).

[0089] It should be explained that the Regenerative Kernel Hilbert Space (RKHS) is a mathematical framework suitable for handling nonlinear relationships and mapping features from different sources to a high-dimensional space for fusion.

[0090] Specifically, the fusion method uses weighted summation:

[0091]

[0092] Where, x i and x j Let x represent the feature vectors of the i-th and j-th samples, respectively; i and s j Let represent the classical stress wave features of the i-th and j-th samples, respectively; SWD represents the slice Wasserstein distance of the stress wave signal; σ represents the activation function.

[0093] It should be noted that after the fusion is completed, support vector machine (SVM) is used to classify the icing state. SVM distinguishes different categories by finding the optimal hyperplane in high-dimensional space, which is very suitable for processing data in RKHS.

[0094] Specifically, a support vector machine is used in the icing state classification model, and aerodynamic constraints are embedded in the support vector machine;

[0095] It needs to be explained that the goal of SVM (Support Vector Machine) is to find a hyperplane that minimizes the classification error while maximizing the margin between classes. This optimization problem can be expressed as:

[0096]

[0097] Where w is the weight vector of the hyperplane, b is the bias term, and ξ i Represented as a slack variable, used to handle inseparable samples, C is a penalty parameter that controls the tolerance of the icing state classification model to errors, φ(x) i ) is a feature mapping function that maps input data to a high-dimensional space, y i Here, η represents the true label of the sample, and η is the physical constraint strength coefficient. This represents the pressure gradient value on the surface of the wind turbine blades.

[0098] It should be noted that by constructing a model-driven data model and considering the physical driving force of the wind turbine blades, the icing state classification model can both fit the data and conform to the aerodynamic laws of the wind turbine blades, thereby improving the accuracy of the classification of the icing state of the wind turbine blades.

[0099] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0100] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0101] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0102] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0103] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0104] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for multimodal identification of acoustic signature and stress wave in the icing state of wind turbine blades, characterized in that, include: A quantum Mel filter bank was constructed to convert the acoustic signature signal of the icing state of the wind turbine blades into quantum state encoding; The process of converting the acoustic signature signal of the icing state of the wind turbine blades into quantum state encoding includes: Based on the icing thickness of the wind turbine blades, the high-frequency and low-frequency changes caused by icing are non-uniformly sampled and assigned to the Mel frequency range corresponding to each quantum state for encoding. The quantum state encoding is subjected to quantum convolution feature extraction, and the result of quantum convolution feature extraction is converted into a classical feature vector. Extracting quantum convolution features from the quantum state encoding includes: Construct a quantum convolution kernel that is sensitive to icing characteristics, and apply a controlled rotation gate to the quantum convolution kernel according to the high-frequency and low-frequency changes caused by icing to obtain the quantum convolution output state; An icing state classification model is constructed, and the classic feature vector is input into the icing state classification model for processing, thereby realizing the identification results of the original wind turbine blade acoustic signature signal and stress wave in the icing state. The construction of the icing state classification model involves inputting the classic feature vector into the icing state classification model for processing, including: Classical feature vectors and classical stress wave features are used as inputs to the icing state classification model, and are fused in the regenerative kernel Hilbert space.

2. The method for multimodal recognition of acoustic signature and stress wave in the icing state of wind turbine blades as described in claim 1, characterized in that, The construction of the quantum Mel filter bank includes: The frequency range of the original wind turbine blade acoustic signature signal is obtained, and the frequency range is regarded as the frequency space. The frequency space is mapped to the quantum state space using the Mel scale, thus obtaining the Mel frequency range corresponding to each quantum state in the quantum state space.

3. The method for multimodal recognition of acoustic signature and stress wave in the icing state of wind turbine blades as described in claim 2, characterized in that, The frequency range of the original wind turbine blade acoustic signature signal includes: Vibration signals of wind turbine blades were acquired using a metasurface acoustic array, and acoustic fingerprint signals ranging from 0.1 to 20 kHz were extracted using a lock-in amplifier circuit.

4. The method for multimodal recognition of acoustic signature and stress wave in the icing state of wind turbine blades as described in claim 1, characterized in that, The results of quantum convolution feature extraction are converted into classical feature vectors, including: The quantum convolution output state is converted into a set of classical probability values ​​through quantum measurement, and the probability values ​​are mapped to obtain a classical feature vector.

5. The method for multimodal recognition of acoustic signature and stress wave in the icing state of wind turbine blades as described in claim 1, characterized in that, Also includes: Support vector machines are used in the icing state classification model, and aerodynamic constraints are embedded in the support vector machines.

6. The method for multimodal recognition of acoustic signature and stress wave in the icing state of wind turbine blades as described in claim 5, characterized in that, The aerodynamic constraints include: the pressure gradient value on the surface of the wind turbine blades.

Images