Off-grid sparse Bayesian partial discharge DOA estimation method based on deep learning

Through the off-grid sparse Bayesian algorithm and LSTM network based on deep learning, the problems of low local discharge positioning accuracy and high computational complexity are solved, and high precision and low complexity are achieved, which improves noise robustness and real-time application capabilities.

CN118918905AActive Publication Date: 2024-11-08CHONGQING UNIV +2
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202410965380.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-18
Publication Date
2024-11-08
Estimated Expiration
2044-07-18

AI Technical Summary

Technical Problem

The prior art has problems of low positioning accuracy, high computational complexity and poor robustness in local discharge positioning, which is difficult to effectively apply under complex working conditions.

Method used

The off-grid sparse Bayesian algorithm based on deep learning is used to extract the target frequency of the acoustic array signal and two sparse Bayesian iterations, combined with the LSTM network, to realize the positioning of the local discharge sound source.

Benefits of technology

It realizes local discharge positioning with high accuracy and low computing complexity, improves noise robustness, and is suitable for real-time online identification and early warning equipment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118918905A_ABST
    Figure CN118918905A_ABST
Patent Text Reader

Abstract

The invention relates to an off-grid sparse Bayesian partial discharge DOA estimation method based on deep learning, and belongs to the field of partial discharge detection. The method comprises the following steps: carrying out two times of iterative operation on a sampled discharge sound array signal by adopting an off-grid sparse Bayesian algorithm to obtain a mean value and a variance of the first two times of off-grid sparse Bayesian algorithms; converting the obtained mean value and variance into normalized real numbers, and inputting the normalized real numbers into a deep learning network LSTM, thereby solving the DOA of an unknown sound source; compared with the traditional off-grid sparse Bayesian algorithm, the method has the advantages that the positioning speed is obviously improved and the frequency fluctuation resistance and the noise robustness are obviously enhanced under the condition that the positioning precision is close to that of off-grid sparse Bayesian.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of partial discharge detection and relates to an off-grid sparse Bayesian partial discharge DOA estimation method based on deep learning. Background Art

[0002] Partial discharge detection plays an important role in judging the degree of insulation degradation, operating conditions and fault warning of power equipment. At present, ultra-high frequency method or ultrasonic method is often used to perform time-frequency analysis on the electromagnetic or acoustic signals collected by the sensor to determine whether partial discharge occurs or the type of partial discharge. Compared with the positioning method based on acoustic array, this method cannot accurately locate the specific location of partial discharge.

[0003] The most mature sound source localization method currently used is the subspace-based positioning method, but the traditional subspace-based positioning method has poor noise robustness, is sensitive to the size of the array and the layout of the array elements, and does not work well under complex working conditions. In recent years, algorithms based on compressed sensing theory can help improve noise robustness, but such algorithms are not easy to obtain global optimal solutions and often require special toolboxes for solving, resulting in poor universality of the algorithms. In contrast, the Bayesian algorithm has been proven to be able to obtain the global optimal solution, but in order to improve the positioning accuracy, the algorithm introduces off-grid operation, which further increases the computational complexity and is difficult to be applied in actual engineering.

[0004] In summary, there is an urgent need for a positioning method with high positioning accuracy, low computational complexity and high robustness, which can be applied in the field of real-time partial discharge detection to improve the safety of power transmission equipment operation. Summary of the invention

[0005] In view of this, the object of the present invention is to provide an off-grid sparse Bayesian local discharge direction of arrival (DOA) estimation method based on deep learning.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] An off-grid sparse Bayesian partial discharge DOA estimation method based on deep learning, the method specifically includes the following steps:

[0008] S1: Extract the target frequency from the collected acoustic array signal;

[0009] S2: The acoustic array signal containing only the target frequency after extraction is subjected to two off-grid sparse Bayesian algorithm iterations to obtain the mean, variance and direction of the discharge sound source;

[0010] S3: Convert the mean and variance of the off-grid sparse Bayes algorithm after two iterations into normalized real numbers;

[0011] S4: Input the normalized real number mean and variance into the deep learning network (Long Short-Term Memory, LSTM) to obtain the amplitude of the local discharge sound source.

[0012] Further, in S1, the target frequency is extracted from the collected acoustic array signal;

[0013] S11: directly use a bandpass filter to filter the acoustic array signal;

[0014] S12: The filter is set to extract a narrowband acoustic array signal with a bandwidth of 1kHz, centered on the target frequency.

[0015] Further, in S2, the acoustic array signal containing only the target frequency after extraction is subjected to two off-grid sparse Bayesian algorithm iterations to obtain the mean, variance and direction of the discharge sound source;

[0016] S21: Define signal model:

[0017] A two-stage hierarchical model is used. The first layer is a signal X of length L that obeys a zero-mean complex Gaussian probability distribution:

[0018]

[0019] In the above formula, Π represents the multiplication operation, P(A|B) represents the conditional probability, that is, the probability of condition A occurring under the premise of condition B, Λ=diag(ζ), diag(·) is the diagonal operation, ζ is the signal accuracy, X .,l represents the lth column of X, Indicates that the variable u obeys a complex Gaussian distribution with mean μ and covariance matrix Σ; the second layer is a gamma distribution of ζ, expressed as:

[0020]

[0021] Where ζ n is the nth element of ζ, N is the number of grids defined by off-grid sparse Bayes, ρ is a fixed prior parameter; and the gamma distribution for any variable u is expressed as Γ(u|a,b)=[Γ(a)] -1 b a u a-1 e -bu , where Γ(·) is the gamma function, and a and b are set close to 0;

[0022] S22: Define the noise model:

[0023] A complex Gaussian distribution with zero mean is used for the noise E, expressed as:

[0024]

[0025] Where α determines the accuracy of the noise. If δ 2 is the variance of the noise, then α=δ -2 ; E .,l represents the lth column of E, I M is an M-dimensional identity matrix; α adopts gamma distribution, expressed as:

[0026] p(α|,c,d)=Γ(α|c,d)

[0027] In the above formula, c and d are constants close to 0;

[0028] S23: Sparse Bayesian algorithm reasoning:

[0029] According to S21 and S22, the likelihood function of the received data Y is obtained as:

[0030]

[0031] In the above distribution A is the transfer matrix between the microphone array and the search plane, where and A in the elevation angle θ and azimuth angle The first derivative on the square; Y .,l represents the lth column of Y, I M is the M-dimensional unit matrix, where M is the number of array elements;

[0032] The sparse Bayesian algorithm model is solved by using the expectation-maximization EM algorithm, and the posterior distribution of X is obtained to obey the complex Gaussian distribution with mean μ and covariance Σ:

[0033]

[0034] where μ .,l is the lth column of μ:

[0035]

[0036] In the above formula, Σ t is the intermediate variable, Σ t =ΦΛΦ H +α -1 I N , I N is the N-dimensional identity matrix; where:

[0037]

[0038] in is the expected calculation operation, ||·||2 represents the two-norm operation, represents the nth row of X, Σ n,n represents the nth element of the diagonal of Σ;

[0039]

[0040] In the formula κ n is an intermediate variable, and κ n =1-[ζ n ] -1 Σ n,n ,||·|| F represents the Frobenius norm

[0041] S24: Off-grid operation

[0042] Assume that the elevation and azimuth deviations of the discharge source in space relative to the preset grid are Δθ and And it satisfies the uniform distribution:

[0043]

[0044] where r θ and Indicates that the search grid is in θ and The resolution in the axial direction, C~U(A,B), means that C follows a uniform distribution from A to B, obtained by the EM algorithm:

[0045]

[0046] where ⊙ is the Kronecker product, To obtain real number operations, [·] * To take the conjugate, and ν are intermediate variables, thus obtaining

[0047] Further, in S3, the mean and variance after two iterations of the sparse Bayesian algorithm are converted into normalized real numbers;

[0048] S31: Take the absolute value of the mean and variance after iterating the off-grid sparse Bayesian algorithm twice:

[0049]

[0050] In the formula, |·| is the absolute value operation, μ abs and Σ abs are the mean and variance after taking the absolute value respectively;

[0051] S32: Normalize the mean and variance after taking the absolute value:

[0052]

[0053] In the formula and Respectively represent μ abs and Σ abs The normalized value of , min(·) and max(·) represent the minimum and maximum values ​​respectively.

[0054] Furthermore, in S4, the normalized real number mean and variance are input into the LSTM deep learning network to obtain the amplitude of the discharge sound source:

[0055] S41: Set 3 layers of LSTM to contain 128, 256, and 128 hidden units respectively, set 4 inputs, and input the two iteration parameters of off-grid sparse Bayes, namely, mean and variance, to obtain the amplitude distribution of the discharge sound source;

[0056] S42: Discharge sound source azimuth calculated by combining off-grid sparse Bayesian algorithm and the elevation angle θ = θ G +Δθ, where θ G and They respectively represent the preset grid and the sound source amplitude predicted by the LSTM network, thereby realizing the DOA estimation of the partial discharge sound source.

[0057] The beneficial effects of the present invention are: it can ensure positioning accuracy, has low computational complexity and high robustness, and can be applied to online identification of partial discharge in early warning equipment.

[0058] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0060] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0061] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0062] Among them, the drawings are only used for illustrative explanations, and they only represent schematic diagrams rather than actual pictures, and should not be understood as limitations on the present invention. In order to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of actual products. For those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0063] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right", "front", "rear", etc. indicate the orientation or position relationship, they are based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation. Therefore, the terms describing the position relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0064] Figure 1 The overall flow chart of the method of the present invention provides an off-grid sparse Bayesian partial discharge DOA estimation method based on deep learning, which specifically includes the following steps:

[0065] S1: Extract the target frequency from the collected acoustic array signal;

[0066] S2: The acoustic array signal containing only the target frequency after extraction is subjected to two off-grid sparse Bayesian algorithm iterations to obtain the mean, variance and direction of the discharge sound source;

[0067] S3: Convert the mean and variance of the off-grid sparse Bayes algorithm after two iterations into normalized real numbers;

[0068] S4: Input the normalized real number mean and variance into the LSTM deep learning network to obtain the amplitude of the local discharge sound source.

[0069] In S1, the target frequency is extracted from the collected acoustic array signal;

[0070] S11: directly use a bandpass filter to filter the acoustic array signal;

[0071] S12: The filter is set to extract a narrowband acoustic array signal with a bandwidth of 1kHz, centered on the target frequency.

[0072] In S2, the acoustic array signal containing only the target frequency after extraction is subjected to two off-grid sparse Bayesian algorithm iterations to obtain the mean, variance and direction of the discharge sound source;

[0073] S21: Define signal model:

[0074] A two-stage hierarchical model is used. The first layer is a signal X of length L that obeys a zero-mean complex Gaussian probability distribution:

[0075]

[0076] In the above formula, Π represents the multiplication operation, P(A|B) represents the conditional probability, that is, the probability of condition A occurring under the premise of condition B, Λ=diag(ζ), diag(·) is the diagonal operation, ζ is the signal accuracy, X .,l represents the lth column of X, Indicates that the variable u follows a complex Gaussian distribution with mean μ and covariance matrix Σ. The second layer is the gamma distribution of ζ, expressed as:

[0077]

[0078] Where ζ n is the nth element of ζ, N is the number of grids defined by off-grid sparse Bayes, and ρ is a fixed prior parameter. And the gamma distribution for any variable u is expressed as Γ(u|a,b)=[Γ(a)] -1 b a u a-1 e -bu , where Γ(·) is the gamma function, and a and b are usually set close to 0;

[0079] S22: Define the noise model:

[0080] A complex Gaussian distribution with zero mean is used for the noise E, expressed as:

[0081]

[0082] Where α determines the accuracy of the noise. If δ 2 is the variance of the noise, then α=δ -2 . E .,l represents the lth column of E, I M is an M-dimensional identity matrix. α uses a gamma distribution, expressed as:

[0083] p(α|,c,d)=Γ(α|c,d)

[0084] In the above formula, c and d are usually constants close to 0.

[0085] S23: Sparse Bayesian algorithm reasoning:

[0086] According to S21 and S22, the likelihood function of the received data Y can be obtained as:

[0087]

[0088] In the above distribution A is the transfer matrix between the microphone array and the search plane, where and A in the elevation angle θ and azimuth direction respectively The first derivative on the square. .,l represents the lth column of Y, I M is the M-dimensional identity matrix, where M is the number of array elements.

[0089] The sparse Bayesian algorithm model can be solved using the expectation-maximization (EM) algorithm, and the posterior distribution of X is obtained to obey the complex Gaussian distribution with mean μ and covariance Σ:

[0090]

[0091] where μ .,l is the lth column of μ:

[0092]

[0093] In the above formula, Σ t is the intermediate variable, Σ t =ΦΛΦ H +α -1 I N , I N is the N-dimensional identity matrix. Among them:

[0094]

[0095] in is the expected calculation operation, ||·||2 represents the two-norm operation, represents the nth row of X, Σ n,n represents the nth element of the diagonal of Σ.

[0096]

[0097] In the formula κn is an intermediate variable, and κ n =1-[ζ n ] -1 Σ n,n ,||·|| F represents the Frobenius norm

[0098] S24: Off-grid operation

[0099] Assume that the elevation and azimuth deviations of the discharge source in space relative to the preset grid are Δθ and And it satisfies the uniform distribution:

[0100]

[0101] where r θ and Indicates that the search grid is in θ and The resolution in the axial direction, C~U(A,B), means that C follows a uniform distribution from A to B, obtained by the EM algorithm:

[0102]

[0103] where ⊙ is the Kronecker product, To obtain real number operations, [·] * To take the conjugate, and ν are intermediate variables, thus obtaining

[0104] In S3, the mean and variance after two iterations of the sparse Bayesian algorithm are converted into normalized real numbers;

[0105] S31: Take the absolute value of the mean and variance after iterating the off-grid sparse Bayesian algorithm twice:

[0106]

[0107] In the formula, |·| is the absolute value operation, μ abs and Σ abs are the mean and variance after taking the absolute value respectively.

[0108] S32: Normalize the mean and variance after taking the absolute value:

[0109]

[0110] In the formula and Respectively represent μ abs and Σ absThe normalized value of , min(·) and max(·) represent the minimum and maximum values ​​respectively.

[0111] In S4, the normalized real number mean and variance are input into the LSTM deep learning network to obtain the amplitude of the discharge sound source:

[0112] S41: Set 3 layers of LSTM to contain 128, 256, and 128 hidden units respectively, set 4 inputs, input the two iteration parameters (mean and variance) of the off-grid sparse Bayes respectively, and obtain the amplitude distribution of the discharge sound source;

[0113] S42: Discharge sound source azimuth calculated by combining off-grid sparse Bayesian algorithm and the elevation angle θ = θ G +Δθ, where θ G and They respectively represent the preset grid and the sound source amplitude predicted by the LSTM network, thereby realizing the DOA estimation of the partial discharge sound source.

[0114] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. Off-grid sparse Bayesian partial discharge DOA estimation method based on deep learning, characterized by: The method specifically comprises the following steps: S1: Extract the target frequency from the collected acoustic array signal; S2: The acoustic array signal containing only the target frequency after extraction is subjected to two off-grid sparse Bayesian algorithm iterations to obtain the mean, variance and direction of the discharge sound source; S3: Convert the mean and variance of the off-grid sparse Bayes algorithm after two iterations into normalized real numbers; S4: Input the normalized real number mean and variance into the LSTM deep learning network to obtain the amplitude of the local discharge sound source.

2. The off-grid sparse Bayesian partial discharge DOA estimation method based on deep learning according to claim 1 is characterized by: In S1, the target frequency is extracted from the collected acoustic array signal; S11: directly use a bandpass filter to filter the acoustic array signal; S12: The filter is set to extract a narrowband acoustic array signal with a bandwidth of 1kHz, centered on the target frequency.

3. The off-grid sparse Bayesian partial discharge DOA estimation method based on deep learning according to claim 1 is characterized by: In S2, the acoustic array signal containing only the target frequency after extraction is subjected to two off-grid sparse Bayesian algorithm iterations to obtain the mean, variance and direction of the discharge sound source; S21: Define signal model: A two-stage hierarchical model is used. The first layer is a signal X of length L that obeys a zero-mean complex Gaussian probability distribution: In the above formula, Π represents the multiplication operation, P(A|B) represents the conditional probability, that is, the probability of condition A occurring under the premise of condition B, Λ=diag(ζ), diag(·) is the diagonal operation, ζ is the signal accuracy, X .,l represents the lth column of X, Indicates that the variable u obeys a complex Gaussian distribution with mean μ and covariance matrix Σ; the second layer is a gamma distribution of ζ, expressed as: Where ζ n is the nth element of ζ, N is the number of grids defined by off-grid sparse Bayes, ρ is a fixed prior parameter; and the gamma distribution for any variable u is expressed as Γ(u|a,b)=[Γ(a)] -1 b a u a-1 e -bu , where Γ(·) is the gamma function, and a and b are set close to 0; S22: Define the noise model: A complex Gaussian distribution with zero mean is used for the noise E, expressed as: Where α determines the accuracy of the noise. If δ 2 is the variance of the noise, then α=δ -2 ; E .,l represents the lth column of E, I M is an M-dimensional identity matrix; α adopts gamma distribution, expressed as: p(α|,c,d)=Γ(α|c,d) In the above formula, c and d are constants close to 0; S23: Sparse Bayesian algorithm reasoning: According to S21 and S22, the likelihood function of the received data Y is obtained as: In the above distribution A is the transfer matrix between the microphone array and the search plane, where and A in the elevation angle θ and azimuth angle The first derivative on the square; Y .,l represents the lth column of Y, I M is the M-dimensional unit matrix, where M is the number of array elements; The sparse Bayesian algorithm model is solved by using the expectation-maximization EM algorithm, and the posterior distribution of X is obtained to obey the complex Gaussian distribution with mean μ and covariance Σ: where μ .,l is the lth column of μ: In the above formula, Σ t is the intermediate variable, Σ t =ΦΛΦ H +α -1 I N , I N is the N-dimensional identity matrix; where: in is the expected calculation operation, ||·||2 represents the two-norm operation, represents the nth row of X, Σ n,n represents the nth element of the diagonal of Σ; In the formula κ n is an intermediate variable, and κ n =1-[ζ n ] -1 Σ n,n ,||·|| F represents the Frobenius norm S24: Off-grid operation Assume that the elevation and azimuth deviations of the discharge source in space relative to the preset grid are Δθ and And it satisfies the uniform distribution: where r θ and Indicates that the search grid is in θ and The resolution in the axial direction, C~U(A,B), means that C follows a uniform distribution from A to B, obtained by the EM algorithm: where ⊙ is the Kronecker product, To obtain real number operations, [·] * To take the conjugate, and ν are intermediate variables, thus obtaining 4. The off-grid sparse Bayesian partial discharge DOA estimation method based on deep learning according to claim 1 is characterized by: In S3, the mean and variance after two iterations of the sparse Bayesian algorithm are converted into normalized real numbers; S31: Take the absolute value of the mean and variance after iterating the off-grid sparse Bayesian algorithm twice: In the formula, |·| is the absolute value operation, μ abs and Σ abs are the mean and variance after taking the absolute value respectively; S32: Normalize the mean and variance after taking the absolute value: In the formula and Respectively represent μ abs and Σ abs The normalized value of , min(·) and max(·) represent the minimum and maximum values ​​respectively.

5. The off-grid sparse Bayesian partial discharge DOA estimation method based on deep learning according to claim 1, characterized in that: In S4, the normalized real number mean and variance are input into the LSTM deep learning network to obtain the amplitude of the discharge sound source: S41: Set 3 layers of LSTM to contain 128, 256, and 128 hidden units respectively, set 4 inputs, and input the two iteration parameters of off-grid sparse Bayes, namely, mean and variance, to obtain the amplitude distribution of the discharge sound source; S42: Discharge sound source azimuth calculated by combining off-grid sparse Bayesian algorithm and the elevation angle θ = θ G +Δθ, where θ G and They respectively represent the preset grid and the sound source amplitude predicted by the LSTM network, thereby realizing the DOA estimation of the partial discharge sound source.

Citation Information

Patent Citations

  • Missing vibration signal recovery method based on variational Bayesian parallel factorization

    CN113704688A

  • Robust DOA estimation method based on sparse Bayesian learning

    CN116125369A

  • Near-far field mixed source estimation method of Bayesian learning method based on sparse array

    CN117113832A

  • Environment map construction method for any continuous irregular rough diffuse reflection scene

    CN118332056A

  • Block sparse compressive sensing based infrared image reconstruction method and system thereof

    WO2018039904A1