Switch cabinet vibration source type analysis method and system in electromagnetic environment, and storage medium

Abstract

The invention discloses a switch cabinet vibration source type analysis method and system in an electromagnetic environment, and a storage medium. The method comprises the steps of obtaining a sound wave signal and a vibration signal of operation of a high-voltage switch cabinet; carrying out preprocessing and feature extraction on the sound wave signal to obtain a voiceprint feature, and carrying out VMD decomposition on the vibration signal to obtain a permutation entropy feature of a modal component; and performing vibration source type judgment according to the voiceprint features and the permutation entropy features. By adopting the technical scheme of the invention, the acoustic wave characteristics of the broadband of the vibration signal of the vibration source are analyzed to obtain the acoustic wave signal with the mechanical loosening characteristic and the frequency range.

Description

Technical Field

[0001] The present invention belongs to the technical field of high-voltage switch cabinets, and in particular relates to a method and system for analyzing vibration source types of switch cabinets under electromagnetic conditions, and a storage medium. Background Art

[0002] High-voltage switchgear is the most widely used switchgear with the largest number of uses. Due to problems of varying degrees in design, manufacturing, installation, operating environment, equipment maintenance, and line channels, the accident rate is relatively high.

[0003] During the operation of high-voltage switchgear, mechanical component failure, current imbalance, voltage fluctuation, and external environmental factors can cause vibration failure in the switchgear. According to the characteristics of the vibration fault spectrum, vibration failure can generally be divided into low-frequency vibration, ordinary forced vibration (mainly at power frequency), and high-frequency vibration. The cause of the vibration source is mostly caused by electromagnetic fields. In the power system, when current passes through the conductor, a magnetic field is generated, and the interaction between the magnetic field and the current may cause electromagnetic force, thereby causing vibration. This vibration may affect the conductors and magnetic components in the switchgear, especially in the case of high current or large current changes. In addition, if there is electrical discharge in the switchgear, such as arc flash, it may be accompanied by vibration. Arc flash is a discharge phenomenon that is usually related to excessive voltage or poor insulation of equipment. In this case, it is necessary to check the insulation condition of the equipment and take corresponding insulation improvement measures. If other equipment or lines in the power system fail, vibration may be transmitted to the switchgear. In this case, in addition to checking the switchgear itself, other parts of the power system need to be checked to determine the exact source of the vibration. According to the characteristics of the vibration fault spectrum, vibration failure can generally be divided into low-frequency vibration, ordinary forced vibration (mainly at power frequency), and high-frequency vibration. In order to determine the source of vibration, there is an urgent need for a method that can analyze the wide-band acoustic characteristics of the vibration signal and distinguish the characteristic acoustic wave signal and frequency range of mechanical looseness. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method and system for analyzing the vibration source type of a switch cabinet under an electromagnetic environment, and a storage medium.

[0005] To achieve the above object, the present invention adopts the following technical solution:

[0006] A method for analyzing the vibration source type of a switch cabinet under an electromagnetic environment, comprising:

[0007] Step S1, obtaining the sound wave signal and vibration signal of the high-voltage switch cabinet operation through the data acquisition module;

[0008] Step S2: preprocessing and feature extraction of the sound wave signal through the data analysis module to obtain the voiceprint feature, and VMD decomposition of the vibration signal to obtain the permutation entropy feature of the modal component;

[0009] Step S3: using an expert analysis module to determine the vibration source type based on the voiceprint features and permutation entropy features.

[0010] Preferably, in step S2, the acoustic wave signal is denoised using a wavelet threshold and then subjected to a generalized S transform, and the voiceprint features of the signal are constructed based on the directionality, box dimension and contrast of the time-frequency graph.

[0011] Preferably, in step S3, the voiceprint features and permutation entropy features are input into a pre-stored sound-vibration joint feature SVM evaluation model, the equipment historical data and expert knowledge are compared, the wide-band acoustic wave characteristics of the vibration signal of the vibration source are analyzed, and the mechanical looseness characteristic acoustic wave signal and frequency range are obtained, thereby identifying whether the transformer has an operating fault and diagnosing the potential fault type of the equipment, while achieving timely warning of abnormal conditions.

[0012] The present invention also provides a switch cabinet vibration source type analysis system under an electromagnetic environment, comprising:

[0013] Data acquisition module, used to obtain the sound wave signal and vibration signal of the high-voltage switchgear operation;

[0014] The data analysis module is used to preprocess and extract features of the sound wave signal to obtain the voiceprint features, and to perform VMD decomposition on the vibration signal to obtain the permutation entropy features of the modal components;

[0015] The expert analysis module is used to determine the vibration source type based on the voiceprint features and permutation entropy features.

[0016] Preferably, the data analysis module uses a wavelet threshold to denoise the acoustic signal and then performs a generalized S transform, and constructs the voiceprint features of the signal based on the directionality, box dimension and contrast of the time-frequency graph.

[0017] Preferably, the expert analysis module is used to input the voiceprint features and permutation entropy features into a pre-stored sound-vibration joint feature SVM evaluation model, compare the equipment historical data and expert knowledge, analyze the wide-band acoustic wave characteristics of the vibration signal of the vibration source, and obtain the mechanical looseness characteristic acoustic wave signal and frequency range, thereby identifying whether the transformer has an operating fault and diagnosing the potential fault type of the equipment, while achieving timely early warning of abnormal conditions.

[0018] The present invention also provides a storage medium, on which a computer program is stored, and when the computer program is running, a method for analyzing the vibration source type of a switch cabinet under an electromagnetic environment is executed.

[0019] The present invention obtains the acoustic wave signal and vibration signal of the high-voltage switch cabinet operation through a data acquisition module; pre-processes and extracts features of the acoustic wave signal through a data analysis module to obtain a soundprint feature, performs VMD decomposition on the vibration signal to obtain a permutation entropy feature of the modal component; and determines the type of vibration source according to the soundprint feature and the permutation entropy feature through an expert analysis module. The present invention can analyze the wide-band acoustic wave characteristics of the vibration signal of the vibration source to obtain a mechanical loosening characteristic acoustic wave signal and a frequency range. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0021] Figure 1 This is a flow chart of a method for analyzing vibration source types of a switch cabinet under an electromagnetic environment according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0024] Embodiment 1:

[0025] like Figure 1 As shown, an embodiment of the present invention provides a method for analyzing the vibration source type of a switch cabinet under an electromagnetic environment, comprising:

[0026] Step S1, obtaining the sound wave signal and vibration signal of the high-voltage switch cabinet operation through the data acquisition module;

[0027] Step S2: preprocessing and feature extraction of the sound wave signal through the data analysis module to obtain the voiceprint feature, and VMD decomposition of the vibration signal to obtain the permutation entropy feature of the modal component;

[0028] Step S3: using an expert analysis module to determine the vibration source type based on the voiceprint features and permutation entropy features.

[0029] As an implementation method of an embodiment of the present invention, the data acquisition module is composed of a microphone, an acceleration sensor and a synchronous data acquisition system, which is used to capture and record sound signals and object motion information in the environment. The microphone is used to convert the sound pressure fluctuations of the sound signal into a corresponding electrical signal output. The acceleration sensor is used to detect vibration signals in the environment when performing sound recognition in a noisy environment to help analyze and filter noise in the sound signal; in addition, in some specific sound source positioning systems, the acceleration sensor may also be used to assist in determining the position and direction of the sound source. The synchronous data acquisition system is used to realize synchronous data acquisition between multiple sensors, ensuring that the data captured by different sensors can be aligned in time for subsequent data analysis and processing.

[0030] As an implementation method of an embodiment of the present invention, the acoustic wave characteristics of the vibration signal can be described by the basic equation of acoustic wave propagation. For one-dimensional acoustic wave propagation, the propagation of the acoustic wave can be represented by the following one-dimensional acoustic wave equation:

[0031]

[0032] Among them, p(x,t) is the sound pressure, x is the spatial coordinate system, t is time, and v is the speed of sound.

[0033] The wave equation of sound waves describes the changes of sound waves in space and time. This equation can be solved by separation of variables or other appropriate numerical methods to obtain the characteristics of sound waves in vibration signals.

[0034] Furthermore, due to the different operating conditions of the switch cabinet, the sound waves and vibrations emitted by the components are different, and the collected sound wave signals will be affected by environmental noise and transmission interference. In order to improve the accuracy of high-voltage switch cabinet status identification, the sound wave signal denoising preprocessing is first performed; the details are as follows:

[0035] The ambient noise around the high-voltage switchgear will cause great interference to the acoustic signal. Wavelet transform (WT) has the characteristics of decorrelation, multi-resolution, low entropy and flexible basis selection. The acoustic signal is filtered out by setting a threshold. After WT decomposition, the wavelet coefficients corresponding to the noise are obtained. After inverse transformation and reconstruction, the useful acoustic signal after noise reduction can be obtained. The following soft threshold denoising is selected:

[0036]

[0037] Where w is the wavelet coefficient; the threshold T is related to the signal denoising effect, and a reasonable threshold T is estimated by the standard deviation σ of the wavelet detail coefficient; N is the signal length, when the maximum noise level is lower than When , the variance of the noise in the subband is obtained by dividing the middle value of the modulus of the wavelet coefficients by 0.6745.

[0038] Furthermore, the feature extraction of the sound wave signal after denoising preprocessing is as follows:

[0039] The STFT time-frequency window is fixed and suitable for analyzing processes with roughly the same characteristic scale, but has poor accuracy for multi-scale and sudden transient processes; the wavelet transform WT is a localized function analysis; the generalized S transform is used to describe the time-frequency diagram of the non-stationary signal x(t).

[0040]

[0041] Among them, τ is the Gaussian window position parameter, t is time, f is frequency, and ω(f,τ-t) is the Gaussian window function. In order to improve the time-frequency energy aggregation, the parameters α and β are introduced.

[0042]

[0043] Among them, α and β are adjustment factors, which are generally positive numbers.

[0044] The box dimension describes the texture roughness of the spatial distribution of the time-frequency graph. The calculation process is as follows: divide the time-frequency graph into equal sub-blocks (boxes), each grid is the range of a column of boxes, and the number of boxes required for the (i, j) grid image is:

[0045] n r (i,j)=k-l+1

[0046] Among them, k and l are the number of image grayscales falling in the maximum and minimum boxes respectively.

[0047] The number of boxes that achieve full coverage of the entire acoustic spectrum is defined as:

[0048] N r =∑n r (i,j)

[0049] The least squares method was used to calculate lg(N r )and The slope, the absolute value is the box dimension:

[0050]

[0051] The orientation can be used to describe the global properties of image texture: ie.

[0052]

[0053] Among them, r is the normalization factor; n p is the number of histogram peaks; w p is the amplitude of the Pth peak; H D The subscript D is the Dth peak of the histogram.

[0054] The contrast reflects the gray level difference of the time-frequency diagram of the acoustic signal. The kurtosis k4 is used to characterize its deviation μ4. The contrast can be calculated by the following formula.

[0055]

[0056] Among them, σ h is the image standard deviation; kurtosis

[0057] Furthermore, the transformer operating environment is complex and the vibration signal presents a non-stationary chaotic state. Variational mode decomposition (VMD), as a set of Wiener filters, has good robustness to noise, can effectively avoid modal aliasing, and is highly targeted at transformer vibration signals.

[0058] VMD decomposition includes the establishment and solution of variational constraint problems. For the vibration signal of transformer acquisition length N, the row constraint column variation expression is:

[0059]

[0060] Where k is the number of modal decompositions; {μ k}{ω k} is the kth modal component and center frequency; δ(t) is the pulse signal; st represents the constraint condition, and f is the constraint condition of the signal sum.

[0061] The alternating direction multiplier method combined with Fourier equidistant change is used to continuously optimize the iteration {μ k}{ω k}, the iteration result is:

[0062]

[0063] in, It is the Fourier transform (frequency domain representation) of the original signal f(t), providing global information of the signal in the frequency domain; is the estimated value of the remaining modes except the k-th mode at the n-th step of the current iteration, in order to ensure that the influence of other modes is considered when updating the k-th mode; is the Lagrange multiplier used to satisfy the constraints The constraint compensation introduced by the optimization method (such as the alternating direction multiplier method ADMM); the denominator 1+2α(ω-ω k ) 2 The term in the denominator reflects the frequency selectivity. The parameters α and (ω-ω k ) controls the center frequency ω of the mode component k When ω is close to ω k When , the smaller the denominator value is, the greater the contribution of this frequency point will be.

[0064] In the formula, represents the frequency-weighted mode energy. k (ω) 2 By integrating the product of , we can get the weighted center of the mode in the frequency distribution. Represents the total energy of the mode (integrated and summed in the frequency domain). The ratio of the numerator to the denominator gives the center frequency of the mode component. This is a weighted average, with the weight being the energy distribution of the mode component.

[0065] The objective function is constructed using modal aliasing density and Pearson correlation coefficient, and it is optimized iteratively. The signal mutation point detection is performed using the characteristics of permutation entropy, which has strong noise resistance and high time resolution. The signal sequence {X(i), i=1,2,···,n} is reconstructed in phase space to obtain a sequence. There are m! permutations of the symbol sequence mapped in the m-dimensional phase space, and the probability of s different symbol sequences appearing. The relationship between the two is:

[0066]

[0067] Normalize Pe(m), that is,

[0068]

[0069] As an implementation method of an embodiment of the present invention, the expert analysis module inputs the voiceprint features and permutation entropy features into a pre-stored sound-vibration joint feature SVM evaluation model, compares the equipment historical data and expert knowledge, analyzes the wide-band acoustic wave characteristics of the vibration signal of the vibration source, and obtains the mechanical looseness characteristic acoustic wave signal and frequency range, thereby identifying whether the transformer has an operating fault and diagnosing the potential fault type of the equipment, while achieving timely warning of abnormal conditions.

[0070] The specific implementation process is as follows:

[0071] First, data collection is performed. Vibration signals are collected through vibration sensors (such as accelerometers) installed on transformers or other equipment. Acoustic sensors or microphones are used to collect acoustic signals emitted by the equipment, especially low-frequency and high-frequency noises related to mechanical looseness, faults, etc. The second step is feature extraction. The soundprint feature is information with equipment characteristics and fault characteristics extracted from the sound wave signal, usually including sound spectrum, instantaneous frequency, power spectrum density, etc. By combining the characteristics of vibration signals and sound wave signals, a joint feature is constructed, and the mechanical vibration and the generated sound wave characteristics are comprehensively considered to obtain a more accurate evaluation of the equipment operation status. The third step is feature fusion. Information such as soundprint features and permutation entropy features are fused into a complete feature vector and input into the support vector machine (SVM) evaluation model. SVM is a commonly used supervised learning algorithm that can effectively classify high-dimensional feature spaces. The fourth step is model training and evaluation. The SVM model is trained through historical data (including normal operation data and fault data). The training data needs to include vibration and sound wave signal features of different fault types and normal states. During the training process, expert experience and knowledge are incorporated into the selection and adjustment of the model to improve the diagnostic accuracy of the model. The fifth step is to identify and diagnose faults. The real-time collected vibration signals and acoustic wave signals are input into the SVM model for evaluation after feature extraction and fusion to determine whether the equipment has faults. At the same time, the type of fault is determined based on the classification results output by the model, such as whether it is mechanical looseness, bearing wear, electrical fault, etc. The output of the SVM can give a classification result to indicate the operating status of the equipment.

[0072] Embodiment 2:

[0073] An embodiment of the present invention further provides a switch cabinet vibration source type analysis system under an electromagnetic environment, comprising:

[0074] Data acquisition module, used to obtain the sound wave signal and vibration signal of the high-voltage switchgear operation;

[0075] The data analysis module is used to preprocess and extract features of the sound wave signal to obtain the voiceprint features, and to perform VMD decomposition on the vibration signal to obtain the permutation entropy features of the modal components;

[0076] The expert analysis module is used to determine the vibration source type based on the voiceprint features and permutation entropy features.

[0077] As an implementation method of the present invention, the data analysis module uses wavelet threshold to denoise the sound wave signal and then performs generalized S transform, and constructs the voiceprint feature of the signal based on the directionality, box dimension and contrast of the time-frequency graph.

[0078] As an implementation method of an embodiment of the present invention, the expert analysis module is used to input the voiceprint features and permutation entropy features into a pre-stored sound-vibration joint feature SVM evaluation model, compare the equipment historical data and expert knowledge, analyze the wide-band acoustic wave characteristics of the vibration signal of the vibration source, and obtain the mechanical looseness characteristic acoustic wave signal and frequency range, thereby identifying whether the transformer has an operating fault and diagnosing the potential fault type of the equipment, and realizing timely warning of abnormal conditions.

[0079] Embodiment 3:

[0080] The present invention also provides a storage medium, on which a computer program is stored, and when the computer program is running, a method for analyzing the vibration source type of a switch cabinet under an electromagnetic environment is executed.

[0081] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for analyzing the vibration source type of a switch cabinet under an electromagnetic environment, characterized in that: include: Step S1, obtaining the sound wave signal and vibration signal of the high-voltage switch cabinet operation through the data acquisition module; Step S2: preprocessing and feature extraction of the sound wave signal through the data analysis module to obtain the voiceprint feature, and VMD decomposition of the vibration signal to obtain the permutation entropy feature of the modal component; Step S3: using an expert analysis module to determine the vibration source type based on the voiceprint features and permutation entropy features.

2. The method for analyzing the vibration source type of a switch cabinet under an electromagnetic environment according to claim 1, characterized in that: In step S2, the acoustic signal is denoised using a wavelet threshold and then subjected to a generalized S transform, and the voiceprint features of the signal are constructed based on the directionality, box dimension and contrast of the time-frequency graph.

3. The method for analyzing the vibration source type of a switch cabinet under an electromagnetic environment according to claim 2, characterized in that: In step S3, the voiceprint features and permutation entropy features are input into the pre-stored sound-vibration joint feature SVM evaluation model, and the equipment historical data and expert knowledge are compared to analyze the wide-band acoustic wave characteristics of the vibration signal of the vibration source to obtain the mechanical looseness characteristic acoustic wave signal and frequency range, thereby identifying whether the transformer has an operating fault and diagnosing the potential fault type of the equipment, and realizing timely warning of abnormal conditions.

4. A switch cabinet vibration source type analysis system under electromagnetic environment, characterized in that: include: Data acquisition module, used to obtain the sound wave signal and vibration signal of the high-voltage switchgear operation; The data analysis module is used to preprocess and extract features of the sound wave signal to obtain the voiceprint features, and to perform VMD decomposition on the vibration signal to obtain the permutation entropy features of the modal components; The expert analysis module is used to determine the vibration source type based on the voiceprint features and permutation entropy features.

5. The switch cabinet vibration source type analysis system under electromagnetic environment according to claim 4, characterized in that: The data analysis module uses wavelet threshold to denoise the acoustic signal and then performs generalized S transform, and constructs the voiceprint features of the signal based on the directionality, box dimension and contrast of the time-frequency graph.

6. The switch cabinet vibration source type analysis system under electromagnetic environment according to claim 5, characterized in that: The expert analysis module is used to input the voiceprint features and permutation entropy features into the pre-stored sound-vibration joint feature SVM evaluation model, compare the equipment historical data and expert knowledge, analyze the wide-band acoustic characteristics of the vibration signal of the vibration source, and obtain the mechanical loose characteristic acoustic wave signal and frequency range, thereby identifying whether the transformer has an operating fault and diagnosing the potential fault type of the equipment, while achieving timely warning of abnormal conditions.

7. A storage medium, characterized in that: The storage medium stores a computer program, which, when running, executes the method for analyzing the type of vibration sources of a switch cabinet under an electromagnetic environment as claimed in any one of claims 1 to 3.

Images