Wireless pulse insulator remote live detection device

Abstract

This application relates to the field of intelligent sensor technology, specifically to a wireless pulse insulator remote live-line inspection device. The device includes: a data acquisition module for acquiring multiple vibration-sound signals at the insulator; an insulator analysis module for determining a first noise characteristic value by comprehensively analyzing the time-domain amplitude fluctuation, peak disorder, and signal similarity of the vibration-sound signals, and determining a second noise characteristic value based on the frequency-domain harmonic amplitude distribution and energy proportion; calculating the disturbance degree based on the first and second noise characteristic values ​​to dynamically optimize the wavelet threshold for adaptive denoising of the vibration-sound signals; and an insulator detection module for performing defect detection on the insulator based on all denoised vibration-sound signals. This application solves the problem of poor denoising effect in insulator detection caused by electromagnetic interference by adaptively optimizing the wavelet denoising threshold, thereby improving the accuracy of insulator defect detection.

Description

Technical Field

[0001] This application relates to the field of intelligent sensor technology, specifically to a remote live-line detection device for wireless pulse insulators. Background Technology

[0002] Insulators, as key components for electrical insulation and mechanical fixation in transmission lines, are susceptible to degradation due to long-term exposure to outdoor environments such as sun and rain. Reduced internal insulation can easily lead to breakdown accidents, seriously threatening power grid safety. While existing live-line testing methods can achieve uninterrupted testing, they pose a risk of high-voltage electric shock. Therefore, achieving accurate insulator defect detection while ensuring safety is crucial.

[0003] Vibration acoustics is a common method for remote live-line detection of wireless pulse insulators. However, this method is susceptible to electromagnetic interference from the line when using smart sensors to collect signals. Existing wavelet threshold denoising algorithms usually set the threshold based only on the noise standard deviation of the wavelet coefficients, ignoring the differences in time-frequency domain noise characteristics, resulting in poor denoising effect and thus affecting the accuracy of insulator defect detection. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a remote live-line testing device for wireless pulse insulators, and the specific technical solution adopted is as follows: This application discloses a wireless pulse insulator remote live-line detection device, the device comprising: The data acquisition module is used to collect multiple vibration and acoustic signals at the insulator. The insulator analysis module is used to evaluate the amplitude fluctuation of each vibration sound signal by analyzing the dispersion of the difference between all adjacent peaks in each vibration sound signal at the insulator, and to measure the disorder of all peaks in each vibration sound signal, as well as the similarity between vibration acoustic signals, in order to quantify the first noise characteristic value of each vibration sound signal. By analyzing the distribution of all harmonic amplitudes of each vibration sound signal in the frequency domain, electromagnetic interference harmonics are screened out from all harmonic amplitudes; the proportion of the energy distribution of all electromagnetic interference harmonics in the total energy under the preset normal vibration frequency band, and the proportion of the energy distribution of all harmonics in the preset low frequency band of each vibration sound signal in the total energy under the low frequency band are measured, so as to derive the second noise characteristic value of each vibration sound signal. By combining the first noise feature value and the second noise feature value, the disturbance degree of each vibration sound signal is determined, so as to optimize the wavelet threshold in the wavelet denoising algorithm, and the optimized wavelet denoising algorithm is used to denoise each vibration sound signal. The insulator detection module is used to detect defects in insulators based on all the noise-reduced vibration and acoustic signals.

[0005] Preferably, the amplitude fluctuation of each vibration sound signal is the coefficient of variation of the difference between all adjacent peak values ​​in each vibration sound signal.

[0006] Preferably, the quantization method for the first noise characteristic value of each vibration sound signal is as follows: Based on the degree of disorder of all peaks in each vibration sound signal, the noise characteristic value of each vibration sound signal is determined. Based on the similarity between vibration sound signals at the insulator, the similarity characteristic value of the insulator is determined; The first noise characteristic value of each vibration sound signal is positively correlated with the amplitude fluctuation and noise characteristic value of each vibration sound signal, and negatively correlated with the similarity characteristic value.

[0007] Preferably, the noise-affected characteristic value of each vibration sound signal is the information entropy of all peaks in each vibration sound signal.

[0008] Preferably, the method for determining the similarity feature values ​​of the insulator is as follows: The characteristic signal is obtained by averaging the amplitudes of all vibration sound signals. The mean of the similarity between all vibration sound signals and the characteristic signal is calculated and denoted as the similarity characteristic value of the insulator.

[0009] Preferably, the step of filtering out electromagnetic interference harmonics from all harmonic amplitudes includes: The normalized values ​​of all harmonic amplitudes of each vibration sound signal in the frequency domain are used as the input of the threshold segmentation algorithm, and the segmentation threshold is output. Harmonics with amplitudes greater than or equal to the segmentation threshold are denoted as electromagnetic interference harmonics.

[0010] Preferably, the expression for the second noise characteristic value of each vibration sound signal is: In the formula, This represents the second noise characteristic value of the i-th group of vibration sound signals; This represents the proportion of the sum of squares of the amplitudes of all electromagnetic interference harmonics of the i-th group of vibration sound signals in the total energy under the preset normal vibration frequency band. This represents the proportion of the sum of squares of all harmonic amplitudes in the preset low-frequency band within the frequency domain of the i-th vibration signal group in the total energy of the low-frequency band.

[0011] Preferably, the degree of disturbance of each vibration sound signal is positively correlated with the first noise characteristic value and the second noise characteristic value, respectively.

[0012] Preferably, the wavelet threshold in the optimized wavelet denoising algorithm includes: The optimized wavelet threshold corresponding to the i-th group of vibration sound signals The expression is: In the formula, This indicates the degree of disturbance to the i-th group of vibration sound signals; This represents the wavelet threshold obtained by applying the wavelet denoising algorithm before optimization to the i-th group of vibration sound signals; This indicates the preset adjustment coefficient.

[0013] Preferably, the defect detection of the insulator includes: If the peak values ​​of a preset number of vibration sound signals in the frequency domain are all within the preset normal vibration frequency band, then the insulator has a defect.

[0014] This application has the following beneficial effects: This application determines amplitude fluctuation by calculating the coefficient of variation of the difference between adjacent peaks to quantify amplitude mutation characteristics. It combines the noise characteristic value determined by permutation entropy to characterize peak arrangement disorder, and the similarity characteristic value determined by the similarity between multiple signal groups to characterize signal repetition stability. A weighted fusion is then used to construct a first noise characteristic value, thus comprehensively reflecting the significance of electromagnetic interference to the vibration sound signal in the time domain. This effectively solves the problem of incomplete characterization by a single feature. Quantifying the noise level provides an accurate basis for subsequent adaptive adjustment of the denoising intensity, improving the accuracy and robustness of time-domain noise identification. Furthermore, this application analyzes the harmonic amplitude distribution and uses a threshold segmentation algorithm to screen out significant electromagnetic interference harmonics, calculating their energy proportion in the normal vibration frequency band and the low-frequency band. A piecewise correction logic is introduced to construct a second noise characteristic value, thereby accurately quantifying the actual intensity of power frequency harmonic interference in the frequency domain. This effectively distinguishes between real defect signals and electromagnetic interference noise, avoiding misjudgments caused by harmonic energy covering defect characteristics, and ensuring that even in strong vibration signals... This application improves the accuracy of insulator defect identification under electromagnetic conditions. Furthermore, it determines the disturbance level by weighted fusion of the first noise feature value in the time domain and the second noise feature value in the frequency domain, and adaptively corrects the general wavelet threshold based on the disturbance level to obtain an optimized wavelet threshold. This achieves adaptive denoising of the vibration sound signal. By establishing a positive correlation between the disturbance level and the denoising intensity, the wavelet threshold is dynamically adjusted according to the severity of signal contamination. When severely interfered with, the denoising intensity is enhanced to eliminate noise; when the signal quality is excellent, the threshold is reduced to retain more detailed features. This effectively solves the problem of poor denoising effect with a fixed threshold, significantly improving the signal-to-noise ratio and purity of the insulator vibration sound signal. Finally, this application determines whether there is a defect in the insulator and triggers an early warning by analyzing whether all peaks of the denoised vibration sound signal in the frequency domain are within a preset normal vibration frequency band. This method utilizes wireless transmission to achieve non-contact remote real-time monitoring, and combined with denoising processing, improves the accuracy and reliability of spectrum analysis, thereby improving the precision of insulator defect detection. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A block diagram of a wireless pulse insulator remote live-line detection device provided in one embodiment of this application; Figure 2 This is a schematic diagram of the disturbance extraction process provided in one embodiment of this application. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the wireless pulse insulator remote live-line detection device proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0019] The specific scheme of the wireless pulse insulator remote live-line detection equipment provided in this application is described in detail below with reference to the accompanying drawings.

[0020] Please see Figure 1 The diagram shows a block diagram of a wireless pulse insulator remote live-line testing device according to an embodiment of this application. The device includes: a data acquisition module 101, an insulator analysis module 102, and an insulator testing module 103.

[0021] The data acquisition module 101 is used to acquire multiple vibration sound signals at the insulator.

[0022] The handheld insulating rod integrates a smart sensor on the top inner side. This smart sensor contacts the bottom flange of the insulator through a vibration sound sensor probe, triggering the probe to emit and receive vibration waves to obtain the vibration sound signal of the insulator. The sampling frequency is set to f and the sampling period is t. Each insulator repeatedly collects a preset number of vibration sound signals and transmits them to the processing end through wireless communication technology.

[0023] It should be noted that the values ​​of sampling frequency f, sampling period t, and preset number are all set manually. In this embodiment, the value of sampling frequency f is 10kHz, the length of sampling period is 5s, and the value of preset number is 5. In actual application, as other implementation methods, implementers can also set them according to specific circumstances. This embodiment does not impose any special restrictions.

[0024] It should be noted that, in order to balance the reliability of the test results with the efficiency of on-site operations, and taking into account both computational and time costs, the number of vibration sound signal acquisition groups was set to 5. This number can effectively average random noise and resist sudden strong interference through multiple samplings. Compared with fewer samples, it improves stability and avoids excessive data processing burden and excessive on-site operation time caused by too many samples, thus achieving the best balance between detection accuracy and practicality.

[0025] The insulator analysis module 102 is used to determine the first noise characteristic value by comprehensively analyzing the time-domain amplitude fluctuation, peak disorder and signal similarity of the vibration sound signal, and to determine the second noise characteristic value based on the frequency domain harmonic amplitude distribution and energy ratio; based on the first noise characteristic value and the second noise characteristic value, the disturbance degree is calculated to dynamically optimize the wavelet threshold and perform adaptive denoising processing on the vibration sound signal.

[0026] S1: By analyzing the differences between all adjacent peaks in each vibration acoustic signal at the insulator, as well as the mean of all peaks, the amplitude fluctuation of each vibration acoustic signal is evaluated, and the disorder of all peaks in each vibration acoustic signal and the similarity between vibration acoustic signals are measured to quantify the first noise characteristic value of each vibration acoustic signal.

[0027] Given the extremely frequent and random fluctuations in power grid frequency, the resulting electromagnetic interference exhibits significant uncertainty and is easily superimposed on the acquired signal, thereby reducing the accuracy of insulator defect detection. In this embodiment, the vibration acoustic signal is an electrical signal converted by the probe and circuit. This signal is susceptible to induced coupling from the power frequency electromagnetic field in a strong electromagnetic environment, exhibiting harmonic noise characteristics including 50Hz and its harmonics. This noise is external circuit noise rather than acoustic vibration of the insulator itself; therefore, feature extraction is crucial to quantify the noise level. Since the causes of insulator defects are long-term, the degree of defect and the generated useful signal remain relatively stable within a short sampling interval. However, random electromagnetic interference causes drastic fluctuations in signal amplitude and significant morphological differences across different sampling periods. Therefore, this embodiment analyzes the differences between all adjacent peaks and the peak range in each vibration acoustic signal at the insulator to evaluate the amplitude fluctuation of each vibration acoustic signal, and measures the disorder of all peaks in each vibration acoustic signal, as well as the similarity between vibration acoustic signals, to quantify the first noise characteristic value of each vibration acoustic signal. The specific process is as follows: First, a peak detection algorithm is used to extract all peak values ​​in each vibration sound signal at the insulator. There are many commonly used peak detection algorithms. In this embodiment, the Automatic Multi-Scale Peak Detection (AMPD) algorithm is used to extract peak values. In practical applications, as other implementation methods, implementers may also use other peak detection algorithms according to specific circumstances. This embodiment does not impose any special restrictions. The Automatic Multi-Scale Peak Detection algorithm is a known technology, and the process of extracting signal peak values ​​using it is also a known technology, which will not be described in detail here.

[0028] Furthermore, this embodiment evaluates the amplitude fluctuation of each vibration sound signal by analyzing the differences between all adjacent peaks in each vibration sound signal at the insulator, as well as the mean of all peaks. Specifically: In this embodiment, the coefficient of variation of the difference between all adjacent peaks in each vibration sound signal is used as the amplitude fluctuation of each vibration sound signal.

[0029] It should be noted that there are many methods to measure the differences between data. In this embodiment, the absolute value of the difference between all adjacent peaks in each vibration sound signal is taken as the difference between all adjacent peaks in each vibration sound signal. In practical applications, as other implementation methods, implementers may also use other methods such as the square or ratio of the difference to measure the differences between data, depending on the specific circumstances. This embodiment does not impose any special restrictions on the selection of methods for measuring the differences between data.

[0030] The calculation process of the coefficient of variation is a well-known technique and will not be described in detail here.

[0031] Based on the amplitude fluctuation of each vibration sound signal, it can be understood that amplitude fluctuation is a key indicator for quantifying the amplitude stability of the insulator vibration sound signal under the action of electromagnetic interference. It is used to characterize the drastic changes between adjacent peaks in the vibration sound signal, reflecting the modulation effect of electromagnetic noise on the signal amplitude and the instantaneous change characteristics. The greater the dispersion of the difference between adjacent peaks, the greater the amplitude fluctuation, reflecting that the signal is subjected to strong random electromagnetic interference, causing the amplitude to fluctuate violently, the noise is highly significant, and the denoising difficulty is increased. Conversely, when the dispersion of the difference between adjacent peaks is small, the amplitude fluctuation is smaller, reflecting that the signal amplitude is relatively stable, the influence of electromagnetic interference is weak, and the signal quality is good.

[0032] Furthermore, this embodiment measures the amplitude fluctuation of each vibration sound signal, the disorder of all peaks in each vibration sound signal, and the similarity between vibration acoustic signals to quantify the first noise characteristic value of each vibration sound signal. Specifically: In this embodiment, the noise-affected characteristic value of each vibration sound signal is determined based on the disorder of all peaks in each vibration sound signal. Specifically, the information entropy of all peaks in each vibration sound signal is used as the noise-affected characteristic value of each vibration sound signal. In this embodiment, the information entropy is calculated using the permutation entropy method. In practical applications, implementers may also use other information entropy calculation methods such as Shannon entropy depending on the specific circumstances. The calculation process of permutation entropy will not be elaborated further.

[0033] Based on the noise-affected characteristic value, it can be understood that the noise-affected characteristic value is an index that measures the randomness and complexity of the peak arrangement in a vibration sound signal. It is used to characterize the disorder of the signal's time-domain waveform and reflects the destructive effect of electromagnetic noise on the signal's temporal structure. Its calculation logic is to extract the peak values ​​of the vibration sound signal and calculate the information entropy of all peak values. The noise-affected characteristic value is affected by the randomness of the peak distribution. When the information entropy of the peaks is more chaotic and the disorder is higher, the noise-affected characteristic value is larger, reflecting that a large amount of random electromagnetic noise has been mixed into the signal, causing the originally regular vibration characteristics to be masked and the signal reliability to decrease. Conversely, when the peak distribution is more regular and the periodicity is stronger, the noise-affected characteristic value is smaller, reflecting that the vibration sound signal mainly maintains the vibration characteristics of the insulator body and is not significantly affected by random interference.

[0034] Furthermore, based on the similarity between the vibration sound signals at the insulator, the similarity feature value of the insulator is determined. Specifically, the amplitude of a preset number of vibration sound signals is averaged to obtain the feature signal, and the mean value of the similarity between all vibration sound signals and the feature signal is calculated and recorded as the similarity feature value of the insulator.

[0035] It should be noted that in this embodiment, the normalized value of the cosine similarity between all vibration sound signals and feature signals is used as the similarity between all vibration sound signals and feature signals. The normalization method used is the maximum-minimum normalization method, and the final normalized result is within [0,1]. In practical applications, implementers may also use other methods to measure the similarity between signals, such as the reciprocal of Euclidean distance, or other normalization methods such as z-score normalization. Implementers may also choose the similarity calculation method and normalization method according to the specific situation. This embodiment does not impose any special restrictions.

[0036] The processes of averaging and fusing signals, calculating cosine similarity, and normalizing data using the maximum-minimum normalization method are all well-known techniques and will not be elaborated further.

[0037] Based on the similarity feature value, it can be understood that the similarity feature value is an index for evaluating the consistency between multiple collected vibration sound signals. It is used to characterize the repeatability stability of the signal in different sampling periods, reflecting the inherent properties of the insulator defect signal and the distinguishability of external random interference. Its calculation logic is to average multiple sets of vibration sound signals of a preset number to obtain a standard feature signal, calculate the normalized cosine similarity value between each set of vibration sound signals and the standard feature signal, and take the average value. The similarity feature value is affected by the difference in shape between each set of vibration sound signals and the standard signal. When the difference between vibration sound signals is smaller and the shape is more consistent, the similarity feature value is larger, reflecting that the influence of external random electromagnetic interference on the signal consistency is small and the collected signal is real and reliable. Conversely, when the difference between vibration sound signals is larger and the shape is more discrete, the similarity feature value is smaller, reflecting that the electromagnetic interference received in different sampling periods is huge, the vibration sound signal has high uncertainty, and the unreliability increases.

[0038] Furthermore, based on the amplitude fluctuation, the noise-affected characteristic value, and the similarity characteristic value, the first noise characteristic value of each vibration sound signal is quantified, specifically: The first noise characteristic value of each vibration sound signal is positively correlated with the amplitude fluctuation and noise characteristic value of each vibration sound signal, and negatively correlated with the similarity characteristic value.

[0039] It should be understood that a positive correlation means that the dependent variable increases as the independent variable increases, and the dependent variable decreases as the independent variable decreases. The specific relationship can be additive or multiplicative, etc., and is determined by the actual application. This application does not impose any special restrictions. A negative correlation means that the dependent variable decreases as the independent variable increases, and the dependent variable increases as the independent variable decreases. The relationship can be subtractive or divisive, etc., and is determined by the actual application.

[0040] Preferably, as one implementation, in this embodiment, the expression for the first noise characteristic value of the vibration sound signal i is: In the formula, This represents the first noise characteristic value of the vibration sound signal i; This represents the amplitude fluctuation of the vibration sound signal i; The noise-sensitive characteristic value of the vibration sound signal; The similarity feature value represents the vibration sound signal i; This indicates a preset constant greater than 0, used to prevent the denominator from being 0. The value is set manually; in this embodiment, The value of is 0.01. Under the premise of ensuring that the denominator is not 0 and does not excessively affect the calculation result, the implementer can also set it according to the specific situation. This embodiment does not impose any special restrictions. Sig() represents the Sigmoid normalization function. The Sigmoid normalization function is a well-known technique. The specific process of using it to normalize data will not be described in detail.

[0041] Based on the first noise characteristic value of each vibration sound signal, it can be understood that the first noise characteristic value is an index that comprehensively quantifies the degree of electromagnetic interference affecting the vibration sound signal in the time domain. It is used to characterize the comprehensive significance of amplitude fluctuation and disorder characteristics of the vibration sound signal, reflecting the level of noise pollution on the signal from the time domain perspective. The first noise characteristic value integrates amplitude fluctuation, noise-affected characteristic value, and similarity characteristic value. It is positively affected by amplitude fluctuation and noise-affected characteristic value and negatively affected by similarity characteristic value. When the amplitude fluctuation and noise-affected characteristic value are larger and the signal similarity is lower, the first noise characteristic value is larger, reflecting that the signal has suffered severe electromagnetic interference in the time domain and must be subjected to strong denoising processing. Conversely, when the amplitude is stable, the arrangement is orderly, and the signal is highly consistent, the first noise characteristic value is smaller, reflecting that the signal's time domain characteristics are pure and the degree of interference is low.

[0042] Thus, this embodiment determines the amplitude fluctuation by calculating the coefficient of variation of the difference between adjacent peaks to quantify the amplitude mutation characteristics. It combines the noise-affected feature value determined by the permutation entropy to characterize the peak permutation disorder, and the similarity feature value determined by the similarity between multiple groups of signals to characterize the signal repetition stability. The first noise feature value is constructed by weighted fusion, thereby comprehensively reflecting the degree of electromagnetic interference on the vibration sound signal in the time domain. This method effectively solves the problem of incomplete characterization by a single feature. By quantifying the noise level, it provides an accurate basis for subsequent adaptive adjustment of the denoising intensity, improving the accuracy and robustness of time-domain noise identification.

[0043] S2: By analyzing the distribution of all harmonic amplitudes of each vibration sound signal in the frequency domain, electromagnetic interference harmonics are screened out from all harmonic amplitudes; the proportion of the energy distribution of all electromagnetic interference harmonics in the total energy under the preset normal vibration frequency band, and the proportion of the energy distribution of all harmonics under the preset low frequency band in the frequency domain of each vibration sound signal in the total energy under the low frequency band, are measured to derive the second noise characteristic value of each vibration sound signal.

[0044] Given that electromagnetic interference in power grid lines mainly originates from power frequency (50Hz) fluctuations, when smart sensors are subjected to severe electromagnetic interference, the induced noise will excite abundant high-order harmonics in the spectrum of the vibration sound signal they collect. This manifests as significant energy peaks at the power frequency and its harmonic frequencies, which can easily lead to spectral aliasing when using power spectral density for defect identification. This can cause the original electromagnetic interference noise characteristics to be misjudged as defects at the bottom of the insulator. This is because, under normal vibration acoustic mechanisms, the spectral distribution has a clear frequency band correspondence. That is, the appearance of independent peaks in the 3kHz-6kHz range indicates that the insulator is in a normal state, while peaks in the 0-3kHz range correspond to defects at the bottom of the insulator, and peaks in the 6kHz-10kHz range correspond to defects at the top of the insulator. Therefore, it is necessary to accurately identify and eliminate harmonics in the frequency domain to prevent low-frequency interference harmonics from covering the true defect characteristics, thereby ensuring the accuracy of defect identification.

[0045] Based on the above analysis, this embodiment analyzes the distribution of all harmonic amplitudes of each vibration sound signal in the frequency domain to filter out electromagnetic interference harmonics from all harmonic amplitudes; it measures the proportion of the energy distribution of all electromagnetic interference harmonics in the total energy under a preset normal vibration frequency band, and the proportion of the energy distribution of all harmonics in the preset low-frequency band of each vibration sound signal in the total energy under the low-frequency band, to derive the second noise characteristic value of each vibration sound signal. The specific process is as follows: First, this embodiment analyzes the distribution of all harmonic amplitudes of each vibration sound signal in the frequency domain to filter out electromagnetic interference harmonics from all harmonic amplitudes. Specifically: The normalized values ​​of all harmonic amplitudes in the frequency domain of each vibration sound signal are used as input to the threshold segmentation algorithm, and the output is the segmentation threshold. Harmonics with amplitudes greater than or equal to the segmentation threshold are denoted as electromagnetic interference harmonics. This is a key parameter used to distinguish the real vibration component in the vibration sound signal from power frequency electromagnetic interference noise. It is used to characterize whether the energy at a specific frequency point originates from grid induction and reflects the distribution characteristics of electromagnetic noise in the frequency domain. By analyzing the harmonic amplitude distribution in the spectrum, the threshold segmentation algorithm is used to screen out electromagnetic interference harmonics with significant energy from all harmonics, thereby eliminating irrelevant frequency components and accurately locking the 50Hz and its harmonic noise generated by strong electromagnetic environment coupling, providing an accurate data basis for subsequent calculation of the second noise characteristic value.

[0046] It should be noted that there are many commonly used threshold segmentation algorithms. In this embodiment, the Otsu's inter-class variance algorithm is used to segment harmonics. In practical applications, as other implementation methods, implementers may also use other threshold segmentation methods according to specific circumstances. This embodiment does not impose any special restrictions.

[0047] It should be noted that there are many commonly used normalization methods. In this embodiment, the maximum-minimum normalization method is used to normalize all harmonic amplitudes. In practical applications, as other implementation methods, implementers may also use other normalization methods according to specific circumstances. This embodiment does not impose any special restrictions on the selection of normalization methods.

[0048] The process of normalizing harmonic amplitude using the maximum-minimum normalization method and the process of classifying harmonics using the maximum inter-class variance are both well-known techniques and will not be elaborated further.

[0049] Furthermore, this embodiment measures the proportion of the energy distribution of all electromagnetic interference harmonics in the total energy under a preset normal vibration frequency band, and the proportion of the energy distribution of all harmonics under a preset low frequency band in the frequency domain of each vibration sound signal in the total energy under the low frequency band, in order to derive the second noise characteristic value of each vibration sound signal. Specifically: The second noise characteristic value of vibration sound signal i The expression is: In the formula, This represents the second noise characteristic value of the i-th group of vibration sound signals; This represents the proportion of the sum of squares of the amplitudes of all electromagnetic interference harmonics of the i-th group of vibration sound signals in the total energy under the preset normal vibration frequency band. This represents the proportion of the sum of squares of all harmonic amplitudes in the preset low-frequency band within the frequency domain of the i-th vibration signal group in the total energy of the low-frequency band.

[0050] It should be noted that the preset normal vibration frequency band is 3kHz-6kHz, and the preset low frequency band is 0-3kHz; the process of obtaining harmonics is a well-known technology and will not be described in detail here.

[0051] Based on the second noise characteristic value of each vibration sound signal, it can be understood that the second noise characteristic value is an index for evaluating the intensity of power frequency harmonic interference to the vibration sound signal in the frequency domain. It is used to characterize the relative level of low-frequency electromagnetic interference energy under the condition of considering the aliasing of real defect characteristics, reflecting the potential threat of frequency domain noise after correction to defect identification. Its calculation logic is based on the selected electromagnetic interference harmonics, calculating their energy ratio in the preset normal vibration frequency band as the benchmark interference intensity, and combining the proportion of harmonic energy in the preset low-frequency band to the total energy for segmented discrimination correction: when the total energy in the low-frequency band is greater than the harmonic energy, it is determined that a real defect exists. The peak is used to lower the characteristic value using a correction coefficient to eliminate the influence of defective signals. When the total energy of the low-frequency band equals the harmonic energy, it is determined that it is only affected by electromagnetic interference, and the characteristic value maintains the baseline strength. The second noise characteristic value is affected by both the proportion of harmonic energy and the segmented correction logic. When the proportion of harmonic energy in the low-frequency band is higher and the correction logic does not determine it as a defect, the second noise characteristic value is larger, reflecting that the spectrum of the vibration sound signal is seriously polluted by electromagnetic interference, which is very easy to lead to misjudgment. Conversely, when the proportion of harmonic energy is lower or it is significantly corrected due to the presence of strong defective characteristics, the second noise characteristic value is smaller, reflecting that the frequency domain characteristics of the vibration sound signal are clear and the influence of electromagnetic interference is weak.

[0052] Thus, this embodiment analyzes the harmonic amplitude distribution and uses a threshold segmentation algorithm to screen out significant electromagnetic interference harmonics, calculates their energy proportion in the normal vibration frequency band and the energy proportion in the low frequency band, and introduces segmented correction logic to construct a second noise feature value, thereby accurately quantifying the actual intensity of power frequency harmonic interference in the frequency domain. This method effectively distinguishes between real defect signals and electromagnetic interference noise, avoids misjudgment caused by harmonic energy covering defect features, and ensures the accuracy of insulator defect identification in strong electromagnetic environments.

[0053] S3: Combine the first noise feature value and the second noise feature value to determine the disturbance degree of each vibration sound signal, so as to optimize the wavelet threshold in the wavelet denoising algorithm, and use the optimized wavelet denoising algorithm to denoise each vibration sound signal.

[0054] Based on the first and second noise characteristic values ​​obtained from S1 and S2, the degree of disturbance of each vibration sound signal is determined, specifically: The degree of disturbance of each vibration sound signal is positively correlated with the first noise characteristic value and the second noise characteristic value, respectively.

[0055] Preferably, as one implementation method, in this embodiment, the disturbance degree of the vibration sound signal i The expression is: In the formula, , These represent the first noise characteristic value and the second noise characteristic value of the vibration sound signal i, respectively; , These represent the preset first weight factor and the preset second weight factor, respectively. In this embodiment, the first weighting factor and the second weighting factor are preset to be 0.5. In actual application, the implementer can also set them according to the importance of the first noise feature value and the second noise feature value. This embodiment does not impose any special restrictions.

[0056] Preferably, the disturbance extraction process provided in this embodiment is illustrated in the following diagram. Figure 2 As shown.

[0057] Based on the perturbation degree of each vibration sound signal, it can be understood that perturbation degree is an index for comprehensively assessing the degree of influence of electromagnetic interference on the vibration sound signal. It is used to characterize the overall noise level of the vibration sound signal in the time and frequency domains, reflecting the severity of signal contamination and the intensity of the need for denoising strategies. Its calculation logic is to perform a weighted fusion and summation of the first and second noise characteristic values. Perturbation degree is jointly affected by the time domain noise characteristics and the frequency domain noise characteristics. The more severe the time domain fluctuations and frequency domain harmonic interference, the greater the perturbation degree, reflecting poor quality of the acquired vibration sound signal, requiring significant adjustment of denoising parameters to remove noise. Conversely, when the time domain is stable and the frequency domain is clean, the perturbation degree is smaller, reflecting excellent signal quality, requiring only slight denoising to retain effective features.

[0058] Furthermore, based on the aforementioned disturbance level, the wavelet threshold in the wavelet denoising algorithm is optimized, specifically: The optimized wavelet threshold corresponding to the i-th group of vibration sound signals The expression is: In the formula, This indicates the degree of disturbance to the i-th group of vibration sound signals; This represents the wavelet threshold obtained by applying the wavelet denoising algorithm before optimization to the i-th group of vibration sound signals; This represents a preset adjustment coefficient, with a value range of 0.1 to 0.5. In this embodiment, it is set to... The value is 0.4. In practical applications, as other implementation methods, implementers can also set it according to specific circumstances. This embodiment does not impose any special restrictions.

[0059] The process of obtaining the wavelet threshold using the wavelet denoising algorithm is a well-known technique and will not be elaborated further.

[0060] Based on the optimized wavelet threshold, it can be understood that the wavelet threshold is a denoising critical value that is dynamically adjusted according to the degree of disturbance of the vibration sound signal. It is used to characterize the balance standard of the adaptive wavelet denoising algorithm in preserving signal details and removing noise, reflecting the adaptive matching ability of denoising intensity and signal characteristics. Its calculation logic is to add the product of the disturbance degree and the adjustment coefficient as an incremental correction term on the basis of the general wavelet threshold. The optimized wavelet threshold is positively affected by the signal disturbance degree. When the disturbance degree is greater, the optimized wavelet threshold is larger, reflecting that the algorithm has enhanced noise suppression and filters out more high-frequency wavelet coefficients to eliminate interference. Conversely, when the disturbance degree is smaller, the optimized wavelet threshold is closer to the general threshold, reflecting that the algorithm tends to retain more vibration sound signal details and avoid excessive denoising that leads to the loss of real features.

[0061] Furthermore, each vibration sound signal is used as the input to the wavelet denoising algorithm, where the wavelet threshold is set to the corresponding optimized wavelet threshold, and the denoised vibration sound signal is output.

[0062] The process of using wavelet denoising algorithm to denoise signals is a well-known technique and will not be elaborated further.

[0063] Thus, this embodiment determines the degree of disturbance by weighted fusion of the first noise feature value in the time domain and the second noise feature value in the frequency domain, and obtains an optimized wavelet threshold by adaptively correcting the general wavelet threshold based on the degree of disturbance, thereby achieving adaptive denoising of the vibration sound signal. This method establishes a positive correlation between the degree of disturbance and the denoising intensity, dynamically adjusts the wavelet threshold according to the severity of signal contamination, enhances the denoising intensity to eliminate noise when there is severe interference, and reduces the threshold to retain more detailed features when the signal quality is good. This effectively solves the problem of poor denoising effect of fixed threshold, and significantly improves the signal-to-noise ratio and purity of the insulator vibration sound signal.

[0064] The insulator detection module 103 is used to detect defects in insulators based on all the noise-reduced vibration sound signals.

[0065] After the microprocessor unit of the intelligent sensor completes the denoising of the vibration sound signal, it transmits the two sets of denoised vibration sound signals to the insulator detection module of the equipment through the wireless communication technology of the communication unit, realizing remote live-line detection of wireless pulse insulators. Specifically: If the peak values ​​of a preset number of vibration sound signals in the frequency domain are all within the preset normal vibration frequency band, then the insulator is not defective; otherwise, the insulator is defective, triggering the early warning device to issue an alarm and reminding staff to inspect and repair it.

[0066] It should be noted that the preset ratio is set manually. In this embodiment, the preset ratio is 95%. Setting 95% as the judgment ratio is to ensure the accuracy of detection while taking into account the fault tolerance and anti-interference capability in actual engineering applications. Specifically, although theoretically the spectral peaks of a normal insulator should all be concentrated in the preset normal frequency band of 3kHz-6kHz, in the actual complex electromagnetic environment or during sensor acquisition, the denoised signal may still have a few random noise points or abnormally high values. If these discrete points caused by non-defects are judged as defects by a 100% strict standard, it will lead to unnecessary false alarms and maintenance costs. Therefore, using 95% as the threshold allows a very small number (within 5%) of abnormal frequency band peaks to exist, which can effectively filter out residual random noise interference, greatly reduce the false judgment rate caused by environmental noise, and at the same time ensure that when the insulator is truly defective, the significant defect characteristic peak (energy much greater than noise) it produces will significantly exceed this fault tolerance range, thereby ensuring the sensitivity of fault detection and achieving the best balance between accurately identifying real defects and tolerating environmental noise.

[0067] Thus, this embodiment determines whether there is a defect in the insulator and triggers an early warning by analyzing whether all peaks of the denoised vibration sound signal in the frequency domain are within the preset normal vibration frequency band. This method uses wireless transmission to achieve non-contact remote real-time monitoring, and the combination of denoising processing improves the accuracy and reliability of spectrum analysis, thereby improving the precision of insulator defect detection.

[0068] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0069] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. The focus of each embodiment is to describe the differences from other embodiments.

[0070] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A wireless pulse insulator live-line detection device, characterized in that, The device includes: The data acquisition module is used to collect multiple vibration and acoustic signals at the insulator. The insulator analysis module is used to evaluate the amplitude fluctuation of each vibration sound signal by analyzing the dispersion of the difference between all adjacent peaks in each vibration sound signal at the insulator, and to measure the disorder of all peaks in each vibration sound signal, as well as the similarity between vibration acoustic signals, in order to quantify the first noise characteristic value of each vibration sound signal. By analyzing the distribution of all harmonic amplitudes of each vibration sound signal in the frequency domain, electromagnetic interference harmonics are screened out from all harmonic amplitudes; the proportion of the energy distribution of all electromagnetic interference harmonics in the total energy under the preset normal vibration frequency band, and the proportion of the energy distribution of all harmonics in the preset low frequency band of each vibration sound signal in the total energy under the low frequency band are measured, so as to derive the second noise characteristic value of each vibration sound signal. By combining the first noise feature value and the second noise feature value, the perturbation degree of each vibration sound signal is determined to optimize the wavelet threshold in the wavelet denoising algorithm. The optimized wavelet denoising algorithm is then used to denoise each vibration sound signal. The perturbation degree of each vibration sound signal is positively correlated with the first noise feature value and the second noise feature value, respectively. The insulator detection module is used to detect defects in insulators based on all the noise-reduced vibration and acoustic signals.

2. The wireless, pulsed insulator remote charging detection device of claim 1, wherein, The amplitude fluctuation of each vibration sound signal is the coefficient of variation of the difference between all adjacent peak values ​​in each vibration sound signal.

3. The wireless, pulsed insulator remote charging detection apparatus of claim 1, wherein, The quantization method for the first noise characteristic value of each vibration sound signal is as follows: Based on the degree of disorder of all peaks in each vibration sound signal, the noise characteristic value of each vibration sound signal is determined. Based on the similarity between vibration sound signals at the insulator, the similarity characteristic value of the insulator is determined; The first noise characteristic value of each vibration sound signal is positively correlated with the amplitude fluctuation and noise characteristic value of each vibration sound signal, and negatively correlated with the similarity characteristic value.

4. The wireless, pulsed insulator remote charging detection apparatus of claim 3, wherein, The noise-affected characteristic value of each vibration sound signal is the information entropy of all peak values ​​in each vibration sound signal.

5. The wireless, pulsed insulator remote charging detection apparatus of claim 3, wherein, The method for determining the similarity characteristic values ​​of the insulator is as follows: The characteristic signal is obtained by averaging the amplitudes of all vibration sound signals. The mean of the similarity between all vibration sound signals and the characteristic signal is calculated and denoted as the similarity characteristic value of the insulator.

6. The wireless pulse insulator remote live-line detection device according to claim 1, characterized in that, The process of filtering out electromagnetic interference harmonics from all harmonic amplitudes includes: The normalized values ​​of all harmonic amplitudes of each vibration sound signal in the frequency domain are used as the input of the threshold segmentation algorithm, and the segmentation threshold is output. Harmonics with amplitudes greater than or equal to the segmentation threshold are denoted as electromagnetic interference harmonics.

7. The wireless pulse insulator remote live-line detection device according to claim 1, characterized in that, The expression for the second noise characteristic value of each vibration sound signal is: In the formula, This represents the second noise characteristic value of the i-th group of vibration sound signals; This represents the proportion of the sum of squares of the amplitudes of all electromagnetic interference harmonics of the i-th group of vibration sound signals in the total energy under the preset normal vibration frequency band. This represents the proportion of the sum of squares of all harmonic amplitudes in the preset low-frequency band within the frequency domain of the i-th vibration signal group in the total energy of the low-frequency band.

8. The wireless pulse insulator remote live-line detection device according to claim 1, characterized in that, The wavelet threshold in the optimized wavelet denoising algorithm includes: The optimized wavelet threshold corresponding to the i-th group of vibration sound signals The expression is: In the formula, This indicates the degree of disturbance to the i-th group of vibration sound signals; This represents the wavelet threshold obtained by applying the wavelet denoising algorithm before optimization to the i-th group of vibration sound signals; This indicates the preset adjustment coefficient.

9. The wireless pulse insulator remote live-line detection device according to claim 1, characterized in that, The defect detection of the insulator includes: If the peak values ​​of a preset number of vibration sound signals in the frequency domain are all within the preset normal vibration frequency band, then the insulator is not defective; otherwise, the insulator is defective.

Images