A method and system for online insulation detection of power transmission and distribution

Abstract

The present application relates to the field of power transmission and distribution insulation detection, and in particular to a method and system for online insulation detection of power transmission and distribution, the method comprising: installing a leakage sensor on the power transmission and distribution circuit to obtain historical sampling signals and real-time sampling signals about time and current; setting a window according to the historical sampling signal; taking the length of the window as a period, calculating the similarity between the real-time sampling signals corresponding to the window in the current time period and the previous time period, and removing the corresponding sampling signal segment according to the similarity; completing the real-time sampling signal according to the interpolation method, and determining the endpoints of the optimal cutoff frequency interval according to the completed real-time sampling signal; deleting the real-time sampling signal outside the optimal cutoff frequency interval to obtain a denoised leakage signal, and performing abnormal detection on the leakage signal. The present application has the effect of setting the cutoff frequency according to the actual situation of power transmission and distribution.

Description

Technical Field

[0001] The present application relates to the field of power transmission and distribution insulation detection, and in particular to a power transmission and distribution online insulation detection method and system. Background Art

[0002] Online insulation testing for power transmission and distribution refers to the real-time or periodic monitoring of insulation materials and insulation equipment in transmission lines and distribution networks during the operation of the power system to evaluate their insulation performance and status.

[0003] For example, the patent document with the authorization announcement number CN114236312B discloses a method and system for online segmented insulation status monitoring and fault location, in which a number of cable health monitoring devices are arranged at intervals on the cable to collect cable monitoring information and obtain the cable monitoring information processed by each cable health monitoring device in real time; if a fault occurs, the leakage current waveforms processed by all cable health monitoring devices are collected, and the leakage current waveforms are timestamped; the leakage current waveforms are sorted on the time axis, and the fault location is performed according to the leakage current waveforms of each cable health monitoring device at the same time.

[0004] When collecting leakage signals, noise signals may appear due to reasons such as loose connections of sampling devices. Existing technologies can use filters to separate signals and noise. The cutoff frequency is one of the key parameters of the filter, which determines the boundary between the passband and stopband of the filter. Selecting a suitable cutoff frequency can ensure that useful signal components pass through while attenuating noise components to a minimum. Existing technologies usually use a fixed cutoff frequency for filtering.

[0005] In the power transmission and distribution system, the characteristics of signals and noise may change with time, environmental conditions, equipment status and other factors. Filters with fixed cutoff frequencies may not be able to adapt to these changes, resulting in the inability to effectively separate signals and noise in some cases. The cutoff frequency needs to comprehensively consider the characteristics of signals and noise, the type and performance requirements of the filter, and the test results in actual applications. Different cables and equipment may have different insulation characteristics and noise levels. Therefore, how to set the filter cutoff frequency according to the actual situation of power transmission and distribution has become an urgent problem to be solved. Summary of the invention

[0006] In order to solve the technical problem of how to set the filter cutoff frequency according to the actual situation of power transmission and distribution, the present application provides a method and system for online insulation detection of power transmission and distribution.

[0007] In a first aspect, the present application provides a method for online insulation detection of power transmission and distribution, which adopts the following technical solution:

[0008] A method for detecting insulation of power transmission and distribution online comprises the following steps: installing a leakage sensor on a power transmission and distribution circuit to obtain historical sampling signals and real-time sampling signals about time and current; setting a window according to the historical sampling signals; taking the length of the window as a period, calculating the similarity between the real-time sampling signals corresponding to the current time period and the previous time period in the window, and when the similarity is less than a preset threshold, deleting the sampling signal corresponding to the current time period, and similarly calculating the similarity between the next time period and the previous time period until the similarity is greater than or equal to the threshold, the deletion is completed; filling the real-time sampling signal according to the interpolation method, and determining the endpoint of the optimal cutoff frequency interval according to the filled real-time sampling signal; and separating the real-time sampling signal at the optimal cutoff frequency. The signal outside the rate interval is deleted to obtain the denoised leakage signal, and the leakage signal is detected for abnormality; the window setting method is as follows: Fourier transform is performed on the historical sampling signal to obtain a frequency spectrum diagram about frequency and amplitude; Gaussian distribution is performed on the amplitude to obtain a Gaussian distribution curve, and the abnormal amplitude interval is obtained according to the Gaussian distribution curve, and the frequency range corresponding to the abnormal amplitude interval is used as the abnormal frequency band; wavelet transform is performed on the historical sampling signal to obtain a time-frequency diagram about time and frequency, and the time period of the abnormal frequency band is used as the initial window size, and the data visibility degree of the historical sampling signal within the initial window size is calculated, and the initial window size corresponding to the maximum value of the data visibility degree is used as the optimal size, and the window is set according to the optimal size.

[0009] In the real-time power transmission and distribution system, the optimal cutoff frequency range is adaptively set according to the actual situation of power transmission and distribution. Setting an adaptive filter can improve the signal-to-noise ratio to a certain extent, and by automatically adjusting the filter parameters to adapt to different noise environments, it can more effectively separate leakage signals and noise signals, and improve the accuracy of online insulation detection of power transmission and distribution.

[0010] For leakage signals and noise signals, leakage signals are usually concentrated in a specific frequency range, with a small amplitude variation. Leakage signals usually have a specific pulse shape and duration, while noise signals are more variable, usually covering a wider frequency band, and with a larger amplitude variation. Noise signals appear as random and irregular signals. Therefore, the amplitude characteristics and frequency characteristics of the sampled signal are found through Fourier transform. After Gaussian distribution of the amplitude characteristics, the abnormal amplitude that does not conform to most of the amplitudes is found. The corresponding abnormal frequency is found through the spectrum diagram, and the abnormal frequency band is divided. However, Fourier transform can only provide frequency information, but not the distribution of these frequency components in time. Therefore, it is necessary to find a frequency signal with a time series through wavelet transform to find the period when the sampled signal is abnormal. Delete the signal segment of the abnormal period, and then fill the signal according to the normal leakage signal, and set the filter bandwidth according to the filled sampling signal.

[0011] Optionally, the data visibility is the entropy value of the historical sampling signal in the window.

[0012] The higher the entropy value, the more frequency components the signal contains and the greater the uncertainty of the signal, that is, the greater the variability of the signal. Conversely, the smaller the variability of the signal.

[0013] Optionally, the method for calculating the degree of data visibility is: for any window, construct a sequence of amplitude and phase of the historical sampling signal; take the Pearson correlation coefficient of the sequence corresponding to any window and the sequence corresponding to the previous adjacent window as the first correlation value, take the Pearson correlation coefficient of the sequence corresponding to any window and the sequence corresponding to the next adjacent window as the second correlation value, and take the average of the first correlation value and the second correlation value as the degree of data visibility.

[0014] The Wilson correlation coefficient is a statistic that measures the degree of correlation between two variables. Its value is between -1 and 1. The closer the value is to 0, the weaker the correlation between the two sequences. The closer it is to 1, the more positive the correlation is. The closer it is to -1, the more negative the correlation is. Therefore, the first correlation value and the second correlation value can quantify the similarity or difference of the signals in adjacent windows. The mean of the first correlation value and the second correlation value is used as the degree of data visibility. The mean can reflect the correlation between the data in the window and the previous and next data.

[0015] Optionally, the data explicitness expression is: ; Display Window The degree of data visibility of the internal historical sampling signal; Display Window The corresponding sequence With window The corresponding sequence The first correlation value of Display Window The corresponding sequence With window The corresponding sequence The second correlation value of

[0016] ; Display Window The sequence corresponding to the historical sampling signal, Display Window The sequence corresponding to the historical sampling signal, Representation sequence No. data, Representation sequence No. data, Representation sequence The mean of the data, Representation sequence The mean of the data in ;

[0017] ; Display Window The sequence corresponding to the historical sampling signal, Representation sequence No. data, Representation sequence The mean of the data, express , or The total amount of data in the sequence , and Equal length.

[0018] Optionally, the similarity is calculated as follows: taking the sequence corresponding to the real-time sampling signal segment contained in the window in the current time period as the first sequence, taking the sequence corresponding to the real-time sampling signal segment contained in the window in the previous time period as the second sequence, and taking the Pearson correlation coefficient between the first sequence and the second sequence as the similarity.

[0019] The Pearson correlation coefficient between the first sequence and the second sequence is close to 1, which means that the more similar the two sequences are, the more similar the sampling signal segments in the windows corresponding to the two sequences are, and the less likely a noise signal will appear.

[0020] Optionally, the similarity degree is calculated as follows: taking the sequence corresponding to the real-time sampling signal segment contained in the window in the current time period as the first sequence, taking the sequence corresponding to the real-time sampling signal segment contained in the window in the previous time period as the second sequence, and normalizing the Euclidean distance between the data points of the first sequence and the second sequence as the similarity degree.

[0021] The smaller the Euclidean distance between the data points in the two signals, the smaller the difference between the data points, the more similar the changes in the two signals, and the smaller the possibility of noise.

[0022] Optionally, the endpoints of the optimal cutoff frequency interval are determined based on the padded real-time sampling signal, including: taking the maximum value or the average of the maximum values ​​of the frequency of the padded real-time sampling signal as the right endpoint of the optimal cutoff frequency interval, and taking the minimum value or the average of the minimum values ​​of the frequency of the padded real-time sampling signal as the left endpoint of the optimal cutoff frequency interval.

[0023] In a second aspect, the present application provides an online insulation detection system for power transmission and distribution, which adopts the following technical solution:

[0024] A power transmission and distribution online insulation detection system, a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned power transmission and distribution online insulation detection method is implemented.

[0025] The beneficial effect is that the above-mentioned online insulation detection method for power transmission and distribution is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a system based on the memory and the processor for easy use.

[0026] This application has the following technical effects:

[0027] 1. In the real-time power transmission and distribution system, the optimal cutoff frequency range is adaptively set according to the actual situation of power transmission and distribution. Setting an adaptive filter can improve the signal-to-noise ratio to a certain extent, and by automatically adjusting the filter parameters to adapt to different noise environments, it can more effectively separate leakage signals and noise signals, and improve the accuracy of online insulation detection of power transmission and distribution.

[0028] 2. For leakage signals and noise signals, leakage signals are usually concentrated in a specific frequency range, with a small amplitude variation. Leakage signals usually have a specific pulse shape and duration. Noise signals are more variable, usually covering a wider frequency band, and have a larger amplitude variation. Noise signals appear as random and irregular signals. Therefore, the amplitude characteristics and frequency characteristics of the sampled signal are found through Fourier transform. After Gaussian distribution of the amplitude characteristics, the abnormal amplitude that does not conform to most of the amplitudes is found. The corresponding abnormal frequencies are found through the spectrum diagram, and the abnormal frequency bands are divided. However, Fourier transform can only provide frequency information, but cannot provide the distribution of these frequency components in time. Therefore, it is necessary to find a frequency signal with a time series through wavelet transform to find the period when the sampled signal is abnormal. Delete the signal segment of the abnormal period, and then fill the signal according to the normal leakage signal, and set the filter bandwidth according to the filled sampling signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present application will become easily understood. In the accompanying drawings, several embodiments of the present application are shown in an exemplary and non-restrictive manner, and the same or corresponding numbers represent the same or corresponding parts.

[0030] Figure 1 It is a method flow chart of an online insulation detection method for power transmission and distribution according to an embodiment of the present application.

[0031] Figure 2 This is a method flow chart of step S2 in an online insulation detection method for power transmission and distribution according to an embodiment of the present application. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0033] It should be understood that when the terms "first", "second", etc. are used in the claims, specification and drawings of the present application, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the specification and claims of the present application indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their collections.

[0034] The embodiment of the present application discloses a method for online insulation detection of power transmission and distribution, which is used to monitor the leakage of power transmission and distribution system. Figure 1 , comprising steps S1 to S4:

[0035] S1: Install leakage sensors on the power transmission and distribution circuits to obtain historical sampling signals and real-time sampling signals about time and current.

[0036] Leakage current is caused by the leakage of current to an unexpected path due to damage to the insulation material or other reasons, such as flowing from the live wire to the ground, forming a loop between the live wire and the ground. Leakage current measurement is usually performed at a high voltage to expose the weakness of the insulation material, and the leakage current value can be directly read by a microammeter. In one embodiment, leakage sensors can be installed at equal intervals on the power transmission and distribution circuit, and leakage sensors can also be installed at key nodes of the power transmission and distribution circuit or areas where insulation problems may exist to detect leakage current.

[0037] The current values ​​are recorded at a certain sampling rate (for example, 60 samples per second). These current values ​​change over time, forming time series data, that is, sampling signals, recorded as The sampling signal contains a timestamp and the corresponding current value. The timestamp marks the specific moment when the current value is sampled, while the current value reflects the current intensity in the circuit at that moment.

[0038] S2: Set the window based on the historical sampling signal. Figure 2 The window setting method includes steps S20 to S22, which are as follows:

[0039] S20: Perform Fourier transform on the historical sampling signal to obtain a frequency spectrum diagram of frequency and amplitude.

[0040] Fourier transform is a mathematical method for converting time domain signals into frequency domain signals. The historical sampling signal (time series data) is converted into a spectrum signal and a spectrum graph is constructed. The horizontal axis of the spectrum graph is frequency and the vertical axis is amplitude, which shows the amplitude of different frequency components. Fourier transform is an existing technology and will not be described in detail here. The spectrum function obtained by Fourier transforming the historical sampling signal is recorded as .

[0041] S21: Performing Gaussian distribution on the amplitude to obtain a Gaussian distribution curve, obtaining an abnormal amplitude interval according to the Gaussian distribution curve, and taking a frequency range corresponding to the abnormal amplitude interval as an abnormal frequency band.

[0042] Gaussian distribution, also known as normal distribution, is an existing technology. The Gaussian distribution curve can filter out amplitudes that do not conform to the distribution of most data.

[0043] In one embodiment, the endpoints of the abnormal amplitude interval are as well as , is the mean of the data, Refers to the standard deviation of the data. Data points outside the abnormal amplitude interval are considered abnormal amplitudes. Three times the standard deviation can also be replaced by two times the standard deviation or one times the standard deviation. The present application prefers three times the standard deviation.

[0044] In one embodiment, the abnormal amplitudes are clustered to obtain multiple clusters, the maximum amplitude and the minimum amplitude in each cluster are used as the interval endpoints of the abnormal amplitude interval, and multiple abnormal amplitude intervals are obtained, and the maximum frequency and the minimum frequency in the cluster are used as the interval endpoints of the abnormal frequency band.

[0045] The clustering method may be an existing clustering algorithm that is applicable to the application scenario of the present application, such as a DBSCAN clustering algorithm or a K-means clustering algorithm. This embodiment is applicable to the case where the outliers are not distributed in a single area but are scattered in multiple areas.

[0046] In one embodiment, clustering may not be performed, and the maximum and minimum values ​​of the abnormal amplitudes may be directly used as the endpoints of the abnormal amplitude interval. This method is suitable for scenarios where data distribution is relatively single.

[0047] For leakage signals and noise signals, leakage signals are usually concentrated in a specific frequency range, with a small amplitude variation. Leakage signals usually have a specific pulse shape and duration, while noise signals are more variable, usually covering a wider frequency band, and with a larger amplitude variation. Noise signals appear as random and irregular signals. Therefore, the amplitude and frequency characteristics of the sampled signal are found through Fourier transform. After Gaussian distribution of the amplitude characteristics, the abnormal amplitude that does not conform to most of the amplitudes is found. The corresponding abnormal frequency is found through the spectrum diagram, and the abnormal frequency band is divided. However, Fourier transform can only provide frequency information, but cannot provide the distribution of these frequency components in time, so it is necessary to find a frequency signal with a time series through wavelet transform to find the time period when the sampled signal is abnormal.

[0048] S22: Perform wavelet transform on the historical sampling signal to obtain a time-frequency diagram of time and frequency, take the time period of the abnormal frequency band as the initial window size, calculate the data visibility of the historical sampling signal within the initial window size, take the initial window size corresponding to the maximum value of the data visibility as the optimal size, and set the window according to the optimal size.

[0049] In one embodiment, the data visualization degree may be the entropy value of the historical sampling signal in the window. The higher the entropy value, the more frequency components contained in the signal, the greater the uncertainty of the signal, that is, the greater the variability of the signal, and vice versa, the smaller the variability of the signal.

[0050] In order to combine the historical sampling signals in the window with the historical sampling signals in the adjacent windows for comprehensive evaluation, in other embodiments, the calculation method of the data visibility degree may also be:

[0051] For any window, a sequence of the amplitude and phase of the historical sampled signal is constructed.

[0052] Specifically, for the historical sampling signal , which is expressed as a complex number, ,in, is the amplitude, is the phase, Indicates a moment, represents an imaginary number. Then the expression of the sequence is: .

[0053] The Pearson correlation coefficient between the sequence corresponding to any window and the sequence corresponding to the previous adjacent window is taken as the first correlation value, and the Pearson correlation coefficient between the sequence corresponding to any window and the sequence corresponding to the next adjacent window is taken as the second correlation value.

[0054] The Pearson correlation coefficient is a statistic that measures the degree of correlation between two variables. Its value is between -1 and 1. The closer the value is to 0, the weaker the correlation between the two sequences. The closer the value is to 1, the more positive the correlation is. The closer the value is to -1, the more negative the correlation is. Therefore, the first correlation value and the second correlation value can quantify the similarity or difference of signals in adjacent windows.

[0055] The mean of the first correlation value and the second correlation value is used as the degree of data visualization. The mean can reflect the correlation between the data in the window and the previous and next data.

[0056] In one embodiment, the data visibility degree expression is:

[0057] ; Display Window The degree of data visibility of the internal historical sampling signal; Display Window The corresponding sequence With window The corresponding sequence The first correlation value of Display Window The corresponding sequence With window The corresponding sequence The second correlation value of .

[0058] ; Display Window The sequence corresponding to the historical sampling signal, Display Window The sequence corresponding to the historical sampling signal, Representation sequence No. data, Representation sequence No. data, Representation sequence The mean of the data, Representation sequence The mean of the data in ;

[0059] ; Display Window The sequence corresponding to the historical sampling signal, Representation sequence No. data, Representation sequence The mean of the data, express , or The total amount of data in the sequence , and The lengths are the same, so the total amount of data in the sequences is the same.

[0060] The size corresponding to the window with the highest degree of data visibility is taken as the optimal size setting window.

[0061] S3: Taking the length of the window as the period, calculate the similarity between the real-time sampling signals corresponding to the current time period and the previous time period in the window. When the similarity is less than the preset threshold, delete the sampling signal corresponding to the current time period. Similarly, calculate the similarity between the next time period and the previous time period until the similarity is greater than or equal to the threshold, and then complete the elimination.

[0062] For example, the window moves one time (1 second) each time. If the length of the window is 10 seconds, then the similarity between the real-time sampling signal in the window at time T and the real-time sampling signal in the window at time T+10 is calculated with a period of 10 seconds.

[0063] In one embodiment, the similarity is calculated by taking the sequence corresponding to the real-time sampling signal segment included in the window in the current period as the first sequence, taking the sequence corresponding to the real-time sampling signal segment included in the window in the previous period as the second sequence, and taking the Pearson correlation coefficient between the first sequence and the second sequence as the similarity. The calculation formula of the Pearson correlation coefficient is not described in detail here.

[0064] The larger the Pearson correlation coefficient between the first sequence and the second sequence, the more similar the two sequences are, and the more similar the sampling signal segments in the windows corresponding to the two sequences are. At this time, the possibility of the occurrence of noise signals is smaller. Conversely, the possibility of the occurrence of noise signals is greater.

[0065] The Euclidean distance between the data points of the first sequence and the second sequence can also be normalized as the degree of similarity, which can be expressed as follows: , where S represents the similarity between the first sequence and the second sequence, Indicates An exponential function with base Perform normalization processing, represents the Euclidean distance, represents the first sequence, represents the second sequence, Indicates the current period. Indicates the previous period.

[0066] The Euclidean distance between two sequences is defined as the square root of the sum of the squares of the distances between corresponding points in the sequences. For example, suppose the corresponding points in the two sequences are and . Then the distance between these two points is:

[0067] , the Euclidean distance of the entire sequence is the square root of the sum of the squares of all these distances.

[0068] The smaller the Euclidean distance between the data points in the two signals, the smaller the difference between the data points, the more similar the changes in the two signals, and the smaller the possibility of noise. Conversely, the greater the possibility of noise.

[0069] For example, the preset threshold is 0.5. Time period, Time period and If the similarity of the signals corresponding to the time period is lower than 0.5, delete Signals within the time window continue to be calculated Time period and If the similarity of the signals corresponding to the time period is still lower than 0.5, delete Repeat the above operation for the signals in the time window until the similarity of the corresponding historical sampling signals in the two windows is greater than 0.5, and the elimination is completed.

[0070] S4: The real-time sampling signal is complemented by the interpolation method, and the endpoints of the optimal cutoff frequency interval are determined according to the complemented real-time sampling signal; the signal outside the optimal cutoff frequency interval of the real-time sampling signal is deleted to obtain a denoised leakage signal, and anomaly detection is performed on the leakage signal.

[0071] The real-time sampling signal can be complemented by an interpolation method in the prior art that is applicable to the present application, such as linear interpolation, polynomial interpolation, spline interpolation, etc.

[0072] The maximum value or the maximum mean value of the frequency of the padded real-time sampling signal is used as the right endpoint of the optimal cutoff frequency interval, and the minimum value or the minimum mean value of the frequency of the padded real-time sampling signal is used as the left endpoint of the optimal cutoff frequency interval.

[0073] In cable insulation testing, improving the signal-to-noise ratio is crucial because it directly affects the detectability of the signal and the accuracy of the analysis. Setting an adaptive filter can improve the signal-to-noise ratio to a certain extent, by automatically adjusting the filter parameters to adapt to different noise environments, thereby more effectively separating the useful signal (leakage signal) and the noise signal.

[0074] The real-time sampling signal outside the optimal cutoff frequency range is deleted to obtain the denoised leakage signal, and then the leakage signal is subjected to subsequent abnormal detection and analysis. The leakage signal is a signal about current and time. When the current value is continuously greater than the preset current threshold for a period of time, it indicates that leakage may occur.

[0075] In one embodiment, the current threshold can be set to 300 mA, and the current threshold can be adjusted according to the actual application scenario. If the current value detected by the system suddenly increases from the normal 50 mA to 350 mA within a period of time, and continues to exceed the 300 mA threshold, the system will consider that a leakage has occurred and issue an alarm, indicating that immediate inspection and processing are required to avoid possible electrical fires.

[0076] An embodiment of the present application also discloses an online insulation detection system for power transmission and distribution, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an online insulation detection method for power transmission and distribution according to the present application is implemented.

[0077] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0078] In the present application, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high bandwidth memory HBM (High Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or accessible or connectable to a device.

[0079] Although this specification has shown and described a plurality of embodiments of the present application, it is obvious to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will think of many changes, modifications and alternatives without departing from the thought and spirit of the present application. It should be understood that in the process of practicing the present application, various alternatives to the embodiments of the present application described herein may be adopted.

[0080] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for online insulation detection of power transmission and distribution, characterized in that: Includes steps: Install leakage sensors on the power transmission and distribution circuits to obtain historical sampling signals and real-time sampling signals about time and current; set windows based on historical sampling signals; Taking the length of the window as the period, calculate the similarity between the real-time sampling signals corresponding to the current time period and the previous time period. When the similarity is less than the preset threshold, delete the sampling signal corresponding to the current time period. Similarly, calculate the similarity between the next time period and the previous time period until the similarity is greater than or equal to the threshold, and then complete the elimination. The real-time sampling signal is padded according to the interpolation method, and the endpoints of the optimal cutoff frequency interval are determined according to the padded real-time sampling signal; the signal of the real-time sampling signal outside the optimal cutoff frequency interval is deleted to obtain the denoised leakage signal, and the leakage signal is detected for abnormality; The window setting method is: perform Fourier transform on the historical sampling signal to obtain a spectrum diagram about frequency and amplitude; The amplitude is Gaussian distributed to obtain a Gaussian distribution curve, an abnormal amplitude interval is obtained according to the Gaussian distribution curve, and the frequency range corresponding to the abnormal amplitude interval is used as the abnormal frequency band, including: setting an initial abnormal amplitude interval, and the data points exceeding the initial abnormal amplitude interval are abnormal amplitudes, clustering the abnormal amplitudes to obtain multiple cluster clusters, and the maximum amplitude and the minimum amplitude in each cluster cluster are used as the interval endpoints of the abnormal amplitude interval, and multiple abnormal amplitude intervals are obtained, and the maximum frequency and the minimum frequency in the cluster cluster are used as the interval endpoints of the abnormal frequency band; performing wavelet transform on the historical sampling signal to obtain a time-frequency diagram about time and frequency, and the abnormal The time period of the constant frequency band is used as the initial window size; the data visibility of the historical sampling signal within the initial window size is calculated, including: for any window, constructing a sequence of the amplitude and phase of the historical sampling signal; taking the Pearson correlation coefficient of the sequence corresponding to the arbitrary window and the sequence corresponding to the previous adjacent window as the first correlation value, taking the Pearson correlation coefficient of the sequence corresponding to the arbitrary window and the sequence corresponding to the next adjacent window as the second correlation value, and taking the average of the first correlation value and the second correlation value as the data visibility; taking the initial window size corresponding to the maximum value of the data visibility as the optimal size, and setting the window according to the optimal size.

2. The method for online insulation detection of power transmission and distribution according to claim 1, characterized in that: The degree of data visibility can also be replaced by the entropy value of the historical sampling signal within the window.

3. The method for online insulation detection of power transmission and distribution according to claim 1, characterized in that: The expression of data visibility is: ; Display Window The degree of data visibility of the internal historical sampling signal; Display Window The corresponding sequence With window The corresponding sequence The first correlation value of Display Window The corresponding sequence With window The corresponding sequence The second correlation value of ; Display Window The sequence corresponding to the historical sampling signal, Display Window The sequence corresponding to the historical sampling signal, Representation sequence No. data, Representation sequence No. data, Representation sequence The mean of the data, Representation sequence The mean of the data in ; ; Display Window The sequence corresponding to the historical sampling signal, Representation sequence No. data, Representation sequence The mean of the data, express , or The total amount of data in the sequence , and Equal length.

4. The method for online insulation detection of power transmission and distribution according to claim 2 or 3, characterized in that: The similarity is calculated as follows: the sequence corresponding to the real-time sampling signal segment contained in the window in the current period is taken as the first sequence, the sequence corresponding to the real-time sampling signal segment contained in the window in the previous period is taken as the second sequence, and the Pearson correlation coefficient between the first sequence and the second sequence is taken as the similarity.

5. The method for online insulation detection of power transmission and distribution according to claim 2 or 3, characterized in that: The similarity is calculated as follows: the sequence corresponding to the real-time sampling signal segment contained in the window in the current period is taken as the first sequence, the sequence corresponding to the real-time sampling signal segment contained in the window in the previous period is taken as the second sequence, and the Euclidean distance between the data points of the first sequence and the second sequence is normalized as the similarity.

6. The method for online insulation detection of power transmission and distribution according to claim 1, characterized in that: Determining the endpoints of the optimal cutoff frequency interval according to the padded real-time sampling signal includes: taking the maximum value or the average of the maximum values ​​of the frequency of the padded real-time sampling signal as the right endpoint of the optimal cutoff frequency interval, and taking the minimum value or the average of the minimum values ​​of the frequency of the padded real-time sampling signal as the left endpoint of the optimal cutoff frequency interval.

7. An online insulation detection system for power transmission and distribution, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the online insulation detection method for power transmission and distribution according to any one of claims 1 to 6 is implemented.

Images