An online monitoring system for insulation defect discharge of distribution network overhead lines

Abstract

The present invention relates to the technical field of measuring electrical variables, and more specifically to an online monitoring system for insulation defect discharges in distribution network overhead lines. The system comprises the following steps: obtaining pulse current signals of several candidate transformers in the current period, determining the degree of similarity in current changes, and screening target transformers based on the degree of similarity in current changes; determining a clustering distance based on the starting time and the degree of similarity in current changes of each two adjacent target transformers connected by a line, thereby obtaining several transformer clusters; sorting target transformers in the same transformer cluster according to the starting time, and obtaining the two optimal transformers in each transformer cluster based on the sorting results and the high-frequency energy proportion of each target transformer; and determining the locations of various insulation defects on the main trunk of the overhead line to be monitored by performing insulation defect monitoring between the two optimal transformers. The present invention effectively improves the accuracy of defect location and avoids the risk of missed detection.

Description

Technical Field

[0001] The present invention relates to the technical field of measuring electrical variables, and in particular to an online monitoring system for insulation defect discharge of a distribution network overhead line. Background Art

[0002] With the widespread use of overhead distribution lines, partial discharge (PD) caused by insulation defects has become a major hidden danger to the safe operation of power grids. The long-term accumulation of PD can lead to insulation degradation, which in turn can cause cable breakdown, tripping, and even line failures, seriously impacting the stability and reliability of the power system. Currently, traditional manual inspections and regular maintenance methods make it difficult to achieve 24 / 7 online monitoring of overhead lines and are unable to effectively detect hidden insulation defects. In response, a PD online monitoring technology based on high-frequency current transformers (HFCTs) has been proposed. Combining signal delay analysis with intelligent data processing, it enables real-time monitoring and precise location of insulation defects in overhead lines, providing a scientific basis for grid operation and maintenance and enhancing the intelligence level of distribution networks.

[0003] However, existing online partial discharge monitoring methods face several challenges in overhead line environments, including the following:

[0004] First, the discharge pulse signal is affected by the line topology, branch interference and attenuation characteristics during propagation, which makes it difficult for traditional single-end or double-end detection methods to accurately locate the defect point.

[0005] Secondly, the difference in discharge intensity means that low-intensity discharge signals may be submerged by noise or attenuated to the point of being undetectable during propagation, thus creating the risk of missed detection. Summary of the Invention

[0006] In order to solve the above technical problems, the purpose of the present invention is to provide an online monitoring system for insulation defect discharge of distribution network overhead lines. The technical solutions adopted are as follows:

[0007] An embodiment of the present invention provides an online monitoring system for insulation defect discharge of a distribution network overhead line, including a memory and a processor, wherein the processor is configured to process instructions stored in the memory to implement the following monitoring process:

[0008] Obtain pulse current signals of several candidate current transformers on the main trunk of the overhead line to be monitored in the current time period, wherein the candidate current transformers are high-frequency current transformers whose high-frequency energy proportion is greater than an initial energy proportion threshold;

[0009] Determining, based on the pulse current signal, a degree of similarity in current variation between each candidate current transformer and its adjacent and line-connected high-frequency current transformer;

[0010] According to the similarity of the current changes, a target current transformer is selected from all high-frequency current transformers on the main trunk of the overhead line to be monitored;

[0011] Determining a clustering distance based on the start time and the similarity of the current change of every two adjacent target transformers connected by lines, and clustering all target transformers according to the clustering distance to obtain a plurality of transformer clusters;

[0012] Sorting the target transformers in the same transformer cluster according to the starting time, and obtaining two optimal transformers in each transformer cluster based on the sorting result and the high-frequency energy proportion of each target transformer;

[0013] Insulation defect monitoring is performed based on the overhead line length, the time difference between the two optimal transformers in the same transformer cluster, and the lag time, to determine the locations of various insulation defects on the main trunk of the overhead line to be monitored.

[0014] Furthermore, obtaining the high-frequency energy ratio includes:

[0015] For any transformer, obtain the IMF components corresponding to the current signal of the transformer;

[0016] Determine the proportion of high-frequency energy in the transformer based on the IMF value of each IMF component at each moment and the frequency of each IMF component;

[0017] The IMF value and the frequency are both positively correlated with the high-frequency energy ratio.

[0018] Furthermore, the determining, based on the pulse current signal, the degree of similarity of current changes between each candidate current transformer and its adjacent and line-connected high-frequency current transformer includes:

[0019] For any candidate transformer, the transformer adjacent to the candidate transformer and connected to the line is taken as the transformer to be analyzed;

[0020] Determining a lag time between the candidate transformer and any transformer to be analyzed, wherein the lag time represents a difference in signal arrival time between the transformers;

[0021] The degree of similarity between the current changes of the candidate transformer and the transformer to be analyzed is determined based on the lag time and the pulse current signal of the candidate transformer.

[0022] Furthermore, determining the lag time between the candidate transformer and any transformer to be analyzed includes:

[0023] The candidate transformer and any transformer to be analyzed generate synchronous high-frequency current signals, so that the candidate transformer captures the high-frequency current signal generated by the transformer to be analyzed, and the transformer to be analyzed captures the high-frequency current signal generated by the candidate transformer, and records the time points when the high-frequency current signal starts to be transmitted and is finally received;

[0024] The difference between two time points of the same high-frequency current signal is calculated to obtain two time differences; and the average of the two time differences is used as the lag time between the candidate mutual inductor and any mutual inductor to be analyzed.

[0025] Furthermore, determining the similarity between the current changes of the candidate transformer and the transformer to be analyzed based on the hysteresis time and the pulse current signal of the candidate transformer includes:

[0026] Determine the respective pending starting times of the current signal collected by the transformer to be analyzed based on the lag time and the starting time corresponding to the pulse current signal of the candidate transformer;

[0027] Obtaining the current signal of the transformer to be analyzed at each pending starting moment through each pending starting moment; wherein the starting moment is the first moment, and the time interval of the current signal is equal to the time interval of the current period;

[0028] According to the current signal of the transformer to be analyzed and the pulse current signal of the candidate transformer at each undetermined starting time, the similarity degree of the current change between the candidate transformer and the transformer to be analyzed is determined.

[0029] Furthermore, the method of selecting a target current transformer from all high-frequency current transformers on the main trunk of the overhead line to be monitored according to the similarity of the current changes includes:

[0030] For any transformer to be analyzed, a normalized negative correlation process is performed on the similarity of the current changes to obtain a correction coefficient for the initial energy proportion threshold;

[0031] The initial energy ratio threshold is corrected using the correction coefficient to obtain the energy ratio threshold of the transformer to be analyzed;

[0032] The high-frequency energy proportion of the transformer to be analyzed is determined, and if the high-frequency energy proportion of the transformer to be analyzed is greater than the energy proportion threshold, the transformer to be analyzed is used as a target transformer.

[0033] Furthermore, the method further includes: taking a candidate mutual inductor whose high-frequency energy proportion is greater than an initial energy proportion threshold as a target mutual inductor.

[0034] Furthermore, the determining of the cluster distance according to the start time of each two adjacent and line-connected target transformers and the similarity of the current change includes:

[0035] For any two adjacent target transformers connected by a line, according to the similarity of current changes between the two adjacent target transformers connected by a line, a negative correlation value of the similarity of the current changes is used as a first distance factor;

[0036] Determining a cluster distance between two adjacent target mutual inductors that are connected by a line according to a comparison result of a time difference between the start times and a lag time between the two adjacent target mutual inductors that are connected by a line, in combination with the first distance factor;

[0037] The lag time represents the signal arrival time difference between the mutual inductors.

[0038] Furthermore, the method of obtaining two optimal transformers in each transformer cluster based on the sorting result and the high-frequency energy proportion of each target transformer includes:

[0039] For each transformer cluster, determining a cumulative value of the ratio of the high-frequency energy proportion of any target transformer in the transformer cluster to each comparison high-frequency energy proportion; wherein the comparison high-frequency energy proportion is the high-frequency energy proportion of the remaining target transformers in the transformer cluster except the any target transformer;

[0040] Performing negative correlation processing on the sorting sequence number of any target mutual inductor to obtain a negative correlation value, and combining the accumulated value of the ratio to obtain a discharge detection reliability index of any target mutual inductor;

[0041] The discharge detection reliability index of each target transformer in the same transformer cluster is obtained, and the two target transformers with the largest discharge detection reliability index are taken as the two optimal transformers in the corresponding transformer cluster.

[0042] Furthermore, the insulation defect monitoring is performed based on the overhead line length, the time difference between the start time and the hysteresis time between the two optimal transformers in the same transformer cluster, and the location of each insulation defect on the main trunk of the overhead line to be monitored is determined, including:

[0043] For the starting times of the two optimal transformers in the same transformer cluster, the location of the optimal transformer with the smaller starting time is taken as the near end of the overhead line, and the difference between the larger starting time and the smaller starting time is taken as the time difference of the starting time;

[0044] determining a ratio of the time difference to the lag time, and performing negative correlation processing on the ratio to obtain a negative correlation value;

[0045] Using the negative correlation value to correct half of the overhead line length, to obtain the distance between the insulation defect and the near end of the overhead line;

[0046] The position of the insulation defect is obtained according to the position of the proximal end of the overhead line and the distance between the insulation defect and the proximal end of the overhead line.

[0047] The present invention has the following beneficial effects:

[0048] The present invention provides an online monitoring system for insulation defects and discharges in distribution network overhead lines. By collaboratively analyzing the pulse current signals of several high-frequency current transformers on the main trunk of the overhead line to be monitored, it achieves real-time monitoring and precise positioning of insulation defects in the overhead line, overcoming the limitation of traditional inspection methods that are difficult to monitor around the clock.

[0049] First, the target transformers with a high-frequency energy ratio greater than the energy ratio threshold are screened out through the similarity of current changes, which can optimize the detection and processing of pulse current signals, improve the identification accuracy of insulation defects, reduce the false alarm and missed detection rates, and enhance the reliability of the system. In addition, adaptive energy threshold adjustment technology is also used when screening target transformers to improve the sensitivity to low-intensity discharge signals and accurately analyze different types of insulation defects. Secondly, all target transformers are clustered through adaptively determined clustering distances, which can effectively distinguish between real insulation defect signals and environmental noise. At the same time, based on the clustering results, line section anomalies can be directly reflected, narrowing the detection range. Then, based on the sorting results and the high-frequency energy ratio of each target transformer, the two optimal transformers with the highest detection reliability are selected for each transformer cluster. Subsequently, only the optimal transformer data needs to be analyzed, which can reduce the amount of calculation to a certain extent and improve the accuracy of insulation defect discharge monitoring.

[0050] Overall, the present invention improves the intelligence level and power supply reliability of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0052] Figure 1 An execution flow chart of an online monitoring system for insulation defect discharge of a distribution network overhead line provided by one embodiment of the present invention;

[0053] Figure 2 Flowchart for implementing step S2 in an embodiment of the present invention;

[0054] Figure 3 Schematic diagram of analysis of hysteresis time in an embodiment of the present invention;

[0055] Figure 4 Schematic diagram of the insulation defect location process in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] To further illustrate the technical means and effects employed by the present invention to achieve its intended objectives, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementations, structures, features, and effects of the technical solutions proposed by the present invention. In the following description, references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0057] Unless defined otherwise, 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 invention belongs.

[0058] The application scenarios targeted by the present invention may be:

[0059] Each high-frequency current transformer on the monitored overhead line trunk line collects current signals in real time and calculates the high-frequency energy ratio for each collected current signal. This means that the high-frequency energy ratio of each high-frequency current transformer is monitored in real time. If the high-frequency energy ratio of at least one high-frequency current transformer exceeds the initial energy ratio threshold, the likelihood of insulation defect discharge on the monitored overhead line trunk line is high, and the insulation defect discharge location needs to be determined.

[0060] However, existing online partial discharge monitoring systems, operating in overhead line environments, are prone to poor insulation defect location accuracy and the risk of missed detections. To accurately determine the location of insulation defect discharges, this paper proposes an online insulation defect discharge monitoring system for distribution network overhead lines. This system comprehensively analyzes multiple high-frequency current transformers used to detect discharge signals and extracts the transformers corresponding to the primary discharge areas, improving the accuracy of insulation defect location determination and preventing missed detections.

[0061] Specifically, an embodiment of the present invention provides an online monitoring system for insulation defect discharge of a distribution network overhead line, including a memory and a processor, wherein the processor is configured to process instructions stored in the memory to implement the following monitoring process:

[0062] Obtain pulse current signals of several candidate current transformers on the main trunk of the overhead line to be monitored in the current period, where the candidate current transformers are high-frequency current transformers whose high-frequency energy ratio is greater than an initial energy ratio threshold;

[0063] Determining the similarity of current changes between each candidate current transformer and its adjacent and line-connected high-frequency current transformer based on the pulse current signal;

[0064] According to the similarity of current changes, the target current transformer is selected from all high-frequency current transformers on the main trunk of the overhead line to be monitored;

[0065] The clustering distance is determined based on the similarity between the starting time and current change of every two adjacent target transformers connected by the line. All target transformers are clustered according to the clustering distance to obtain several transformer clusters.

[0066] Sort the target transformers in the same transformer cluster according to the starting time, and obtain the two optimal transformers in each transformer cluster based on the sorting results and the high-frequency energy proportion of each target transformer;

[0067] Insulation defect monitoring is performed by measuring the overhead line length, starting time difference and lag time between the two optimal transformers in the same transformer cluster, and the locations of various insulation defects on the main trunk of the overhead line to be monitored are determined.

[0068] The following is a detailed explanation of each of the above steps:

[0069] refer to Figure 1 , shows an execution flow chart of an insulation defect discharge online monitoring system for distribution network overhead lines of the present invention.

[0070] S1, obtaining pulse current signals of several candidate transformers on the main trunk of the overhead line to be monitored in the current period.

[0071] Here, the duration of the current period can be set to The candidate transformer is a high-frequency current transformer whose high-frequency energy ratio is greater than the initial energy ratio threshold. The horizontal axis of the pulse current signal is the acquisition time, and the vertical axis is the pulse current value.

[0072] In this embodiment, candidate current transformers deployed on overhead lines are used to collect current signals for the current period. The current signals of these candidate transformers are affected by insulation defect discharges, and the current signals fluctuate during this period. Therefore, the current signals collected by these candidate transformers can be used as pulse current signals. These pulse current signals can be used to analyze the primary discharge area. Each candidate transformer has a corresponding pulse current signal. This pulse current signal can be used to determine discharge-related information, such as the start time of the transformer's current signal affected by insulation defects, the magnitude of the current change, and so on.

[0073] Since the current signal generated by partial discharge contains high-frequency energy, the current signal generated by internal insulation defects will show a transient high-frequency energy accumulation area. In addition, the current pulse rise time generated by partial discharge is extremely short, and its Fourier transform contains rich high-frequency components. Therefore, the possibility of local insulation defect discharge can be analyzed by analyzing the proportion of high-frequency energy.

[0074] For any high-frequency current transformer, the sub-steps of obtaining the high-frequency energy ratio include:

[0075] The first sub-step is to obtain the corresponding time of the high-frequency current transformer for any high-frequency current transformer. are the IMF components with a starting time and a time length of T.

[0076] In this embodiment, the high frequency current transformer corresponding to the time is obtained. The current signal with the starting time and time length T is decomposed by EMD (Empirical Mode Decomposition) to obtain the IMF components corresponding to the current signal.

[0077] In the process of implementing EMD decomposition, the higher the IMF component is ranked, the more likely it is to be a high-frequency component, and the IMF value at each moment corresponding to the high-frequency component is more likely to be high-frequency energy; is the first moment of the current signal, i.e. the starting moment, and the time length T is the preset period. The time length can also be called the time interval, which can be taken as an empirical value. .

[0078] The second sub-step is to determine the high-frequency current transformer at time according to the IMF value of each IMF component at each moment and the frequency of each IMF component. The proportion of high-frequency energy under .

[0079] When calculating the proportion of high-frequency energy, the calculation can be performed in real time or after a period of time. If the proportion of high-frequency energy at a certain moment is larger, it indicates that the possibility of abnormal fluctuation of the pulse current signal collected at that moment is greater, and further indicates that the possibility of insulation defects in the overhead line at that moment is greater.

[0080] As an example, calculate the rth high-frequency current transformer at time The calculation formula for the high-frequency energy ratio under can be:

[0081] Where, Indicates that the rth high-frequency current transformer is at time The high-frequency energy ratio under the condition of θ is θ, I represents the number of high-frequency components, which can be taken as 3, i.e. the first three IMF components, i represents the sequence number of the IMF component, and norm represents the linear normalization function. represents the frequency of the i-th IMF component corresponding to the r-th high-frequency current transformer, represents the energy of the i-th IMF component corresponding to the r-th high-frequency current transformer, and N represents the total number of IMF components.

[0082] in, , Represents the energy of the i-th IMF component, T represents the time length, T can be adjusted with the sampling frequency of the mutual inductor and is not specifically limited, t represents each moment in the IMF component, represents the IMF value at time t in the i-th IMF component.

[0083] In the calculation formula of high-frequency energy ratio, Represents the high-frequency energy corresponding to the r-th high-frequency current transformer, Represents the total energy corresponding to the rth high-frequency current transformer. The greater the proportion of high-frequency energy in the energy synthesis, the greater the rth high-frequency current transformer at time The higher the ratio of high-frequency energy, the greater the

[0084] With reference to the above-mentioned process of obtaining the high-frequency energy ratio, the high-frequency energy ratio of each high-frequency current transformer can be obtained.

[0085] On the main trunk of the overhead line to be monitored, a high-frequency current transformer (HFCT) can be deployed every 2 km to ensure continuous signal acquisition and coverage, enabling the acquisition and location of discharge signals on the main trunk of the overhead line to be monitored. The number of HFCTs is determined by the spacing between adjacent HFCTs and the length of the main trunk of the overhead line to be monitored, and is not specifically limited here. The sampling frequency of the HFCT can be 500MHz, and the storage depth can be 500k. The sampling frequency and storage depth can be set by the implementer based on specific circumstances.

[0086] It should be noted that all high-frequency current transformers located on the same overhead line support high-precision clock synchronization, such as PTPv2. This helps solve the time consistency problem in distributed monitoring systems, avoids data distortion caused by clock drift or transmission delays, and ensures that the pulse current signals collected by different transformers can be compared and analyzed over time.

[0087] So far, this embodiment has obtained the pulse current signal of each high-frequency current transformer.

[0088] S2, determining, based on the pulse current signal, a degree of similarity between current changes of each candidate current transformer and its adjacent high-frequency current transformer connected to the line.

[0089] Here, the degree of similarity of current variation indicates the similarity between the current signals detected by the two transformers, which can be used to adjust the initial energy ratio threshold of the corresponding high-frequency current transformer.

[0090] Common insulation defects in overhead lines include insulators, tree barriers, transformers, and circuit breakers. Discharge signals generated by different types of insulation defects have distinct characteristics. For example, internal transformer discharges correspond to higher pulse amplitudes, while contaminated or damaged insulators correspond to relatively smaller pulse amplitudes, which are significantly affected by humidity. Therefore, traditional current change detection methods based on energy thresholds can miss detection of some low-amplitude pulse signals or result in a high incidence of false detections.

[0091] To avoid missed detections or false detections, this embodiment considers combining the connection relationships between the lines of multiple different transformers. First, high-frequency current transformers whose high-frequency energy ratio is greater than the initial energy ratio threshold are selected. Second, transformers adjacent to the selected high-frequency current transformers and connected to the lines are selected as transformers to be analyzed. Subsequently, by analyzing the similarity of current changes between the selected high-frequency current transformers and the transformers to be analyzed, the energy ratio threshold corresponding to each transformer to be analyzed can be adaptively determined. All transformers to be analyzed are then screened again to select all transformers that may be affected by partial discharge defects. First, it is necessary to determine one of the key factors for transformer screening, namely the degree of similarity in current changes.

[0092] As an exemplary embodiment, the above step S2 can be performed by Figure 2 The steps shown are as follows:

[0093] S21: For any candidate transformer, a transformer adjacent to the candidate transformer and connected to the transformer by a line is used as a transformer to be analyzed.

[0094] In this embodiment, the transformer to be analyzed is a high-frequency current transformer whose provisional current signal does not contain partial discharge defect characteristics.

[0095] It should be noted that among all the transformers to be analyzed, there may be high-frequency current transformers whose high-frequency energy ratio is greater than the initial energy ratio threshold. They can continue to be used as transformers to be analyzed. However, in order to reduce the amount of calculation, only the transformers to be analyzed whose high-frequency energy ratio is less than or equal to the initial energy ratio threshold can be used as the final transformers to be analyzed for the subsequent steps of current change similarity analysis. No specific limitation is made here.

[0096] For the convenience of analysis and description, this embodiment takes a candidate mutual inductor and any mutual inductor to be analyzed as an example to perform subsequent steps S22 to S23.

[0097] S22, determining the lag time between the candidate transformer and any transformer to be analyzed.

[0098] Here, the lag time refers to the signal arrival time difference between transformers.

[0099] Regarding hysteresis time, if two transformers detect the same partial discharge signal, the time lag will lead to signal matching errors, affecting correlation analysis. Furthermore, by analyzing the time difference between the signals arriving at the transformers, the fault location can be calculated. Therefore, for transformers under analysis where it is currently uncertain whether they are affected by partial insulation defect discharge, hysteresis time is required to determine the time range within which the transformer under analysis receives the partial discharge current signal.

[0100] In this embodiment, the time difference between the start times of the current signal arriving at any two transformers may not be the same due to differences in impedance matching and attenuation of the corresponding lines. The candidate transformer and any transformer to be analyzed are two adjacent transformers connected by a line.

[0101] As an exemplary embodiment, the above step S22 can be implemented through the following process:

[0102] The candidate transformer and any transformer to be analyzed generate synchronous high-frequency current signals, so that the candidate transformer captures the high-frequency current signal generated by the transformer to be analyzed, and the transformer to be analyzed captures the high-frequency current signal generated by the candidate transformer, and records the time points when the high-frequency current signal starts to be transmitted and is finally received;

[0103] The difference between two time points of the same high-frequency current signal is calculated to obtain two time differences; the average of the two time differences is used as the lag time between the candidate transformer and any transformer to be analyzed.

[0104] In this embodiment, the corresponding positions of the candidate transformer and the transformer to be analyzed are end a and end b respectively. Synchronous pulses are generated at the modules at end a and end b respectively. A high-frequency pulse signal generator is used to artificially apply a known high-frequency pulse signal on the overhead line so that the transformer captures the high-frequency pulse signal. The time when the pulse current reaches end b and end a is recorded. The corresponding time differences are recorded as and , and then get the lag time.

[0105] The analysis diagram of the lag time is as follows: Figure 3 As shown, in Figure 3 middle, Indicates the time point when end a transmits the signal. Indicates the time point when end B receives the signal. , Indicates the time point when the B end transmits the signal. Indicates the time point when end a receives the signal, , each V represents the current signal at a different position.

[0106] As an example, the lag time between the rth candidate transformer and the lth transformer to be analyzed is The calculation formula can be: .

[0107] With reference to the above calculation process of the lag time between the rth candidate transformer and the lth transformer to be analyzed, the lag time between each candidate transformer and its corresponding transformer to be analyzed can be obtained.

[0108] S23 , determining the similarity between the current changes of the candidate transformer and the transformer to be analyzed based on the lag time and the pulse current signal of the candidate transformer.

[0109] As an exemplary embodiment, the above step S23 can be implemented through the following sub-steps:

[0110] In the first sub-step, various pending starting times for collecting current signals of the transformers to be analyzed are determined according to the lag time and the starting times corresponding to the pulse current signals of the candidate transformers.

[0111] In this embodiment, the lag time between the rth candidate transformer and the lth transformer to be analyzed is , if the starting time of the rth high-frequency current transformer is , you can set the start time range In the analysis The current change detected by the first transformer to be analyzed is The similarity of the current changes detected by the candidate transformers can be used to define the starting time range. Each moment in is taken as the pending starting time.

[0112] In the second sub-step, the current signal of the transformer to be analyzed at each pending starting moment is obtained through each pending starting moment.

[0113] Here, the time interval of the current signal is equal to the time interval of the current period.

[0114] In this embodiment, the pending starting moment is the first moment of obtaining the current signal. The current data at the pending starting moment is used as the starting point to obtain the current signal with the same length as the current period, thereby obtaining the current signal of the transformer to be analyzed at each pending starting moment.

[0115] The third sub-step is to determine the similarity between the current changes of the candidate transformer and the transformer to be analyzed based on the current signal of the transformer to be analyzed and the pulse current signal of the candidate transformer at each undetermined starting time.

[0116] For current or pulse current signals, EMD decomposition is used to decompose the signal and extract different frequency components to facilitate subsequent analysis of the instantaneous energy changes of the intrinsic mode functions (IMFs). The higher the ranked IMF components after decomposition, the more likely they are high-frequency components. The implementation process of EMD decomposition is prior art and is beyond the scope of this invention, so it will not be elaborated here.

[0117] In this embodiment, since the discharge pulse is mainly in the high-frequency part, the similarity of the current change is quantified by the first three IMF components of the current change at the candidate transformer and the transformer to be analyzed.

[0118] As an example, the calculation formula for the similarity between the current changes of the rth candidate transformer and the lth transformer to be analyzed can be:

[0119] ;

[0120] Where, Indicates the starting time is time The similarity between the current changes of the rth candidate transformer and the lth transformer to be analyzed, max represents the maximum value function, represents the lag time between the rth candidate transformer and the lth transformer to be analyzed, Represents the time interval variable, I represents the number of high-frequency components, the empirical value of which is 3, i represents the IMF component number, and the IMF component with the higher component number is the higher the high-frequency IMF component. represents the function for finding the correlation coefficient, Indicates the starting time is time The i-th IMF component corresponding to the r-th candidate mutual inductor with a time interval of T is, Indicates the starting time is time And the time interval is T, corresponding to the i-th IMF component of the l-th mutual inductor to be analyzed.

[0121] In the calculation formula of the similarity of current change, Indicates that the starting time corresponding to the rth candidate transformer is time The IMF value sequence corresponding to the i-th IMF component under the time interval T, and is the IMF value sequence with insulation defect discharge characteristics, is the IMF value sequence corresponding to the i-th IMF component with the starting time corresponding to the l-th transformer to be analyzed constantly changing and the time interval being T.

[0122] make With different starting times Perform correlation analysis, such as calculating the Pearson correlation coefficient, and average the Pearson correlation coefficients corresponding to each IMF component to obtain the mean of each Pearson correlation coefficient, and use the maximum Pearson correlation coefficient mean as the degree of similarity of current change; the larger the mean of the Pearson correlation coefficient, the stronger the similarity between the high-frequency data of the rth candidate transformer and the lth transformer to be analyzed, the more likely the lth transformer to be analyzed is a missed transformer affected by partial discharge defects, and the greater the possibility that the lth transformer to be analyzed belongs to the target transformer.

[0123] In addition, get the maximum Pearson correlation coefficient corresponding to the mean , at this time Recorded as , as the starting time corresponding to the lth mutual inductor to be analyzed.

[0124] Referring to the above-mentioned process of determining the similarity of the current changes between the rth candidate transformer and the lth transformer to be analyzed, the similarity of the current changes between the rth candidate transformer and each transformer to be analyzed can be obtained, and then the similarity of the current changes between each high-frequency current transformer whose high-frequency energy proportion is greater than the initial energy proportion threshold and its corresponding transformers to be analyzed can be obtained.

[0125] It is worth noting that the degree of similarity of current changes can be calculated and analyzed multiple times for the transformer to be analyzed, that is, there are multiple high-frequency current transformers whose high-frequency energy proportion is greater than the initial energy proportion threshold corresponding to the same transformer to be analyzed, that is, the transformer to be analyzed is adjacent to the multiple high-frequency current transformers whose high-frequency energy proportion is greater than the initial energy proportion threshold and the lines are connected.

[0126] Thus, this embodiment obtains the degree of similarity of current changes between each candidate transformer having a high-frequency energy proportion greater than the initial energy proportion threshold and its corresponding transformers to be analyzed.

[0127] S3, based on the similarity of current changes, the target current transformer is selected from all high-frequency current transformers on the main trunk of the overhead line to be monitored.

[0128] When the similarity between the high-frequency component of the current of the candidate transformer and its corresponding transformer to be analyzed is higher, it means that the transformer to be analyzed is more likely to be affected by partial discharge defects. At this time, the initial energy proportion threshold should be lowered and adjusted to obtain the corrected energy proportion threshold of the transformer to be analyzed. This helps to detect PDs with relatively small discharge pulse amplitudes, that is, it can ensure that the pulse current signal of the transformer to be analyzed exhibits insulation defect characteristics. The insulation defect characteristic refers to the high-frequency energy proportion determined by the current signal being greater than the energy proportion threshold.

[0129] Based on the similarity between the current changes of the rth candidate transformer and its corresponding lth transformer to be analyzed, and the initial energy ratio threshold, the improved initial energy ratio threshold corresponding to the lth transformer to be analyzed is determined. The improved initial energy ratio threshold is the energy ratio threshold recorded in step S3 above. Specifically,

[0130] For any transformer to be analyzed, a normalized negative correlation process is performed on the similarity of current changes to obtain a correction coefficient for the initial energy ratio threshold.

[0131] The initial energy ratio threshold is corrected using the correction coefficient to obtain the energy ratio threshold of the transformer to be analyzed, that is, the corrected energy ratio threshold;

[0132] The high-frequency energy proportion of the transformer to be analyzed is determined. If the high-frequency energy proportion of the transformer to be analyzed is greater than an energy proportion threshold, the transformer to be analyzed is used as a target transformer.

[0133] As an example, the calculation formula for the energy proportion threshold corresponding to the lth transformer to be analyzed can be:

[0134] Where, Indicates time is the energy proportion threshold corresponding to the lth transformer to be analyzed at the starting time, norm represents the linear normalization function, Indicates the starting time is time The similarity of the current changes corresponding to the rth high-frequency current transformer and the lth transformer to be analyzed, represents the correction factor, Indicates the initial energy percentage threshold.

[0135] In the calculation formula of the energy proportion threshold, The larger the time, the The more similar the current change characteristics of the lth transformer to be analyzed at the starting time are to the rth candidate transformer whose high-frequency energy ratio is greater than the initial energy ratio threshold, the greater the possibility that the lth transformer to be analyzed is affected by the partial discharge defect. At this time, the energy ratio threshold corresponding to the lth transformer to be analyzed is Should be smaller.

[0136] Compare the high-frequency energy proportion of the lth transformer to be analyzed with its corresponding energy proportion threshold, and take the transformer to be analyzed whose high-frequency energy proportion is greater than the energy proportion threshold as the target transformer.

[0137] Specifically, the high-frequency energy ratio of the lth transformer to be analyzed is If the energy ratio is greater than the energy ratio threshold corresponding to the lth transformer to be analyzed, it means that the lth transformer to be analyzed can also detect PD and can be used as the target transformer. Otherwise, it means that the discharge node is too far away from the lth high-frequency current transformer, and the pulse current of the lth high-frequency current transformer is not affected by the abnormality of the discharge node. Local insulation defect discharge detection and location can be carried out through the remaining transformers to be analyzed that meet the high-frequency energy ratio conditions.

[0138] By referring to the calculation process of the energy proportion threshold corresponding to the first transformer to be analyzed, the energy proportion threshold corresponding to each transformer to be analyzed can be obtained, and then the target transformer that meets the high-frequency energy proportion condition can be screened out from all the transformers to be analyzed.

[0139] At the same time, for the convenience of unified description, this embodiment also takes the candidate transformers whose high-frequency energy proportion is greater than the initial energy proportion threshold as target transformers.

[0140] So far, this embodiment has obtained all target transformers representing approximate discharge locations on the overhead line.

[0141] S4, determining a clustering distance according to the similarity between the starting time and current change of every two adjacent target transformers connected by the line, clustering all target transformers according to the clustering distance to obtain a number of transformer clusters.

[0142] To analyze situations where multiple target transformers are affected by the same discharge event—that is, a single partial discharge can be detected by multiple target transformers—we cluster all target transformers based on the current change data at the moment of discharge. This clustering effectively distinguishes true insulation defect signals from environmental noise and roughly groups target transformers affected by the same discharge node into a single cluster. Furthermore, the clustering results directly reflect line section anomalies, narrowing the detection range.

[0143] As an exemplary embodiment, the above step S4 can be implemented by the following steps:

[0144] In the first step, for any two adjacent target transformers that are connected in line, based on the similarity of current changes between the two adjacent target transformers that are connected in line, a negative correlation value of the similarity of current changes is used as a first distance factor.

[0145] In this embodiment, the starting time is determined to be the time The mth target transformer and the starting time is time The similarity of the current change corresponding to the nth target transformer is recorded as , the inverse of the similarity of current changes can be used as the first distance factor.

[0146] It should be noted that the high-frequency energy proportion of each target transformer is greater than the energy proportion threshold or initial energy proportion threshold of the corresponding target transformer, which means that each target transformer can collect the pulse current signal of the partial discharge defect characteristic, and then the starting time of collecting the pulse current signal can be determined.

[0147] In the second step, based on the comparison result of the time difference between the starting moments and the hysteresis time between two adjacent target transformers connected by a line and combined with the first distance factor, the cluster distance between the two adjacent target transformers connected by a line is determined.

[0148] As an example, the starting time is time The mth target transformer and the starting time are time The clustering distance between the nth target mutual inductors The calculation formula can be:

[0149] ;

[0150] Where, Indicates the starting time is time The mth target transformer and the starting time is time The clustering distance between the nth target mutual inductors, represents the first distance factor, Represents the absolute value function, when hour, ,on the contrary, , To avoid the hyperparameter from having a denominator of zero, an empirical value of 0.01 is used.

[0151] ;

[0152] In the formula, the starting time is the time The mth target transformer and the starting time is time The similarity of the current change corresponding to the nth target transformer, and The corresponding time length is T, Indicates the high frequency distribution of the mth target transformer, Indicates the high frequency distribution of the nth target transformer; and All of them are high-frequency distributions with discharge defect characteristics. and The larger the correlation coefficient, the higher the similarity of the current changes of two adjacent target transformers connected by lines, and the more likely they are to be classified into the same cluster. In other words, the more likely two adjacent target transformers connected by lines are to receive the pulse current signal generated by the same discharge defect.

[0153] It represents the difference between the lag time and the discharge time (time difference). When D is equal to 0, the denominator is the smallest and the clustering distance can reach the maximum. The reason is that when When , it means that the time difference is greater than the lag time. The power data segments collected by the two target transformers cannot come from the same partial discharge. Therefore, the maximum difference is taken here to make the cluster distance The larger the value is, the greater the risk is to avoid classifying the m-th target transformer and the n-th target transformer into the same cluster.

[0154] Refer to the start time as the time The mth target transformer and the starting time is time The clustering distance between each two adjacent target transformers that are connected by lines can be obtained by calculating the clustering distance between the nth target transformers.

[0155] It is worth noting that for two high-frequency current changes with large time differences, the local discharge signals corresponding to the two target transformers may be caused by different insulation defects, so the difference between the two current changes will be set to the maximum; considering that the current change data segments corresponding to multiple local discharges detected by different target transformers may be caused by different discharge nodes, when the starting times at two different positions are close, the time difference between the local discharge signals received by the two target transformers is small, which will affect the subsequent determination of insulation defect positioning.

[0156] After obtaining the clustering distance, all target transformers are clustered based on the clustering distance between every two adjacent target transformers that are connected by lines, to obtain transformer clusters.

[0157] Specifically, DBSCAN cluster analysis is performed based on the cluster distance between any two adjacent target transformers connected by lines to obtain transformer clusters. Here, transformers within the same transformer cluster can collect discharge signals of the same insulation defect, that is, the current changes caused by the PD signal reaching each transformer within the same transformer cluster.

[0158] Among them, the minimum number of samples in the cluster can be set to 5, and the cluster radius can be set to 1.3 kilometers; the implementation process of DBSCAN clustering is a prior art and is not within the scope of protection of the present invention, and will not be elaborated here.

[0159] Thus, this embodiment obtains transformer clusters that can characterize insulation defects at different locations.

[0160] S5, sorting the target transformers in the same transformer cluster according to the starting time, and obtaining two optimal transformers in each transformer cluster based on the sorting result and the high-frequency energy proportion of each target transformer.

[0161] Insulation defects are typically located on a line connected by two transformers, and each transformer cluster corresponds to one insulation defect. Taking any transformer cluster as an example, the detection reliability of each target transformer is calculated based on the discharge intensity and discharge time, thereby determining the discharge location of the local insulation defect.

[0162] As an exemplary embodiment, the above step S5 can be implemented by the following steps:

[0163] In the first step, the target transformers in the same transformer cluster are sorted according to the time when the target transformer collects the discharge signal, and the sorting sequence number of each target transformer in the same transformer cluster is obtained.

[0164] In this embodiment, the closer the starting moment of the pulse current signal is to the current moment, the more likely the target transformer is to be at the end of the sequence, that is, the target transformer with a larger sorting number collects the discharge signal closer to the current moment, which means that the target transformer takes a longer time to collect the discharge signal and the reliability of its discharge detection is worse.

[0165] In the second step, for each transformer cluster, the cumulative value of the ratio of the high-frequency energy proportion of any target transformer in the transformer cluster to the high-frequency energy proportions of each comparison is determined.

[0166] In this embodiment, the comparison high-frequency energy proportion is the high-frequency energy proportion of the remaining target transformers in the transformer cluster except any target transformer.

[0167] In the third step, negative correlation processing is performed on the sorting sequence number of any target transformer to obtain a negative correlation value, and combined with the accumulated value of the ratio, the discharge detection reliability index of any target transformer is obtained.

[0168] In this embodiment, the current change data detected by the pth target transformer is recorded as , and its corresponding starting time is When the insulation defect begins to discharge, the target transformer closer to the insulation defect location receives a shorter current change time and has a higher PD signal strength, that is, the high-frequency energy accounts for a larger proportion. The discharge detection reliability index of the p-th target transformer is determined based on this.

[0169] As an example, the calculation formula for the discharge detection reliability index of the p-th target transformer can be:

[0170] Where, represents the discharge detection reliability index of the pth target transformer, represents the total number of target transformers in the transformer cluster to which the p-th target transformer belongs, Indicates the starting time is time The high-frequency energy proportion of the p-th target transformer, Indicates the starting time is time The high-frequency energy proportion of the qth target transformer, Indicates the time sequence number of the p-th target transformer in the transformer cluster to which it belongs.

[0171] With reference to the calculation process of the discharge detection reliability index of the p-th target transformer, the discharge detection reliability index of each target transformer can be obtained.

[0172] The fourth step is to obtain the discharge detection reliability index of each target transformer in the same transformer cluster, and select the two target transformers with the largest discharge detection reliability index as the two optimal transformers in the corresponding transformer cluster.

[0173] Here, the two optimal transformers refer to the two target transformers in the transformer cluster that are most affected by insulation defect discharge.

[0174] In this embodiment, the larger the discharge detection reliability index of a target transformer, the greater the possibility that the partial discharge location is near the location of the target transformer. Generally, the insulation defect is located on the connection line between two transformers, so the two largest target transformers are used as the two optimal transformers in the corresponding transformer cluster.

[0175] Thus, this embodiment obtains two optimal mutual inductors for determining the locations of insulation defects.

[0176] S6, monitoring insulation defects is performed based on the overhead line length, the time difference between the start time and the lag time between the two optimal transformers in the same transformer cluster, and the locations of each insulation defect on the main trunk of the overhead line to be monitored are determined.

[0177] Here, insulation defects are generally located on the overhead line between the transformers, so it is necessary to obtain the length of the overhead line to preliminarily determine the distance between the insulation defect and the near end of the overhead line. The time difference and lag time can be used to correct the preliminarily determined distance between the insulation defect and the near end of the overhead line. The larger the time difference, the earlier the pulse current signal with insulation defect characteristics is collected at the near end of the overhead line, and the closer the insulation defect is to the near end of the overhead line.

[0178] As an exemplary embodiment, taking the two best transformers in any transformer cluster as an example to determine the insulation defect location, the above step S6 can be implemented by the following steps:

[0179] In the first step, for the starting times of the two optimal transformers in the same transformer cluster, the location of the optimal transformer with the smaller starting time is taken as the near end of the overhead line, and the difference between the larger starting time and the smaller starting time is taken as the time difference of the starting time.

[0180] In this embodiment, when When , the location of the pth target transformer is recorded as the near end of the overhead line, otherwise the location of the qth transformer is recorded as the near end of the overhead line. When , the time difference is calculated as .in, Indicates the starting moment when the qth target transformer collects the pulse current signal, Indicates the starting time when the pth target transformer collects the pulse current signal.

[0181] The second step is to determine the ratio of the time difference to the lag time, and perform negative correlation processing on the comparison value to obtain a negative correlation value.

[0182] In the third step, half of the overhead line length is corrected using the negative correlation value to obtain the distance between the insulation defect and the near end of the overhead line.

[0183] As an example, when the pth target transformer is located near the overhead line, the distance between the insulation defect and the near end of the overhead line can be calculated as follows:

[0184] Where, Indicates the distance between the insulation defect and the near end of the overhead line, represents the length of the overhead line between the pth target transformer and the qth target transformer, Indicates the starting moment when the qth target transformer collects the pulse current signal, Indicates the starting moment when the pth target transformer collects the pulse current signal, It represents the time difference between the starting time of the qth target transformer and the pth target transformer. It represents the lag time between the p-th target transformer and the q-th target transformer. There is no possibility that the lag time is zero.

[0185] In the calculation formula of the distance between the insulation defect and the near end of the overhead line, It is equivalent to the total time duration of the current passing through the pth and qth target transformers, which can also be used to normalize the time difference. represents the difference between the starting time when the qth target transformer collects the partial discharge signal and the starting time when the pth target transformer collects the partial discharge signal; The bigger, The smaller it is, the shorter the time it takes for the pth target transformer to collect the pulse current signal with insulation defect characteristics, and the smaller the distance between the pth target transformer and the insulation defect, that is, the closer the insulation defect is to the near end of the overhead line.

[0186] The fourth step is to obtain the insulation defect position through the position of the near end of the overhead line and the distance between the insulation defect and the near end of the overhead line.

[0187] In this embodiment, after obtaining the position of the near end of the overhead line and the distance between the insulation defect and the near end of the overhead line, because the near end of the overhead line is the location of the pth target transformer, and the pth target transformer is located to the left of the qth target transformer, the position of the near end of the overhead line and the distance are added to obtain a new position, and the new position is used as the insulation defect position. The insulation defect positioning process is shown in the figure below. Figure 4 As shown, Figure 4 Each V in the numeral represents a current signal at a different position.

[0188] So far, this embodiment has obtained the locations of various insulation defects on the overhead lines of the distribution network to be monitored.

[0189] The present invention provides an online monitoring system for insulation defects and discharges in overhead power lines of a distribution network. Through collaborative analysis using multiple transformers, it enables real-time monitoring and precise location of insulation defects in overhead power lines, overcoming the limitations of traditional inspection methods that hinder all-weather monitoring. By optimizing the detection and processing of discharge signals, the accuracy of identifying discharge events is improved, the false alarm and missed detection rates are reduced, and the reliability of the system is enhanced. Furthermore, the present invention utilizes adaptive energy threshold adjustment technology to increase sensitivity to low-intensity discharge signals and accurately analyze different types of insulation defects. Overall, the present invention enhances the intelligence level and power supply reliability of the distribution network.

[0190] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. An online monitoring system for insulation defect discharge of distribution network overhead lines, characterized in that: The system comprises a memory and a processor, wherein the processor is configured to process instructions stored in the memory to implement the following monitoring process: Obtain pulse current signals of several candidate current transformers on the main trunk of the overhead line to be monitored in the current time period, wherein the candidate current transformers are high-frequency current transformers whose high-frequency energy proportion is greater than an initial energy proportion threshold; Determining, based on the pulse current signal, a degree of similarity in current variation between each candidate current transformer and its adjacent and line-connected high-frequency current transformer; According to the similarity of the current changes, a target current transformer is selected from all high-frequency current transformers on the main trunk of the overhead line to be monitored; Determining a clustering distance based on the start time and the similarity of the current change of every two adjacent target transformers connected by lines, and clustering all target transformers according to the clustering distance to obtain a plurality of transformer clusters; Sorting the target transformers in the same transformer cluster according to the starting time, and obtaining two optimal transformers in each transformer cluster based on the sorting result and the high-frequency energy proportion of each target transformer; Insulation defect monitoring is performed by using the overhead line length, the time difference between the start time and the lag time between the two optimal transformers in the same transformer cluster, and determining the locations of various insulation defects on the main trunk of the overhead line to be monitored; The determining of the cluster distance according to the start time of each two adjacent target transformers connected by a line and the similarity of the current change includes: For any two adjacent target transformers connected by a line, according to the similarity of current changes between the two adjacent target transformers connected by a line, a negative correlation value of the similarity of the current changes is used as a first distance factor; Determining a cluster distance between two adjacent target mutual inductors that are connected by a line according to a comparison result of a time difference between the start times and a lag time between the two adjacent target mutual inductors that are connected by a line, in combination with the first distance factor; Wherein, the lag time represents the signal arrival time difference between mutual inductors; The method of obtaining two optimal transformers in each transformer cluster based on the sorting result and the high-frequency energy proportion of each target transformer includes: For each transformer cluster, determining a cumulative value of the ratio of the high-frequency energy proportion of any target transformer in the transformer cluster to each comparison high-frequency energy proportion; wherein the comparison high-frequency energy proportion is the high-frequency energy proportion of the remaining target transformers in the transformer cluster except the any target transformer; Performing negative correlation processing on the sorting sequence number of any target mutual inductor to obtain a negative correlation value, and combining the accumulated value of the ratio to obtain a discharge detection reliability index of any target mutual inductor; Obtain the discharge detection reliability index of each target transformer in the same transformer cluster, and select the two target transformers with the largest discharge detection reliability index as the two optimal transformers in the corresponding transformer cluster; The insulation defect monitoring is performed based on the overhead line length, the time difference between the start time and the hysteresis time between the two optimal transformers in the same transformer cluster, and the locations of each insulation defect on the main trunk of the overhead line to be monitored are determined, including: For the starting times of the two optimal transformers in the same transformer cluster, the location of the optimal transformer with the smaller starting time is taken as the near end of the overhead line, and the difference between the larger starting time and the smaller starting time is taken as the time difference of the starting time; determining a ratio of the time difference to the lag time, and performing negative correlation processing on the ratio to obtain a negative correlation value; Using the negative correlation value to correct half of the overhead line length, to obtain the distance between the insulation defect and the near end of the overhead line; The position of the insulation defect is obtained according to the position of the proximal end of the overhead line and the distance between the insulation defect and the proximal end of the overhead line.

2. The online monitoring system for insulation defect discharge of distribution network overhead lines according to claim 1, characterized in that: Obtain the high-frequency energy ratio, including: For any transformer, obtain the IMF components corresponding to the current signal of the transformer; Determine the proportion of high-frequency energy in the transformer based on the IMF value of each IMF component at each moment and the frequency of each IMF component; The IMF value and the frequency are both positively correlated with the high-frequency energy ratio.

3. The online monitoring system for insulation defect discharge of distribution network overhead lines according to claim 1 is characterized in that: The determining, based on the pulse current signal, the degree of similarity of current changes between each candidate current transformer and its adjacent and line-connected high-frequency current transformer includes: For any candidate transformer, the transformer adjacent to the candidate transformer and connected to the line is taken as the transformer to be analyzed; Determining a lag time between the candidate transformer and any transformer to be analyzed, wherein the lag time represents a difference in signal arrival time between the transformers; The degree of similarity between the current changes of the candidate transformer and the transformer to be analyzed is determined based on the lag time and the pulse current signal of the candidate transformer.

4. The online monitoring system for insulation defect discharge of distribution network overhead lines according to claim 3 is characterized in that: The determining of the lag time between the candidate transformer and any transformer to be analyzed includes: The candidate transformer and any transformer to be analyzed generate synchronous high-frequency current signals, so that the candidate transformer captures the high-frequency current signal generated by the transformer to be analyzed, and the transformer to be analyzed captures the high-frequency current signal generated by the candidate transformer, and records the time points when the high-frequency current signal starts to be transmitted and is finally received; The difference between two time points of the same high-frequency current signal is calculated to obtain two time differences; and the average of the two time differences is used as the lag time between the candidate mutual inductor and any mutual inductor to be analyzed.

5. The online monitoring system for insulation defect discharge of distribution network overhead lines according to claim 3 is characterized in that: The determining, based on the lag time and the pulse current signal of the candidate transformer, a degree of similarity in current change between the candidate transformer and the transformer to be analyzed includes: Determine the respective pending starting times of the current signal collected by the transformer to be analyzed based on the lag time and the starting time corresponding to the pulse current signal of the candidate transformer; Obtaining the current signal of the transformer to be analyzed at each pending starting moment through each pending starting moment; wherein the starting moment is the first moment, and the time interval of the current signal is equal to the time interval of the current period; According to the current signal of the transformer to be analyzed and the pulse current signal of the candidate transformer at each undetermined starting time, the similarity degree of the current change between the candidate transformer and the transformer to be analyzed is determined.

6. The online monitoring system for insulation defect discharge of distribution network overhead lines according to claim 1, characterized in that: The method of selecting a target current transformer from all high-frequency current transformers on the main trunk of the overhead line to be monitored according to the similarity of the current changes includes: For any transformer to be analyzed, a normalized negative correlation process is performed on the similarity of the current changes to obtain a correction coefficient for the initial energy proportion threshold; The initial energy ratio threshold is corrected using the correction coefficient to obtain the energy ratio threshold of the transformer to be analyzed; The high-frequency energy proportion of the transformer to be analyzed is determined, and if the high-frequency energy proportion of the transformer to be analyzed is greater than the energy proportion threshold, the transformer to be analyzed is used as a target transformer.

7. The online monitoring system for insulation defect discharge of distribution network overhead lines according to claim 6, characterized in that: Also includes: The candidate transformers whose high-frequency energy ratio is greater than the initial energy ratio threshold are also used as target transformers.

Images