Power distribution line tree line discharge fault detection method based on primary and secondary fusion switch

Abstract

The application relates to the technical field of tree line discharge fault detection, in particular to a power distribution line tree line discharge fault detection method based on primary-secondary fusion switches, which comprises the following steps: when a high-resistance ground fault occurs in a single signal collection cycle of a to-be-detected medium-voltage power distribution line, tree line discharge fault detection is carried out on the line connected with a single primary-secondary fusion switch in the to-be-detected medium-voltage power distribution line, specifically: in the single signal collection cycle, trend characteristic values of zero sequence voltage signals and zero sequence current signals are respectively acquired, and then a first tree line discharge fault characteristic value is comprehensively constructed; a second tree line discharge fault characteristic value is constructed by analyzing the reverse correlation degree between the voltage effective value of a fault phase voltage signal and the voltage effective value of a zero sequence voltage signal in time sequence; and whether a tree line discharge fault occurs in the line connected with the single primary-secondary fusion switch is judged. The application can significantly improve the accuracy and robustness of tree line discharge fault detection.

Description

Technical Field

[0001] This application relates to the field of tree-line discharge fault detection technology, specifically to a method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary integrated switch. Background Technology

[0002] With the development of power resources, more and more medium-voltage power distribution lines need to cross forests and grasslands. However, the conductors on the power distribution lines are prone to coming into contact with tree branches and leaves due to external factors, which can lead to arc discharge, or tree discharge. The high temperature generated by tree discharge can ignite trees and surrounding vegetation, thus causing forest and grassland fires.

[0003] Although tree-line discharge faults in medium-voltage distribution lines are single-phase high-resistance grounding faults, the fault current generated by tree-line discharge faults is relatively small and usually cannot reach the detection threshold of fault current in traditional grounding fault protection devices, making it difficult to take timely measures against tree-line discharge faults in medium-voltage distribution lines. Summary of the Invention

[0004] The purpose of this application is to provide a method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary integrated switch, so as to solve the problems existing in the above-mentioned background technology.

[0005] The above-mentioned technical objective of this application is achieved through the following technical solution:

[0006] One embodiment of this application provides a method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary fusion switch. The method includes the following steps:

[0007] When a high-resistance grounding fault occurs in the medium-voltage distribution line under test within a single signal acquisition cycle, tree-line discharge fault detection is performed on the line containing a single primary and secondary fusion switch in the medium-voltage distribution line under test. Specifically:

[0008] Within a single signal acquisition cycle, by analyzing the rising trend characteristics of the transient DC component of the zero-sequence voltage signal and the transient DC component of the zero-sequence current signal over time, as well as the decreasing trend characteristics of the proportion of low-order odd harmonic frequency amplitude of the zero-sequence voltage signal and the proportion of low-order odd harmonic frequency amplitude of the zero-sequence current signal over time, the trend characteristic values ​​of the zero-sequence voltage signal and the zero-sequence current signal are obtained respectively, and then the first tree-line discharge fault characteristic value is comprehensively constructed.

[0009] By analyzing the degree of inverse temporal correlation between the effective voltage values ​​of the fault phase voltage signal and the effective voltage values ​​of the zero-sequence voltage signal, the characteristic values ​​of the second tree-line discharge fault are constructed.

[0010] By combining the first tree-line discharge fault feature value and the second tree-line discharge fault feature value, and using a pre-trained tree-line discharge fault detection model, it is determined whether a tree-line discharge fault has occurred in the line where a single primary and secondary fusion switch is located.

[0011] In one embodiment, the analysis process for the upward trend feature is as follows:

[0012] For zero-sequence voltage signals, a sliding window is used to divide the zero-sequence voltage signal within a single signal acquisition period into multiple non-overlapping time windows. The DC component of the zero-sequence voltage signal in each time window is extracted and the absolute value of the DC component is taken. The time windows are numbered according to the time sequence.

[0013] The correlation coefficient between the absolute value of the DC component and its number in all time windows within a single signal acquisition period is denoted as the rising correlation coefficient; the rising correlation coefficient reflects the rising trend characteristic of the zero-sequence voltage signal.

[0014] The upward trend characteristics of the zero-sequence current signal are analyzed according to the analysis process of the upward trend characteristics of the zero-sequence voltage signal.

[0015] In one embodiment, the characteristic value of the rising trend of the zero-sequence voltage signal is the rising correlation.

[0016] In one embodiment, the analysis process for the downward trend feature is as follows:

[0017] The percentage of the amplitude of the preset low-order odd harmonic frequency of the zero-sequence voltage signal in all harmonic frequencies within each time window is denoted as the low-order odd harmonic percentage.

[0018] The negative of the correlation coefficient between the proportion of low-order odd harmonics and their numbers in all time windows within a single signal acquisition period is denoted as the decreasing correlation; the decreasing correlation reflects the decreasing trend of the zero-sequence voltage signal.

[0019] The downward trend characteristics of the zero-sequence current signal are analyzed according to the analysis process of the downward trend characteristics of the zero-sequence voltage signal.

[0020] In one embodiment, the characteristic value of the downward trend of the zero-sequence voltage signal is the downward correlation.

[0021] In one embodiment, the trend characteristic value of the zero-sequence voltage signal is the average of the characterization values ​​of the rising trend characteristic of the zero-sequence voltage signal and the characterization values ​​of the falling trend characteristic of the zero-sequence voltage signal; the trend characteristic value of the zero-sequence current signal is the average of the characterization values ​​of the rising trend characteristic of the zero-sequence current signal and the characterization values ​​of the falling trend characteristic of the zero-sequence current signal.

[0022] In one embodiment, the first tree-line discharge fault characteristic value is the average of the trend characteristic values ​​of the zero-sequence voltage signal and the zero-sequence current signal.

[0023] In one embodiment, the process of constructing the second tree-line discharge fault characteristic value is as follows:

[0024] Using the same sliding window as the zero-sequence voltage signal, the fault phase voltage signal within a single signal acquisition cycle is divided into multiple time windows;

[0025] Calculate the effective voltage value of the zero-sequence voltage signal and the effective voltage value of the fault phase voltage signal for each time window;

[0026] The correlation coefficient in time between the effective voltage value of the zero-sequence voltage signal and the effective voltage value of the fault phase voltage signal within a single signal acquisition cycle is denoted as the trend correlation.

[0027] The second tree-line discharge fault characteristic value is negatively correlated with the changing trend.

[0028] In one embodiment, the second treeline discharge fault characteristic value is the inverse of the correlation of the change trend.

[0029] In one embodiment, after determining that a tree-line discharge fault has occurred, the single primary and secondary fusion switch is controlled to operate itself to perform a grounding protection operation on the circuit.

[0030] Compared with the prior art, the beneficial effects of this application are as follows:

[0031] (1) By analyzing the transient DC component of the zero-sequence voltage signal and the transient DC component of the zero-sequence current signal of the line where the single primary and secondary fusion switch is located, as well as the proportion of the low-order odd harmonic frequency amplitude of the zero-sequence voltage signal and the proportion of the low-order odd harmonic frequency amplitude of the zero-sequence current signal over time, the first tree-line discharge fault characteristic value is comprehensively constructed. It can make full use of the physical characteristics of the tree branch temperature rising during the tree-line discharge fault, which leads to the decrease of the transition resistance, accurately capture the rising trend of the transient DC component of the zero-sequence voltage and zero-sequence current, as well as the decreasing trend of the proportion of the low-order odd harmonic, and improve the comprehensive extraction capability of the weak characteristics of the tree-line discharge fault.

[0032] (2) By analyzing the degree of reverse correlation between the effective voltage value of the fault phase voltage signal and the effective voltage value of the zero sequence voltage signal in time, the second tree-line discharge fault characteristic value is constructed. This can make full use of the physical law that the fault phase voltage and the zero sequence voltage show an inverse trend in the neutral point ungrounded system, which helps to enhance the reliability of tree-line discharge fault identification.

[0033] (3) By combining the first tree-line discharge fault characteristic value and the second tree-line discharge fault characteristic value, it is determined whether a tree-line discharge fault has occurred in the line where a single primary and secondary fusion switch is located. It can integrate multi-dimensional information to achieve comprehensive judgment of tree-line discharge fault, significantly improve the accuracy and robustness of tree-line discharge fault detection, thereby effectively overcoming the technical defects of traditional ground fault protection devices based on fault current threshold that are difficult to detect tree-line discharge faults, timely and accurately identifying tree-line discharge faults in medium-voltage distribution lines, and providing a reliable decision basis for subsequent grounding protection operations on the circuit. Attached Figure Description

[0034] Figure 1 A flowchart illustrating the steps of the power distribution line tree discharge fault detection method based on a primary and secondary fusion switch provided in this application;

[0035] Figure 2 This is a schematic diagram illustrating the process of obtaining the tree-line discharge fault detection vector. Detailed Implementation

[0036] The following description, in conjunction with the accompanying drawings, details the specific scheme of the power distribution line tree discharge fault detection method based on primary and secondary integrated switches provided in this application.

[0037] Figure 1 The flowchart of the method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary integrated switch provided in this application is as follows: The method includes the following steps:

[0038] Since most power distribution lines crossing forest areas use a neutral point ungrounded system, this application takes a medium-voltage power distribution line using a neutral point ungrounded system as an example.

[0039] The primary and secondary fusion switch installed in the medium-voltage distribution line under test integrates a tree-line discharge fault detection function to determine whether a tree-line discharge fault has occurred in the line where each primary and secondary fusion switch is located. After determining that a tree-line discharge fault has occurred, the primary and secondary fusion switch is controlled to perform grounding protection operation on the circuit and report maintenance information.

[0040] Taking a single primary and secondary integrated switch installed in a medium-voltage distribution line under test as an example, since tree-line discharge faults in medium-voltage distribution lines are high-resistance grounding faults, a distribution network high-resistance grounding fault detection method based on the fault component principle is used to determine whether a high-resistance grounding fault has occurred in the medium-voltage distribution line under test. The distribution network high-resistance grounding fault detection method based on the fault component principle is a known technology and will not be described in detail in this application.

[0041] When a high-resistance grounding fault occurs in a single signal acquisition cycle of the medium-voltage distribution line under test, tree-line discharge fault detection is performed on the line where the single primary and secondary fusion switch of the medium-voltage distribution line under test is located.

[0042] In this embodiment, the length of the signal acquisition period is 12s. The length of the signal acquisition period can be flexibly adjusted by the implementer according to the actual situation, and this application does not impose any restrictions on it.

[0043] Step 1: Analyze the rising trend characteristics of the transient DC component of the zero-sequence voltage signal and the transient DC component of the zero-sequence current signal of the line where a single primary and secondary fusion switch is located over time.

[0044] Generally, when a conductor in a medium-voltage power distribution line under test comes into contact with a tree branch due to wind deflection or sag, tree arcing occurs. At this time, the arc acts as a conductive channel, allowing current from the conductor to flow into the branch. This causes the branch's temperature to gradually increase during the tree arcing process due to the Joule heating effect of the current. Since the resistivity of the branch is negatively correlated with temperature, the branch resistance decreases as temperature increases. The transition resistance of the conductor depends on the branch resistance, and therefore, the transition resistance also decreases accordingly.

[0045] In a neutral-point ungrounded system, the transient DC components of zero-sequence voltage and zero-sequence current typically increase as the transition resistance decreases. Therefore, when a tree-line discharge fault occurs in the medium-voltage distribution line under test, the transient DC components of zero-sequence voltage and zero-sequence current in the medium-voltage distribution line under test will show an upward trend.

[0046] Based on the above analysis, the zero-sequence voltage signal and zero-sequence current signal of a single primary and secondary integrated switch line are obtained in a single signal acquisition cycle; wherein, the zero-sequence voltage signal and zero-sequence current signal are acquired in real time through a zero-sequence voltage transformer and a zero-sequence current transformer, respectively.

[0047] In this embodiment, the sampling frequency of both the zero-sequence voltage signal and the zero-sequence current signal is 100kHz. The sampling frequency can be flexibly adjusted by the implementer according to the actual situation, and this application does not impose any restrictions on it.

[0048] Furthermore, the increasing trend characteristics of the transient DC component of the zero-sequence voltage signal and the transient DC component of the zero-sequence current signal of the line where a single primary and secondary integrated switch is located are analyzed over time. The specific process is as follows:

[0049] For zero-sequence voltage signals, a sliding window is used to divide the zero-sequence voltage signal within a single signal acquisition period into multiple non-overlapping time windows. The DC component of the zero-sequence voltage signal in each time window is extracted and the absolute value of the DC component is taken. The time windows are numbered according to the time sequence.

[0050] The correlation coefficient between the absolute value of the DC component and its number in all time windows within a single signal acquisition period is denoted as the rising correlation coefficient; the rising correlation coefficient reflects the rising trend characteristic of the zero-sequence voltage signal.

[0051] The upward trend characteristics of the zero-sequence current signal are analyzed according to the analysis process of the upward trend characteristics of the zero-sequence voltage signal.

[0052] In this embodiment, the rising correlation is used as a characterization value of the rising trend feature of the zero-sequence voltage signal.

[0053] In this embodiment, the length and step size of the sliding window are equal, both being integer multiples of the period corresponding to the fundamental frequency of the zero-sequence voltage signal. This ensures that each time window can contain at least one complete period of waveform data and that adjacent time windows are continuous and do not overlap. It also prevents the problem of spectral range leakage in the fast Fourier transform operation. The fundamental frequency is 50Hz, so the length of the period corresponding to the fundamental frequency is 0.02s, and the length of the sliding window is 0.06s. The length of the sliding window can be flexibly adjusted by the implementer according to the actual situation, and this application does not impose any restrictions on it.

[0054] In this embodiment, the DC component is extracted using Fast Fourier Transform (FFT). The extraction of the DC component using FFT is a well-known technique and will not be described in detail in this application.

[0055] In this embodiment, when calculating the temporal correlation coefficient between the absolute value of the DC component and its number, the absolute values ​​of the DC components in all time windows within a single signal acquisition period are first arranged in temporal order to form a DC component sequence. Simultaneously, the numbers of all time windows within a single signal acquisition period are arranged in temporal order to form a number sequence. Then, the correlation coefficient between the DC component sequence and the number sequence is calculated. Specifically, the correlation coefficient is the Spearman correlation coefficient. The Spearman correlation coefficient is a well-known technique and will not be described in detail in this application. As other implementation methods, based on the ability to measure the temporal correlation between the absolute value of the DC component and its number, the implementer may choose other existing feasible techniques. This application does not impose any restrictions on this.

[0056] It should be added that when calculating the temporal correlation coefficient between the absolute value of the DC component and its number in all time windows within a single signal acquisition period, if the absolute value of the DC component is the same in all time windows within a single signal acquisition period, the correlation coefficient should be set to 0.

[0057] It should be noted that: the larger the value of the upward trend characteristic of the calculated zero-sequence voltage signal, the more obvious the upward trend of the transient DC component of the zero-sequence voltage signal over time; the larger the value of the upward trend characteristic of the calculated zero-sequence current signal, the more obvious the upward trend of the transient DC component of the zero-sequence current signal over time.

[0058] Step 2: Analyze the decreasing trend of the proportion of low-order odd harmonic frequency amplitude of the zero-sequence voltage signal and the proportion of low-order odd harmonic frequency amplitude of the zero-sequence current signal over time in the circuit where a single primary and secondary fusion switch is located.

[0059] In the initial stage of a tree-line discharge fault, an arc discharge has just occurred between the conductor and the tree branch in the medium-voltage distribution line under test. At this time, the harmonics of the zero-sequence voltage and zero-sequence current in the medium-voltage distribution line under test are dominated by low-order odd harmonics, such as the 3rd, 5th, and 7th harmonics. As the tree-line discharge fault continues to develop, the middle and high-order harmonics in the zero-sequence voltage and zero-sequence current increase, causing the proportion of the frequency amplitude of low-order odd harmonics to gradually decrease.

[0060] Based on the above analysis, the following analysis examines the decreasing trend of the proportion of low-order odd-order harmonic frequencies in the zero-sequence voltage signal and the proportion of low-order odd-order harmonic frequencies in the zero-sequence current signal over time. The specific process is as follows:

[0061] Extract the spectrum of the zero-sequence voltage signal in each time window, and record the amplitude ratio of the preset low-order odd harmonic frequency of the zero-sequence voltage signal in each time window among all harmonic frequencies as the low-order odd harmonic ratio.

[0062] The negative of the correlation coefficient between the proportion of low-order odd harmonics and their numbers in all time windows within a single signal acquisition period is denoted as the decreasing correlation; the decreasing correlation reflects the decreasing trend of the zero-sequence voltage signal.

[0063] The downward trend characteristics of the zero-sequence current signal are analyzed according to the analysis process of the downward trend characteristics of the zero-sequence voltage signal.

[0064] In this embodiment, the decreasing correlation is used as a characterization value of the decreasing trend feature of the zero-sequence voltage signal.

[0065] In this embodiment, the spectrum of the zero-sequence voltage signal is extracted using Fast Fourier Transform. The extraction of the spectrum of the zero-sequence voltage signal using Fast Fourier Transform is a well-known technique and will not be described in detail in this application. As other implementation methods, implementers may choose other existing feasible techniques at their own discretion, and this application does not impose any restrictions on this.

[0066] In this embodiment, the preset low-order odd harmonics refer to the 3rd, 5th, and 7th harmonics.

[0067] In this embodiment, when calculating the temporal correlation coefficient between the proportion of low-order odd harmonics and their numbering, the proportions of low-order odd harmonics in all time windows within a single signal acquisition cycle are first arranged chronologically to form a low-order odd harmonic proportion sequence. Then, the correlation coefficient between the low-order odd harmonic proportion sequence and the numbering sequence is calculated. Specifically, the correlation coefficient is the Spearman correlation coefficient. As another implementation method, based on the ability to measure the temporal correlation between the proportion of low-order odd harmonics and their numbering, the implementer may choose other existing feasible technologies, and this application does not impose any restrictions in this regard.

[0068] It should be added that when calculating the temporal correlation coefficient between the proportion of low-order odd harmonics and their numbers in all time windows within a single signal acquisition period, if the proportion of low-order odd harmonics is the same in all time windows within a single signal acquisition period, the correlation coefficient should be set to 0.

[0069] It should be noted that: the larger the value of the descending trend characteristic of the calculated zero-sequence voltage signal, the more obvious the descending trend of the proportion of the low-order odd harmonic frequency amplitude of the zero-sequence voltage signal over time; the larger the value of the descending trend characteristic of the calculated zero-sequence current signal, the more obvious the descending trend of the proportion of the low-order odd harmonic frequency amplitude of the zero-sequence current signal over time.

[0070] Step 3: Obtain the trend characteristic values ​​of the zero-sequence voltage signal and the zero-sequence current signal respectively, and then comprehensively construct the first tree-line discharge fault characteristic value.

[0071] By analyzing the upward trend characteristics and downward trend characteristics of the zero-sequence voltage signal, the trend characteristic value of the zero-sequence voltage signal is obtained, specifically as follows:

[0072] The average of the characterization values ​​of the upward trend feature and the downward trend feature of the zero-sequence voltage signal is taken as the trend feature value of the zero-sequence voltage signal.

[0073] It should be noted that: the larger the trend characteristic value of the calculated zero-sequence voltage signal, the more significant the upward trend of the transient DC component of the zero-sequence voltage signal over time, and the more significant the downward trend of the proportion of the low-order odd harmonic frequency amplitude of the zero-sequence voltage signal over time; the more the zero-sequence voltage signal in a single signal acquisition cycle has the trend change characteristics under tree-line discharge fault conditions.

[0074] By analyzing the upward trend characteristics and downward trend characteristics of the zero-sequence current signal, the trend characteristic value of the zero-sequence current signal is obtained, specifically as follows:

[0075] The average of the characterization values ​​of the upward trend feature and the downward trend feature of the zero-sequence current signal is taken as the trend feature value of the zero-sequence current signal.

[0076] It should be noted that: the larger the trend characteristic value of the calculated zero-sequence current signal, the more significant the upward trend of the transient DC component of the zero-sequence current signal over time, and the more significant the downward trend of the proportion of the low-order odd harmonic frequency amplitude of the zero-sequence current signal over time; the more the zero-sequence current signal in a single signal acquisition cycle has the trend change characteristics under tree-line discharge fault conditions.

[0077] Furthermore, the average of the trend characteristic values ​​of the zero-sequence voltage signal and the zero-sequence current signal is used as the first tree-line discharge fault characteristic value to evaluate whether the zero-sequence voltage signal and the zero-sequence current signal both have the trend change characteristics caused by the occurrence of a tree-line discharge fault within a single signal acquisition period. The larger the calculated first tree-line discharge fault characteristic value, the more likely the zero-sequence voltage signal and the zero-sequence current signal are to simultaneously have the trend change characteristics caused by the occurrence of a tree-line discharge fault within a single signal acquisition period.

[0078] Step 4: By analyzing the degree of reverse correlation in time between the effective voltage value of the fault phase voltage signal and the effective voltage value of the zero-sequence voltage signal, the second tree-line discharge fault characteristic value is constructed.

[0079] In medium-voltage distribution lines of neutral-point ungrounded systems, when a tree-line discharge fault occurs, the effective value of the fault phase voltage and the effective value of the zero-sequence voltage in the medium-voltage distribution line under test usually show opposite trends: when the effective value of the zero-sequence voltage increases, the effective value of the fault phase voltage decreases; conversely, when the effective value of the zero-sequence voltage decreases, the effective value of the fault phase voltage increases.

[0080] Based on the above analysis, the fault phase voltage signal of a single line containing a primary and secondary fusion switch is obtained within a single signal acquisition cycle. The identification of the fault phase in a high-resistance grounding fault is a well-known technique and will not be elaborated upon in this application. The fault phase voltage signal is acquired in real time using a three-phase voltage transformer.

[0081] In this embodiment, the fault phase voltage signal uses the same sampling frequency as the zero-sequence voltage signal and the zero-sequence current signal.

[0082] Furthermore, by analyzing the degree of inverse temporal correlation between the effective voltage values ​​of the fault phase voltage signal and the effective voltage value of the zero-sequence voltage signal, a second tree-line discharge fault characteristic value is constructed. The specific process is as follows:

[0083] Using the same sliding window as the zero-sequence voltage signal, the fault phase voltage signal within a single signal acquisition cycle is divided into multiple time windows;

[0084] Calculate the effective voltage value of the zero-sequence voltage signal and the effective voltage value of the fault phase voltage signal for each time window;

[0085] The correlation coefficient in time between the effective voltage value of the zero-sequence voltage signal and the effective voltage value of the fault phase voltage signal within a single signal acquisition cycle is denoted as the trend correlation.

[0086] The second tree-line discharge fault characteristic value is negatively correlated with the aforementioned trend. The calculation of the effective voltage value is a well-known technique and will not be elaborated upon in this application.

[0087] It should be noted that negative correlation refers to variables changing in opposite directions, with one increasing and the other decreasing.

[0088] In this embodiment, the second tree-line discharge fault characteristic value is the inverse of the correlation of the change trend.

[0089] In this embodiment, when calculating the temporal correlation coefficient between the effective voltage value of the zero-sequence voltage signal and the effective voltage value of the fault phase voltage signal, the effective voltage values ​​of the zero-sequence voltage signal in all time windows within a single signal acquisition period are first arranged in temporal order to form a first effective voltage value sequence. Simultaneously, the effective voltage values ​​of the fault phase voltage signal in all time windows within a single signal acquisition period are arranged in temporal order to form a second effective voltage value sequence. Then, the correlation coefficient between the first and second effective voltage value sequences is calculated. Specifically, the correlation coefficient is the Spearman correlation coefficient. As another implementation method, based on the ability to measure the temporal correlation between the effective voltage values ​​of the zero-sequence voltage signal and the fault phase voltage signal, the implementer may choose other existing feasible technologies, and this application does not impose any limitations in this regard.

[0090] It should be added that when calculating the time-series correlation coefficient between the effective voltage value of the zero-sequence voltage signal and the effective voltage value of the fault phase voltage signal, if the effective voltage value of the zero-sequence voltage signal is the same in all time windows within a single signal acquisition period, or if the effective voltage value of the fault phase voltage signal is the same in all time windows within a single signal acquisition period, the correlation coefficient should be set to 0.

[0091] It should be noted that the larger the calculated second tree-line discharge fault characteristic value, the more likely the fault phase voltage signal and the zero-sequence voltage signal are to have the opposite trend of signal distribution characteristics caused by the occurrence of tree-line discharge fault within a single signal acquisition cycle.

[0092] Step 5: Combining the first tree-line discharge fault feature value and the second tree-line discharge fault feature value, and using the pre-trained tree-line discharge fault detection model, determine whether a tree-line discharge fault has occurred in the line where a single primary and secondary fusion switch is located.

[0093] The vector composed of the first tree-line discharge fault feature value and the second tree-line discharge fault feature value within a single signal acquisition cycle is used as the tree-line discharge fault detection vector for a single signal acquisition cycle. This vector is then used to detect whether a tree-line discharge fault has occurred in the line where the single primary and secondary fusion switch is located within a single signal acquisition cycle. Figure 2 This is a schematic diagram illustrating the process of obtaining the tree-line discharge fault detection vector.

[0094] Before performing treeline discharge fault detection, a treeline discharge fault detection model needs to be trained in advance to detect treeline discharge faults.

[0095] In this embodiment, the training method for the tree-line discharge fault detection model is as follows: the fault phase voltage signal, zero-sequence voltage signal, and zero-sequence current signal of a preset number of medium-voltage distribution lines belonging to a neutral point ungrounded system are collected within a preset number of signal acquisition cycles. The tree-line discharge fault detection vector of each medium-voltage distribution line in each signal acquisition cycle is obtained according to the method for obtaining the tree-line discharge fault detection vector of a single signal acquisition cycle. This includes the signal acquisition cycle in which no tree-line discharge fault has occurred and the signal acquisition cycle in which a tree-line discharge fault has just occurred.

[0096] A Support Vector Machine (SVM) model is trained using all the obtained tree-line discharge fault detection vectors. The labels of tree-line discharge fault detection vectors in the signal acquisition period where no tree-line discharge fault has occurred are set to 0, and the labels of tree-line discharge fault detection vectors in the signal acquisition period where a tree-line discharge fault has just occurred are set to 1. The performance evaluation metric for the SVM model is the F1-score. The SVM model uses the Radial Basis Function (RBF) kernel function. The training of the SVM model is a well-known technique and will not be described in detail in this application.

[0097] In this embodiment, the preset quantity and the preset number are 10 and 500 respectively. The preset quantity and the preset number can be flexibly adjusted by the implementer according to the actual situation. For example, when the medium voltage power distribution line is in a complex and changeable operating environment, the value of the preset quantity can be increased. This application does not impose any restrictions on this.

[0098] Furthermore, the tree-line discharge fault detection vector of a single signal acquisition cycle is used as the input of the tree-line discharge fault detection model. If the output value of the tree-line discharge fault detection model is 1, it indicates that a tree-line discharge fault has occurred in the line where the single primary and secondary fusion switch of the medium-voltage distribution line under test is located within a single signal acquisition cycle. In this case, the primary and secondary fusion switch is controlled to operate to perform grounding protection operation on the circuit and report maintenance information. If the output value of the tree-line discharge fault detection model is 0, it indicates that no tree-line discharge fault has occurred in the line where the single primary and secondary fusion switch of the medium-voltage distribution line under test is located within a single signal acquisition cycle.

[0099] In summary, (1) by analyzing the transient DC component of the zero-sequence voltage signal and the transient DC component of the zero-sequence current signal of the line where the single primary and secondary fusion switch is located, as well as the proportion of the low-order odd harmonic frequency amplitude of the zero-sequence voltage signal and the proportion of the low-order odd harmonic frequency amplitude of the zero-sequence current signal over time, the first tree-line discharge fault characteristic value is constructed in a comprehensive manner. This can make full use of the physical characteristics of the decrease in transition resistance caused by the increase in tree branch temperature during the tree-line discharge fault, accurately capture the upward trend of the transient DC component of the zero-sequence voltage and zero-sequence current, as well as the downward trend of the proportion of low-order odd harmonics, and improve the comprehensive extraction capability of the weak characteristics of the tree-line discharge fault.

[0100] (2) By analyzing the degree of reverse correlation between the effective voltage value of the fault phase voltage signal and the effective voltage value of the zero sequence voltage signal in time, the second tree-line discharge fault characteristic value is constructed. This can make full use of the physical law that the fault phase voltage and the zero sequence voltage show an inverse trend in the neutral point ungrounded system, which helps to enhance the reliability of tree-line discharge fault identification.

[0101] (3) By combining the first tree-line discharge fault characteristic value and the second tree-line discharge fault characteristic value, it is determined whether a tree-line discharge fault has occurred in the line where a single primary and secondary fusion switch is located. It can integrate multi-dimensional information to achieve comprehensive judgment of tree-line discharge fault, significantly improve the accuracy and robustness of tree-line discharge fault detection, thereby effectively overcoming the technical defects of traditional ground fault protection devices based on fault current threshold that are difficult to detect tree-line discharge faults, timely and accurately identifying tree-line discharge faults in medium-voltage distribution lines, and providing a reliable decision basis for subsequent grounding protection operations on the circuit.

Claims

1. A method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary integrated switch, characterized in that, The method includes the following steps: When a high-resistance grounding fault occurs in the medium-voltage distribution line under test within a single signal acquisition cycle, tree-line discharge fault detection is performed on the line containing a single primary and secondary fusion switch in the medium-voltage distribution line under test. Specifically: Within a single signal acquisition cycle, by analyzing the rising trend characteristics of the transient DC component of the zero-sequence voltage signal and the transient DC component of the zero-sequence current signal over time, as well as the decreasing trend characteristics of the proportion of low-order odd harmonic frequency amplitude of the zero-sequence voltage signal and the proportion of low-order odd harmonic frequency amplitude of the zero-sequence current signal over time, the trend characteristic values ​​of the zero-sequence voltage signal and the zero-sequence current signal are obtained respectively, and then the first tree-line discharge fault characteristic value is comprehensively constructed. By analyzing the degree of inverse temporal correlation between the effective voltage values ​​of the fault phase voltage signal and the effective voltage values ​​of the zero-sequence voltage signal, the characteristic values ​​of the second tree-line discharge fault are constructed. By combining the first tree-line discharge fault feature value and the second tree-line discharge fault feature value, and using a pre-trained tree-line discharge fault detection model, it is determined whether a tree-line discharge fault has occurred in the line where a single primary and secondary fusion switch is located. The analysis process for the aforementioned downward trend characteristics is as follows: For zero-sequence voltage signals, a sliding window is used to divide the zero-sequence voltage signal within a single signal acquisition period into multiple non-overlapping time windows, and the time windows are numbered according to the timing sequence. The percentage of the amplitude of the preset low-order odd harmonic frequency of the zero-sequence voltage signal in all harmonic frequencies within each time window is denoted as the low-order odd harmonic percentage. The negative of the correlation coefficient between the proportion of low-order odd harmonics and their numbers in all time windows within a single signal acquisition period is denoted as the decreasing correlation; the decreasing correlation reflects the decreasing trend of the zero-sequence voltage signal. The downward trend characteristics of the zero-sequence current signal are analyzed according to the analysis process of the downward trend characteristics of the zero-sequence voltage signal.

2. The method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary integrated switch as described in claim 1, characterized in that, The analysis process for the upward trend feature is as follows: Extract the DC component of the zero-sequence voltage signal in each time window and take the absolute value of the DC component; denot the correlation coefficient between the absolute value of the DC component and the number in all time windows within a single signal acquisition period as the rising correlation; the rising correlation reflects the rising trend characteristic of the zero-sequence voltage signal. The upward trend characteristics of the zero-sequence current signal are analyzed according to the analysis process of the upward trend characteristics of the zero-sequence voltage signal.

3. The method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary integrated switch as described in claim 2, characterized in that, The characteristic value of the upward trend of the zero-sequence voltage signal is the upward correlation.

4. The method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary integrated switch as described in claim 1, characterized in that, The characteristic value of the downward trend of the zero-sequence voltage signal is the downward correlation.

5. The method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary integrated switch as described in claim 1, characterized in that, The trend characteristic value of the zero-sequence voltage signal is the average of the characterization value of the upward trend characteristic of the zero-sequence voltage signal and the characterization value of the downward trend characteristic of the zero-sequence voltage signal. The trend characteristic value of the zero-sequence current signal is the average of the characterization values ​​of the upward trend characteristic of the zero-sequence current signal and the characterization values ​​of the downward trend characteristic of the zero-sequence current signal.

6. The method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary integrated switch as described in claim 1, characterized in that, The first tree-line discharge fault characteristic value is the average of the trend characteristic values ​​of the zero-sequence voltage signal and the zero-sequence current signal.

7. The method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary integrated switch as described in claim 2, characterized in that, The process of constructing the second tree-line discharge fault characteristic value is as follows: Using the same sliding window as the zero-sequence voltage signal, the fault phase voltage signal within a single signal acquisition cycle is divided into multiple time windows; Calculate the effective voltage value of the zero-sequence voltage signal and the effective voltage value of the fault phase voltage signal for each time window; The correlation coefficient in time between the effective voltage value of the zero-sequence voltage signal and the effective voltage value of the fault phase voltage signal within a single signal acquisition cycle is denoted as the trend correlation. The second tree-line discharge fault characteristic value is negatively correlated with the changing trend.

8. The method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary integrated switch as described in claim 7, characterized in that, The second tree-line discharge fault characteristic value is the inverse of the correlation of the changing trend.

9. The method for detecting tree-line discharge faults in power distribution lines based on a primary and secondary integrated switch as described in claim 1, characterized in that, After determining that a tree-line discharge fault has occurred, the single primary and secondary fusion switch is controlled to operate to perform grounding protection operation on the circuit.

Images