A Distribution Network Fault Early Warning Method Based on Repetitive Fault Identification

Abstract

The present invention discloses a distribution network fault early warning method based on repetitive fault identification, comprising the following steps: S1, detecting and collecting instantaneous fault data information and establishing an instantaneous fault database; S2, integrating the faults that can be self-recovered by insulation according to the database, and extracting instantaneous single-phase grounding fault events therefrom; S3, extracting single-phase grounding faults occurring on the same feeder and the same phase; S4, integrating a repetitive fault sequence library through manual identification of waveform similarity; S5, obtaining evaluation indexes of single fault events and fault events in a time series; S6, evaluating the severity of the faults and analyzing the fault development trend; S7, implementing early warning according to the evaluation and analysis results of the faults. The evaluation indexes proposed by the present invention have good adaptability to single-phase grounding faults, and the more the number of fault samples, the higher the accuracy of the proposed evaluation scheme.

Description

Technical Field

[0001] The present invention relates to the field of electricity, and particularly to a distribution network fault early warning method based on repetitive fault identification. Background Art

[0002] Distribution network faults can be divided into transient faults and permanent faults according to the nature of the faults. Although the duration of transient faults is short, the arcs formed by them will cause a certain degree of damage to the insulation device. Repeated occurrence will lead to the accumulation of damage and ultimately result in permanent faults. Among transient faults, there is a more special single-phase grounding fault, which occurs within a short time scale, and also occurs in the same feeder and the same phase, and the voltage and current waveforms within the fault time range have extremely high similarity, and are very likely to be caused by the same fault reason in the actual environment. Such faults are called repetitive faults, and feeders with frequent repetitive faults are more likely to have permanent grounding faults.

[0003] Existing technologies for analyzing repetitive faults usually rely on waveform data, and generally rely on waveform recording devices with low output voltage to record interference waveforms. After the waveform recording device detects an interference signal, the characteristics of the interference signal are further extracted to achieve the measurement and identification of the interference. In existing methods, the monitoring and analysis of repetitive faults are often realized by using the waveform data monitored when the voltage sags. However, due to the differences between the extracted characteristics and the actual characteristics, some faults cannot be detected, resulting in missed detections, with low detection accuracy and sensitivity, and poor adaptability in the actual monitoring process. Summary of the Invention

[0004] In order to solve the problems of low accuracy and poor adaptability of conventional repetitive fault identification methods, the present invention proposes a distribution network fault early warning method based on repetitive fault identification to solve the above problems.

[0005] The method includes the following steps:

[0006] S1. Detect and collect transient fault data information, and establish a transient fault database containing waveform recording files;

[0007] S2. Integrate the faults that can be self-recovered by insulation according to the database, and extract transient single-phase grounding fault events therefrom;

[0008] S3. Extract single-phase grounding faults that occur in the same feeder and the same phase;

[0009] S4. Integrate a repetitive fault sequence library by manually identifying the waveform similarity;

[0010] S5. Obtain the evaluation indexes of single fault events and fault events in the time series;

[0011] S6. Evaluate the severity of the fault and analyze the fault development trend using the analytic hierarchy process;

[0012] S7. Implement early warning according to the evaluation and analysis results of the fault.

[0013] Preferably, the extraction of instantaneous single-phase grounding faults in S2 includes the following steps:

[0014] Screen according to the characteristics of the reduced voltage of the fault phase, the increased voltage of the non-fault phase, the increased zero-sequence voltage, and the increased zero-sequence current of the fault feeder in single-phase grounding faults, and screen out waveform recording events of interphase faults, three-phase voltage oscillations, and non-fault disturbances;

[0015] Regard the disturbance events that have not achieved self-clearance of the fault until the end record time of the waveform recording file as permanent fault events, and screen out this part of permanent fault events.

[0016] Preferably, in S3, the polarities of the zero-sequence currents of single-phase grounding faults on the same feeder and of the same phase are opposite, and the phase voltage of the fault phase decreases.

[0017] Preferably, the evaluation indexes for a single fault event in S5 include:

[0018] Zero-sequence voltage peak value U 0i , where i represents the i-th instantaneous fault: the higher the zero-sequence voltage peak value U 0i , the greater the impact of this fault on the original voltage level of the line, the greater the damage to the tolerance of the feeder, and the higher the severity of the fault;

[0019] Fault energy P i : The greater the fault energy P i , the higher the severity of the fault and the greater the impact on the line. The calculation formula for the fault energy P i is:

[0020] P i = ∫ t i f (t)u f (t)dt;

[0021] where t is the duration of a single fault, i f is the fault current, and u f is the fault voltage;

[0022] Fault initial phase angle ω 0i : Take the phase angle difference between the fault initial phase angle ω 0i and the phase angle at the adjacent nearest peak as an evaluation index. The greater the phase angle difference, the lower the insulation tolerance;

[0023] The single - fault duration \(t\). The longer the single - fault duration \(t\), the lower the self - recovery ability of the line, and the greater the damage to the line insulation.

[0024] Preferably, the evaluation indexes of the fault events in the time series in S5 include:

[0025] The total fault energy \(P\) borne by the line within the time series:

[0026] \(P=\sum P\) i ;

[0027] Using the total fault energy to reflect the cumulative effect of the fault on the line, the degree of line deterioration increases with the increase of the cumulative effect;

[0028] The total fault time \(T\) of the line within the time series:

[0029] \(T = \sum t\);

[0030] That is, within the time series, the time when the line is actually in the fault state can represent the severity of the fault and the self - clearing fault recovery ability of the line;

[0031] The total number of faults \(N\) of the line within the time series: The more the total number of faults \(N\), the more obvious the deterioration trend of the line's tolerance to faults;

[0032] The time - interval sequence \(\{a\) n \} of \(N\) fault occurrences within the time series:

[0033] \(\{a\) n \} = \{t2 - t1,t3 - t2,\cdots,t\) N - t N-1 \};

[0034] If the frequency of fault occurrence is getting higher and higher, that is, the time interval between two faults is getting smaller and smaller, it proves that the insulation level of the line has a deterioration trend and the tolerance to faults decreases.

[0035] Advantages of the present invention:

[0036] The detection indexes proposed in the present invention are divided into two different categories for single events and time series, and the severity of the fault and the insulation level of the feeder are evaluated. The evaluation indexes proposed in the present invention have good adaptability to single - phase grounding faults, and the more the number of fault samples, the higher the accuracy of the proposed evaluation scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a flowchart of the distribution network fault early - warning method based on repetitive fault identification according to an embodiment of the present invention;

[0038] Figure 2Statistical chart of zero-sequence voltage peak value and arc energy during transient faults in the embodiments of the present invention;

[0039] Figure 3 Schematic diagram of evaluation indexes for single fault events and fault events in time series in the embodiments of the present invention;

[0040] Figure 4 Zero-sequence voltage change waveform of the first sequence in the embodiments of the present invention;

[0041] Figure 5 Zero-sequence voltage change waveform of the second sequence in the embodiments of the present invention;

[0042] Figure 6 Zero-sequence voltage change waveform of the third sequence in the embodiments of the present invention;

[0043] Figure 7 Schematic diagram of parameter change of single fault event in the first sequence in the embodiments of the present invention;

[0044] Figure 8 Schematic diagram of parameter change of single fault event in the second sequence in the embodiments of the present invention;

[0045] Figure 9 Schematic diagram of parameter change of single fault event in the third sequence in the embodiments of the present invention;

[0046] Figure 10 Schematic diagram of time series change in the first sequence in the embodiments of the present invention;

[0047] Figure 11 Schematic diagram of time series change in the second sequence in the embodiments of the present invention;

[0048] Figure 12 Schematic diagram of time series change in the third sequence in the embodiments of the present invention;

[0049] Figure 13 Schematic diagram of time interval between faults in each sequence in the embodiments of the present invention. Detailed implementation manners

[0050] To make the objectives, technical solutions and advantages of the present application clearer and more understandable, the following takes examples with reference to the accompanying drawings and further elaborates on the present application in detail.

[0051] The embodiments of the present application disclose a distribution network fault early warning method based on repetitive fault identification, including the following steps:

[0052] S1. Detect and collect transient fault data information, and establish a transient fault database including waveform record files;

[0053] S2. Based on database integration, identify the faults with self - recovering insulation, and extract the instantaneous single - phase grounding fault events from them;

[0054] S3. Extract the single - phase grounding faults that occur on the same feeder and of the same phase;

[0055] S4. Through manual identification of waveform similarity, integrate a repetitive fault sequence library;

[0056] S5. Obtain the evaluation indexes of single - fault events and fault events in time series;

[0057] S6. Evaluate the severity of the faults according to the evaluation indexes, and analyze the fault development trend by using the analytic hierarchy process;

[0058] S7. Implement early warning according to the evaluation and analysis results of the faults.

[0059] The extraction of the instantaneous single - phase grounding faults in S2 includes the following steps:

[0060] Ensure that the disturbance event is a single - phase grounding fault event. When a single - phase grounding fault occurs, the voltage and current have obvious characteristics, including the reduction of the voltage of the fault phase, the increase of the voltage of the non - fault phase, the increase of the zero - sequence voltage, and the increase of the zero - sequence current of the fault feeder (the increase degree depends on the system grounding type). According to these characteristics, screen out the waveform record events of inter - phase faults, three - phase voltage oscillations, non - fault disturbances, etc.

[0061] Ensure that the disturbance event is an instantaneous fault event. The main purpose of data analysis is to realize the estimation of the fault development trend and the insulation early warning of relevant feeders. Therefore, select the instantaneous fault events, analyze the change trend of each index when they occur repeatedly, so as to evaluate their development status and achieve early warning before evolving into permanent faults. In one embodiment, the disturbance events that have not achieved self - clearing of the fault by the end record time of the waveform record file are regarded as permanent fault events, and such events are screened out.

[0062] In each repetition of the faults in the embodiments of the present application, it should occur at the same spatial position. In one embodiment, in S3, ignore the influence of the fault distance at the time of fault occurrence on the evaluation, use the opposite polarity of the zero - sequence current as the discriminant condition for the fault feeder, and use the significant reduction of the phase voltage as the discriminant condition for the fault phase.

[0063] In S4, in order to reflect the repetitive characteristics of the fault events, in the embodiments of the present application, use the high similarity between the recorded disturbance waveforms as the basis for the events being caused by the same interference. If several fault events have highly similar waveform characteristics, they are regarded as faults caused by the same actual interference, such as the situation of a tree branch touching the line at the same position, etc. This step is the last step of data screening and is realized by manual identification.

[0064] In one embodiment, based on the measured data of a certain substation, 24 instantaneous fault events occurring on a certain feeder were screened, and data such as the peak value of zero-sequence voltage and arc energy during the fault were extracted, obtaining the statistical chart as shown in Figure 2 which, where Figure 2 (a) shows the distribution of the peak value of zero-sequence voltage, Figure 2 (b) shows the distribution of fault arc energy. Under normal operating conditions, the amplitude of the secondary-side phase voltage of this line is about 86V. As can be seen from Figure 2 , in instantaneous faults, the peak value of zero-sequence voltage of fault events is close to the phase voltage value and shows an approximately normal distribution centered around it, and the fault arc energy also has a very large amplitude change. Therefore, both of these are parameter indicators suitable for fault detection.

[0065] The fault evaluation index is as shown in Figure 3 . The assessment of the severity of the fault can be transformed into the diagnosis of the change in line insulation strength during the fault, which is specifically reflected in two aspects: the ability of the feeder to withstand the fault and the ability of each index to return to the normal level after the fault occurs. When using the instantaneous fault parameters to evaluate the fault intensity and insulation strength in S5, the influence of the following parameters is mainly considered. Taking the i-th instantaneous fault as an example:

[0066] The peak value of zero-sequence voltage U of the bus where the instantaneous ground fault occurs 0i . In a single-phase ground fault, the zero-sequence voltage can reflect the unbalance degree of the three phases of the line to a certain extent. The higher its peak value, the greater the impact of this fault on the original voltage level of the line, the greater the damage to the tolerance of the feeder, and the higher the severity of the fault.

[0067] The arc energy P brought by the fault i :

[0068] P i =∫ t i f (t)u f (t)dt;

[0069] where t is the duration of a single fault, i f is the fault current, and u f is the fault voltage;

[0070] At the moment when a single-phase ground fault occurs, a fault current will flow through the fault point. If the grounding impedance is large, there will also be a large zero-sequence voltage. The calculation of arc energy includes three important fault parameters: fault voltage, fault current, and fault duration. If the arc energy generated at the fault point is large, it means that the severity of the fault is high and the impact on the line is greater;

[0071] Initial fault phase angle ω 0i : A single-phase grounding fault generally occurs near the moment before the voltage of the fault phase reaches its peak. If the initial fault phase angle is small, it means that the normal operating voltage level that the line can withstand decreases, and the insulation strength of the line drops. Taking the phase angle difference between the initial fault phase angle and the nearest peak as an evaluation index, the larger the phase angle difference, the lower the insulation tolerance.

[0072] Single fault duration t. The longer the fault duration, the lower the ability of the line to self-recover to the normal operating level after the fault occurs, and the insulation strength of the line drops. Also, the longer the arc burns, the greater the damage to the line insulation. The start and end times of the fault can be measured by setting a threshold voltage U set to measure the start and end times of the fault.

[0073] Since the insulation level of the line will also decrease due to the cumulative effect of faults as time goes by and the fault frequency increases, the statistical parameters such as the number of faults, frequency, and duration of the line in a time series can reflect the insulation state of the line. In S5, these parameters are used to comprehensively evaluate the insulation level of the line:

[0074] Total fault energy P that the line withstands in the time series:

[0075] P = ∑P i ;

[0076] Each fault will have different degrees of impact on the line, and the degree of line deterioration will increase due to the cumulative effect. Then the total fault energy that the line withstands in a given period is an index that can characterize this cumulative effect.

[0077] Total fault time T of the line in the time series:

[0078] T = ∑t;

[0079] That is, in the time series, how much time the line is actually in the fault state, which characterizes the severity of the fault of the line and the line's ability to self-clear the fault from the time scale.

[0080] Total number of faults N of the line in the time series: The insulation level of the line will decrease to a certain extent after multiple repetitive faults. The more the total number of faults, the more obvious the deterioration trend of the line's tolerance to faults. At the same time, according to the actual acquisition and recording of waveform detection devices, some lines will repeatedly occur multiple faults with similar characteristics in a short time, which can be considered as a continuation of a serious fault. In this case, the insulation state of the line is continuously deteriorated, having a greater impact on the insulation level of the entire line.

[0081] Time interval sequence {a of N faults occurring in the time series n}}:

[0082] {a n}} = {t2 - t1, t3 - t2, … t N -t N-1}};

[0083] If the frequency of faults occurring is getting higher and higher, that is, the time interval between two consecutive faults is getting smaller and smaller, it proves that the insulation level of the line has a tendency to deteriorate and the tolerance to faults has decreased.

[0084] In a specific embodiment, experimental simulation is carried out on the method of the present application. The selected data is divided into 3 groups according to the spatial location where they occur, as shown in Table 1, and arranged in the order of occurrence of events as the first sequence, the second sequence, and the third sequence. The waveform characteristics of the elements within the sequence have high similarity. Among them, the bus grounding methods corresponding to the fault feeders in the first sequence and the second sequence are ungrounded, and the third sequence is resonant grounding. The event span of the first sequence is relatively long, while all events in the second sequence and the third sequence occur within one day.

[0085] Table 1: Detailed information of event sequence data

[0086]

[0087]

[0088] As Figure 4 shown, in the first sequence, the disturbances mainly present as single-phase grounding faults lasting for 3 - 7 cycles. There are obvious signs of arc combustion in the fault phase, and the zero-sequence voltage is square-wave shaped.

[0089] As Figure 5 shown, in the second sequence, the fault waveform shows obvious single-phase grounding fault characteristics. The voltage of the non-fault phase increases significantly, the zero-sequence voltage is a stable sine wave, and the duration is very long. It realizes self-clearing after about a dozen cycles.

[0090] As Figure 6 shown, in the third sequence, the fault waveform mainly presents that the voltage of the fault phase decreases significantly and is accompanied by arc combustion, the voltage of the non-fault phase increases significantly, and it lasts for about eight cycles. Since in a resonant grounding system, the transient process after arc extinction is an underdamped series resonance process between the arc suppression coil inductance and the capacitance to ground, after the fault realizes self-clearing, the recovery of the three-phase voltage is oscillatory, and the zero-sequence voltage is also damped oscillation.

[0091] Based on the differences in duration, waveform characteristics, and grounding methods among the three groups of sequences, the single-fault events and time-series fault event index analyses are respectively carried out for each group of sequences. The severity of the fault is evaluated according to the changes and developments of the fault parameter characteristics, providing data support for the trend estimation of the insulation level of this feeder.

[0092] Analysis of the changes in the evaluation indexes for single-fault events:

[0093] As Figure 7 shown in (a), in the first sequence, the peak value of the zero-sequence voltage is relatively stable, fluctuating smoothly in the range of about 8 kV - 10 kV, and there is a slight upward trend starting from the 4th event, indicating that the damage caused by the fault to the line has slightly increased; as Figure 7 (b) and Figure 7 shown in (c), both the fault duration and the fault energy have a relatively obvious upward trend starting from the 4th event, indicating that the recovery ability of the line after suffering a fault has been significantly weakened, and there is a great correlation between their trends, indicating that when the zero-sequence current is small (in a small-current system) and the zero-sequence voltage is relatively stable, the magnitude of the fault energy may largely depend on the arc combustion time; as Figure 7 shown in (d), the fault phase angle difference fluctuates relatively smoothly between 0 - 20° in the previous events, but suddenly increases significantly in the last event, which may contain certain accidental factors.

[0094] As Figure 8 (a) and Figure 8 shown in (b), in the second sequence, the peak value of the zero-sequence voltage during the fault and the fault duration are relatively stable. The zero-sequence voltage is not high, but the fault duration is generally long, and the ability of the line to recover to its original operating level is low; as Figure 8 shown in (c), the arc energy of the fault fluctuates within a small range, and there is a trend of energy reduction in the last 3 events, indicating that the fault destructive power has slightly decreased; as Figure 8 shown in (d), the phase angle difference between the fault occurrence moment and the peak moment fluctuates greatly as a whole, and it is difficult to see an obvious pattern. The 16 disturbance waveform recording events in the second sequence all occurred on the same day. Although no obvious insulation deterioration trend was seen from the extracted fault parameters, the fault duration is generally long and the number of faults is large, and there are factors that continuously affect the line operation.

[0095] As Figure 9 shown in (a) - 9(c), in the third sequence, the changes of the three indexes have no obvious pattern, but the first event is significantly higher than the other events, with a higher fault severity. The subsequent events have a weaker impact on the feeder. For the continuous interference of the same actual fault, the small fluctuations indicate that the overall fault tolerance and recovery ability of the feeder are relatively strong.Figure 9 As shown in (d), the phase angle difference between the fault occurrence time and the peak time decreases to a certain extent in the fifth event, but shows an overall upward trend, indicating that the feeder insulation is more likely to be broken down and there is already a certain degree of deterioration trend. Similar to the first feeder, there is a strong correlation between the trends of the fault duration and the fault energy, further indicating that under the state where the zero-sequence current is small (small current system) and the zero-sequence voltage is relatively stable, the magnitude of the fault energy depends to a large extent on the arc combustion time.

[0096] Analysis of the change of evaluation indexes for fault events in time series:

[0097] As Figures 10 - 12 shown, the cumulative index of the whole sequence will obviously show an upward trend, but its slope can represent the growth rate between events, and can reflect to a certain extent the impact of the fault on the line insulation level.

[0098] As Figure 10 (a) and Figure 10 (b) shown, in the first sequence, the growth trends of the cumulative fault occurrence time and the cumulative energy are relatively similar, and both can be fitted to a power function curve of a cubic four-term polynomial. Moreover, both have an obvious increase in slope after the 4th event, indicating that the severity of the fault has increased significantly.

[0099] As Figure 11 (a) and Figure 11 (b) shown, in the second sequence, the change trends of the cumulative fault time and the cumulative energy are close to linear growth, that is, the growth rate between events is stable and there is no obvious deterioration trend. Therefore, the 16 disturbances recorded within this day in the second sequence are very likely to be multiple continuations of a specific fault.

[0100] As Figure 12 (a) and Figure 12 (b) shown, in the third sequence, the cumulative fault energy can be fitted to a power function curve of a cubic four-term polynomial, while the cumulative fault time is close to linear growth. During the occurrence of the third to the fourth event, the slope of the broken line increases significantly, indicating that the 4th fault event causes a large fluctuation in the stability of the feeder and the fault is more serious.

[0101] In a set of time series data, the time interval between the occurrences of repetitive fault events can indicate the insulation deterioration speed of its feeder. If the time interval gradually shortens, the tolerance of the insulation device to disturbances or faults will decrease significantly, and the insulation level is in the process of accelerating deterioration, and finally leads to a permanent fault. Table 2 shows the time intervals of the events in the first sequence and the third sequence, and their time units are days and seconds respectively. Table 3 shows the event intervals of the events in the second sequence, and its time unit is seconds.

[0102] Table 2: Time interval between the occurrences of the first sequence and the third sequence

[0103]

[0104] Table 3: Time interval between the occurrences of the second sequence

[0105]

[0106] Plot the data of the three groups of sequences respectively as Figure 13 shown in the figure, where Figure 13 (a) is the time interval of the first sequence when a fault occurs, 13(b) is the time interval of the third sequence when a fault occurs, and 13(c) is the time interval of the second sequence when a fault occurs. From Figure 13 (a), it can be seen that the overall time span of the first sequence is relatively long, which is 22 days. The time intervals between its various events do not show a trend of getting shorter and shorter. Instead, there is an obvious increasing trend. And combined with the analysis of other fault indicators of the first sequence, it is found that the fault evolution of this sequence does not have the characteristics of being intensive and multiple. Instead, the fault interval is long, but the severity of the fault increases significantly each time. This kind of fault is not easy to detect because of its long interval, but the energy at the time of occurrence may be very large, so its potential threat is relatively large.

[0107] Both the second sequence and the third sequence are multiple repetitive faults that occur within one day. The change of their time intervals is more in line with the expected gradual shortening. Remove the first interval of the second sequence and the last interval of the third sequence. As Figure 13 (b) and Figure 13 (c) show, among them, the third sequence reaches the shortest time interval value when the 4th event occurs, but then it recovers, indicating that the feeder insulation level still has good fault tolerance ability; in the second sequence, taking the 6th event as the demarcation point, the time intervals before that are still in a fluctuating state, and the time intervals from then to the 9th event are continuously decreasing. Even if there is a slight recovery at the 10th and 11th events, it can be clearly seen that it is lower than the previous time interval level. This shows that after the 6th repetitive fault, the insulation level of the second feeder has further decreased, and it is necessary to remove the potential fault hazards in time to avoid more serious situations.

[0108] Estimation of the development trend of the time series:

[0109] By analyzing the changes in the evaluation indexes of the fault events in the time series, it can be seen that among the data changes of the three groups of series, only the first series has an obvious upward trend, and the possibility of the accelerated deterioration of the feeder insulation level is relatively large. Therefore, for the data of the first series, the historical data of the first 6 events of the series are used for trend estimation, and the estimation result is compared with the parameter level of the actual seventh event to further analyze its fault evolution process.

[0110] Two methods, namely the arithmetic mean method and the analytic hierarchy process (AHP), are used for trend estimation of the data of the first series. The arithmetic mean method takes the arithmetic mean of the historical data as the parameter estimation value for the planning period. The analytic hierarchy process assigns weights according to the distance from the planning period. According to the magnitude of each data item in the first series of single fault events and time series fault events, the judgment matrix is constructed as follows. After calculation, the consistency ratio of this judgment matrix is less than 0.1, meeting the requirements of consistency test.

[0111]

[0112] The data are weighted by the arithmetic mean method and the analytic hierarchy process respectively to estimate the parameter level of the 7th event. The arithmetic mean method is not very applicable to the estimation of the severity of fault development with an obvious change trend. The analytic hierarchy process assigns weights to each event respectively, making the events closer to the estimation date have greater weights, which better reflects the process of fault development. Its estimation result is significantly better than that of the arithmetic mean method and is relatively closer to the actual fault level.

[0113] Regarding the gap between the estimation results of the two methods and the actual level, the error comparison results shown in Table 4 are calculated. It can be clearly seen that the estimation result error under the weighting of the analytic hierarchy process is smaller than that of the arithmetic mean method. Especially for the estimation of fault time and fault energy (including both single events and time series), the relative error of the result obtained by the analytic hierarchy process is more than 20% lower than that of the arithmetic mean method.

[0114] Table 4: Comparison of relative errors of two estimation methods

[0115]

[0116] In summary, through simulation analysis and case verification, it is shown that the evaluation indexes proposed in this application have good adaptability to single-phase grounding faults, and the more the number of fault samples, the higher the accuracy of the proposed evaluation scheme.

[0117] The basic principles, main features and advantages of the present invention have been shown and described above. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A distribution network fault warning method based on repetitive fault identification, characterized in that, It includes the following steps: S1. Detect and collect instantaneous fault data information, and establish an instantaneous fault database containing waveform record files; S2. Integrate the faults that can be self-recovered for insulation according to the database, and extract the instantaneous single-phase grounding fault events therefrom; S3. Extract the single-phase grounding faults occurring on the same feeder and of the same phase; S4. Integrate a repetitive fault sequence library by manually identifying waveform similarities; S5. Obtain the evaluation indexes of single fault events and fault events in time series; The evaluation indexes of single fault events include: Zero-sequence voltage peak value U 0i , where i represents the i-th instantaneous fault: the zero-sequence voltage peak value U 0i The higher it is, the greater the impact of this fault on the original voltage level of the line, the greater the damage to the tolerance of the feeder, and the higher the severity of the fault; Fault energy P i : Fault energy P i The larger the value, the higher the severity of the fault and the greater the impact on the line. The calculation formula for the fault energy P i is as follows: P i = ∫ t i f (t) u f (t) dt; Among them, t is the single fault duration, and i f is the fault current, and u f is the fault voltage; Initial fault phase angle ω 0i : Take the phase angle difference between the initial fault phase angle ω 0i and the phase angle at the nearest adjacent peak as the evaluation index. The larger the phase angle difference, the lower the insulation withstand capacity; The single fault duration t. The longer the single fault duration t is, the lower the ability of the line to self-recover, and the greater the damage to the line insulation; The evaluation indexes of fault events in time series include: The total fault energy P borne by the line within the time series: P = ∑P i ; Use the total fault energy to reflect the cumulative effect of the fault on the line. The degree of line deterioration increases with the increase of the cumulative effect; The total fault time T of the line within the time series: T = ∑t; That is, within the time series, the time when the line is actually in the fault state can represent the severity of the line fault and the recovery ability to self-clear the fault; The total number of faults N of the line within the time series: The more the total number of faults N is, the more obvious the deterioration trend of the line's tolerance to faults; Time interval sequence {a of the occurrence of N faults within the time series n}: {a n} = {t2 - t1, t3 - t2, … t N -t N-1}; If the frequency of faults is getting higher and higher, that is, the time interval between two faults is getting smaller and smaller, it proves that the insulation level of the line has a deterioration trend and the tolerance to faults decreases; S6. Evaluate the severity of the fault, and analyze the fault development trend by using the analytic hierarchy process; S7. Implement early warning according to the evaluation and analysis results of the fault.

2. The method for warning of distribution network faults based on repetitive fault identification according to claim 1, wherein, The extraction of the instantaneous single-phase grounding fault in S2 includes the following steps: Screen according to the characteristics of the reduced voltage of the fault phase, the increased voltage of the non-fault phase, the increased zero-sequence voltage, and the increased zero-sequence current of the fault feeder in the single-phase grounding fault, and screen out the waveform record events of interphase faults, three-phase voltage oscillations, and non-fault disturbances; Regard the disturbance events that have not achieved self-clearing of the fault until the end record moment of the waveform record file as permanent fault events, and screen out this part of permanent fault events.

3. A distribution network fault warning method based on repetitive fault identification according to claim 2, characterized in that, In S3, the polarities of the zero-sequence currents of the single-phase grounding faults on the same feeder and of the same phase are opposite, and the phase voltage of the fault phase is reduced.

Images