JP2009017637A - System, method, and program for inferring cause of distribution line fault - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To grasp the aspects of a grounding fault that vary with time, when the fault occurs in correspondence with individual concrete electrical events that differ according to the causes of faults, and thereby makes it possible to accurately infer the causes of the fault in detail. <P>SOLUTION: A zero-phase current classifying means 12 carries out frequency analysis to grasp temporal changes in zero-phase current, computes the rate of harmonic content of each representative order, each time to classify the waveform of zero-phase current into multiple waveform patterns, and computes the selected time occurrence ratio of each waveform pattern. A coincidence-determining means 13 determines the coincidence between the selected time occurrence ratio of each waveform pattern and the sample data of zero-phase current prepared, beforehand, in correspondence with individual different electrical events with respect to each fault cause. A zero-phase voltage classifying means 14 classifies the waveform of zero-phase voltage, according to the magnitude of zero-phase voltage. A fault cause final determination means 15 estimates the ultimate determined fault cause, based on the result of zero-phase current waveform classification and on the result of the coincidence determination thereof and the result of the zero-phase voltage waveform classification. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a technique for estimating the cause of an accident by analyzing electric quantity data such as a zero-phase current and a zero-phase voltage when a ground fault occurs in a distribution line.

  With the recent progress of information processing technology and communication technology and the improvement of communication environment, it has become possible to easily capture various data of the power distribution system that was difficult to obtain by the optical IP remote control system. For this reason, it is expected that various operations and maintenance operations of the power distribution system will be advanced and efficient. In particular, the use of waveform data (earth fault data) is expected to greatly improve the efficiency of operations that have been conducted with a great deal of labor and time, such as analysis and inspection.

In addition, as a conventional technique for analyzing (determining) the cause of the accident by analyzing the ground fault data of the distribution line, for example, the zero-phase voltage waveform and the zero-phase current waveform at the time of occurrence of the ground fault of the distribution line are detected. A technique has been proposed in which waveform analysis is performed and those analysis values are input to a neural network to specify a failure phase and to estimate and output the cause of an accident (see Patent Document 1). In addition, the transient part of the zero-phase current waveform and zero-phase voltage waveform at the time of the ground fault occurrence is detected, frequency analysis is performed by wavelet transform, the analysis result is displayed as a contour image, and pattern recognition (neural network) A technique for identifying the cause of an accident by repeating the number of times of learning is proposed (see Patent Document 2).
JP-A-11-142466 JP2000-50488

  However, the conventional technique for estimating (determining) the cause of the accident as described above has the following problems.

  First, the technique of Patent Document 1 is a method for analyzing a zero-phase voltage waveform and a zero-phase current waveform at the time of occurrence of a ground fault, but a typical waveform used for comparison for each cause of an accident is a simple pattern. Therefore, the cause of the accident cannot be estimated based on (captured) the accident aspect that changes over time. In other words, the actual zero-phase voltage waveform and zero-phase current waveform that occur at the time of a ground fault do not repeat a simple pattern regularly, but cause large changes in time. Unless time changes are captured, there is a limit to the accuracy of accident cause estimation.

  Moreover, although the technique of patent document 2 captures the temporal change of the transient part of the zero phase current waveform and zero phase voltage waveform at the time of the occurrence of the ground fault, the waveform used for comparison is It is difficult to estimate the cause of the accident based on (capturing) specific electrical events that occur at the time of the ground fault because the waveforms are classified only according to the difference in the ground fault path. . In other words, electrical events that occur in the event of a ground fault include, for example, metal ground faults due to equipment failures and contact with materials, current leakage, discharge, etc. Depending on the specific cause of the accident including, etc., specific individual electrical events will be different. Therefore, there is a limit to the accuracy of the accident cause estimation unless the waveform classification corresponding to such specific individual electrical events is captured.

  The present invention has been proposed in order to solve the above-described problems of the prior art. The purpose of the present invention is to change the aspect of the accident that changes over time when a ground fault occurs, depending on the cause of the accident. A system for estimating the cause of distribution line accidents that enables detailed and highly accurate estimation of accident causes by associating them with individual electrical events, and is effective in improving the efficiency and efficiency of distribution system operation and maintenance operations And its method and program.

  The distribution line accident cause estimation system of the present invention is a distribution line accident cause estimation system that estimates the cause of an accident by analyzing the amount of electricity data at the time of occurrence of a ground fault in the distribution line. Means, zero phase current classification means for classifying the waveform of the zero phase current, coincidence determination means for determining the degree of coincidence between the waveform classification result of the zero phase current and the sample data of the zero phase current prepared in advance, Accident cause final determination means for finally determining the cause of the expected accident based on the waveform classification result of the zero-phase current and the coincidence determination result, and an output means for outputting the final determination result of the accident cause It is said. Here, the zero-phase current classifying means performs frequency analysis that captures temporal changes in the zero-phase current, calculates frequency content means for each representative order at each time, and each time of the zero-phase current. A waveform that classifies the waveform of the zero-phase current into multiple waveform patterns based on the harmonic content for each representative order in, and calculates the appearance ratio of the time when each waveform pattern is selected as the selection time appearance ratio of each waveform pattern It has a classification means. The coincidence determination means is a sample of the zero phase current prepared in advance in association with the selection time appearance ratio of each waveform pattern obtained as a waveform classification result of the zero phase current, and individual electrical events that differ for each cause of the accident. It is configured to determine the degree of coincidence with the data.

  The distribution line accident cause estimation method and the distribution line accident cause estimation program according to the present invention are obtained by grasping the above distribution line accident cause estimation system from the viewpoints of the method and the computer program.

  In the present invention, the frequency analysis of the zero-phase current is performed, the harmonic content rate is obtained for each order, the waveform is classified, and the selection time appearance ratio of each waveform pattern as the classification result matches the sample data for each cause of the accident By determining the degree of accident, the cause of the accident is estimated after grasping the accident aspect that changes in time at the time of the occurrence of the ground fault in association with specific individual electrical events that differ according to the cause of the accident. be able to. Therefore, it is possible to provide a distribution line accident cause estimation system, a method and a program thereof that can estimate the cause of the accident in detail and with high accuracy, and are effective for the advancement and efficiency of the operation and maintenance of the distribution system.

[1. Configuration of Embodiment]
FIG. 1 is a block diagram showing a functional configuration of one embodiment of a distribution line accident cause estimation system according to the present invention together with a target distribution system. As shown in FIG. 1, the distribution system 1 is configured by connecting each switch 4 to the distribution substation 2 via the bus 3, and the distribution substation 2 is connected to the power system 5. It is connected via an electric wire 6.

  Then, among the switches 4 of the distribution system 1, the zero-phase current, the zero-phase voltage, etc. are transmitted from the switch 4 having a function of taking in data of various distribution systems by an optical IP remote control system or the like through the communication line 7. A communication path for taking in various data into the ground fault cause estimation system 10 according to the present invention provided in the distribution system monitoring control system 8 is configured.

  The ground fault accident cause estimation system 10 includes a data acquisition unit 11, a zero phase current classification unit 12, a coincidence degree determination unit 13, a zero phase voltage classification unit 14, an accident cause final determination unit 15, an output unit 16, and a storage unit 17. It has. This ground fault cause estimation system 10 realizes each means 11-17 having the following functions by a computer system comprising a computer such as a personal computer or a workstation, a program for controlling the computer, and peripheral devices.

  The data acquisition means 11 is means for taking in various data such as zero-phase current and zero-phase voltage from the switch 4 of the distribution system 1 via the communication line 7, and is basically an input device or communication control device that the computer has. Etc., or a combination of these hardware and input software or communication software.

  The zero phase current classifying means 12 is a means for classifying the waveform of the zero phase current, and includes a frequency analyzing means 121 and a waveform classifying means 122. Here, the frequency analysis unit 121 is a unit that performs frequency analysis that captures temporal changes in the zero-phase current and calculates the harmonic content for each representative order at each time. Further, the waveform classification unit 122 classifies the waveform of the zero-phase current into a plurality of waveform patterns based on the harmonic content for each representative order at each time of the zero-phase current, and the appearance of the time when each waveform pattern is selected. This is means for calculating the ratio as the selection time appearance ratio of each waveform pattern.

  The degree of coincidence determination means 13 calculates the selection time appearance ratio of each waveform pattern obtained as the waveform classification result of the zero-phase current and the zero-phase current prepared in advance in association with each individual electric event for each cause of the accident. It is a means for determining the degree of coincidence with the sample data, and has a reference parameter calculation means 131. The reference parameter calculation unit 131 calculates a reference parameter for each waveform pattern of the zero phase current for each cause of the accident based on the waveform classification result by the actually measured zero phase current, and calculates the calculated reference parameter for each waveform pattern of the zero phase current. This is means for setting each zero cause as sample data of zero-phase current.

  Accordingly, the coincidence degree determination means 13 selects the appearance time ratio of each waveform pattern obtained as the waveform classification result of the zero phase current and the waveform pattern of the zero phase current preset for each cause of the accident by the reference parameter calculation means 131. The degree of coincidence with each reference parameter is determined.

  The zero-phase voltage classifying unit 14 calculates the maximum value of the amplitude of the zero-phase voltage (that is, the maximum value of the absolute value of the amplitude of the zero-phase voltage), and classifies the waveform of the zero-phase voltage according to the magnitude of the amplitude. Means.

  The accident cause final determination means 15 is a means for estimating a finally assumed accident cause based on the waveform classification result of the zero-phase current, the coincidence determination result, and the waveform classification result of the zero-phase voltage.

  The zero-phase current classifying means 12, the coincidence degree determining means 13, the zero-phase voltage classifying means 14, and the accident cause final determining means 15 as described above include arithmetic processing hardware such as a CPU basically included in the computer, It is realized by a combination of programs specialized for distribution line accident cause estimation.

  The output unit 16 stores various processing results including the final determination result of the accident cause obtained as the processing result in the ground fault accident cause estimation system 10 in the storage unit 17 in the ground fault cause estimation system 10. And a function of outputting to a display device, a printer, a storage device and the like of the distribution system monitoring control system 8.

  The storage means 17 stores in advance data such as distribution line connection configuration information, line impedance, load data, installation positions of each switch as data related to the distribution system 1, and in advance for each cause of the accident by the reference parameter calculation means 131. Sample data of zero phase current prepared in advance in association with individual electrical events that differ for each cause of the accident, such as a reference parameter for each set waveform pattern of zero phase current, is stored. The storage unit 17 also has a function of storing various processing results including the final determination result of the accident cause obtained in the ground fault accident cause estimation system 10. The storage means 16 is realized by various types of memory, auxiliary storage devices, or other various storage media that the computer basically has.

[2. Overview of accident cause estimation algorithm]
The accident cause estimation algorithm by the ground fault cause estimation system 10 of the present embodiment as described above classifies waveforms based on zero phase current (I0 data) and zero phase voltage (V0 data) in ground fault data. To estimate the possible cause of the accident. In addition, the final determination result of the cause of the accident is not limited to a single cause of the accident, but is displayed as a result in which priorities are set for several possible causes of the accident.

  FIG. 2 is a flowchart showing an example of an outline of the accident cause estimation algorithm by the ground fault accident cause estimation system 10 of the present embodiment. As shown in FIG. 2, in Step 10, as data acquisition processing by the data acquisition means 11, measurement data of zero phase current (I0 data) and zero phase voltage (V0 data) at the time of a ground fault is taken.

  In Step 20, as the zero-phase current classification process by the zero-phase current classification means 12, the frequency analysis means 121 performs frequency analysis that captures the temporal change of the zero-phase current, and generates harmonics for each representative order at each time. The content rate is calculated, and the waveform classification unit 122 classifies the waveform of the zero-phase current into a plurality of waveform patterns, and calculates the selection time appearance ratio of each waveform pattern.

  In Step 30, as the coincidence determination process by the coincidence determination means 13, the selection time appearance ratio of each waveform pattern obtained as the waveform classification result of the zero phase current based on the waveform classification result of the zero phase current in Step 20, and the accident A candidate for the cause of the accident is determined by determining the degree of coincidence with the sample data of the zero-phase current prepared in advance in association with individual electrical events that differ for each cause.

  In Step 40, as the zero-phase voltage classification process by the zero-phase voltage classification means 14, the maximum value of the amplitude of the zero-phase voltage is calculated, and the zero-phase voltage is classified according to the magnitude of the amplitude.

  In Step 50, as the accident cause final determination processing by the accident cause final determination means 15, the zero-phase current waveform classification result in Step 20, the zero-phase current coincidence determination result in Step 30, and the zero-phase voltage waveform in Step 40 are based thereon. From the classification results, the possible cause of the accident is finally determined.

  In Step 60, as the output processing by the output means 16, various processing results including the final determination result of the accident cause in Step 50 are stored in the storage means 17 in the ground fault accident cause estimation system 10, and the distribution system monitoring control system 8 display device, printer, storage device and the like.

  Below, the detail of the process of each Step10,20,30,40,50 is demonstrated sequentially.

[3. Data acquisition process]
In the data acquisition process by the data acquisition means 11 of Step 10 shown in FIG. 2, the measurement data is acquired from the switch 4 (FIG. 1) having a function of acquiring data of various distribution systems by an optical IP remote control system or the like. This is realized by fetching various data of the distribution system 1 such as zero-phase current and zero-phase voltage via the communication line 7.

[4. Zero-phase current classification process]
The zero-phase current classification process by the zero-phase current classification means 12 at Step 20 shown in FIG. This is a process of classifying into a plurality of waveform patterns and calculating the selection time appearance ratio of each waveform pattern.

  That is, in the zero-phase current classification process, first, the frequency analysis unit 121 performs frequency analysis that captures the temporal change of the zero-phase current, and calculates the harmonic content for each representative order at each time. Next, the waveform classification means 122 classifies the waveform of the zero-phase current into a plurality of waveform patterns based on the harmonic content for each representative order at each time of the zero-phase current, and the time at which each waveform pattern is selected. The appearance ratio is calculated as the selection time appearance ratio of each waveform pattern. Hereinafter, the zero-phase current classification process will be described in more detail.

[4-1. Classification example of zero-phase current waveform by electrical event]
First, as a result of classifying the waveform of the zero-phase current in the measured ground fault data in association with individual electrical events that differ for each cause of the accident, for example, it can be classified as shown in Table 1 below. confirmed. In the example shown in Table 1, the waveform of the zero-phase current can be classified into eight types of waveforms (I to VIII) when associated with individual electrical events.

[4-2. Outline of Zero Phase Current Classification Algorithm]
3 is a flowchart showing an example of a zero-phase current classification algorithm in the zero-phase current classification process by the zero-phase current classification means 12 in Step 20 shown in FIG. 2, and FIG. 4 shows the zero-phase current classification algorithm of FIG. It is a figure shown by the flow of.

  As shown in FIG. 3 and FIG. 4, in Step 21, the frequency analysis unit 121 performs wavelet conversion of the zero-phase current and performs scaling conversion to extract a waveform for each representative order. Match the coefficient of the waveform to the magnitude of the zero-phase current. In the next Step 22, the frequency analysis means 121 converts the instantaneous value into the signal magnitude (effective value, etc.), and in the subsequent Step 23, converts the waveform of the representative order other than the fundamental wave at the ratio to the fundamental wave. In this case, the harmonic content at each time is calculated.

  Then, in Step 24, the waveform classification means 122 performs a threshold value determination of the harmonic content rate for each representative order at each time. 0 ”. The waveform classification means 122 classifies the waveform into a plurality of waveform patterns using a logical operation expression in the next Step 25, and then, in the subsequent Step 26, calculates the appearance ratio (selection time appearance ratio) at which each waveform pattern is selected. calculate.

[4-3. Example of zero-phase current classification]
Hereinafter, as examples of zero-phase current classification by the zero-phase current classification algorithm shown in FIGS. 3 and 4, “wavelet transform method”, “representative order determination method”, “waveform classification result”, “threshold determination method” , “Selected time appearance ratio calculation method” will be sequentially described.

[4-3-1. Wavelet transform method]
Wavelet analysis is an analysis technique that can simultaneously capture the time and frequency of a signal. Here, the “wavelet” is a unit for cutting out a signal portion in time-frequency analysis in which a certain signal is simultaneously captured from both sides of time and frequency. This is called a mother wavelet and is represented by a function “Ψ”. “Wavelet transform” refers to a process of transforming this “mother wavelet Ψ” with a scale parameter and a translation parameter, and represents the size of each part of the extracted signal.

The scale parameter is used for expanding / contracting the width of the mother wavelet (frequency adjustment), and the translation parameter is used for translating the mother wavelet (time adjustment). This can be expressed by the following equation (1).

There are various functions (Haar, Meyer, Daubechies, Gabor, etc.) as the mother wavelet. In this embodiment, the analysis is performed using the Daubechies function as an example. In this Daubechies function, the natural number N must be used, and the 0th to (N−1) th order moments must be set to zero. Therefore, the following equation (2) is obtained.

[4-3-2. How to determine the representative order]
As a result of wavelet analysis of zero-phase current using some measured ground fault data, the following characteristics (1) to (5) were confirmed as an overall trend.

Characteristic (1): The fundamental wave component is clearly extracted corresponding to the magnitude of the fundamental wave component of the zero-phase current.
Characteristic (2): In the low-order group, the third-order and fifth-order components are relatively clearly extracted. The difference between the 7th order and the 5th order is not clear.
Feature (3): In the intermediate order group, there is not much difference among the 9th, 11th and 13th orders. However, the waveform is clearly different from the low-order group.
Characteristic (4): In the high-order group, there is almost no difference between the 15th, 17th, and 19th orders. However, the waveform is clearly different from the low-order and middle-order groups.
Characteristic (5): In the second order, the fundamental wave and the third order component are mixed. There is no significant difference between the odd-order conversion results before and after other even-orders.

  From the above features (1) to (5), in this embodiment, the representative orders for classification determination are fundamental wave, third order, fifth order (low order), eleventh order (intermediate order), 19th order ( Higher order).

[4-3-3. Waveform classification result]
Step 21 and Step 22 (FIGS. 3 and 4) show the waveforms of the representative orders of the fundamental wave, third order, fifth order (low order), 11th order (intermediate order), and 19th order (high order) determined as described above. In Step 23, the harmonic content at each time is calculated for the waveform other than the fundamental wave in Step 23, the threshold is determined in Step 24, and the waveform is classified into a plurality of waveform patterns in Step 25 using a logical operation expression. Here, the plurality of waveform patterns are a sine wave (large), a sine wave (small), a harmonic, a triangular wave, a needle wave (large), and a needle wave (small).

The following general tendency was confirmed in each waveform pattern as a feature of the waveform classified in this way from the viewpoint of shape.
Sine wave (large): Most of the fundamental wave components are not seen, and no harmonic components are seen (fundamental wave: 10A or more)
Sine wave (small): Most of the fundamental wave component is not seen, and no harmonic component is seen (fundamental wave: less than 10A and 0.5A or more)
Harmonic: Fundamental wave component + Harmonic component in a wide range Triangular wave: Fundamental wave component + low-order harmonic wave component needle wave (large): Fundamental wave component + high-order harmonic wave component needle wave (small): Single component without fundamental component

[4-3-4. Threshold judgment method]
When classifying accident waveforms according to the threshold determination result, the threshold setting method is important. Generally, as described above with respect to Step 24 of the zero-phase current classification algorithm shown in FIG. 3 and FIG. 4, the ratio of the harmonics at each representative order at each time relative to the fundamental wave (harmonic content) The accident waveform can be appropriately classified by setting the threshold and determining the threshold.

[4-3-5. Calculation method of selection time appearance ratio]
In Step 26 of the zero-phase current classification algorithm shown in FIGS. 3 and 4, the following two methods are conceivable as methods for calculating the appearance ratio (selection time appearance ratio) at which each waveform pattern is selected. .

Method (1): A method of calculating the ratio of the time when each waveform pattern is selected with respect to the evaluation period, using the entire detection period as the evaluation period.
Method (2): Cut out the evaluation period from the time when each waveform pattern was first selected within the detection period to the time when each waveform pattern was last selected within the detection period, not the entire time of the detection period, A method of calculating a ratio of time when each waveform pattern is selected with respect to the evaluation period.

  For example, as shown in FIG. 5, the waveform pattern detection period is 2.0 seconds, and the appearance times of the waveform patterns are sine wave (large): 0.8 seconds and sine wave (small): 0, respectively. 0.0 second, harmonic wave: 0.0 second, triangular wave: 0.6 second, needle wave (large): 0.4 second, needle wave (small): 0.0 second. In the following, with respect to the example of FIG. 5, a case where the selection time appearance ratio of each waveform pattern is calculated by the method (1) and the method (2) will be described.

  First, in the case of method (1), in the example of FIG. 5, since the evaluation period is 2.0 seconds with respect to the detection period: 2.0 seconds, the selection time appearance ratio of each waveform pattern is calculated as follows. Is done.

Appearance ratio of sine wave (large) = 0.8 seconds ÷ 2.0 seconds × 100 = 40.0%
Appearance ratio of sine wave (small) = 0.0 seconds / 2.0 seconds × 100 = 00.0%
Harmonic appearance ratio = 0.0 seconds ÷ 2.0 seconds x 100 = 00.0%
Appearance ratio of triangular wave = 0.6 seconds / 2.0 seconds x 100 = 30.0%
Appearance ratio of needle wave (large) = 0.4 seconds ÷ 2.0 seconds × 100 = 20.0%
Appearance ratio of needle wave (small) = 0.0 seconds ÷ 2.0 seconds × 100 = 00.0%

  In the case of method (2), in the example of FIG. 5, since the detection period is 2.0 seconds, the evaluation period is 1.0 second, so the selection time appearance ratio of each waveform pattern is calculated as follows. Is done.

Appearance ratio of sine wave (large) = 0.8 seconds ÷ 1.0 seconds × 100 = 80.0%
Appearance ratio of sine wave (small) = 0.0 seconds ÷ 1.0 seconds × 100 = 00.0%
Harmonic appearance ratio = 0.0 seconds ÷ 1.0 seconds x 100 = 00.0%
Appearance ratio of triangular wave = 0.6 seconds ÷ 1.0 seconds x 100 = 60.0%
Appearance ratio of acicular wave (large) = 0.4 seconds ÷ 1.0 seconds x 100 = 40.0%
Appearance ratio of needle wave (small) = 0.0 seconds ÷ 1.0 seconds x 100 = 00.0%

[5. Match determination process]
In the coincidence determination processing by the coincidence determination means 13 in Step 30 shown in FIG. 2, a basic reference parameter is set in advance for the selection time appearance ratio of each waveform pattern, and using this reference parameter, This is a process of selecting a possible cause of an accident by determining the degree of coincidence (pattern matching) with the result of the selection time appearance ratio of each waveform pattern obtained as a waveform classification result of the zero-phase current.

[5-1. Prior work (setting of standard parameters)]
In this coincidence degree determination process, first, as a preliminary work, the reference parameter calculation means 131 classifies each accident according to the cause of the accident (for example, eight waveforms I to VIII shown in Table 1). The basic reference parameters are calculated and set for each waveform pattern, and stored in the storage means 17. As specific reference parameters, the average of the results of the selection time appearance ratio of each waveform pattern based on the zero-phase current classification results by the zero-phase current classification algorithm shown in FIGS. A value can be used. It is also conceivable to perform parameter tuning (parameter update) by a learning function according to accumulation of actually measured ground fault data.

[5-2. Selection of accident cause candidates by calculating the degree of coincidence with reference parameters]
The degree of coincidence determination means 13 calculates the degree of coincidence with the target zero-phase current using the reference parameters set in advance and stored in the storage means 17, and the cause of the accident that is assumed according to the degree of coincidence. Select candidates. In this case, the degree of coincidence is calculated by calculating the difference between the reference parameters for each waveform set for each cause of the accident and the selection time appearance ratio of each waveform pattern obtained by the zero-phase current classification algorithm shown in FIGS. Is possible. The following formula (3) shows an example of such a coincidence determination formula.

  The degree of coincidence determination means 13 sorts the degrees of coincidence obtained by such a degree of coincidence determination formula in the descending order, so that “first candidate, second candidate, third candidate” , ... "etc.

[6. Zero-phase voltage classification process]
The zero-phase voltage classification process by the zero-phase voltage classification means 14 in Step 40 shown in FIG. 2 is a process for calculating the maximum value of the amplitude of the zero-phase voltage and classifying the zero-phase voltage according to the magnitude of the amplitude as described above. It is. That is, when a ground fault occurs, the zero-phase voltage is difficult to classify characteristically compared to the zero-phase current.

FIG. 6 is a diagram showing the relationship between the zero-phase voltage and the magnitude of the fundamental wave of the zero-phase current plotted according to the cause of the accident. In FIG. 6, the zero-phase voltage and the zero-phase current have a proportional relationship that almost passes through the origin, and can be said to be in a correlation. From this, the cause of the accident can be picked up within a somewhat promising range depending on the magnitude of the zero-phase voltage or the zero-phase current. In FIG. 6, as an example, the magnitude of the zero-phase voltage is classified into four sections of 6 kV or more, 2 kV to 6 kV, 300 V to 2 kV, 300 V or less, and the cause of the accident in each section is shown. Yes. Table 2 below shows possible accident cause candidates in each section.

[7. Accident cause final judgment process]
The accident cause final determination process by the accident cause final determination means 15 of Step 50 shown in FIG. 2 is based on the waveform classification result of the zero-phase current and its coincidence determination result and the waveform classification result of the zero-phase voltage as described above. This is a process of estimating the cause of the accident that is ultimately assumed. In the accident cause final determination process in Step 50, the accident cause final determination means 15 determines the accident cause candidate based on the waveform classification result of the zero-phase current in Step 20 and Step 30 and the coincidence determination result, and then the zero in Step 40 is determined. Based on the waveform classification result of the phase voltage, the possible cause of the accident is finally determined (a method for narrowing the determination condition by screening).

[8. Example of accident cause estimation]
Hereinafter, an example in which the cause of the accident is estimated using the measured zero-phase current and the ground fault data of the zero-phase current shown in FIG. 7 by the accident cause estimation algorithm of the present embodiment will be described.

[8-1. Zero-phase current classification process]
First, as the zero-phase current classification process by the zero-phase current classification means 12 in Step 20 shown in FIG. 2, the following series of processes is performed by the zero-phase current classification algorithm shown in FIGS.

  In Step 21, the frequency analysis unit 121 performs wavelet transform of the zero-phase current in FIG. 7 and performs scaling conversion to obtain the first order, third order, fifth order (low order), and eleventh order (as shown in FIG. 8). Each waveform having a representative order of (intermediate order) and 19th order (high order) is extracted. The frequency analysis means 121 converts the scaled representative order waveform from the instantaneous value to the signal magnitude (effective value, etc.) at Step 22, and then at Step 23, the fundamental wave is applied to the waveform of the representative order other than the fundamental wave. The harmonic content rate at each time is calculated when converted to a ratio with respect to. In Step 23, the waveforms of the representative orders at the time of converting the waveforms of the representative orders other than the fundamental wave by the ratio to the fundamental wave are as shown in FIG.

  In Step 24, the waveform classification unit 122 performs the threshold determination of the harmonic content rate for each waveform at each representative order at each time, and “1” and “ By setting any one of “0”, each waveform for each representative order is converted as shown in FIG. In Step 25, the waveform classification unit 122 classifies the waveform shown in FIG. 10 into a plurality of waveform patterns using a logical operation expression, so that a sine wave (large), a sine wave (small), as shown in FIG. Data indicating temporal appearance states of a plurality of waveform patterns such as a harmonic wave, a triangular wave, a needle wave (large), and a needle wave (small) are obtained.

From the data shown in FIG. 11, the waveform classification unit 122 calculates an appearance ratio (selection time appearance ratio) at which each waveform pattern is selected in Step 26. In the present embodiment, the method (2) described above is used as a method for calculating the selection time appearance ratio, and as a result, each waveform pattern as shown in Table 3 below is obtained from the data shown in FIG. Appearance ratio is obtained.

[8-2. Match determination process]
Subsequently, as the coincidence determination process by the coincidence determination unit 13 in Step 30 shown in FIG. 2, using the above-described coincidence determination formula (3), the selection time appearance ratio of each waveform pattern shown in Table 3 and the cause of the accident The degree of coincidence with the reference parameter for each waveform set every time is calculated. For example, the reference parameter for each waveform in the coincidence determination formula (3) is set as shown in Table 4 below, and the coincidence is calculated using 2.0 as a penalty coefficient, and the coincidence calculation result is large. When the first candidate, the second candidate, and the third candidate are ranked in order, the results shown in Table 5 below are obtained. “I” to “VIII” in Table 5 correspond to the eight waveforms shown in Table 1.

[8-3. Zero-phase voltage classification process]
Subsequently, as the zero-phase voltage classification process by the zero-phase voltage classification means 14 in Step 40 shown in FIG. 2, classification is performed based on the magnitude of the zero-phase voltage. Here, in the measured ground fault data shown in FIG. 7, the maximum value of the zero-phase voltage is about 8700 V. When this zero-phase voltage is classified based on Table 2, it is classified into the section “large V0”. .

[8-4. Accident cause final judgment process]
Finally, as the accident cause final determination process by the accident cause final determination means 15 of Step 50 shown in FIG. 2, as described below, the waveform classification results of Step 20 and Step 30 and the coincidence determination result thereof are used. After determining the candidate for the cause of the accident, the assumed cause of the accident is finally determined based on the waveform classification result of the zero-phase voltage in Step 40 (a method for narrowing the determination condition by screening).

First, regarding the waveform classification result of the zero-phase current and its coincidence determination result, from the coincidence determination result shown in Table 5 and the zero-phase current classification (I to VIII) shown in Table 1, as the accident cause candidates, Such a determination result is obtained.
First candidate: Complete ground fault 99.8%
Second candidate: switching surge, inundation agreement degree 70.5%
3rd candidate: Tree (with discharge), defective coconut match 52.1%

Subsequently, since the waveform classification result of the zero-phase voltage becomes the section “V0 large” as described above, there is a possible cause of the accident from the zero-phase voltage classification result and the zero-phase voltage classification shown in Table 2. As a range, the following determination result is obtained.
Complete ground fault, insulated wire, defective insulator, tree (with discharge)

Finally, from the first to third candidate accident causes obtained as a result of the zero-phase current classification, by selecting the zero-phase voltage with “V0 large”, the possible cause of the accident is finally determined judge. In the case of this example, in Table 2, “switching surge” and “flooding” are excluded from the candidates because the zero-phase voltages are “V0 none” and “V0 medium”. Result can be obtained.
First candidate: Complete ground fault 99.8%
Second candidate: Tree (with discharge), defective coconut match 52.1%

  In the above description, as an example, the zero-phase current classification result up to the third candidate is shown as a candidate, but this is only an example, and in the present invention, the zero-phase current classification result or the final cause of the accident is shown. The range of candidate ranks to be adopted as the determination result can be freely selected. In addition, as a modification, without determining the accident cause candidates, the final determination is made by quantifying the probability of each accident as the result of coincidence calculation and various index values calculated based on it. As a result, it may be adopted.

[9. Modified example of accident cause estimation algorithm]
In the above embodiment, as the accident cause estimation algorithm, as shown in FIG. 2, after performing the zero-phase current classification process and the coincidence degree determination process of Step 20 and Step 30, the zero-phase voltage classification process of Step 40 is performed. Although the case has been described, the accident cause estimation algorithm in the present invention is not limited to this.

For example, as shown in FIG. 12, an algorithm for performing the zero-phase voltage classification process in Step 40 in parallel with the zero-phase current classification process and the coincidence degree determination process in Step 20 and Step 30, and as shown in FIG. On the contrary, an algorithm or the like that performs the zero-phase voltage classification process in Step 40 and then performs the zero-phase current classification process and the matching degree determination process in Step 20 and Step 30 can be considered.
[10. Modified example of accident cause final judgment process]
Further, in the above embodiment, as the accident cause final determination process, after determining the cause of the accident based on the waveform classification result of the zero-phase current and the coincidence determination result, based on the waveform classification result of the zero-phase voltage In the present invention, the case of applying the “method for narrowing down judgment conditions by screening” for finally determining the cause of an accident has been described. Can be applied.

[10-1. Narrowing criteria for screening by another screening method]
As the accident cause final determination process, contrary to the above embodiment, after determining the cause of the accident based on the waveform classification result of the zero phase voltage, based on the waveform classification result of the zero phase current and the coincidence determination result It is also possible to apply a screening condition narrowing method by another screening that finally determines the cause of the assumed accident.

  When this method is applied, only the accident cause final determination process of Step 50 in FIG. 2 or FIG. 12 may be changed as the accident cause estimation algorithm. However, as shown in FIG. After performing the phase voltage classification process, it is conceivable to apply an algorithm for performing the zero-phase current classification process and the matching degree determination process in Step 20 and Step 30.

[10-2. A technique for determining the degree of coincidence for the classification results of both zero-phase current and zero-phase voltage]
As the accident cause final determination process, it is also possible to apply a method of determining the degree of coincidence with respect to the classification results of both the zero-phase current and the zero-phase voltage. Such a technique can be realized, for example, by incorporating a determination formula based on the classification result of the zero phase voltage into the coincidence determination formula (3) based on the classification result of the zero phase current. The following formula (4) shows an example of the coincidence determination formula based on the classification result of both the zero-phase current and the zero-phase voltage.

  In this coincidence determination formula (4), the “reference parameter of the zero phase voltage for each cause of the accident” is “1.0”, and the “classification based on the magnitude of the zero phase voltage” is the zero phase shown in Table 2. It corresponds to each section of the voltage, the corresponding voltage is set as “1.0”, and the non-corresponding voltage is set as “0.0”.

  When this method is applied, only the accident cause final determination process at Step 50 in FIG. 2, FIG. 12 or FIG. 13 may be changed as an accident cause estimation algorithm. 2, 12, and 13, the zero-phase current coincidence determination process at Step 30 is omitted, and after performing the zero-phase current classification process at Step 20 and the zero-phase voltage classification process at Step 40, as Step 50 a, It is conceivable to perform an accident cause final determination process including a coincidence degree determination for both the zero-phase current and the zero-phase voltage classification results. In this case, as the system configuration, for example, as shown in FIG. 17, a configuration using an accident cause final determination unit 15a which also functions as a coincidence determination unit is conceivable.

[10-3. Accident cause estimation method based only on zero-phase current classification results]
As the accident cause final determination process, it is also possible to apply a method of estimating the cause of the accident based only on the zero-phase current classification result without using the zero-phase voltage classification result. Such a technique is realized, for example, by rearranging the accident cause candidates obtained as a result of the coincidence determination formula (1) based on the above-described zero-phase current classification result in descending order of coincidence.

  When this method is applied, as the accident cause estimation algorithm, for example, as shown in FIG. 18, the zero-phase voltage classification process of Step 40 in FIGS. 14 to 16 is omitted, and the zero-phase current classification process of Step 20 is performed. After performing, it is possible to perform the accident cause final determination process by step 50b by the coincidence determination with respect to the zero-phase current classification result. In this case, since the zero-phase voltage is not used, the zero-phase voltage classification means 14 in FIG. 17 can be omitted in the system configuration as shown in FIG.

[11. Effects of the embodiment]
According to the present embodiment as described above, the frequency analysis of the zero-phase current is performed, the harmonic content rate is obtained for each order, the waveform is classified, and the selection time appearance ratio of each waveform pattern as the classification result and each accident cause By determining the degree of coincidence with the sample data, the accident aspect that changes over time at the time of the occurrence of the ground fault accident is considered in association with specific individual electrical events that differ depending on the cause of the accident. The cause of the accident can be estimated.

  Accordingly, it is possible to provide a distribution line accident cause estimation system and method capable of estimating the cause of the accident in detail and with high accuracy, and effective for improving the efficiency and efficiency of the operation and maintenance of the distribution system.

  In particular, in this embodiment, since wavelet transform is used, it is possible to perform highly accurate frequency analysis that sufficiently captures the temporal change of the zero-phase current. In the waveform classification after frequency analysis, the threshold value of the harmonic content at each representative order at each time with respect to the fundamental wave is determined, and the waveform of the zero-phase current is classified by a logical operation based on this threshold determination. Waveform classification can be performed efficiently.

  Also, as the degree of coincidence determination of the waveform classification result of the zero phase current, the selection time appearance ratio of each waveform pattern obtained as the waveform classification result of the zero phase current and the waveform pattern of the zero phase current set in advance for each cause of the accident By calculating the degree of coincidence with each reference parameter, it is possible to efficiently perform quantitative coincidence evaluation for each cause of the accident. And by calculating and setting such reference parameters from measured data, it becomes possible to set statistically more accurate reference parameters according to the amount of measured data, resulting in a more accurate accident. Cause estimation is possible.

[12. Other Embodiments]
It should be noted that the present invention is not limited to the above-described embodiments, and various other variations can be implemented within the scope of the present invention. For example, the system configuration illustrated in the drawings is merely an example, and a specific functional configuration, hardware configuration, software configuration, and the like can be selected as appropriate. A program specialized to realize the function of the distribution line accident cause estimation system by computer hardware is also an aspect of the present invention.

The block diagram which shows the function structure of the distribution line accident cause estimation system which concerns on one Embodiment of this invention. The flowchart which shows an example of the outline | summary of the accident cause estimation algorithm by the ground fault accident cause estimation system of FIG. The flowchart which shows an example of the zero phase current classification algorithm in the zero phase current classification process shown in FIG. The figure which shows the zero phase electric current classification algorithm of FIG. 3 with the flow of data. The figure which shows an example of the detection period and appearance time of each waveform pattern obtained by the zero phase current classification algorithm of FIG. 3 and FIG. The figure which plots the relationship between the magnitude | size of the fundamental wave of a zero phase voltage and a zero phase current according to the cause of an accident in order to demonstrate the zero phase voltage classification | category process shown in FIG. The wave form diagram which shows an example of the ground fault accident data of the measured zero phase current and zero phase current used for accident cause estimation by the accident cause estimation algorithm of FIG. The wave form diagram which shows each waveform of the representative order obtained as a result of performing wavelet transformation of the zero phase current shown in FIG. 7, and performing scaling conversion. FIG. 9 is a waveform diagram showing each waveform of the representative order at the time when each waveform of the representative order shown in FIG. 8 is converted into a signal magnitude and then converted by a ratio to the fundamental wave. The wave form diagram which shows the result of having converted each waveform of the representative order shown in FIG. 9 by the threshold value determination of a harmonic content rate. FIG. 11 is a waveform diagram showing a result of classifying the waveform shown in FIG. 10 by a logical operation expression. The flowchart which shows an example of the outline | summary of the accident cause estimation algorithm different from FIG. 2 as a modification which concerns on this invention. The flowchart which shows an example of the outline | summary of the accident cause estimation algorithm different from FIG. 2 as a modification which concerns on this invention. The flowchart which shows an example of the outline | summary of the accident cause estimation algorithm different from FIG. 2 as a modification which concerns on this invention. The flowchart which shows an example of the outline | summary of the accident cause estimation algorithm different from FIG. 2 as a modification which concerns on this invention. The flowchart which shows an example of the outline | summary of the accident cause estimation algorithm different from FIG. 2 as a modification which concerns on this invention. The block diagram which shows the function structure of the ground fault accident cause estimation system different from FIG. 1 as a modification which concerns on this invention. The flowchart which shows an example of the outline | summary of the accident cause estimation algorithm different from FIG. 2 as a modification which concerns on this invention. The block diagram which shows the function structure of the ground fault accident cause estimation system different from FIG. 1 as a modification which concerns on this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Distribution system 2 ... Distribution substation 3 ... Bus 4 ... Switch 5 ... Electric power system 6 ... Transmission line 7 ... Communication line 8 ... Distribution system monitoring control system 10 ... Distribution line accident cause estimation system 11 ... Data acquisition means 12 ... Zero phase current classifying means 121 ... Frequency analyzing means 122 ... Waveform classifying means 13 ... Matching degree judging means 131 ... Standard parameter calculating means 14 ... Zero phase voltage classifying means 15 ... Accident cause final judging means 16 ... Output means 17 ... Storage means

Claims (13)

In a distribution line accident cause estimation system that analyzes the amount of electricity data at the time of occurrence of a ground fault in the distribution line and estimates the cause of the accident,
Data acquisition means for capturing zero-phase current measurement data;
Zero phase current classification means for classifying waveform of zero phase current;
A degree of coincidence determination means for determining a degree of coincidence between the waveform classification result of the zero phase current and the sample data of the zero phase current prepared in advance;
Based on the waveform classification result of the zero-phase current and the coincidence determination result, an accident cause final determination means for finally determining the assumed cause of the accident,
Having an output means for outputting the final judgment result of the cause of the accident;
The zero-phase current classification means includes
Frequency analysis means that performs frequency analysis that captures temporal changes in zero-phase current and calculates the harmonic content for each representative order at each time point;
Based on the harmonic content of each representative order at each time of the zero phase current, the waveform of the zero phase current is classified into multiple waveform patterns, and the appearance ratio of the time when each waveform pattern is selected is the selection time of each waveform pattern Having waveform classification means for calculating the appearance ratio,
The coincidence degree determination means includes a selection time appearance ratio of each waveform pattern obtained as a waveform classification result of the zero-phase current, and zero-phase currents prepared in advance in association with individual electrical events that are different for each cause of the accident. A distribution line accident cause estimation system configured to determine the degree of coincidence with sample data. The frequency analysis means is configured to perform frequency analysis that captures the temporal change of the zero-phase current by wavelet transform, and to calculate the harmonic content with respect to the fundamental wave with respect to the representative order at each time. The distribution line accident cause estimation system according to claim 1. The waveform classification means is configured to perform a threshold determination of the harmonic content rate for each representative order at each time with respect to the fundamental wave, and classify the waveform of the zero-phase current by a logical operation based on the threshold determination. The distribution line accident cause estimation system according to claim 1. The degree of coincidence determination means is a degree of coincidence between the selection time appearance ratio of each waveform pattern obtained as a waveform classification result of the zero phase current and a reference parameter for each waveform pattern of the zero phase current preset for each cause of the accident. The distribution line accident cause estimation system according to claim 1, wherein the distribution line accident cause estimation system is configured to calculate the distribution line accident. The coincidence degree judging means
Based on the measured waveform classification results based on the zero-phase current, calculate the reference parameter for each waveform pattern of the zero-phase current for each cause of the accident, and use the calculated reference parameter for each zero-phase current waveform pattern as sample data for the zero-phase current. 5. The distribution line accident cause estimation system according to claim 4, further comprising reference parameter calculation means that is set for each cause of the accident. The accident cause final determination means ranks accident causes according to the coincidence determination result of the selection time appearance ratio of each waveform pattern of zero phase current by the coincidence determination means, or the possibility of each accident cause The distribution line accident cause estimation system according to claim 1, wherein the system is configured to quantify the distribution line accident. The zero-phase voltage classification means for calculating the maximum value of the amplitude of the zero-phase voltage and classifying the waveform of the zero-phase voltage according to the magnitude of the amplitude,
The data acquisition means is configured to capture measurement data of zero phase current and zero phase voltage,
The accident cause final determination means is configured to finally determine an assumed accident cause based on the waveform classification result of the zero-phase current, the coincidence determination result thereof, and the waveform classification result of the zero-phase voltage. The distribution line accident cause estimation system according to claim 1, wherein The coincidence determination means is configured to determine the coincidence between the waveform classification result of the zero-phase current and the zero-phase voltage and the sample data of the prepared zero-phase current and the zero-phase voltage,
The accident cause final determination means is configured to finally determine an assumed cause of an accident based on a coincidence determination result of waveform classification results of a zero-phase current and a zero-phase voltage. 7. Distribution line accident cause estimation system according to 7. The accident cause final determination means determines an accident cause candidate based on the waveform classification result of the zero phase voltage, and then determines the assumed accident cause based on the waveform classification result of the zero phase current and the coincidence determination result. The distribution line accident cause estimation system according to claim 7, wherein the system is configured to finally determine. The accident cause final determination means, after determining the accident cause candidate based on the waveform classification result of the zero-phase current and the coincidence determination result, and based on the waveform classification result of the zero-phase voltage, The distribution line accident cause estimation system according to claim 7, wherein the system is configured to finally determine. 2. The accident cause final determination means is configured to finally determine an assumed accident cause based on a waveform classification result of a zero-phase current and a coincidence determination result thereof. Distribution line accident cause estimation system. In the distribution line accident cause estimation method to analyze the amount of electricity data at the time of ground fault occurrence in the distribution line and estimate the cause of the accident,
Using data acquisition means, zero-phase current classification means, coincidence determination means, accident cause final determination means, and output means,
A data acquisition process for acquiring measurement data of zero-phase current by the data acquisition means;
Zero phase current classification means for classifying the waveform of the zero phase current by the zero phase current classification means;
A coincidence degree determination process for determining the degree of coincidence between the waveform classification result of the zero-phase current and the sample data of the zero-phase current prepared in advance by the coincidence degree determining unit,
Accident cause final determination processing for finally determining the assumed cause of the accident based on the waveform classification result of the zero phase current and the coincidence determination result by the accident cause final determination means,
The output means has an output process for outputting the final determination result of the cause of the accident,
The zero-phase current classification process is:
A frequency analysis that captures the temporal change of the zero-phase current, and a process for calculating the harmonic content for each representative order at each time,
Based on the harmonic content of each representative order at each time of the zero phase current, the waveform of the zero phase current is classified into multiple waveform patterns, and the appearance ratio of the time when each waveform pattern is selected is the selection time of each waveform pattern Has a process to calculate the appearance ratio,
The degree of coincidence determination process includes the selection time appearance ratio of each waveform pattern obtained as a waveform classification result of the zero-phase current, and the zero-phase current prepared in advance in association with each individual electrical event for each cause of the accident. A distribution line accident cause estimation method comprising a process of determining a degree of coincidence with sample data. In the distribution line accident cause estimation program that analyzes the amount of electricity data at the time of occurrence of a ground fault in the distribution line and estimates the cause of the accident,
On the computer,
A data acquisition function that captures zero-phase current measurement data,
Zero-phase current classification function to classify waveform of zero-phase current,
A degree of coincidence determination function for determining the degree of coincidence between the waveform classification result of the zero phase current and the sample data of the zero phase current prepared in advance;
Based on the waveform classification result of the zero-phase current and the coincidence determination result, the accident cause final determination function that finally determines the assumed cause of the accident,
An output function that outputs the final judgment result of the cause of the accident is realized.
The zero-phase current classification function is
Performs frequency analysis that captures temporal changes in zero-phase current, calculates the harmonic content for each representative order at each time, and
Based on the harmonic content of each representative order at each time of the zero phase current, the waveform of the zero phase current is classified into multiple waveform patterns, and the appearance ratio of the time when each waveform pattern is selected is the selection time of each waveform pattern Has a function to calculate the appearance ratio,
The coincidence determination function is a function of selecting a zero-phase current prepared in advance in association with a selection time appearance ratio of each waveform pattern obtained as a waveform classification result of the zero-phase current and an individual electrical event that is different for each cause of the accident. A distribution line accident cause estimation program characterized by having a function of determining the degree of coincidence with sample data.

Images