JP5401503B2 - Power system fault waveform data search device and recording medium - Google Patents

Description

  The present invention relates to, for example, a power system fault waveform database obtained by collecting voltage / current waveform data of a power system, a power system fault waveform data search device using the fault waveform database, and the like. A recording device (general recording device or automatic oscillograph device or simply oscilloscope) installed in a power transmission and distribution system that measures the voltage and current of electric wires, buses, and transformers and records the voltage waveform and current waveform when an electrical accident occurs It relates to the transmission / distribution system fault waveform database system that collects the recorded voltage / current waveform data.

  The transmission / distribution system is equipped with a recording device (general recording device, automatic oscillograph device, or simply oscilloscope) that collects the waveform of the accident at the time of the power system failure as waveform data. Data collection location name such as data and substation name, date and time of data collection, line name such as which transmission / distribution line data, distinction between voltage waveform and current waveform for each channel data, data sampling frequency, data The number of samples, the VT ratio of each line voltage, the CT ratio of each line current, etc. are recorded and transmitted to the server device via the network line (these data are hereinafter referred to as “record data” or simply “data ").

  On the other hand, the person in charge in the transmission / distribution line maintenance department looks at the above “recorded data” to see what kind of accident (earth fault or short circuit), what phase the accident is, and how long the accident lasts. After confirming whether the accident was successful or where the accident was, we went to the site where the accident occurred and went to inspection.

  Power system accidents can occur for various reasons such as lightning strikes, contact with other objects, dielectric breakdown due to deterioration, snow damage, equipment failures due to earthquakes, storms and floods. The voltage and current waveforms at the time of the accident vary depending on the system conditions such as the cause of the accident, the voltage class, and the neutral point grounding method. The skilled engineer traces past memories and estimates the cause of the accident and considers countermeasures. It is usual to go to patrol.

  On the other hand, in the field of load equipment to which power is supplied from the upper power system, the harmonic component analysis results of the voltage and current waveforms of the power supply supplied to the load facility are stored in a database and similar harmonic components appear. In such a case, a method has been proposed in which an accident sign is detected in such a case so that it can be quickly dealt with (see, for example, Patent Document 1).

  Separately from this, in the field of rotating machine systems, there has been proposed a method for storing shaft vibration waveform data of a rotating machine and creating a database to quickly detect an accident sign (for example, see Patent Document 2).

JP 2010-29003 A Japanese Patent No. 4373350

  However, the work of estimating the cause of the accident in the conventional power system and finding the countermeasures can be performed only by skilled engineers. When searching for necessary data from the past accident data, the search work is work. Even if a skilled engineer does it, it takes a lot of time to manually search the corresponding data on the server, determine the type of the accident waveform, and estimate the cause of the accident. It had the subject that it took.

  Therefore, if it is attempted to automate the work without relying on the skill of the worker as described above, data can be collected even with existing technology. If existing database software is used, date and data collection are possible. Although it is easy to search using information such as location as a key (cue), in order to search data with ambiguous requirements such as waveform similarity, it is digitized in some form and the range that the numerical value can take is the key It is necessary to search as (cue).

  Therefore, these cannot be realized only by a general-purpose database system.

  In addition, in the method of calculating harmonic components described in Patent Document 1, fast Fourier transform processing is used, which corresponds only to a failure phenomenon of a specific load facility, and this is assumed to be an accident in the transmission and distribution system. If it is going to be applied to the cause estimation work, the amount of data to be processed is much larger and cannot be used as it is.

  In addition, the method described in Patent Document 2 discloses a technique for reliably detecting an accidental abnormal vibration phenomenon at the initial stage of occurrence of an abnormal shaft vibration of a rotating machine even if the vibration amplitude is small. In a power transmission / distribution system accident that is rich in nature, this cannot be directly applied to the estimation of the cause of the accident.

  An object of the present invention is to provide a power system fault waveform data search device and a recording medium capable of quickly searching for fault data in consideration of the above-mentioned problems in such a conventional power system fault.

  The invention of the present application is “information on physical conditions of transmission / distribution system” such as information on installation location, voltage class, neutral grounding system, system capacity, line length, etc. from accident waveform data collected in transmission / distribution system, data “Information on environmental conditions at the time of data recording” such as recording date and time, weather conditions, etc., and whether the accident is a ground fault or short circuit accident, whether the accident is a 1-wire, 2-wire, or 3-wire (hereinafter referred to as the accident aspect), “Accident waveform analysis results” such as voltage values, current values, and waveform graphic analysis results at the time of an accident are stored as a database together with waveform data, and past similar accidents based on those parameters Search the waveform, classify and organize “information that can be obtained after an accident” such as accident cause and recovery method recorded together, search and display them, and infer the cause of the accident from new accident data. It is possible to understand the contents of accident response and its results from past similar accident cases, and analyze newly analyzed inexperienced accident data and quantify similar results. By making it possible to search waveform data quickly, we will build a system that can be used for accident response.

  The first essential point of the present invention for solving the problem is that an accident in which an accident waveform in the electric power system is classified as a search condition based on the classification of the accident waveform data for each accident based on a predetermined rule. Is distinguished based on the physical properties of the power system (for example, the difference in the short-circuit capacity of the power system and the difference in the neutral grounding method) corresponding to the phenomenon (for example, whether or not a short-circuit current component is included), A data storage unit that is stored, and at least the classification result of the accident waveform data stored in the data storage unit as a search target, those that match the classification result of the accident waveform data specified as a search condition When searching, according to the specified accident phenomenon, the search target is narrowed down using the distinction based on physical properties (for example, if the fault is a short-circuit current, a short circuit occurs) Searching for systems with almost the same amount, and for example, if there is an accident in which no short-circuit current flows, the neutral point grounding system searches for the same system as the target, and the search unit is provided. Is a point.

  The inventor of the present application classifies the data in advance according to the short-circuit capacity of the system, which is a physical property of the power system, or the neutral point grounding method, reduces the number of data to be searched at once, and then performs the instantaneous change amount. We noticed that the method of examining waveform data based on the value change and low-frequency fluctuation and extracting the characteristics of the waveform is more effective than FFT and harmonic analysis for all data. He devised a method for analyzing the characteristics of individual waveforms.

The second essential point of the present invention is that the sample value of the waveform data in the graphical feature analysis of the waveform, the difference for each sample calculated from the waveform data (hereinafter referred to as instantaneous change), and the moving average of the data for one cycle of AC Change in effective value calculated from statistical analysis value (hereinafter referred to as effective value change) and change in offset amount calculated from data for one period of AC (hereinafter referred to as DC component or low frequency fluctuation component of waveform) From the information, what characteristic such as pulse waveform, rectangular wave, sawtooth wave, half-wave rectified waveform, etc. from the judgment result of whether or not they have changed beyond the threshold set separately within a specific timing range This is because it is judged whether or not a proper waveform is included.

  For the feature analysis of the waveform, Fourier analysis (FFT) is generally used as disclosed in Patent Document 2. However, in most cases, electrical accidents such as ground faults and short circuits are accompanied by sudden changes in voltage and current waveforms, and the duration is usually on the order of several commercial frequency cycles (about 100 ms). Since it is recorded from several hundred ms before the accident to several hundred ms after the accident, it is useless to perform FFT on the entire data. Also, the FFT result and the waveform shape do not necessarily have a fixed relationship, and it is difficult to distinguish, for example, a rectangular wave and a mere distortion wave from only the FFT result.

  An apparatus according to the invention of the present application (hereinafter referred to as an apparatus of the present application) includes a data storage unit for storing collected data, a feature extraction unit for extracting features of the stored data, and a data classification unit for classifying data based on the extracted features A search unit that searches the classified data, a terminal device that inputs search conditions, and a display device that displays the search results.

  The data storage unit of the device of the present application includes information such as installation location information, voltage class, neutral grounding method, short circuit capacity of the system, line length, etc. recorded along with the collected accident waveform data. Information on the physical conditions of the data ”,“ Information on the environmental conditions at the time of data recording ”such as the data recording date and weather conditions, and the analysis of the accident aspect, voltage value, current value at the time of the accident, graphic characteristics of the waveform, etc. “Information obtained by analyzing the accident waveform” and “information that can be obtained by investigating after the accident” such as the accident situation, the cause of the accident, and the recovery method are stored for each accident waveform data.

  In addition, the “information that can be obtained after investigation after the accident” further includes at least one of the section where the accident occurred, the section of the power outage due to the accident, the success or failure of the reclosing by the protection relay due to the accident, and the presence or absence of the progress accident. An included configuration may be used.

  Here, the section of the power failure due to the accident means a section where the circuit breaker is opened and no voltage is generated due to the accident.

  In addition, the success or failure of reclosing by a protection relay due to an accident will be described below. An accident in the power transmission system means that the circuit breaker of the substation was opened during power transmission and a power outage section occurred. The accident included only arc discharge due to lightning strikes, etc., in addition to accidents involving damage to electric works requiring repair. There is also an accident. In the case of an accident caused by air breakdown such as a lightning strike, the accident can be eliminated by opening the circuit breaker in the substation, and repair is often unnecessary. Even in such a case, when the protective relay is activated and the circuit breaker is opened, the power is automatically turned on again after a predetermined time. This is called reclosing. At this time, if the accident has already been removed, the power transmission is resumed as it is, but this is called reclosing success, and when the circuit breaker is opened again due to overcurrent or the like, it is called reclosing failure. In the case of overhead tracks, about 80% are lightning strike accidents, and reclosing is often successful, but in rare cases there are accidents that require repairs such as contact with other objects. In this case, reclosing fails. In other words, since the case of reclosing failure is a real accident, the record of the result of success or failure of reclosing becomes important data for recognizing the extent of the accident.

  In addition, the meaning of “the presence or absence of progress accident” will be explained. For example, as a result of a one-line ground fault, if the ground voltage of the accident phase is close to zero V and the ground voltage of the healthy phase rises to √3 times, dielectric breakdown also occurs in that phase, and a two-wire ground fault can occur. is there. Of course, there can be a three-wire ground fault. A case where a one-line ground fault progresses to a two-line ground fault or a three-line ground fault is called a progress accident. The investigation of the cause of the accident is important as to what was the first accident, but the damage assumption of the degree of accident damage requires information analysis such as how much the short-circuit current was in the final state. Whether it is a complex accident including one aspect of accident or a simple accident with only one aspect is important information in accident analysis.

  Each time new data is added to the database, the feature extraction unit of the device of the present application converts the instantaneous waveform data from the waveform into a pulse waveform component, sawtooth component, rectangular wave component, harmonic component, low frequency vibration component, Determine whether a characteristic waveform component such as a half-wave rectified waveform component is included, and whether the accident is a one-wire ground fault, a two-wire ground fault, a three-wire ground fault, a two-wire short-circuit, or a three-wire short-circuit Determine the accident aspect.

  The data classification unit of the device of the present application classifies the data based on the already stored data and the feature data determined by the feature extraction unit, and performs data selection and rearrangement.

  The data search unit of the device of the present application automatically collects new accident waveform data, or instructs the terminal device to instruct the above-mentioned “information on physical conditions of the transmission / distribution system” and “ Search past data based on "information on environmental conditions" and "information obtained by analyzing the accident waveform" above, sort the data in descending order of the number of matching information elements, and attach each The above-mentioned “information that can be investigated and acquired after the accident” is output to the display device.

  According to the above configuration of the device of the present application, by analyzing and organizing the accident waveform of the complicated and diverse power transmission and distribution system and the data recorded along with the accident waveform, the power transmission and distribution A database system that can be used immediately even by inexperienced people in systematic accident analysis and can provide information that can be handled by accidents can be constructed.

  In addition, according to the above configuration, it is possible to quickly search for accident data necessary for clues such as voltage / current waveforms and relay response characteristics from a large amount of accident data. A database system that can easily search and display similar accident data can be provided.

  Thereby, the estimation work of the cause of the accident can be automated. That is, for example, a database is constructed from recorded data stored in the server device, the data stored in the database is searched, compared, and classified, and past accident waveform data is searched and the accident situation at that time , Accident causes, accident responses and their results can be easily extracted.

  According to the present invention, it is possible to provide a power system fault waveform data search device, a power system fault waveform data search method, and the like that can quickly search for accident data.

The block diagram for demonstrating the structure of the electric power system accident waveform data search device of one embodiment of this invention The figure which shows the example of a typical accident waveform (waveform pattern) classified by this Embodiment Flow chart for explaining waveform determination processing of the present embodiment The conceptual diagram which shows one construction example of the database in the data storage part of the accident waveform data in which the determination of the waveform pattern of this Embodiment was implemented An example of a sawtooth waveform stored in the data storage unit of the present embodiment Schematic diagram of a sawtooth wave for explaining a method for judging a sawtooth wave according to the present embodiment. Flow chart for explaining the sawtooth determination method of the present embodiment The figure which shows the waveform of the sawtooth wave of this Embodiment, and the difference waveform which plotted the difference for every sample on the same graph The figure which shows the normal waveform (sine wave) in this Embodiment, and the difference waveform which plotted the difference for every sample on the same graph One waveform example of a single pulse wave stored in the data storage unit of the present embodiment (A): One waveform example of a single pulse wave for explaining the single pulse wave determination method of the present embodiment, (b): differential data of the single pulse wave shown in FIG. 11 (a) and the absolute value of the difference Figure showing Flow chart for explaining a method of determining a single pulse wave of the present embodiment The figure which shows the waveform of the single pulse wave of this Embodiment, and the difference waveform which plotted the difference for every sample on the same graph One waveform example of a periodic pulse wave stored in the data storage unit of the present embodiment Schematic diagram of periodic pulse wave for explaining the method for determining periodic pulse wave of the present embodiment Flow chart for explaining the periodic pulse wave determination method of the present embodiment (A): One waveform example of intermittent oscillation harmonics for explaining the determination method of intermittent oscillation harmonics of this embodiment, (b): Difference data and difference of intermittent oscillation harmonics shown in FIG. Figure showing the absolute value of Flow chart for explaining the method for determining an intermittent oscillation harmonic according to the present embodiment One waveform example of partial harmonics stored in the data storage unit of the present embodiment Schematic diagram of partial harmonics for explaining the method of determining partial harmonics of the present embodiment Flow chart for explaining the method of determining partial harmonics of the present embodiment The figure which shows the waveform of the partial harmonic of this Embodiment, and the difference waveform which plotted the difference (not absolute value) for every sample on the same graph (A) to (d): Waveform examples of distortion waves stored in the data storage unit of the present embodiment Schematic diagram of a distorted wave for explaining a distorted wave determination method of the present embodiment Flow chart for explaining the distortion wave determination method of the present embodiment Waveform example of a low frequency component superimposed waveform stored in the data storage unit of the present embodiment Flow chart for explaining the determination method of the low frequency component superimposed waveform of the present embodiment Waveform example of half-wave rectified waveform stored in data storage unit of this embodiment Flow chart for explaining a method of determining a half-wave rectified waveform of the present embodiment (A): A diagram showing a waveform example of a half-wave rectified waveform for explaining the method for determining a half-wave rectified waveform of the present embodiment and a differential waveform in which the difference values are plotted, (b): FIG. The figure which plotted effective value / √2 and correlation value calculated from the differential waveform shown in (a) Flow chart for explaining a method for determining a half-wave voltage missing waveform according to the present embodiment Waveform example of waveform with minute Vo / Io before accident stored in data storage unit of embodiment Schematic diagram of a periodic pulse wave for explaining a method of determining a waveform with minute Vo / Io before an accident according to the present embodiment Flow chart for explaining a method of determining a waveform with a minute Vo / Io before an accident according to the present embodiment The figure which shows an example of the search screen of the display part (illustration omitted) of the electric power system accident waveform data search device of this Embodiment. (A), (b): The figure which shows an example of the search item displayed on the display part of the electric power system fault waveform data search apparatus of this Embodiment. The figure which shows an example of the search screen of the display part (illustration omitted) of the electric power system accident waveform data search device of this Embodiment. Accident waveform example at a recent substation that was searched by the power system fault waveform data search device of this embodiment Past accident waveform example 1 retrieved and extracted as being similar to the waveform example of FIG. 38 in the power system accident waveform data search device of the present embodiment (cause of accident is contact of birds and beasts) Past accident waveform example 2 retrieved and extracted as being similar to the waveform example of FIG. 38 in the power system accident waveform data retrieval apparatus of the present embodiment (cause of accident is bird contact) The past accident waveform example 3 retrieved and extracted as similar to the waveform example of FIG. 38 in the power system accident waveform data search device of the present embodiment (cause of an accident is contact of birds and beasts) The figure which shows the example of the accident waveform put on the search in the electric power system accident waveform data search device of this Embodiment

  Embodiments of the present invention will be described in detail below.

  First, the configuration of a power system fault waveform data search device as an embodiment of the power system fault waveform data search device of the present invention will be described with reference to the drawings.

  The power system fault waveform data search device of the present embodiment is also an embodiment of the power system fault waveform database construction device of the present invention.

  FIG. 1 is a block diagram for explaining a configuration of a power system fault waveform data search device according to the present embodiment.

  As shown in FIG. 1, the waveform recording devices 10a, 10b, and 10c, for example, in a power system such as a substation, a series of voltage / current waveform data from the state before the occurrence of the system fault to the end of the system fault. This is an automatic recording device (general recording device or automatic oscillograph device or simply oscilloscope), and is connected to a predetermined network 1 (11).

  The power system fault waveform data search device 20 classifies, for example, past accident waveform data in the power system according to a predetermined rule described later, and constructs various related data including the classification result as a database. Is a device that searches and extracts data that matches at least the waveform data of past accidents or new accident waveform data, and is connected to the network 1. Yes.

  Here, the configuration of the power system fault waveform data search device 20 will be further described.

  The power system fault waveform data search device 20 extracts data characteristics from the data storage unit 21 for storing unclassified data 21a, classified data 21b, and search result data 21c, and unclassified data 21a. Using the feature extraction unit 22 that outputs to the unit 23, the data classification unit 23 that classifies data according to the characteristics of the data extracted by the feature extraction unit 22 and sends the data to the classified data 21b, and the classified data 21b A data search unit 24 that performs data search and sends the search result to the search result data 21c is provided. The data classification unit 23 includes a waveform classification program 23a, and the data search unit 24 includes a data search program 24a.

  The power system fault waveform data device 20 receives oscilloscope data transmitted from a data server (not shown) connected to the network 1 (11) according to a prescribed protocol, and waveform data and other related data. Is stored in the data storage unit 21 as unclassified data 21a.

  When constructing a database of classified data 21b, the data classification unit 23 classifies unclassified accident waveform data sent from the feature extraction unit 22 using the waveform classification program 23a, and classifies the classification. A means for storing the result in the data storage unit 21 as classified data 21b.

  In addition, when performing data search, the data search unit 24 uses the data search program 24a as a search target for accident data already stored in the data storage unit 21 as the classified data 21b. It is a means for efficiently extracting accident data that matches a predetermined condition and storing the extraction result in the data storage unit 21 as search result data 21c.

  In addition, the Web application 26 searches the database of the classified data 21b stored in the data storage unit 21 in response to a request such as data search by the terminal devices 13a and 13b connected to the network 2 (12). The data search unit 24 performs a search process, and the extracted search result data 25 is returned to the terminal devices 13a and 13b.

  Under the above configuration, the operation of the power system fault waveform data search device 20 will be described next with reference to the drawings, and an embodiment of the power system fault waveform data search method of the present invention will be described at the same time. .

  Here, first, (1) a scene for constructing a power system fault waveform database will be described, and then (2) a scene for utilizing the constructed database will be described.

(1) Database construction:
1-1) First, a user of the power system fault waveform data search device 20 is connected to the device 20 while viewing a screen (not shown) of a display unit provided in the power system fault waveform data search device 20. By operating an input terminal unit (not shown) such as a keyboard or a mouse, a construction of a new database based on all past accident data already collected by the waveform recording devices 10a to 10c is instructed.

  In accordance with the above instruction from the user, the power system fault waveform data search device 20 is connected to the network 1 (11) via a data server (not shown) connected to the network 1 (11). Receive "various data about accidents" collected for each accident by the recording devices 10a to 10c, and "information obtained after investigation after accident" such as the cause and recovery method of accidents recorded for each accident. The waveform data and other data included in the received data are stored in the data storage unit 21 as unclassified data 21a.

  Here, as various data relating to the accident, for example, “information on physical conditions of the transmission / distribution system” such as information on the installation location of the waveform recording device 10, voltage class, neutral point grounding system, system capacity, and line length, etc. "Information on environmental conditions at the time of data recording" such as the date and time of data recording, and weather conditions, as well as aspects of accidents such as ground faults or short-circuit accidents, whether the accident is a 1-wire, 2-wire, or 3-wire Waveform data including the voltage value and current value at the time of the accident.

  1-2) When the reception and storage of accident data and the like are completed, next, the feature extraction unit 22 extracts features from the data obtained from the unclassified data 21a, and the data classification unit 23 loads the waveform classification program 23a. Using each waveform data sent from the feature extraction unit 22, a classification process for determining which waveform pattern belongs is performed by paying attention to the graphical feature of the waveform. As shown in FIG. 2, typical waveform patterns according to the above classification of the present embodiment are sawtooth waves, single-shot pulses, periodic pulses, intermittent oscillation harmonics, partial harmonics, distortion waveforms, and low-frequency components. There are 10 types of waveforms with superposition, half-wave rectification waveform, missing half-wave voltage, and waveform with minute Vo (zero-phase voltage) and Io (zero-phase current) before the accident.

  Here, FIG. 2 is a diagram showing a typical accident waveform example (waveform pattern) classified according to the present embodiment.

  The analysis of the graphic characteristics of the waveform cannot be realized only with ordinary database system software. Separately, a graphic recognition method is examined, and application software (waveform classification program 23a) that operates in conjunction with the database software is examined accordingly. Need to create.

  Therefore, in the waveform classification program 23a of the present embodiment, as shown in FIG. 3, in the graphical feature analysis of the waveform, a difference for each sample (hereinafter also referred to as an instantaneous change) calculated from the waveform data is obtained. A first waveform determination process (step S101) focused on and a second waveform determination process focused on a change in offset amount calculated from data for one period of alternating current (hereinafter also referred to as a waveform low frequency variation) ( Step S102) and an AC constant period (in FIG. 29 and FIG. 31, the effective value and correlation value for 1/2 period are calculated, and in FIG. 34, the effective value for 1/4 period is calculated). A third waveform determination process (step S103) focusing on the change in the effective value calculated from the moving average analysis value of the data (hereinafter also referred to as the effective value change) is performed for each waveform data, As shown in FIG. The determination result (shown as “analysis result of waveform graphic feature” in FIG. 4) is associated with other data (also simply referred to as other data) included in one accident data. It records as the classified data 21b in the storage unit 21 (step S104).

  The above steps S101 to S104 are repeated until the waveform determination processing for all the accident waveform data is completed (step S105). When the waveform determination process for all the accident waveform data is completed, the process proceeds to the next processing operation.

  Here, FIG. 3 is a flowchart for explaining the waveform determination processing of the present embodiment. The processing contents at each step shown in FIG. 3 will be further described later.

  FIG. 4 is a conceptual diagram showing a construction example of a database stored as classified data 21b of accident waveform data in which waveform pattern determination is performed on past accident data.

  By the way, one of the major features of the present embodiment (first feature) is that, in the database construction process and the data search process described later, the power grid fault waveform recognition software (the waveform pattern described above) is used as the database software. Software for the determination process) was created and linked based on empirical knowledge analyzing a large number of accident waveforms.

  On the other hand, conventionally, Fourier analysis (FFT) is generally used for the feature analysis of the waveform as disclosed in Patent Document 1 described above. However, when the number of data reaches 10,000 to tens of thousands as in the present embodiment, there is a drawback that it takes too much time to perform FFT on every time section individually. Also, the FFT result and the waveform shape do not necessarily have a fixed relationship, and it is difficult to distinguish, for example, a rectangular wave and a mere distortion wave from only the FFT result.

  Furthermore, another feature (second feature) is that in order to reduce useless searches when the number of data increases to several tens of thousands, in the construction of a database, accident waveform data is converted into physical characteristics of the power system. It is classified into a plurality of types in advance by the difference in the short-circuit capacity of each system and the difference in the neutral point grounding method. That is, as shown in FIG. 4, the accident waveform database as the classified data 21b constructed in the data storage unit 21 of the present embodiment has the short-circuit capacity of each system (1000 MVA or more, less than 1000 MVA to 100 MVA or more, less than 100 MVA). -10 MVA, less than 10 MVA), and classified in advance according to the difference in neutral point grounding method (direct grounding method, resistance grounding method, non-grounding method, etc.). In this classification, the phenomenon at the time of a short-circuit fault of each system (including ground faults of two or more wires through which a short-circuit current flows) depends on the short-circuit capacity (also referred to as the back impedance) of each system, and each system It is noted that the phenomenon at the time of the ground fault changes depending on the ground resistance value of each system.

  That is, when searching for similar accident waveform data, it is a waste of time to search for all the data stored in the database. This classification is effective for shortening the search processing time to be described later.

  In this way, the inventor of the present application classifies the data in advance according to the difference in the short-circuit capacity of the system, which is the physical property of the power system, and the difference in the neutral point grounding method, for example, For newly generated accident waveform data, when extracting data that matches the past data stored in the database, we considered a configuration to process as follows.

  That is, first, with respect to new accident waveform data, an instantaneous change (first waveform determination process), an effective value change (third waveform determination process), and a low-frequency fluctuation (second waveform determination). The waveform pattern is determined by examining the waveform data based on the processing), and if the short-circuit current component is flowing in the new accident, the data of the same classification is searched focusing on the difference in the short-circuit capacity of the system In the case of only ground fault current, we considered a configuration in which data of the same classification is targeted for retrieval, focusing on the difference in the neutral point system of the system.

  For example, the AB phase two-wire ground fault caused by lightning strikes an arc flash between both the A phase and the B phase with a steel tower, etc., but a large short-circuit current flows between the AB phases in spite of the ground fault. The waveform is retrieved from all the accident waveform data stored in the database as the search target for the fault waveform data of the system in the range where the short circuit capacity of the system is similar. In this way, even if the aspect of the accident is classified as a ground fault, the search target is determined according to the difference in the physical properties of each power system depending on whether or not a short-circuit current component is included. Yes. In addition to the two-wire ground fault accident, the three-wire ground fault accident, the two-wire ground fault accident, or the two-wire ground fault accident developed from the one-line ground fault accident is also a ground fault in terms of classification. Current flows.

  As a result, the number of data to be searched in the database can be greatly reduced. And according to such a configuration, we noticed that it is more efficient than conventional FFT and harmonic analysis in waveform feature analysis as in the past, so we built a method to individually analyze waveform features is there. This is due to the characteristics of electrical accidents in which, in most cases, electrical accidents such as ground faults and short circuits are accompanied by sudden changes in voltage and current waveforms and each presents a unique waveform pattern.

  1-3) Next, the above-described waveform determination process will be further described.

1-3-1. Regarding the first waveform determination process:
The accident waveform data sent from the waveform recording devices 10a to 10c is digital data obtained by A / D converting voltage / current data (analog data) sampled at a predetermined sampling period in the waveform recording devices 10a to 10c. is there. Therefore, an example of the accident waveform data stored in the data storage unit 21 as the unclassified data 21a is a c-phase voltage waveform as shown in FIG. FIG. 6 is a diagram schematically showing a sawtooth wave for explaining a sawtooth wave determination method. In FIG. 6, the vertical axis indicates a voltage in any of the phases a to c, and the horizontal axis indicates time. Here, the black circle in the figure represents the voltage level at the time of sampling.

  Hereinafter, No. 2 in FIG. 1-No. A method for determining each waveform pattern of the sawtooth wave to the distortion waveform shown in FIG.

i) How to determine sawtooth wave:
As shown in FIG. 6, the sawtooth wave has a point in time when the value of the sample value before one sampling changes suddenly (corresponding to the interval between t1 and t2 in the figure) (in the figure, In the period of about ¼ cycle immediately after the sudden change (corresponding to about half of the interval between t1 and t2 in the figure), the voltage value is almost zero and changes. Absent.

  Therefore, focusing on such change characteristics of the sawtooth wave, it is classified whether or not it is a sawtooth wave.

  In FIG. 6, when the absolute value of the difference between sample values adjacent in time first exceeds a predetermined threshold L (time t1 in FIG. 6), the threshold L is ½ cycle after the time t1. For example, less than 10% of the amplitude value before the threshold value L (immediately before the time point t1 in the figure) before the threshold value L is exceeded. Sampling points indicating sample values are extracted, and if the interval width Tw on the horizontal axis is greater than 1/8 cycle, it is determined that the sawtooth wave (or a periodic pulse with no input at steady state) is obtained.

  The sawtooth determination method is determined from the above viewpoint. Next, the flow of the determination process will be further described with reference to FIG. FIG. 7 is a flowchart for explaining a sawtooth determination method.

  That is, as shown in FIG. 7, for each accident waveform data, for the digital data of the waveform read from the data storage unit 21, the sample value of the data whose sampling number is k indicating the sampling order and one of them The absolute value of the difference from the sample value of the data sampled before (k−1), that is, | kth sample value− (k−1) th sample value | is obtained (step S201), and the obtained difference Is determined to be greater than a threshold value L preset in the waveform classification program 23a (step S202). If so, the process proceeds to step S203, and if not, the process proceeds to step S206.

  In step S200, 0 is set to the detection number C of a sudden change point described later.

  In step S203, the absolute value of the difference measured immediately after the time when the absolute value of a difference exceeds the threshold L (time t1 in FIG. 6) exceeds the threshold L (at the time t1 in FIG. 6). It is determined whether or not sampling points that are, for example, less than 10% of the amplitude value immediately before) continue for a period of 1/8 cycle or more. If they are consecutive, the process proceeds to step S204.

  Here, how to determine the threshold value L will be described.

  That is, the threshold value L is the maximum difference value when the maximum amplitude sine wave on the recording data is input + the detection margin. The maximum difference value is obtained by taking a derivative of a sine wave.

  When fs is the sampling frequency and Vp is the maximum value of the recording data, the following equation A is established.

d / dk (Vp × sin (ωk / fs)) = Vp × ω / fs × cos (ωk / fs)
This maximum value is Vp × ω / fs.

  Here, since ω is ω = 2πfc when the system frequency is fc (50 or 60 Hz), this maximum value is also expressed as Vp × 2πfc / fs.

That is, the threshold value L can be determined as Vp × 2πfc / fs + detection margin.

  In step S204, it is further determined whether or not there is a sampling point that exceeds the threshold L again after a ½ cycle period from the time when the absolute value of the difference exceeds the threshold L (time t1 in FIG. 6). If it exists, the process proceeds to step S205, and if not, the process proceeds to step S206.

  In step S205, 1 is added to the detection number C of the sudden change point, the sampling number to be sampled next is added to the sampling number of data in the ½ cycle period, and the process proceeds to step S206, where the sampling number k is set. Increase by one.

  In step S207, it is determined whether or not the sampling number k has reached the total number of data (corresponding to all sampling numbers) constituting the waveform data to be determined in this determination flow. Returning to step S201, the processing from step S201 onward is repeated again.

  In step S208, it is determined whether or not the detected number C of sudden change points exceeds 3. If it exceeds, the process proceeds to step S209, and the waveform data to be determined is classified as “sawtooth wave”. If not, the process proceeds to step S210, where it is determined that it is not a “sawtooth wave”, and the process proceeds to the next waveform determination process described later.

  In FIG. 8, a sawtooth waveform 101 and a difference waveform 101d obtained by plotting a difference (not an absolute value) for each sample on the same graph are shown. For reference, FIG. 9 shows a differential waveform 102d in which a normal AC waveform (sine wave) 102 and the difference for each sample are plotted on the same graph.

ii) How to determine a single pulse:
Next, a method for determining a single pulse wave as shown in FIG. 10 will be described with reference to FIGS.

  Here, FIG. 10 is a waveform example of a single pulse wave in which the vertical axis represents Vo voltage and the horizontal axis represents time, and FIG. 11A illustrates a single pulse wave in order to explain a single pulse wave determination method. It is the figure which represented the wave typically. The vertical axis in FIG. 11A represents, for example, Vo voltage, and the horizontal axis represents time, and the vertical axis in FIG. 11B represents sample values that are temporally adjacent in the waveform data shown in FIG. It is the difference value (voltage) between them, and the horizontal axis represents time. In FIG. 11B, the difference data is indicated by a dot and connected by a broken line, and the absolute value of the difference is indicated by a dot and connected by a solid line. In FIG. 11B, the time point at which the difference value is plotted is displayed at the sampling point immediately before the time point at which the difference is generated, but it is needless to say that both time points may be represented in the same manner. .

  FIG. 12 is a flowchart for explaining a method of determining a single pulse wave.

As shown in FIG. 12, the single pulse waveform detection method is performed according to the following procedure.
1) In step S301, in the waveform data read from the unclassified data 21a, for example, as shown in FIG. 11A, all the sampling data in advance according to the sampling order, that is, in order from the head of the waveform data. Is obtained in advance (see FIG. 11B).
2) In step S302, it is checked whether or not the absolute value of the difference is greater than a specific threshold value (L) in the order of sampling from the beginning of the data. If it is larger, the process proceeds to step S303, and if it is not larger, the process proceeds to step S309.

  In FIG. 12, the times (or time points) corresponding to the sampling numbers k and k ′ are expressed as tk and tk ′ (k and k ′ are integers of 1 or more, k <k ′).

In FIG. 11B, the threshold value L is set to about 3.8 kV, and after detecting that the absolute value of the difference is larger than the threshold value L at the time point tk, the process proceeds to step S303.
3) In step S303, after the absolute value of the difference exceeds a specific threshold value L, a point where the sign of the difference value changes later from that time (time tk in FIG. 11B) is found.

In FIG. 11B, the sign of the difference value is changed to a minus value at the time tk ′ after the time tk when the sign of the difference value is a positive value.
4) In step S304, the average amplitude value of the voltage value of the waveform data at the time points tk and tk ′ obtained in steps S302 and S303 is obtained (position where the change is 50%).

In FIG. 11A, the average amplitude value is about 12 kV.
5) In step S305, after the time point tk (see FIG. 11B) when the absolute value of the difference exceeds the threshold value L (see FIG. 11B), the amplitude value of the waveform data is the “change 50% position” obtained in step S303. A crossing point (corresponding to the time point tc in FIG. 11A) is obtained, and a time difference tw (ms) from the time point tk ′ obtained in step S303 is obtained.

In FIG. 11A, it is shown as the time point tc.
6) In step S306, it is determined whether or not the time difference tw obtained in step S305 is equal to or greater than a specific threshold (T (ms)). Then, the process proceeds to step S309, and if it is less than T (ms), the process proceeds to step S307.
7) In step S307, from the threshold value L in a period longer than the preset setting cycle after the point (corresponding to the time point tc in FIG. 11A) crossing the “50% change position” obtained in step S304. If there is no large difference value, the process proceeds to step S308 if there is not, and the process proceeds to step S309 if there is no difference value.
8) In step S308, the waveform data to be determined is classified as “single pulse”, and then the process proceeds to the next waveform determination process described later.
9) In step S309, the sampling number k is incremented by one.
10) In step S310, it is determined whether or not the sampling number k has reached all the data numbers (corresponding to all the sampling numbers) constituting the waveform data to be determined in this determination flow. For example, the process returns to step S302, and the process from step S302 is repeated again. If the process has been reached, the process proceeds to step S311 to determine that it is not a “single pulse” and to the next waveform determination process described later.

  In FIG. 13, a waveform 301d of a single pulse wave and a difference waveform 301d obtained by plotting the difference (not absolute value) for each sample on the same graph are shown.

iii) Regarding the method for determining periodic pulses:
Next, a method for determining a periodic pulse wave as shown in FIG. 14 will be described with reference to FIGS.

  The periodic pulse waveform is a waveform in which a pulse waveform is periodically superimposed on the original AC waveform at regular intervals. Since the input signal of the power system is AC and arc discharge is likely to occur when the voltage becomes high, the periodic pulse waveform can be seen in synchronization with the power supply waveform every 1 cycle or 0.5 cycle of the AC waveform. There are many cases.

  Here, one cycle corresponds to one cycle time of the system frequency to be measured. If it is 50 Hz, 1/50 seconds = 0.02 seconds = 20 ms, if it is 60 Hz, 1/60 seconds≈0.016666 ... second = 16.666666 ... corresponds to ms.

  FIG. 14 shows a waveform example in which a periodic pulse-like waveform is superimposed at one cycle interval, and the vertical axis represents Vc voltage and the horizontal axis represents time.

  FIG. 15 is a diagram schematically showing a periodic pulse wave in which a periodic pulse waveform is superimposed at ½ cycle intervals in order to explain a method for determining a periodic pulse wave. The vertical axis represents, for example, Vc voltage, and the horizontal axis represents time. FIG. 16 is a flowchart for explaining a method for determining a periodic pulse wave.

As shown in FIG. 16, the periodic pulse waveform detection method is performed according to the following procedure.
1) In step S401, as in step S301, in the waveform data read from the unclassified data 21a, for example, as shown in FIG. The difference value is obtained for the sampling data.
2) In step S402, it is checked whether or not the absolute value of the difference is greater than a specific threshold value (L) in the order of sampling from the beginning of the data. If it is larger, the process proceeds to step S403, and if it is not larger, the process proceeds to step S410.

In FIG. 16, the times (or time points) corresponding to the sampling numbers k, k ′, k ″ are tk, tk ′, tk ″ (k, k ′, k ″ are integers of 1 or more. , K <k ′ ≦ k ″).
3) In step S403, the time when the absolute value of the difference exceeds the threshold value L is t1, and in the section [t1, t1 + 1/4 cycle period]
i) obtaining a time t1 ′ at which the sign of the difference value changes;
ii) Find a time t1 ″ that is a difference value of a sign opposite to the sign of the difference value at time t1 and whose absolute value exceeds the threshold value L.

Note that the waveform data shown in FIG. 15 shows an example of the waveform when t1 ′ = t1 ″.
4) In step S404, the absolute value of the difference between the sampling value of the waveform data at time t1 ′ and the sampling value of the waveform data at time t1 ″ is obtained, and the obtained absolute value is the predetermined set value L ′. It is determined whether or not the condition that it is larger and the time t1 ″ exists is satisfied. If the condition is satisfied, the process proceeds to step S405. If not satisfied, the process proceeds to step S410.

  The difference between the threshold value L and the set value L ′ in the present specification will be described.

The threshold value L is a value common to each detection method, and is a term used to mean a value used when a waveform different from a sine wave is desired to be detected. On the other hand, the set value L ′ is used to compare the difference between the “pulse peak value” and the “pulse detection point value”, and is a term meaning a threshold used to determine the pulse peak value.
5) In step S405, after the ½ cycle period has elapsed from time t1, it is determined whether or not there is a difference value whose absolute value exceeds the threshold value L. If there is, the process proceeds to step S406 and exists. If not, the process proceeds to step S410.
6) In step S406, the time corresponding to the difference value confirmed in step S405 is t2, and in the section [t2, t2 + 1/4 cycle period]
i) obtaining a time t2 ′ at which the sign of the difference value changes;
ii) Find a time t2 ″ that is a difference value of a sign opposite to the sign of the difference value at time t2 and whose absolute value exceeds the threshold value L.

In the waveform data shown in FIG. 15, a waveform example in the case of t2 ′ = t2 ″ is shown.
7) In step S407, as in step S404, the absolute value of the difference between the sampling value of the waveform data at time t2 ′ and the sampling value of the waveform data at time t2 ″ is obtained, and the obtained absolute value is It is determined whether or not the condition that the time t2 ″ is greater than the predetermined set value L ′ and that the time t2 ″ exists is satisfied. If the condition is satisfied, the process proceeds to step S408; otherwise, the process proceeds to step S410. move on.
8) In step S408, it is checked whether or not there is a difference value exceeding the predetermined threshold L2 in the section [t1 ″, t2]. If there is no difference value, the process proceeds to step S409, and if it exists, the process proceeds to step S410.
9) In step S409, the waveform data to be determined is classified into “periodic pulses (1/2 cycle interval)”, and then the process proceeds to the next waveform determination process described later.
10) In step S410, the sampling number k is incremented by one.
11) In step S411, it is determined whether or not the sampling number k has reached the total number of data (corresponding to all the sampling numbers) constituting the waveform data to be determined in this determination flow. For example, returning to step S402, the processing from step S402 onward is repeated, and if it has been reached, the process proceeds to step S412. In step S412, it is determined that it is not a “periodic pulse”.
12) Further, the same determination processing as in steps S401 to S412 is performed except that “1/2 cycle” which is the determination condition in step S405 of the determination flow shown in FIG. 16 is replaced with “1 cycle”. The result obtained by ORing the obtained determination results is used as the final determination result. Thereby, the waveform example in which the periodic pulse waveform is superimposed at one cycle interval shown in FIG. 14 can be correctly determined as the “periodic pulse wave”. Thereafter, the process proceeds to the next waveform determination process described later.

  Note that the value of the threshold value L is the maximum difference value that can be theoretically taken by a sine wave whose amplitude is the maximum value of the element in the normal state.

iv) Method for determining intermittent harmonics (periodic oscillation):
Next, a method for determining intermittent harmonics as shown in FIG. 17A will be described with reference to FIGS. 17A, 17B, and 18. FIG.

  Here, FIG. 17A is an example of a waveform for explaining a method for determining an intermittent oscillation harmonic, and FIG. 17B is a sample value that is temporally adjacent in the waveform data of FIG. It is a figure showing the difference value (voltage) between each other. In FIG. 17B, the difference data is indicated by a dot and connected by a broken line, and the absolute value of the difference is indicated by a dot and connected by a solid line.

  FIG. 18 is a flowchart for explaining a method of determining intermittent oscillation harmonics.

As shown in FIG. 18, the intermittent oscillation harmonic waveform detection method is performed according to the following procedure.
1) In step S501, in the waveform data read from the unclassified data 21a, for example, as shown in FIG. 17A, all the sampling data are previously stored according to the sampling order, that is, in order from the head of the waveform data. Is obtained in advance (see FIG. 17B). Also, 0 is set for the sampling number k, and 0 is set for the detection number C.
2) In step S502, it is checked whether the absolute value of the difference is larger than a specific threshold value (L) in the order of sampling from the top of the data. If so, the process proceeds to step S503, and if not, the process proceeds to step S506.

In FIG. 18, the time (or time) corresponding to the sampling number k is represented as tk (k is an integer of 1 or more).
3) In step S503, the time when the absolute value of the difference exceeds the threshold value L is t1, and in the section [t1, t1 + 1/2 cycle period], after the absolute value of the difference exceeds the threshold value L, the sign of the difference value is Count the changing point (n).

  Further, the sign final change point of the difference value is defined as time te (or time te). Here, the sign last change point of the difference value is determined under the condition that the sign of the difference value remains unchanged for a period longer than 1/10 cycle period after time te.

In FIG. 17B, the threshold value L is set to 4 in the vertical axis memory, and the time point t1, the time point of t1 + 1/2 cycle period, and the time point te when the absolute value of the difference exceeds the threshold value L are shown. .
4) In step S504, it is checked from the value obtained in step S503 whether or not the period of te-t1 is longer than several cycles of the AC waveform (the number of cycles set separately). If larger, the process proceeds to step S505. If it is less than that, it is determined not as intermittent but as a normal harmonic, and the process proceeds to step S506.
5) In step S505, the detection number C is incremented by 1, and the sampling number corresponding to the time te is substituted for the sampling number k.
6) In step S506, the sampling number k is incremented by one.
7) In step S507, it is determined whether or not the sampling number k has reached all the data numbers (corresponding to all the sampling numbers) constituting the waveform data to be determined in this determination flow. For example, the process returns to step S502, and the processes in and after step S502 are repeated. If the process has been reached, the process proceeds to step S508.
8) In step S508, if the number of detections C is established more than the number determined by the setting (for example, it may be determined that it is not a single noise such as 3 or 5 times in one accident waveform). Then, the waveform data to be determined is determined to be “intermittent oscillation harmonic”. On the other hand, if not established, the process proceeds to step S510, and it is determined that it is not “intermittent oscillation harmonic”.

  Thereafter, the process proceeds to the next waveform determination process described later.

v) Partial harmonic determination method:
Next, a method of determining partial harmonics generated at the beginning of an accident as shown in FIG. 19 will be described with reference to FIGS.
Generally speaking, harmonics are generally obtained by obtaining a high frequency in a frequency domain such as DFT (Discrete Fourier Transform), but in the present application, they are detected using the characteristics of the waveform. In order to detect the 5th harmonic (the frequency 5 times the fundamental wave) that is most often observed in the harmonics observed in the Japanese power system, the A cycle in FIG. 19 (the commercial frequency cycle of the harmonic generation part). Number) and B times (the number of changes in the sign of the slope of the waveform in section A) are values that satisfy B / A = 5 × 2.

  Here, FIG. 20 is a diagram schematically illustrating an example of a waveform for explaining a method of determining partial harmonics, where the vertical axis represents, for example, the amplitude of waveform data, and the horizontal axis represents time. . FIG. 21 is a flowchart for explaining a method of determining partial harmonics.

As shown in FIG. 21, the partial harmonic detection method is performed in the following procedure.
1) In step S601, as in step S301, for waveform data as shown in FIG. 20, for example, read from the unclassified data 21a, difference values are obtained in advance for all sampling data.
2) In step S602, it is determined whether or not the sampling number k has reached all the data numbers (corresponding to all the sampling numbers) constituting the waveform data to be determined in this determination flow. If so, the process proceeds to step S603. On the other hand, if it has reached, the process proceeds to step S609, where it is determined whether there is no harmonic or partial oscillation, and the process proceeds to the next waveform determination process described later.
3) In step S603, it is checked whether or not the absolute value of the difference is larger than the threshold value L. If it is larger, the process proceeds to step S605, and if not larger, the process proceeds to step S608.
4) In step S604, from the time t1 when the absolute value of the difference exceeds the threshold L (corresponding to the kth sample from the beginning of the waveform data), within the A cycle period (for example, 1 to 2 cycles of the waveform data to be determined) In the period), the number of changes of the sign of the difference is counted, and the process proceeds to step S605.
5) In step S605, the number of code changes counted in step S604 is B times (for example, 20 times) or more, and the interval between code change points is within Tc (ms) (for example, 50 ms). If the condition is satisfied, the process proceeds to step S606, and if not, the process proceeds to step S607.
6) In step S606, a determination process is performed to determine that the waveform data to be determined is “partial harmonic”, and then the process proceeds to a waveform determination process to be described later.
7) In step S607, for the sampling number k, the sampling number corresponding to the time when the sign last changed is substituted within the set cycle period (within Tc (ms)) from time t1, and the process proceeds to step S608. .
8) In step S608, the sampling number k is incremented by 1, and the process returns to step S602.

  In FIG. 22, a partial harmonic waveform 601 and a difference waveform 601d in which a difference (not an absolute value) for each sample is plotted on the same graph are shown.

vi) About the judgment method of a distortion waveform:
Next, a method for determining a distortion waveform as shown in FIGS. 23A to 23D will be described with reference to FIGS.

  Here, FIG. 24 is a diagram schematically showing a waveform example for explaining a distortion waveform determination method, where the vertical axis represents, for example, the amplitude of waveform data, and the horizontal axis represents time. FIG. 25 is a flowchart for explaining a distortion waveform determination method.

  The distortion waveform is extracted using a distortion wave extraction filter. The distorted wave extraction filter is a combination of the front part of the rectangular wave extraction filter and the part that removes the distorted wave.

As shown in FIG. 25, the distorted wave detection method is performed according to the following procedure.
1) In step S701, it is determined whether or not the sampling number k has exceeded the total number of data (corresponding to the total number of samplings) constituting the waveform data to be determined in this determination flow. The process proceeds to step S702. On the other hand, if it exceeds, the determination process ends, and the process proceeds to the next waveform determination process described later.
2) In step S702, a time t1 at which the absolute value of the difference exceeds the threshold L and a time t2 at which the absolute value of the difference after the ½ cycle period has elapsed from the time t1 exceeds the threshold L are obtained. The process proceeds to S703.
3) In step S703, it is determined in step S702 whether both time t1 and time t2 exist. If they exist, the process proceeds to step S704, and if not, the process proceeds to step S707 and the sampling number k is incremented by 1. Return to step S701.
4) In step S704, a time tmax at which the maximum amplitude value of the waveform data is obtained between times t1 and t2, and the time between tmax and t1 is equal to or greater than a predetermined threshold value (1/8 cycle period). If the condition is satisfied, the process proceeds to step S705; otherwise, the process proceeds to step S707.
5) In step S705, determination processing is performed that the waveform data to be determined is a “distortion waveform”, and the process proceeds to step S706.
6) In step S706, the sampling number corresponding to time t2 is substituted for sampling number k, and the process returns to step S701.

1-3-2) Second waveform determination process:
Here, in FIG. A method for determining a waveform pattern when the low frequency component shown in FIG.

i) About the determination method of the low frequency component superimposed waveform:
Next, a determination method for waveform examples 801 to 803 on which low frequency components as shown in FIG. 26 are superimposed will be described with reference to FIG.

  Here, FIG. 27 is a flowchart for explaining a method of determining a low-frequency component superimposed waveform.

The low frequency component superimposed waveform is detected by the following method.
First, an average value for one cycle of the AC waveform is obtained. Speaking of the average value of alternating current, it generally means the average value after being converted to an absolute value, but here it is a simple average value because the purpose is to detect the superimposed low frequency components. That is, a low frequency component is obtained by simply taking a moving average for one cycle, and it is determined that there is a low frequency component when there are continuously two or more cycles exceeding a certain threshold value L.

More specifically, the low frequency component superimposed waveform detection method is performed according to the procedure shown in FIG.
1) In step S801, variables such as a counter C and a maximum continuous counter Cmax described later are initialized, and the process proceeds to step S802.
2) In step S802, is the sampling number k smaller than the number of data obtained as a result of subtracting the number of data per cycle period from the total number of data (corresponding to all sampling numbers) constituting the waveform data to be determined? If not, the process proceeds to step S803. If not, the process proceeds to step S810.
3) In step S803, a simple average value M [k, k + 1 cycle] of the amplitude value of the waveform data from the time corresponding to the sampling number k to one cycle period is obtained, and the process proceeds to step S804.
4) In step S804, it is checked whether or not the absolute value | M | of the average value M [k, k + 1 cycle] exceeds the threshold value L. If so, the process proceeds to step S805. Advances to step S807.
5) In step S805, the counter C that counts the number of times the threshold value L is exceeded is incremented by 1, and the process proceeds to step S806.
6) In step S806, the sampling number k is incremented by 1, and the process returns to step S802.
7) In step S807, it is checked whether or not the value of the counter C is larger than the value of the maximum continuous counter Cmax. If larger, the process proceeds to step S808, and if not larger, the process proceeds to step S809. Here, the maximum continuous counter Cmax can record the maximum number of counts when the count C continuously increases by 1 by the operation of steps S804 to S805 and steps S808 to S809.
8) In step S808, the value of the counter C is substituted for the maximum continuous counter Cmax, and the process proceeds to step S809.
9) In step S809, 0 is substituted into the counter C that counts exceeding the threshold value, and the process proceeds to step S806.
10) In step S810, it is checked whether or not the condition that the value of the maximum continuous counter Cmax corresponds to N cycles (for example, 2 cycles) or more is satisfied. If satisfied, the process proceeds to step S811, and the waveform data to be determined is A determination process of “low frequency component superimposed waveform” is performed, and the process proceeds to the next waveform determination process described later. On the other hand, if not satisfied, the process proceeds to step S812, where a determination process is performed that the waveform data to be determined is not a “low frequency component superimposed waveform”, and the process proceeds to the next waveform determination process described later.

  An example of classification based on the “offset amount change information” of the present invention is whether or not the value of the maximum continuous counter Cmax corresponds to N cycles (for example, 2 cycles) or more in step S810. It corresponds to the operation of determining and determining the presence or absence of low frequency components.

1-3-3) Regarding the third waveform determination process:
Here, in FIG. 8-No. A method of determining each waveform pattern of the half-wave rectified waveform shown in FIG.

i) Method for determining half-wave rectified waveform:
Here, a determination method of waveform examples 901 to 903 of the half-wave rectified waveform as shown in FIG. 28 will be described with reference to FIGS. 29, 30 (a), and (b).

  FIG. 29 is a flowchart for explaining a half-wave rectified waveform determination method. FIG. 30A is a diagram showing a half-wave rectified waveform example 904 that is a determination target, and a differential waveform graph 905 in which the difference values of the waveform data are plotted with squares, and FIG. It is a figure used for description of FIG. In FIG. 30 (b), the effective value graph 906 obtained by plotting the effective value / √2 with ◆ and the value obtained by averaging the sum of the difference data × cosωt in the 1/2 cycle period and plotting the absolute value are plotted with ■. A correlation value graph 907 is shown. Here, ω is an angular velocity of a waveform to be determined (for example, waveform example 904).

As shown in FIG. 29, the half-wave rectified waveform is detected in the following procedure.
1) In step S901, in the waveform data read from the unclassified data 21a of the data storage unit 21, for example, as shown in FIG. 28 (see the waveform example 904 plotted with ◆ in FIG. 30A), All sampling data is processed to obtain a difference with the previous data (such as differentiation) (see the differential waveform graph 905 plotted with (3) in FIG. 30A).
2) In step S902, as in step S802 of FIG. 27, the sampling number k subtracts the number of data per cycle period from the total number of data (corresponding to the total number of samplings) constituting the waveform data to be determined. It is determined whether or not it is smaller than the number of data obtained as a result. If smaller, the process proceeds to step S903. If not smaller, the process proceeds to step S913, and the determination target waveform data is not “half-wave rectified waveform”. It progresses to the following waveform determination process mentioned later.
3) In step S903, an effective value for ½ cycle of the difference data calculated in step S901 is calculated. (See the first graph 906 in FIG. 30B)
That is, for example, a case will be described in which the effective value of the point 906 [t1] at the leftmost end (corresponding to the difference data of the sampling number k = 1) shown in the first graph 906 in FIG. . In this case, in the difference waveform graph 905 in FIG. 30A, the leftmost point 905t1 on the horizontal axis and the point 905 [t1 + 1 / It is obtained by calculating the square root of the mean value of the squares of all the difference values between [2 cycles]. As the sampling number k increases sequentially, the period of 1/2 cycle as the period for calculating the effective value sequentially moves to the right of the horizontal axis in FIG.
4) In step S904, in the ½ cycle period of the difference data calculated in step S901, the sum of difference data × cosωt is taken and averaged to obtain an absolute value. (See the correlation value graph 907 in FIG. 30B)
5) In step S905, it is checked whether or not the difference between the effective value / √2 obtained in step S903 and “difference data × cosωt” (correlation value) obtained in step S904 is less than a predetermined value A. If less, the process proceeds to step S906, and if not less, the process proceeds to step S912.

That is, in FIG. 30B, the timing at which the effective value / √2 and “difference data × cosωt” are substantially the same is searched for by each sampling, and is shown as five regions Ra to Re surrounded by circles.
6) In step S906, a half cycle period before the sampling number k when it is determined in step S905 that the difference between them is less than a predetermined value A (substantially 0 in FIG. 30B), and k + 1 The effective value of the difference data in the 1/2 cycle period after the / 2 cycle period is calculated.
7) In step S907, the effective value of the difference data in the previous 1/2 cycle period and the effective value of the difference data in the subsequent 1/2 cycle period obtained in step S906 are both predetermined values B (see FIG. In 30 (b), it is checked whether or not it is less than approximately 0). If less, the process proceeds to step S908, and if not, the process proceeds to step S912.

In FIG. 30 (b), the point indicating the effective value of the difference data in the half cycle period before the point 906 [tk] (the point satisfying the condition of step S905) corresponding to the difference data of the sampling number k ( Since both the area Rf in the figure and the point indicating the effective value of the difference data in the subsequent half cycle period (see the area Rg in the figure) are almost 0 (an example of the predetermined value B), It can be seen that the condition of step S907 is satisfied.
8) In step S908, the half-wave rectification detection counter is incremented by 1, and the process proceeds to step S909.
9) In step S909, it is checked whether or not the value of the half-wave rectification detection counter is 2 or more, and if it is 2 or more, the process proceeds to step S910, and the determination target waveform data is “half-wave rectification waveform”. The process proceeds to the next waveform determination process described later. On the other hand, if it is not 2 or more, the process proceeds to step S911.
10) In step S911, the sampling number k is increased by ½ cycle, and the process proceeds to step S912.
11) In step S912, the sampling number k is incremented by 1, and the process returns to step S902.

ii) About the method for determining missing half-wave voltage:
Here, in FIG. A method of determining a half-wave voltage missing waveform with respect to a waveform as shown in the column of waveform example 8 will be described with reference to FIG.

  FIG. 31 is a flowchart for explaining a method of determining a half-wave voltage missing waveform. Note that the waveform to be determined is only a voltage element.

As shown in FIG. 31, the half-wave voltage missing waveform is detected by the following procedure.
1) In step S1001, similarly to step S901 in FIG. 29, in the waveform data read from the unclassified data 21a, a process is performed in advance to obtain a difference from the previous data for all sampling data. .
2) In step S1002, as in step S902 of FIG. 29, the sampling number k subtracts the number of data per cycle period from the total number of data (corresponding to the total number of samplings) constituting the waveform data to be determined. It is determined whether it is smaller than the number of data obtained as a result. If smaller, the process proceeds to step S1003, and if not smaller, the process proceeds to step S1010.
3) In step S1003, the same processing as that described in steps S903 and S904 in FIG. 29 is performed.
4) In step S1004, the same processing as that described in step S905 of FIG. 29 is performed.
5) In step S1005, in step S1004, one cycle period before the sampling number k when it is determined that the difference between the effective value and the correlation value is less than the predetermined value A, and after the k + 1/2 cycle period The effective value of the difference data in one cycle period is calculated.
6) In step S1006, the effective value of the difference data in the ½ cycle period after sampling number k obtained in step S1005 is less than a predetermined value B, and the difference in one cycle period before sampling number k. It is checked whether the effective value of the data is equal to or greater than the predetermined value C and the effective value of the difference data in one cycle period after the k + 1/2 cycle period is equal to or greater than the predetermined value C. If YES in step S1007, the flow advances to step S1007; otherwise, the flow advances to step S1009.

  Here, the predetermined values A, B, and C will be described.

That is, the predetermined value A is 5% of the steady voltage / √2 (a value for detecting that there is almost no difference between the effective value / √2 and the difference data × cosωt at the steady voltage / √2). The value B is 5% of the steady voltage (a value for confirming whether the effective value before and after the 1/2 cycle is approximately 0), and the predetermined value C is 95% of the steady voltage (almost the steady value). It is a value for checking whether or not there is).
7) In step S1007, the half-wave loss detection counter is incremented by 1, and the process proceeds to step S1009.
8) In step S1008, the sampling number k is increased by ½ cycle, and the process proceeds to step S1009.
11) In step S1009, the sampling number k is incremented by 1, and the process returns to step S1002.

iii) Regarding the method for determining the waveform with minute Vo and Io before the accident:
Here, as shown in FIG. 33 and FIG. 34, a method for determining a waveform with a small Vo · Io before an accident in which slight changes in Io (zero phase current) and Vo (zero phase voltage) are observed before the accident as shown in FIG. The description will be given with reference.

  FIG. 33 is a diagram schematically showing an example of a waveform for explaining a method of determining a waveform with a small Vo / Io before an accident, and the vertical axis represents, for example, the amplitude (Vo, Io) of the waveform data, and the horizontal axis Represents time. In FIG. 33, from the beginning to t (p) corresponds to the total number of samples before the accident. FIG. 34 is a flowchart for explaining a method for determining a waveform with minute Vo / Io before an accident.

  As shown in FIG. 34, the waveform with minute Vo and Io before the accident is detected by performing an effective value calculation for the element data of Vo and Io with a calculation width of ¼ cycle (see step S1101), and the value is a set value. When it is larger (see step S1102), it is determined that the pre-accident waveform is Vo and Io (see step S1103). However, the determination is made from the time of the first sampling (sampling number k1) (see time t (k1) in FIG. 33) to the time before the start of the oscilloscope data minus 2 cycles (see time t (ke) in FIG. 33). (See step S1105).

  On the other hand, when it is determined that the effective value is not larger than the set value (see step S1102), the sampling number k is incremented by 1 (see step S1104), and the calculation range of the 1/4 cycle period from the sampling number k is the time when the accident occurs ( 33 (see time t (p) in FIG. 33), it is determined whether or not it is two cycles or more before (see step S1105). If it is determined that there is, it proceeds to step S1106 and is not a waveform with minute Vo and Io before the accident. (See step S1106). On the other hand, if it is not determined in step S1105, the process returns to step S1102.

  The above is a method for analyzing the characteristics of individual waveforms.

In general, in the case of a power system fault waveform, the waveform duration of the fault portion is only for several cycles of the AC waveform. Also, in the case of a short-circuit accident or a ground fault, there are many cases of arc flashing, which is accompanied by a sudden change in voltage and current, so that the harmonic components are extracted by performing FFT calculation over the entire waveform. Compared to the case, the case where the change point is detected from the difference between the waveforms can detect the accident waveform portion at a much higher speed. This can be said only in the case of a power system fault waveform, and is not a general idea. For example, a waveform with a long duration such as an earthquake waveform does not hold.
(2) Utilization of constructed database:
2-1) When a new accident occurs:
Here, the newly generated accident waveform data matches the past data stored in the database using the data search program 24a in at least one item (for example, an item such as a waveform pattern). A case where search processing for searching for data is performed will be described.

  i) First, a user of the power system fault waveform data search device 20 or a user of the terminal devices 13a to 13b connected to the network 2 (12) can receive an input from an input terminal unit such as a keyboard or a mouse. The waveform classification program 23a is activated with newly generated accident data as a classification target. As a result, the data classification unit 23 performs a waveform pattern determination process on newly generated accident data according to the same procedure as the processing method described in item (1) “Database construction”. As shown in FIG. 4, the determination result and other data are stored in the database of the data storage unit 21 as the classified data 21b. Here, the new accident has a typical distortion waveform (see Fig. 2), does not include the short-circuit current component, and the neutral point grounding system of the power system is changed to the direct grounding system. Assume that it is analyzed that it belongs. Therefore, the search target in item iii) described later is focused on the difference in the neutral point grounding system of the system, and directly as accident waveform data of a similar system among the three grounded systems. The data is limited to data of the same classification classified as the grounding system (see FIG. 4). .

  The newly generated accident data to be classified is already transmitted from any one of the waveform recording devices 10a to 10c to the power system accident waveform data search device 20 before starting the above classification operation. It is assumed that the data storage unit 21 stores the unclassified data 21a.

  ii) Next, the user searches for past accident data already stored in the data storage unit 21 as the classified data 21b by the input from the input terminal unit such as a keyboard or a mouse as a search condition. In order to extract data having high similarity to the new accident data having the waveform pattern designated as one of the data patterns, the data search program 24a is started.

  Thereby, the screen (refer FIG. 35) for inputting search conditions is displayed on the display part of the electric power grid accident waveform data search device 20 or terminal device 13a, 13b. FIG. 35 is an example of a screen for inputting search conditions and displaying search results displayed on the display unit of power system accident waveform data search device 20 or terminal devices 13a and 13b according to the present embodiment. .

  Here, the search condition is a relationship with new accident data to be searched by the user from past accident data to be searched, in order to narrow down past accident data having more similarities. Examples of necessary input items and search items are shown in FIGS. FIGS. 36A and 36B are diagrams illustrating examples of search items displayed on the display unit of the power system fault waveform data search device 20 or the terminal devices 13a and 13b according to the present embodiment.

In the present embodiment, the search condition input method includes the above-mentioned “information on physical conditions of transmission / distribution system”, the above-mentioned “information on environmental conditions at the time of data recording”, and the above-mentioned “analysis of accident waveform”. Check all applicable data items using the check boxes displayed on the input screen, except for undetermined information, and the information that can be acquired after an accident. Insert. For example, since the waveform pattern of new accident data is a typical distortion waveform as described above, the check boxes displayed in the “waveform pattern” column 3501 in FIG. Check only the “Waveform” checkbox.

  Note that an example of “classification result of accident waveform data specified as a search condition” of the present invention is input to the check box of the column 3501 of “waveform pattern” in FIG. It is a “waveform pattern” for data.

  Also, in order to be able to search the same or similar accident waveform data without omission, it is necessary to cover a wide range of search objects. In FIG. 35, for example, “occurrence time: year” covers all data before 2009, “occurrence time: month” designates January to December, and “accident aspect”. , “Voltage class”, “Accident elimination time”, “Weather” and “Cause of accident (eg, natural degradation, landslide / snowfall, wind / rain, ice / snow, thunder, earthquake, fire, water damage, contact with birds and beasts, etc.)” In the item column, all check boxes are checked.

  iii) When the input of the search condition is completed, the data search unit 24 searches the past accident data stored in the data storage unit 21 as the classified data 21b, and stores the data of the same classification directly classified as the ground contact system. As a search target, a search for data similar to the new accident data is started. Here, instead of searching all past accident data stored in the database as a search target, it is limited to data of the same classification that is classified as a ground contact system directly in the accident waveform data group. This is because, as described above, it has already been found that the above new accident data does not contain a short-circuit current component and that the neutral point grounding system of the power system belongs to the direct grounding system. . As a result, the search time can be shortened.

  As a result of the search, the data search unit 24 sorts past accident data in descending order of the number of matching data items (information elements) (that is, in descending order of similarity), and records them in association with each. The “information that can be investigated and acquired after the accident” (see FIG. 4) is displayed on the output screen (see FIG. 37), and the search result is stored in the data storage unit 21 as search result data 21c. The data search unit 24 transmits search result data 25 in response to a request from the user of the terminal devices 13a and 13b via the network 2 (12), for example.

  Here, FIG. 37 is a diagram showing a search screen list 3710 and a search screen lower column 3720 displayed on the power system accident waveform data search device 20 or the terminal devices 13a and 13b. In the search screen listing 3710, an example of input of each search condition and the number of hits when searching for each search condition alone are displayed as a search result 3710a, and search is performed using AND (product set) of each search condition. The number of hits is displayed as the final search result 3710b (the final search result is 5 in the figure). Further, the lower column 3720 of the search screen shows an example in which five accident data hit as the final search result are extracted and displayed. Here, display conditions can be switched between display and non-display as necessary. FIG. 35 shows an example in which all the search conditions are displayed, but FIG. 37 shows an example in which only the accident occurrence time, accident aspect, and voltage class are displayed as part of the search conditions.

  As described above, according to the power system fault waveform data search device 20 of the present embodiment, even a non-expert engineer can quickly search for necessary accident data using clues such as voltage / current waveforms and relay response characteristics. It is also possible to easily search and display past similar accident data for inexperienced accidents. For this reason, when a new accident occurs, the cause of the accident and the recovery method can be determined quickly.

  FIG. 38 shows an example of an accident waveform recorded recently. The cause of this was unknown at first, but by using the power system fault waveform data search device 20 of the present embodiment, the fault waveforms of FIG. 39, FIG. 40, and FIG. It was immediately retrieved that the features were similar.

  39, 40, and 41, it was inferred that the cause of the accident in the accident waveform was bird and animal contact, but the cause of the accident in the accident waveform in FIG. 38 would be the same. It was confirmed that the cause was contact with birds and animals.

2-2) When specifying accident data to be searched from the classified data 21b as a database storing past accident data:
In this case, the accident data to be searched by the user is not a newly occurred accident but an accident that has occurred in the past, and is one of the already stored data 21b as the already classified data 21b. In a certain point, it differs from the case of said 2-1).

  Therefore, the operation of the power system fault waveform data search device 20 here is basically the same except that the waveform classification operation of the newly generated accident described in item 2-1 above is not necessary. is there.

  In addition, the accident status, the cause of the accident, the recovery method, etc. of “Information that can be investigated and acquired after the accident” are already stored in the database, but if you want to broaden the search target, it is necessary to narrow down the search conditions For example, all the check boxes regarding the cause of the accident may be checked.

  FIG. 42 shows a waveform (half-wave rectification waveform) of the inrush current of the transformer when the circuit breaker is turned on. When it is desired to investigate how many such waveforms have been in the past, the person in charge in the past has been examining past accident waveform recording files (paper files) one page at a time. According to the system fault waveform data search device 20, the corresponding waveform can be searched immediately.

  Moreover, in the said embodiment, the data storage part 21 which is an example of the data storage part of this invention is accident waveform data, information on the installation place of the waveform recording apparatus 10a etc., a voltage class, a neutral point grounding system, a system | strain "Information on physical conditions of power transmission and distribution system" such as short-circuit capacity, line length, etc., "Information on environmental conditions at the time of data recording" such as date and time of data recording, weather conditions, etc., accident aspect, voltage value at the time of accident "Information obtained by analyzing the accident waveform" such as the classification result (waveform pattern) of the current characteristics, graphic characteristics of the accident waveform data, and "post-accident investigation and acquisition" such as the accident situation, the cause of the accident, and the recovery method Although the case where the information is configured for each accident waveform data has been described, the present invention is not limited to this. For example, in addition to the above classification result of accident waveform data, information on accident waveform data and transmission / distribution Systematic All or part of the information on the physical condition, all or part of the information on the environmental condition at the time of data recording, all or part of the information obtained by analyzing the accident waveform, and can be obtained by investigating after the accident A configuration in which at least one of the information in all or a part of the information is stored may be used.

  Moreover, in the said embodiment, the component containing the feature extraction part 22, the data classification part 23, and the waveform classification program 23a which is an example of the feature extraction part and data classification part of this invention is made into one accident waveform data. On the other hand, a classification process using the difference information obtained based on the sampling value of the accident waveform data (first waveform determination process), and a classification using the offset amount information obtained based on the accident waveform data. When performing all the processing (second waveform determination processing) and classification processing (third waveform determination processing) using information on the effective value of the accident waveform data to determine the waveform pattern for each accident waveform data However, the present invention is not limited to this. For example, the waveform pattern is determined from the result of at least one of the first waveform determination process, the second waveform determination process, and the third waveform determination process. It may be formed. In this configuration, there is a possibility that the number of accident data extracted from the search target may increase on average as compared to the case where three types of determination processing are performed. It is possible to search for similar accident data.

  In the above embodiment, the classification result (waveform pattern) of the accident waveform data specified as the search condition is a component including the data classification unit 23 and the waveform classification program 23a (an example of the data classification unit of the present invention). However, the present invention is not limited to this, and for example, the user himself / herself can determine whether new accident waveform data or past data is stored in the data storage unit 21 as classified data 21b. It is stored in the data storage unit 21 as the classified data 21b by determining which waveform pattern it belongs to regardless of the waveform data of the accident and inputting the classification result based on the determination from the input terminal unit. It may be configured.

  In the above embodiment, the data classification unit 23 sequentially processes the first waveform determination process S101, the second waveform determination process S102, and the third waveform determination process S103 in this order by the waveform classification program 23a. However, the present invention is not limited to this. For example, the order of waveform determination processing may be any order.

  In the above embodiment, the case where the items shown in FIG. 36 are employed as the search condition has been described. However, the present invention is not limited to this. For example, which of the items other than the waveform pattern is employed as the search condition is: It may be selectable as appropriate, or may be changed as appropriate.

  In the above embodiment, the power system fault waveform data search apparatus 20 classifies the accident waveform data as an example of the power system fault waveform data search apparatus of the present invention, and stores the database in the data storage unit 21 as the classified data 21b. Although the case where it has the function to construct | assemble was demonstrated, it is not restricted to this, For example, the structure which the electric power system accident waveform data search device 20 does not have the said function may be sufficient. In other words, in this case, the power system fault waveform data search device as another example of the power system fault waveform data search device 20 of the present invention described above is described with reference to FIG. 1 has a waveform data classification program 23a but not a waveform classification program 23a. Other configurations are basically the same as those in FIG. Become. However, in this case, it is assumed that a predetermined database in which the waveform pattern for each accident waveform data is already classified is stored in the data storage unit 21 in advance, and the predetermined database is stored in the power system. For each accident, at least the accident waveform data is based on at least one of the first waveform determination process, the second waveform determination process, and the third waveform determination process described above. As shown in FIG. 4, for example, the classified result (waveform pattern) classifies the accident data in advance by the difference in the short-circuit capacity of the system, which is the physical property of the power system, and the difference in the neutral point grounding method. It is sufficient if the configuration is stored.

  Moreover, in the said embodiment, as an example of the electric power system accident waveform data search device of this invention, the electric power system provided with the function to extract a similar accident waveform using the constructed database with the function to construct a database Although the accident waveform data search device 20 has been described, the present invention is not limited to this. For example, a configuration that does not have a function of extracting a similar accident waveform using a database may be used. That is, in this case, the waveform database construction device as another example of the power system fault waveform data search device of the present invention will be described with reference to FIG. The configuration has no data search function and has the waveform classification program 23a, but does not have the data search program 24a, and the other configurations are basically the same as those in FIG.

  Further, in the above embodiment, in order to narrow down the accident data to be searched, the data for each subsequent event, when constructing the database, the difference in the short-circuit capacity of the system, which is the physical property of the power system, and the neutrality Although the case where it classify | categorizes beforehand by the difference in a point grounding system was demonstrated, it is not restricted to this, For example, the structure which groups according to the classification result of accident waveform data, and stores in a database may be sufficient. Even in this case, accident data to be searched can be narrowed down, so that efficient search can be performed.

As is clear from the above description, the first aspect of the present invention is
Data of the power system fault waveform data collection system that is installed in the power transmission / distribution system, records the voltage / current waveform at the time of the grid fault from the front of the fault for a certain period of time, and collects the data on the server via the communication line. A power system fault waveform data search device comprising a collection,
A data storage unit that stores the collected data, a feature extraction unit that extracts features of the stored data, a data classification unit that classifies data according to the extracted features, a search unit that searches the classified data, A terminal device for inputting search conditions; and a display device for displaying search results.
The data storage unit includes at least one of installation location information, voltage class, neutral grounding method, system short-circuit capacity, and line length recorded along with the collected accident waveform data. “Information on physical conditions of transmission and distribution systems”
“Information on environmental conditions at the time of data recording” including at least one of the data recording date and weather conditions, and
`` Information obtained by analyzing the accident waveform '' including at least one of the accident aspect, the voltage value at the time of the accident, the current value, and the analysis result of the graphic characteristics of the waveform,
`` Information that can be investigated and acquired after the accident '' including at least one of the accident situation, the cause of the accident, and the recovery method,
Is stored for each accident waveform data,
Each time new data is added to the database, the feature extraction unit converts a pulse waveform component, a sawtooth component, a rectangular component, a harmonic component, a low frequency vibration component, a half wave from the instantaneous value data of the waveform to the waveform. Determine whether or not at least one of the rectified waveform components is included, and the aspect of the accident such as whether the accident is a one-wire ground fault, a two-wire ground fault, a three-wire ground fault, a two-wire short-circuit, or a three-wire short-circuit Determine
The data classification unit classifies data based on already stored data and the feature data determined by the feature extraction unit, and performs data selection and rearrangement,
Automatically when new accident waveform data is collected or by the user instructing from the terminal device, the "information on physical conditions of the power transmission and distribution system" and the "information on environmental conditions at the time of data recording", The past data is searched based on the “information obtained by analyzing the accident waveform”, the data is sorted in descending order of the number of matching information elements, and the “ A power system accident waveform data search device characterized by outputting information that can be investigated and acquired after an accident to a display device,
When searching for data based on the analysis result of the graphic characteristics of the above waveform, if the data is short-circuit accident data including data caused by ground faults with two or more wires through which a short-circuit current flows, search within a data group with the same system short-circuit capacity. In the case of ground fault data, the power system fault waveform data search device is characterized in that the neutral point grounding method is searched within the same data group.

The second aspect of the present invention
Specify one of the already collected accident waveform data, and the information related to the physical conditions of the power transmission / distribution system that is recorded along with it, and the information related to the environmental conditions at the time of data recording ”And“ information obtained by analyzing the accident waveform ”and“ information that can be obtained by investigating after the accident ”by the user specifying at least one or more of the above information from the terminal device, The power system fault waveform data search device according to the first aspect of the present invention is characterized in that other data with matching information is searched and the information regarding the data extracted as a result is output.

The third aspect of the present invention
The `` information that can be obtained by investigating after an accident '' further includes at least one of an accident occurrence section, a section of a power outage due to an accident, success or failure of a reclosing by a protection relay due to an accident, and presence or absence of a progress accident. It is the electric power system fault waveform data search apparatus of the said 1st or 2nd this invention.

The fourth aspect of the present invention is
For accidents in the power system, a data storage unit in which at least accident waveform data for each accident is classified based on information on differences obtained based on sampling values of the accident waveform data is stored;
A search unit that searches for at least the classification result of the accident waveform data stored in the data storage unit, and searches for a match with the classification result of the accident waveform data specified as a search condition;
A power system fault waveform data search device comprising:

The fifth aspect of the present invention provides
The information relating to the difference is information including whether or not a difference between the accident waveform data and a sampling value immediately before in time exceeds a predetermined threshold value at least. This is a waveform data search device.

The sixth aspect of the present invention provides
For an accident in a power system, at least an accident waveform data for each accident, a data storage unit in which a classification result is stored that is classified based on information on an effective value obtained based on the accident waveform data;
A search unit that searches for at least the classification result of the accident waveform data stored in the data storage unit, and searches for a match with the classification result of the accident waveform data specified as a search condition;
A power system fault waveform data search device comprising:

The seventh aspect of the present invention
The information on the effective value is the electric power system accident waveform data search device according to the sixth aspect of the present invention, which is information on a change in an effective value calculated from data for a certain period of the accident waveform data.

In addition, the eighth aspect of the present invention
Data storage in which at least accident waveform data for each accident in the electric power system stores classification results classified based on information on offset amounts or information on DC components obtained based on the accident waveform data And
A search unit that searches for at least the classification result of the accident waveform data stored in the data storage unit, and searches for a match with the classification result of the accident waveform data specified as a search condition;
A power system fault waveform data search device comprising:

The ninth aspect of the present invention provides
The information related to the offset amount is the power system fault waveform data search device according to the eighth aspect of the present invention, which is information about a change in the offset amount calculated from data for a certain period of the fault waveform data.

The tenth aspect of the present invention is
A recording medium that records a program that causes a computer to execute the function of the search unit of the power system fault waveform data search device according to any one of the first to ninth aspects of the present invention, and is a recording medium that can be processed by a computer. is there.

  An example of the program of the present invention is the function of at least the data search unit of the power system fault waveform data search apparatus of the present invention described above, or the operation of at least the search processing step of the processing method by the power system fault waveform data search apparatus. Is a program that is executed by a computer and operates in cooperation with the computer.

  An example of the recording medium of the present invention includes all or part of the functions of the data search unit of the power system fault waveform data search apparatus of the present invention described above, or each of the processing methods by the power system fault waveform data search apparatus. A recording medium that records a program that causes a computer to execute all or part of the operations of the steps. The recording medium is readable by a computer, and the read program cooperates with the computer to execute the operation. .

  Further, the “operation of each step” of the present invention means the operation of all or a part of the steps.

  Further, one use form of the program of the present invention may be an aspect in which the program is recorded on a recording medium such as a ROM readable by a computer and operates in cooperation with the computer.

  Also, one use form of the program of the present invention is an aspect in which the program is transmitted through a transmission medium such as the Internet, a transmission medium such as light, radio wave, and sound wave, read by a computer, and operates in cooperation with the computer. Also good.

  The computer of the present invention described above is not limited to pure hardware such as a CPU, but may include firmware, an OS, and peripheral devices.

  As described above, the configuration of the present invention may be realized by software or hardware.

  According to the power system fault waveform data search device and the recording medium according to the present invention, the power system fault waveform database construction device has an effect that the search of the fault data can be quickly performed, and the voltage / current waveform data is aggregated. It is useful as a power system fault waveform data search device using the fault waveform database, a power system fault waveform data search method, and the like.

10a to 10c Waveform recording devices 1 to 3
11 Network 1
12 Network 2
13a, 13b Terminal device 20 Power system fault waveform data search device 21 Data storage unit 22 Feature extraction unit 23 Data classification unit 24 Data search unit 23a Waveform classification program 24a Data search program 25 Search result data 26 Web application

Claims (5)

Data of the power system fault waveform data collection system that is installed in the power transmission / distribution system, records the voltage / current waveform at the time of the grid fault from the front of the fault for a certain period of time, and collects the data on the server via the communication line. A power system fault waveform data search device comprising a collection,
A data storage unit that stores the collected data, a feature extraction unit that extracts features of the stored data, a data classification unit that classifies data according to the extracted features, a search unit that searches the classified data, A terminal device for inputting search conditions; and a display device for displaying search results.
The data storage unit includes at least the neutral point among the information on the installation location recorded along with the collected accident waveform data, the voltage class, the neutral point grounding method, the short circuit capacity of the system, and the line length. "Information on physical conditions of power transmission and distribution system" including grounding method and short-circuit capacity of the system,
“Information on environmental conditions at the time of data recording” including at least one of the data recording date and weather conditions, and
Accident aspect, the voltage value at the time of the accident, a current value, and the "information obtained by analyzing the accident waveform" including the analysis results of the graphical features of the waveform,
`` Information that can be investigated and acquired after the accident '' including at least one of the accident situation, the cause of the accident, and the recovery method,
Is stored for each accident waveform data,
Each time new data is added to the database, the feature extraction unit converts a pulse waveform component, a sawtooth component, a rectangular component, a harmonic component, a low frequency vibration component, a half wave from the instantaneous value data of the waveform to the waveform. Determine whether or not at least one of the rectified waveform components is included, and the aspect of the accident such as whether the accident is a one-wire ground fault, a two-wire ground fault, a three-wire ground fault, a two-wire short circuit, or a three-wire short circuit Determine
The data classification unit classifies data based on already stored data and the feature data determined by the feature extraction unit, and performs data selection and rearrangement,
Automatically when new accident waveform data is collected or by the user instructing from the terminal device, the "information on physical conditions of the power transmission and distribution system" and the "information on environmental conditions at the time of data recording", The past data is searched based on the “information obtained by analyzing the accident waveform”, the data is sorted in descending order of the number of matching information elements, and the “ A power system accident waveform data search device characterized by outputting information that can be investigated and acquired after an accident to a display device,
When searching for data based on the analysis result of the graphical characteristics of the waveform, if the data is short-circuit accident data including data caused by ground faults with two or more wires through which a short-circuit current flows, the short-circuit capacity of the system will be searched within the same data group A power system fault waveform data search device characterized by narrowing down and searching, and in the case of ground fault data , search is performed by narrowing down search targets within the same data group with the same neutral point grounding method. Specify one of the already collected accident waveform data, and the information related to the physical conditions of the power transmission / distribution system that is recorded along with it, and the information related to the environmental conditions at the time of data recording and "the in the" the information obtained by analyzing the accident waveform "," information that can be retrieved investigated after the accident ", by the user to specify at least one of the information from the terminal device, the 2. The power system fault waveform data search device according to claim 1, wherein other data with matching information is searched, and the information related to the data extracted as a result is output.   The `` information that can be obtained by investigating after an accident '' further includes at least one of an accident occurrence section, a section of a power outage due to an accident, success or failure of a reclosing by a protection relay due to an accident, and presence or absence of a progress accident. The power system fault waveform data search device according to claim 1 or 2. The analysis result of the graphical feature of the waveform includes the sample value of the waveform data, the difference for each sample calculated from the waveform data, the change in the effective value calculated from the moving average analysis value of the data for one period of AC, and From the information of each change in the offset amount calculated from the data for one cycle of alternating current, the pulse-shaped waveform component from the determination result of whether or not they have changed beyond a threshold separately determined within a specific timing range The power system fault according to claim 1, wherein the power system fault is a result of determining whether at least one of a sawtooth component, a rectangular wave component, a harmonic component, a low-frequency vibration component, and a half-wave rectified waveform component is included. Waveform data search device. A recording medium that records a program that causes a computer to execute the function of the search unit of the power system fault waveform data search device according to any one of claims 1 to 4 , and that can be processed by the computer.