Distribution line accident cause estimation device and distribution line accident cause estimation program

Abstract

To assist in quickly and accurately identifying contents of a distribution line accident.SOLUTION: A device estimates, when a distribution line accident occurs, contents of the distribution line accident. The device includes: means for acquiring accident status information and accident characteristic information (blackout information acquisition task 151); and means for estimating the contents of the distribution line accident based on the accident status information and the accident characteristic information (accident cause estimation task 152). The accident status information includes a line form, district characteristics, an output voltage meter value, a ground-fault indication meter value, and a zero-phase voltage value. The accident characteristic information includes an aspect and a dynamic state of the distribution line accident.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a power distribution line accident cause estimating device and a power distribution line accident cause estimating program, and more particularly to a technique for estimating the cause and location of a power distribution line accident when the accident occurs. [Background technology]

[0002] A conventional technology to support facility management is a facility management support system that allows the inspection content to be changed depending on the condition of the on-site facility, and that aims for more convenient operation by reviewing the content of initial plan templates for inspection and maintenance work on power transmission and distribution facilities based on the work knowledge of veteran operators (see Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2016-189088 A Summary of the Invention [Problem to be solved by the invention]

[0004] In order to analyze and investigate the cause of power distribution line accidents, visual inspections and measurements using accident point detection devices have been conducted in the past. However, there are many cases where the cause of a power distribution line accident cannot be identified immediately, and recovery takes time. Furthermore, if it takes a long time to investigate the cause of a power distribution line accident, there is also the problem that a similar accident will occur again.

[0005] In addition, the values ​​of the output voltage meter and earth fault indicator meter of the fault location detection device used when detecting the fault location of a distribution line fluctuate depending on rainy weather or the zero-phase voltage value, and show various variations / movements depending on the type of accident cause. Therefore, the analysis and investigation of the cause of distribution line accidents is actually done blindly based on past experience and knowledge, which is not efficient and requires a great deal of effort and time.

[0006] Therefore, an object of the present invention is to provide a power distribution line accident cause estimating device and a power distribution line accident cause estimating program that are capable of supporting rapid and accurate identification of the details of a power distribution line accident. [Means for solving the problem]

[0007] In order to solve the above problems, the invention of claim 1 is a device for estimating the details of a distribution line accident when the accident occurs, comprising: a means for acquiring accident situation information and accident characteristic information; and a means for estimating at least one of the cause, location, and type of distribution equipment of the distribution line accident based on the accident situation information and the accident characteristic information, wherein the accident situation information includes a line configuration, a district characteristic, an output voltage meter value, a ground fault indication meter value, and a zero-phase voltage value, and the accident characteristic information includes a state of the distribution line accident and a dynamic state regarding the change in the state. The types of the behavior of the power distribution line accident include: multiple consecutive power outages, a micro-ground fault that transitions to a short-term power outage, a short-term power outage that is once recovered and then immediately transitions to a long-term power outage, spontaneous resolution, and no change. The present invention relates to an apparatus for estimating the cause of an accident in a power distribution line.

[0008] The invention of claim 2 is characterized in that, in the power distribution line accident cause estimation device described in claim 1, the manner in which the power distribution line accident occurs includes, as types of manner, the occurrence of a long-term power outage, the occurrence of a short-term power outage, and the occurrence of a micro-earth fault.

[0010] Claim 3 The invention of claim 1 OR 2In the distribution line accident cause estimation device described in the above, the process of estimating at least one of the cause, location, and type of distribution equipment of the distribution line accident is performed using a trained model that has been machine-learned based on past performance data so that, when the accident situation information and the accident characteristic information are input, the at least one of the cause, location, and type of distribution equipment of the distribution line accident is output.

[0011] Claim 4 The present invention is a program for estimating the details of a power distribution line accident when the accident occurs, and causes a computer to function as a means for acquiring accident situation information and accident characteristic information, and a means for estimating at least one of the cause, location, and type of power distribution equipment of the power distribution line accident based on the accident situation information and the accident characteristic information, the accident situation information including a line configuration, a district characteristic, an output voltage meter value, a ground fault indication meter value, and a zero-phase voltage value, and the accident characteristic information including a state of the power distribution line accident and a state of change of the state of the accident. The types of the behavior of the power distribution line accident include: multiple consecutive power outages, a micro-ground fault that transitions to a short-term power outage, a short-term power outage that is once recovered and then immediately transitions to a long-term power outage, spontaneous resolution, and no change. The present invention relates to a power distribution line accident cause estimation program.

[0012] Claim 5 The invention is as follows: 4 In the power distribution line accident cause estimating program described in the above, the mode of the power distribution line accident includes, as types of modes, the occurrence of a long-term power outage, the occurrence of a short-term power outage, and the occurrence of a micro-earth fault.

[0014] Claim 6 The invention is as follows: 4 or 5 In the distribution line accident cause estimation program described in the above, the process of estimating at least one of the cause, location, and type of distribution equipment of the distribution line accident is performed using a trained model that has been machine-learned based on past performance data, so that when the accident situation information and the accident characteristic information are input, the at least one of the cause, location, and type of distribution equipment of the distribution line accident is output. Effect of the Invention

[0015] Claim 1 and Claim 4 According to the invention, the line configuration and area characteristics are taken into consideration as the accident situation information, and the type and dynamics of the distribution line accident are taken into consideration as the accident characteristic information, so it is possible to accurately estimate the content of the distribution line accident, and in turn, it is possible to improve the reliability of the estimation technology for the content of the distribution line accident. 4 According to the present invention, the details of a power distribution line accident can be accurately estimated, which supports rapid and accurate identification of the details of the power distribution line accident. This in turn makes it possible to significantly reduce and shorten the effort and time required to analyze and investigate the details of the power distribution line accident, as well as to significantly shorten the time required to restore the power distribution line from the accident, thereby minimizing the impact on consumers.

[0016] Claim 2 and Claim 5 According to the invention, specific types of distribution line accidents include (in other words, take into consideration) the occurrence of long-term power outages, short-term power outages, and micro-earth faults, making it possible to more accurately estimate the content of the distribution line accident.

[0017] Also, Claim 1 or claims 4 According to the invention, the conductors of a power distribution line accident specifically include (in other words, take into consideration) the occurrence of multiple consecutive power outages, a transition from a micro-ground fault to a short-term power outage, a short-term power outage that is once recovered and then immediately transitions to a long-term power outage, natural resolution, and no change, making it possible to more accurately estimate the content of the power distribution line accident.

[0018] Claim 3 or claims 6 According to the invention, the content of a power distribution line accident is estimated according to a trained model constructed by machine learning using accident situation information and accident characteristic information as input and the content of the power distribution line accident (specifically, the cause, location, and type of distribution equipment of the power distribution line accident) as output, making it possible to estimate a more realistic content of the power distribution line accident that reflects actual information on past power distribution line accidents. [Brief description of the drawings]

[0019] [Figure 1] 1 is a functional block diagram showing a schematic configuration of a power distribution line accident cause estimating device according to an embodiment of the present invention; [Diagram 2] 2 is a diagram showing an example of a data layout of an accident situation database of the power distribution line accident cause estimating device of FIG. 1. [Diagram 3] 2 is a diagram showing an example of a data layout of an accident characteristics database of the power distribution line accident cause estimating device of FIG. 1. [Figure 4] 2 is a diagram showing an example of a data layout of an accident cause database of the power distribution line accident cause estimating device of FIG. 1. [Diagram 5] FIG. 2 is a diagram showing an example of information output by a trained model of the power distribution line accident cause estimation device of FIG. 1. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] Hereinafter, the present invention will be described based on the illustrated embodiment.

[0021] 1 is a functional block diagram showing a schematic configuration of a power distribution line accident cause estimating device 1 according to an embodiment of the present invention. The power distribution line accident cause estimating device 1 is a mechanism for estimating the cause of an accident based on the situation at the time of the accident in a power distribution line, and mainly includes an input unit 11, a display unit 12, a storage unit 13, a memory 14, a main task 15, a communication unit 16, and a central processing unit 17 that controls these units.

[0022] The power distribution line accident cause estimation device 1 is configured by installing and executing a control program for the entire device and an application (power distribution line accident cause estimation program 18) for performing various processes related to estimating the details of the power distribution line accident on, for example, a tablet terminal, a personal digital assistant (PDA), or various types of computers such as notebook or desktop computers.

[0023] The input unit 11 is an interface having a function of receiving commands from a user and inputting information to the power distribution line accident cause estimating device 1, and is configured by, for example, a keyboard, a mouse, and a touch panel.

[0024] The display unit 12 has a function of displaying information input via the input unit 11, displaying the processing results of the power distribution line accident cause estimating device 1, and the like, and is configured by, for example, a liquid crystal display.

[0025] The storage unit 13 is a storage area / storage device having a function of storing various information, programs, data, and the like, and is configured by, for example, a hard disk (HDD: abbreviation for Hard Disk Drive) and a ROM (abbreviation for Read Only Memory).

[0026] The storage unit 13 stores a control program for the entire distribution line accident cause estimation device 1 and a distribution line accident cause estimation program 18. The storage unit 13 also stores a trained model 19 for estimating the contents of a distribution line accident (specifically, the cause, location, and type of distribution equipment) based on information related to the distribution line accident when the distribution line accident occurs (referred to as "occurred accident information"), and an accident situation database 131, an accident characteristic database 132, and an accident cause database 133 in which teacher data used for machine learning of the trained model 19 are recorded and accumulated.

[0027] The above databases 131 to 133 may be stored in an external storage device such as a server. In this case, the power distribution line accident cause estimation device 1 may access the external storage device via the communication unit 16 to acquire various data and information, or may be provided with a connection interface for transmitting / receiving / inputting / outputting signals such as data and control commands to / from the external storage device via various signal lines, and may access the external storage device via the connection interface to acquire various data and information.

[0028] The memory 14 is a storage area / storage device that has the function of serving as a working area for temporarily storing data and information generated when the central processing unit 17 executes calculations related to estimating the details of a power distribution line accident, and is composed of, for example, RAM (short for Random Access Memory).

[0029] The main task 15 is a group of tasks for executing various processes related to estimating the details of a power distribution line accident, which are realized by executing a power distribution line accident cause estimation program 18 stored in the storage unit 13.

[0030] The communication unit 16 is a communication interface that has the function of transmitting / receiving / inputting / outputting signals and information transmitted via various wireless / wired communication networks, including, for example, LAN (short for Local Area Network) and WAN (short for Wide Area Network).

[0031] The central processing unit 17 has a function of coordinating and controlling each part that constitutes the power distribution line accident cause estimation device 1, and is configured to include, for example, a central processing unit (CPU: abbreviation for Central Processing Unit), and realizes each function in accordance with the control program stored in the memory unit 13 and the power distribution line accident cause estimation program 18.

[0032] The power distribution line accident cause estimation device 1 is a device that estimates the content of a power distribution line accident when the accident occurs, and includes a means for acquiring accident situation information and accident characteristic information (power outage information acquisition task 151) and a means for estimating the content of the power distribution line accident based on the accident situation information and accident characteristic information (accident cause estimation task 152), in which the accident situation information includes the line configuration, area characteristics, output voltage meter value, earth fault indication meter value, and zero-phase voltage value, and the accident characteristic information includes the nature of the power distribution line accident and the dynamics related to how the nature is changing.

[0033] In addition, the distribution line accident cause estimation program 18 is a program for estimating the contents of a distribution line accident when the accident occurs, and causes the computer to function as a means for acquiring accident situation information and accident characteristic information (power outage information acquisition task 151), and as a means for estimating the contents of the distribution line accident based on the accident situation information and accident characteristic information (accident cause estimation task 152), in which the accident situation information includes the line configuration, area characteristics, output voltage meter value, earth fault indication meter value, and zero-phase voltage value, and the accident characteristic information includes the nature of the distribution line accident and the dynamics regarding how the nature is changing.

[0034] In addition, the distribution line accident cause estimation device 1 and the distribution line accident cause estimation program 18 perform the process of estimating the content of the distribution line accident using a learned model 19 that has been machine-learned based on past performance data so that when accident situation information and accident characteristic information are input, the content of the distribution line accident is output.

[0035] In order to execute various processes related to estimating the details of a power distribution line accident, a power distribution line accident cause estimation program 18 stored in the storage unit 13 is executed to configure a main task 15. The main task 15 includes a power outage information acquisition task 151, an accident cause estimation task 152, and a learning task 153.

[0036] The power outage information acquisition task 151 acquires information related to a power distribution line accident (i.e., accident information) when a power distribution line accident occurs, which is information input by a worker via the input unit 11 or information transmitted from equipment installed in the power distribution system and input via the communication unit 16.

[0037] Then, the power outage information acquisition task 151 transfers the occurred accident information to the accident cause estimation task 152 .

[0038] The accident cause estimation task 152 estimates the details of the power distribution line accident based on the accident information transferred from the power outage information acquisition task 151.

[0039] In the accident cause estimation task 152, when information about an accident that has occurred is input, a trained model 19 that has been trained using machine learning (including techniques also known as deep learning) is used to output information about the details of the power distribution line accident.

[0040] The trained model 19 is created by the learning task 153. The learning task 153 uses as training data the performance information / performance data group on past power distribution line accidents recorded and accumulated in the accident situation database 131, the accident characteristics database 132, and the accident cause database 133 stored in the storage unit 13, and creates the trained model 19 by machine learning various situations that may be related to or affect a power distribution line accident at the time of the accident, and the relationship between the specific characteristics of the power distribution line accident and the details of the power distribution line accident.

[0041] In machine learning, it is considered that a machine learning technique that is particularly suitable for classification may be used, such as a decision tree, an extreme learning machine (ELM), a neural network / deep neural network (DNN), or a support vector machine (SVM).

[0042] As inputs for machine learning, performance information relating to various situations that are considered to be related to or affect a power distribution line accident at the time of the occurrence of the power distribution line accident is recorded and accumulated in an accident situation database 131, performance information relating to specific characteristics of the power distribution line accident is recorded and accumulated in an accident characteristics database 132, and performance information relating to the contents of the power distribution line accident is recorded and accumulated in an accident cause database 133.

[0043] In the accident situation database 131, the accident characteristics database 132, and the accident cause database 133, an accident number common to the databases 131 to 133 is used as identification information (in other words, an ID) for distinguishing and individually identifying each past power distribution line accident from another, and the databases are configured so that the information recorded and accumulated in each of the databases 131 to 133 can be combined and used via the accident number.

[0044] In addition, the various information recorded and accumulated in the accident situation database 131, the accident characteristics database 132, and the accident cause database 133 may be converted into a code, which is a sequence of numbers or letters defined corresponding to each item, and used, or may be represented by a predetermined keyword / character sequence and used.

[0045] In the accident situation database 131, information on various situations that may be related to or affect a past distribution line accident at the time of the occurrence of the past distribution line accident (called "accident situation information") is recorded and accumulated for each past distribution line accident. Specifically, for example, the following information is recorded and accumulated in the accident situation database 131 as the accident situation information (see FIG. 2).

[0046] a) Date This is information about the date on which the power distribution line accident occurred. This item is for reflecting the causal relationship between the season as the background of the power distribution line accident and the accident content in the estimation of the content of the power distribution line accident, and may be only the month or the four seasons, etc.

[0047] b) Time This is information about the time when a power distribution line accident occurred. This item is for reflecting the causal relationship between the time period as the background of the power distribution line accident and the details of the accident in the estimation of the details of the power distribution line accident, and may be a time period divided into about three hours, or a day divided into morning, afternoon, night, late night, early morning, etc.

[0048] c) Weather This is information related to the weather when the power distribution line accident occurred. Specific examples of the types of weather include sunny, cloudy, rain, thunderstorm, and snow. The weather may also include temperature, humidity, and wind speed.

[0049] d) Distribution line This is information for identifying the distribution line in which the accident occurred (for example, an identification number / ID that is given in advance to each distribution line or each section of the distribution line). As the information for this item, for example, information on the power outage section specified by the power distribution automation system may be used. The power distribution automation system remotely monitors and controls the sectioning switches installed in the power distribution system to perform planned switching of the power distribution system and automatic switching of the power distribution system so as to transmit power to a section other than the accident section when a sudden power outage occurs due to a power distribution line accident or the like. When a power outage occurs due to a power distribution line accident, the power distribution automation system transmits high-voltage power distribution line power outage information including information on the power outage section to the power distribution line accident cause estimation device 1.

[0050] e) Line configuration This is information about the form of the power distribution line where the accident occurred. Specific examples of the types of line forms include mountainous areas, coastal areas, urban areas (residential areas, commercial areas), factory areas, and farmland areas.

[0051] f) Line type This is information about the type of power distribution line in which the accident occurred. Specific examples of the type of line include overhead lines, underground lines (cables), and underwater lines (cables).

[0052] g) District characteristics This is information about the characteristics of the natural environment of the area where the power distribution line where the accident occurred is located. Specific types of area characteristics include, for example, salt damage area, lightning damage area, strong wind area, and no characteristics.

[0053] h) Output voltage meter reading This is combined data of the measurement point ID and the output voltage meter value in the fault point detection device when investigating a distribution line accident. The measurement point ID is identification information for distinguishing and individually identifying each measurement point from each other. Note that when multiple output voltage meter values ​​are measured for one distribution line accident, for example, a predetermined number of output voltage meter values ​​are recorded from the measurement point closest to the location of the distribution line accident.

[0054] The fault point detection device is a device used when a ground fault or the like occurs in a distribution line, and specifically, for example, the fault section (in other words, the blackout section) is identified by a distribution automation system that manages the distribution line where the fault occurred, and then a worker goes to the identified section, applies a voltage to the distribution line (specifically, injects a high-voltage DC pulse), and uses a current detector to repeatedly detect the current flowing between the injection point and the fault point in sequence from the high-voltage pulse injection point to the fault point, thereby detecting the fault point. The fault point detection device is provided with an output voltage display meter that displays the pulse voltage value, and a ground fault display meter that displays an indication value proportional to the magnitude of the current flowing between the injection point and the fault point.

[0055] Since the fault point exploration device is usually connected to a utility pole in the outage section to explore the fault point, specifically, identification information for distinguishing and individually identifying each utility pole is used as the measurement point ID.

[0056] i) Earth fault display meter value This is a combination of measurement point IDs and the values ​​of the earth fault meter in the fault point detection device when investigating a distribution line accident. If multiple earth fault meter values ​​are measured for one distribution line accident, a predetermined number of earth fault meter values ​​are recorded from measurement points close to the location of the distribution line accident, for example.

[0057] j) Zero-phase voltage value This is a combination of the measurement device ID and the measurement value of the zero-phase voltage in the power distribution line when a power distribution line accident occurs. The measurement device ID is identification information for individually identifying each measurement device by distinguishing them from each other. The zero-phase-sequence voltage is measured by a measuring device installed on a utility pole, for example. The measurement data of the zero-phase-sequence voltage is periodically transmitted to the power distribution line accident cause estimation device 1 via various wireless / wired communication networks including, for example, a LAN and a WAN.

[0058] k) Overcurrent value This is a combination of the measurement device ID and the measurement value of the overcurrent (i.e., a current greater than the current supplied to the load) when a distribution line accident occurs. The overcurrent is measured by a device for detecting the overcurrent, such as various relays including an overcurrent relay, etc. The measurement data of the overcurrent is periodically transmitted to the power distribution line accident cause estimation device 1 via various wireless / wired communication networks including, for example, a LAN and a WAN.

[0059] Information relating to specific characteristics of each past power distribution line accident (referred to as "accident characteristic information") is recorded and accumulated in the accident characteristic database 132. Specifically, for example, the following information is recorded and accumulated as the accident characteristic information in the accident characteristic database 132 (see FIG. 3).

[0060] l) Aspects This is information on the type of distribution line accident. Specific examples of types of types include situations caused by a distribution line accident, such as a long-term power outage, a short-term power outage, and a micro-ground fault.

[0061] m) dynamics This is information on changes in the nature of power distribution line accidents. Specific types of behavior include, for example, multiple consecutive power outages, a transition from a micro-ground fault to a short-term power outage, a short-term power outage that is restored and then immediately transitions to a long-term power outage, spontaneous resolution, and no change.

[0062] Here, the "state" in l) above is information regarding the initial state / situation when a power distribution line accident occurs, and the "dynamics" is information regarding the change from the initial state / situation.

[0063] Regarding "status," if the initial state / situation at the time the distribution line accident occurred remains unchanged, and the accident is naturally resolved / recovered before any specific restoration measures are taken for the distribution line accident, it is considered to have been "naturally resolved." Conversely, if the initial state / situation at the time the distribution line accident occurred remains unchanged, and the accident is resolved by taking restoration measures for the distribution line accident (i.e., it is not naturally resolved), it is considered to have been "no change."

[0064] The above-mentioned accident situation information and accident characteristic information correspond to information on accidents that have occurred in the past power distribution line accidents.

[0065] Information on the details of each past power distribution line accident is recorded and accumulated in the accident cause database 133. Specifically, for example, the following information is recorded and accumulated in the accident cause database 133 as information on the details of the power distribution line accident (see FIG. 4).

[0066] X)Cause This is information about the causes of power distribution line accidents, and is the cause of the accident confirmed by workers at the accident site when the accident occurred. Specific types of causes include, for example, wind, rain, lightning, corrosion, natural deterioration, maintenance and poor repairs, overload, contact with trees, contact with birds and animals, and man-made accidents (for example, worker negligence or contact with a vehicle).

[0067] Y) Location This is information about the accident point in a power distribution line, and is the accident point confirmed by a worker at the accident site when a power distribution line accident occurs. Specific ways of identifying the location include, for example, the location of the power distribution line where the accident occurred (for example, a combination of pole identification information to identify the power distribution line section divided by two adjacent utility poles) and the facility or equipment where the malfunction occurred (for example, identification information to identify the equipment installed on the utility pole or power distribution system).

[0068] Z) Type This is information about the type of distribution facility involved in the accident, and is the type of faulty distribution facility or equipment that was identified by workers at the accident site when the distribution line accident occurred. Specific examples of types of distribution facilities include overhead lines, underground and underwater wires (cables), utility poles, insulators, transformers, switches, and circuit breakers.

[0069] The trained model 19 is constructed, for example, based on a multi-class classification model, and is created to estimate and output the cause, location, and type of distribution equipment of a power distribution line accident recorded and stored in the accident cause database 133 based on multiple items of input data recorded and stored in the accident situation database 131 and the accident characteristics database 132.

[0070] In other words, the learning task 153 uses the performance information / performance data group on past distribution line accidents recorded and accumulated in the accident situation database 131, the accident characteristics database 132, and the accident cause database 133 as training data to execute a learning process on the cause, location, and type of distribution line accident according to various situations that are thought to be related to or that may affect the distribution line accident (in other words, accident situation information) and the specific characteristics of the distribution line accident (in other words, accident characteristic information), and constructs a learned model 19 based on the performance information / performance data group, so that an accurate estimation result on the cause, location, and type of distribution equipment of the distribution line accident can be obtained when accident information that has occurred is input.

[0071] The learning model used by the learning task 153 uses, for example, machine learning / deep learning using a neural network / deep neural network to create a neural network / deep neural network including an intermediate layer, in which various situations that are considered to be related to or that may affect a distribution line accident (in other words, accident situation information) and specific characteristics of the distribution line accident (in other words, accident characteristic information) are used as an input layer based on historical data on distribution line accidents obtained from the historical information recorded and accumulated in the databases 131 to 133, and information on the cause, location, and type of distribution equipment of the distribution line accident according to the various situations and characteristics of the accident is used as an output layer.

[0072] In this case, a neural network / deep neural network may be created in which the cause, location, and type of distribution equipment of the power distribution line accident are each an output layer, or a neural network / deep neural network may be created in which a combination of two of the cause, location, and type of distribution equipment of the power distribution line accident are each an output layer, or a neural network / deep neural network may be created in which a combination of the cause, location, and type of distribution equipment of the power distribution line accident are each an output layer.In light of the above, the cause of the accident in this invention includes, in addition to the cause of the power distribution line accident itself, the accident point itself in the power distribution line and the power distribution equipment itself that was the subject of the accident.

[0073] The learning model used by the learning task 153 creates an intermediate layer using the causes, locations, and types of distribution equipment of the power distribution line accidents according to the various situations and characteristics of the accidents as actual results as learning data (teaching data), and learns various parameters in the intermediate layer. The learning model used by the learning task 153 learns various parameters in the intermediate layer so as to minimize the error between the output (i.e., the estimated result) of the causes, locations, and types of distribution equipment of the power distribution line accidents according to the various situations and characteristics of the accidents, and the causes, locations, and types of distribution equipment of the power distribution line accidents according to the various situations and characteristics of the accidents contained in the learning data (teaching data) (i.e., actual data).

[0074] The trained model 19 may output the probability of each of the cause, location, and type of distribution equipment of a power distribution line accident as the likelihood of each of the cause, location, and type of distribution equipment of a power distribution line accident. Specifically, the trained model 19 may output information such as that shown in FIG. 5. In FIG. 5, (A) is an example of a case where the cause, location, and type of distribution equipment of a power distribution line accident are each set as an output layer, (B) is an example of a case where two combinations of the cause, location, and type of distribution equipment of a power distribution line accident are set as an output layer (note that both <Estimation result 1> and <Estimation result 2> are output), and (C) is an example of a case where the combination of the cause, location, and type of distribution equipment of a power distribution line accident are set as an output layer.

[0075] Then, the learning task 153 stores the created trained model 19 in the memory unit 13.

[0076] The accident cause estimation task 152 uses the trained model 19 created by the learning task 153 to estimate the content of a newly occurring power distribution line accident.

[0077] Specifically, when a new power distribution line accident occurs, the power outage information acquisition task 151 first acquires information and data on items similar to those recorded and accumulated in the accident situation database 131 and the accident characteristics database 132 (in other words, items adopted by the trained model 19) as accident information.

[0078] The power outage information acquisition task 151 acquires, for example, information on the above items a) date, b) time, c) weather, e) line configuration, f) line type, and g) area characteristics according to the contents input by an operator via the input unit 11. Also, the above item d) information on the distribution line is acquired according to the contents transmitted from the power distribution automation system and input via the communication unit 16 (specifically, high-voltage distribution line power outage information). Also, the above item h) information on the output voltage meter value and i) information on the ground fault display meter value are acquired according to the output voltage meter value and the ground fault display meter value obtained by an (other) operator going to the site and operating a fault point exploration device input by the operator via the input unit 11. Also, the above item j) information on the zero-phase voltage value and k) information on the overcurrent value are acquired according to the contents transmitted from a predetermined device provided in the power distribution system and input via the communication unit 16. Also, the above item l) information on the aspect and m) information on the status are acquired according to the contents input by an operator via the input unit 11.

[0079] Then, the accident cause estimation task 152 estimates and outputs the cause, location, and type of distribution equipment of the power distribution line accident (specifically, for example, the probability that each of them applies) by inputting the occurred accident information transferred from the power outage information acquisition task 151 to the trained model 19. At this time, the accident cause estimation task 152 may list multiple causes, locations, and types of distribution equipment of the power distribution line accident in descending order of the estimated probability.

[0080] Then, the accident cause estimation task 152 causes the display unit 12 to display the estimation result (see FIG. 5; it is optional to display the probability together).

[0081] As described above, according to the distribution line accident cause estimation device 1 and the distribution line accident cause estimation program 18, the accident situation information and the accident characteristic information are input, and the content of the distribution line accident (specifically, the cause, location, and type of distribution equipment of the distribution line accident) is output, and the content of the distribution line accident is estimated according to the trained model 19 constructed by machine learning, so that it is possible to estimate the more realistic content of the distribution line accident reflecting the actual information on the past distribution line accidents, and it is possible to support the rapid and accurate identification of the content of the distribution line accident, which in turn makes it possible to significantly reduce and shorten the effort and time required for analyzing and investigating the content of the distribution line accident, and to significantly shorten the time until the accident is restored, thereby suppressing the impact on consumers. According to the distribution line accident cause estimation device 1 and the distribution line accident cause estimation program 18, in particular, the line configuration and district characteristics are taken into account as the accident situation information, and the form and dynamics of the distribution line accident are taken into account as the accident characteristic information, so it is possible to more accurately estimate the content of the distribution line accident, which in turn makes it possible to improve the reliability of the estimation technology for the content of the distribution line accident.

[0082] According to the distribution line accident cause estimation device 1 and the distribution line accident cause estimation program 18, when a distribution line accident of unknown cause occurs, for example, when a micro-ground fault occurs and then resolves / recovers naturally, the content of the distribution line accident can be estimated. Therefore, it is possible to narrow down the points that should be inspected intensively in a future patrol, and it becomes possible to assist in finding the cause of the distribution line accident, which in turn makes it possible to operate the distribution line stably.

[0083] Although the embodiment of the present invention has been described above, the specific configuration is not limited to the above embodiment, and even if there are design changes and the like within the scope of the gist of the present invention, they are included in the present invention.

[0084] For example, in the above embodiment, the accident cause estimation task 152 (specifically, the trained model 19) estimates three items of the cause, location, and type of distribution equipment as the details of the power distribution line accident, but the accident cause estimation task 152 (trained model 19) may estimate only one or two items of the cause, location, and type of distribution equipment as the details of the power distribution line accident. By estimating at least one of the cause, location, and type of distribution equipment of the power distribution line accident, it is possible to narrow down the points to be checked when a power distribution line accident occurs, so that the analysis and investigation of the cause of the accident on the power distribution line can be efficiently performed, and it is possible to support the rapid and accurate identification of the details of the power distribution line accident. [Explanation of symbols]

[0085] 1 Distribution line accident cause estimation device 11 Input section 12 Display section 13 Storage section 131 Accident Status Database 132 Accident Characteristics Database 133 Accident Cause Database 14. Memory 15 Main Task 151 Power outage information acquisition task 152 Accident Cause Estimation Task 153 Learning Tasks 16 Communications Department 17 Central Processing Unit 18 Power distribution line accident cause estimation program 19 Pre-trained models

Claims

1. A device for estimating details of a power distribution line accident when the power distribution line accident occurs, A means for acquiring accident situation information and accident characteristic information; means for estimating at least one of a cause, a location, and a type of the power distribution facility of the power distribution line accident based on the accident situation information and the accident characteristic information; The fault condition information includes a line configuration, a district characteristic, an output voltage meter value, a ground fault indication meter value, and a zero-phase voltage value, The fault characteristic information includes a state of the power distribution line fault and a dynamic state regarding a change in the state of the fault, The dynamic state of the power distribution line accident includes, as types of dynamic state, a power outage occurring multiple times in succession, a transition from a micro-ground fault to a short-term power outage, a short-term power outage recovering once and then immediately transitioning to a long-term power outage, a natural resolution, and no change; A power distribution line accident cause estimating device characterized by the above.

2. The type of the power distribution line accident includes a long-term power outage, a short-term power outage, and a micro-ground fault, 2. The power distribution line accident cause estimating device according to claim 1 .

3. The process of estimating at least one of the cause, location, and type of the distribution equipment of the power distribution line accident is performed using a trained model that has been machine-learned based on past performance data so that, when the accident situation information and the accident characteristic information are input, the at least one of the cause, location, and type of the power distribution equipment of the power distribution line accident is output.

3. The power distribution line accident cause estimating device according to claim 1 or 2.

4. A program for estimating details of a power distribution line accident when the power distribution line accident occurs, Computer, A means for acquiring accident situation information and accident characteristic information; and and functioning as a means for estimating at least one of a cause, a location, and a type of the power distribution facility of the power distribution line accident based on the accident situation information and the accident characteristic information; The fault condition information includes a line configuration, a district characteristic, an output voltage meter value, a ground fault indication meter value, and a zero-phase voltage value, The fault characteristic information includes a state of the power distribution line fault and a dynamic state regarding a change in the state of the fault, The dynamic state of the power distribution line accident includes, as types of dynamic state, a power outage occurring multiple times in succession, a transition from a micro-ground fault to a short-term power outage, a short-term power outage recovering once and then immediately transitioning to a long-term power outage, a natural resolution, and no change; A power distribution line accident cause estimation program comprising:

5. The type of the power distribution line accident includes a long-term power outage, a short-term power outage, and a micro-ground fault, 5. The power distribution line accident cause estimating program according to claim 4.

6. The process of estimating at least one of the cause, location, and type of the distribution equipment of the power distribution line accident is performed using a trained model that has been machine-learned based on past performance data so that, when the accident situation information and the accident characteristic information are input, the at least one of the cause, location, and type of the power distribution equipment of the power distribution line accident is output.

6. The power distribution line accident cause estimating program according to claim 4 or 5.

Images