Machine learning device and opening / closing operation procedure determination device

Abstract

To provide a machine learning device capable of constructing a learning model for more efficiently determining a switching operation procedure of dividing switches in the event of a distribution line accident, and a switching operation procedure determination device using the learning model.SOLUTION: A machine learning device 1 that builds a learning model to support restoration work in the event of a distribution line accident includes an input data acquisition unit 11 that acquires, as input data, distribution system information including position information of a distribution line in a blackout section of the distribution line accident and position information of a plurality of dividing switches that divide the distribution line, a label acquisition unit 12 that acquires, as a label, the switching operation procedure of the plurality of dividing switches, and a learning model construction unit 14 that constructs the learning model for determining the switching operation procedure of the plurality of dividing switches in the blackout section by performing supervised learning using pairs of input data and labels as teacher data.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present invention relates to a machine learning device and an opening / closing operation procedure determination device.

Background Art

[0002] Conventionally, techniques for assisting restoration work during power distribution line accidents have been known. Patent Documents 1 to 3 describe this type of technology. Patent Document 1 describes a technique for searching for and extracting similar past power distribution line accident data, and obtaining the probability of occurrence of each event represented by information on the root cause, occurrence location, and damage status of a power distribution line accident with respect to the extracted specific power distribution line accident data. Patent Document 2 describes a technique for prioritizing the pole numbers in an accident section where a power distribution line accident has occurred based on repair tickets, weather information, past repair status, etc., and setting a patrol route in the order of the prioritized poles. Patent Document 3 describes a technique for extracting the characteristics of each accident cause using pattern recognition from past quantified accident situation data with known accident causes, and obtaining the accident cause with the highest similarity in light of the characteristics of each accident cause for which the quantified accident situation data has been extracted in advance using pattern recognition.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Patent Document 3

Summary of the Invention

Problems to be Solved by the Invention

[0004] By the way, when a power distribution line accident occurs in a power distribution system where the power distribution line is divided by a plurality of sectional switches, the sectional switches in the power outage section may be opened and closed by remote operation or the like to identify the accident location while gradually advancing the power supply. In this case, it is desired to determine the opening and closing operation procedure of the sectional switch that can quickly detect the accident location and resume power supply to the power outage section other than the section where the accident location exists. However, depending on the operator during the restoration work in the event of a power distribution line accident, it may take time and effort to determine the opening and closing operation procedure due to lack of years of experience and knowledge. The techniques described in Patent Documents 1 to 3 can predict the cause of the accident and the like, but do not describe the technique for determining specific restoration procedures such as the opening and closing operation procedure of the sectional switch.

[0005] An object of the present invention is to provide a machine learning device capable of constructing a learning model for more efficiently determining the opening and closing operation procedure of a sectional switch during a power distribution line accident, and an opening and closing operation procedure determination device using the learning model.

Means for Solving the Problems

[0006] The present invention relates to a machine learning device for constructing a learning model for assisting restoration work during a power distribution line accident, comprising: an input data acquisition unit that acquires, as input data, power distribution system information including the position information of the power distribution line in the power outage section of the power distribution line accident and the position information of a plurality of sectional switches that divide the power distribution line; a label acquisition unit that acquires the opening and closing operation procedures of the plurality of sectional switches as labels; and a learning model construction unit that constructs a learning model for determining the opening and closing operation procedures of the plurality of sectional switches in the power outage section by performing supervised learning using the pair of the input data and the labels as teacher data.

[0007] The input data acquisition unit further acquires the cause of the power distribution line accident as the input data.

[0008] The present invention also relates to an opening / closing operation procedure determination device using the learning model constructed by the machine learning device, the device comprising: a determination data acquisition unit that acquires determination data including at least the power distribution system information; and an opening / closing operation procedure determination unit that determines the opening / closing operation procedures of a plurality of the sectional switches in a power outage section of a new power distribution line accident based on the determination data and the learning model.

Effects of the Invention

[0009] According to the present invention, it is possible to provide a machine learning device capable of constructing a learning model for more efficiently determining the opening / closing operation procedures of sectional switches during a power distribution line accident, and an opening / closing operation procedure determination device using the learning model.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Modes for Carrying Out the Invention

[0011] Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing the overall configuration of the opening / closing operation procedure determination system 100. FIG. 2 is a schematic diagram showing an example of a power outage section 7 of the power distribution system 9 used by the opening / closing operation procedure determination system 100 according to the present embodiment.

[0012] The opening / closing operation procedure determination system 100 is a system for determining the opening / closing operation procedure of the sectionalizer 6 in the power outage section 7 of a distribution line accident. First, before describing the opening / closing operation procedure determination system 100, an example of the power distribution system 9 used by the opening / closing operation procedure determination system 100 will be described with reference to FIG. 2.

[0013] The power distribution system 9 includes, for example, as shown in FIG. 2, a distribution line 3 provided between the first substation 4 and the second substation 5, a feeder breaker 41 provided in the first substation 4, a feeder breaker 51 provided in the second substation 5, and sectionalizers 6 for dividing the distribution line 3.

[0014] The distribution line 3 shown in FIG. 2 is composed of a single main line 31 extending from the first substation 4 to the second substation 5 and two branch lines 32 and 33 branching from the main line 31.

[0015] The power distribution system 9 shown in FIG. 2 is provided with sectionalizers 61 to 65 which are five sectionalizers 6. Specifically, the sectionalizers 61, 62, and 63 are provided on the main line 31 at predetermined intervals in this order from the first substation 4 toward the second substation 5. The sectionalizer 64 is provided on the branch line 32, and the sectionalizer 65 is provided on the branch line 33. That is, the distribution line 3 is divided into a first section 71 between the feeder breaker 41 and the sectionalizer 61 by the sectionalizers 61 to 65, a second section 72 between the sectionalizers 61, 62, and 64, a third section 73 between the sectionalizers 62, 63, and 65, a fourth section 74 between the sectionalizer 63 and the feeder breaker 51, a fifth section 75 on the branch line 32 on the side opposite to the main line 31 from the sectionalizer 64, and a sixth section 76 on the branch line 33 on the side opposite to the main line 31 from the sectionalizer 65.

[0016] Next, an example of the opening and closing operation procedure of sectional switches 61 to 65 in the power outage section 7 during a distribution line accident will be described with reference to Fig. 2. Note that the opening and closing of feeder breakers 41 and 51 and sectional switches 61 to 65 in the power distribution system 9 shown in Fig. 2 are controlled by a distribution automation system (not shown).

[0017] First, when a distribution line accident occurs in the power distribution system 9 shown in Fig. 2, the feeder breaker 41 is opened by the distribution automation system. As a result, the entire section from the first section 71 to the sixth section 76 of the distribution line 3 is de-energized, and a power outage section 7 occurs.

[0018] When the power outage section 7 occurs, only the sectional switch 63 is opened by the distribution automation system, and the feeder breaker 41 is closed. In the example of Fig. 2, since the accident location X of the distribution line accident is in the sixth section 76, the feeder breaker 41 is opened again. By this operation, it can be confirmed that the accident location X does not exist in the fifth section 75. Then, the sectional switch 63 is opened by the distribution automation system, and the feeder breaker 51 is closed. Thereby, the power supply from the second substation 5 to the fifth section 75 is resumed.

[0019] Next, in addition to the sectional switch 63, the sectional switch 62 is opened by the distribution automation system, and the feeder breaker 41 is closed. Since there is no accident location X in the first section 71, the second section 72, and the fifth section 75 where power is supplied from the first substation 4, the feeder breaker 41 is not opened and the power transmission is successful. By this operation, the sections where the accident location X exists are narrowed down to the third section 73 and the sixth section 76. In a state where the sections where the accident location X exists are narrowed down to the third section 73 and the sixth section 76, the operator goes to the site to identify the location of the accident location X and attaches the accident point exploration device 8 to the third section 73 and the sixth section 76. Then, the accident location X existing in the sixth section 76 is detected by the accident point exploration device 8.

[0020] The section where the accident occurrence location X exists is narrowed down to the third section 73 and the sixth section 76, and the opening / closing operation of the sectional disconnector 6 may be further performed. For example, in addition to the sectional disconnector 63, the sectional disconnector 65 is opened and the sectional disconnector 62 is closed by the distribution automation system. Since there is no accident occurrence location X in the third section 73, the feeder breaker 41 is not opened and power transmission is successful. By this operation, the section where the accident occurrence location X exists is narrowed down to the sixth section 76. Then, the worker goes to the site and attaches the accident point detection device 8 to the branch line 33 of the sixth section 76. In this way, by the opening / closing operation of the sectional disconnector 6, the accident occurrence location X can be specified while gradually advancing the power supply to the power outage section 7.

[0021] There are multiple patterns of the opening / closing operation procedures of the sectional disconnector 6 of the distribution system 9 shown in FIG. 2 in addition to the above-described opening / closing operation procedures. Also, a distribution line accident may occur in the distribution system 9 provided at any location. There are various patterns and combinations of the positions of the main line 31 and the branch lines 32 and 33 of the distribution line 3 within the power outage section 7, the surrounding terrain, etc. For example, when the distribution line 3 exists in an urban area, a mountainous area, along a coast, etc., there are various patterns such as when there is one branch line 32 or 33 branching from the main line 31 or three or more branch lines. In order to discover the accident occurrence location X after a distribution line accident and supply power to the power outage section 7 at an early stage, it is important to select an appropriate pattern of the opening / closing operation procedure for each power outage section 7 occurring in various distribution systems 9.

[0022] Regarding this point, the opening / closing operation procedure determination system 100 of the present embodiment was devised for the purpose of enabling an operator in the restoration work at the time of a distribution line accident to more efficiently determine the opening / closing operation procedure of the sectional disconnector 6 in the power outage section 7.

[0023] Next, the configuration of the opening / closing operation procedure determination system 100 will be described. As shown in FIG. 1, the opening / closing operation procedure determination system 100 includes a machine learning device 1 and an opening / closing operation procedure determination device 2.

[0024] Here, the machine learning device 1 and the opening / closing operation procedure determination device 2 are paired one-to-one and are communicably connected. Although not shown in FIG. 1, the machine learning device 1 and the opening / closing operation procedure determination device 2 may be connected to each other via a network. The network is, for example, a LAN (Local Area Network), the Internet, a public telephone network, or a combination thereof. The specific communication method in the network, whether it is a wired connection or a wireless connection, etc. is not particularly limited. Alternatively, the machine learning device 1 and the opening / closing operation procedure determination device 2 may be directly connected via a connector instead of communicating using a network.

[0025] The machine learning device 1 will be described with reference to FIG. 3. FIG. 3 is a diagram showing the functional blocks of the machine learning device 1.

[0026] The machine learning device 1 constructs a learning model used in the opening / closing operation procedure determination device 2 by supervised learning. The learning model constructed by the machine learning device 1 is for assisting in the restoration work in the power distribution line accident, and is used to determine the opening / closing operation procedure of the sectionalizer 6 in the power outage section 7 of the power distribution line accident. As shown in FIG. 3, the machine learning device 1 includes an input data acquisition unit 11, a label acquisition unit 12, a storage unit 13, and a learning model construction unit 14.

[0027] The input data acquisition unit 11 acquires various information related to the power outage section 7 of the power distribution line accident as input data. Examples of the input data acquired by the input data acquisition unit 11 include power distribution system information in the power outage section 7 of the power distribution line accident and the cause of the power distribution line accident.

[0028] As the distribution system information, it includes the location information of the distribution line 3 and the location information of a plurality of sectional switches 6 that divide the distribution line 3. Examples of the location information of the distribution line 3 include the location of the branch line branching from the main line of the distribution line 3, the terrain information around the distribution line 3, etc. Examples of the terrain information around the distribution line 3 include the type of the environment around the distribution line 3 such as urban areas, mountainous areas, along the coastline, and the positional relationship with objects other than the facilities of the distribution system 9 such as trees. Examples of the location information of the sectional switch 6 include the location information of the sectional switch 6 provided on the distribution line 3. The input data acquisition unit 11 may acquire, for example, the distribution system information of the power outage section 7 of past distribution line accidents from a management system of the distribution system 9 such as a distribution automation system.

[0029] Examples of the cause of the distribution line accident include tree contact, damage by birds and beasts, equipment failure, etc. The cause of the distribution line accident acquired by the input data acquisition unit 11 may be the cause of past distribution line accidents stored in a management system of the distribution system 9 such as a distribution line automation system, or may be the cause of the accident predicted at the time of the distribution line accident.

[0030] The label acquisition unit 12 acquires, as labels, the opening and closing operation procedures of a plurality of sectional switches 6 in the power outage section 7 of the distribution line accident. The label acquisition unit 12 may acquire, for example, the opening and closing operation procedures of the sectional switches 6 in the power outage section 7 of past distribution line accidents as labels from a management system of the distribution system 9 such as a distribution automation system. Note that the opening and closing operation procedure of the sectional switch 6 is information regarding the procedure of the opening and closing operation of the sectional switch 6, that is, which sectional switch 6 among the plurality of sectional switches 6 is turned on and off at what timing.

[0031] Here, the relationship between the distribution system information, which is the input data, and the cause of the distribution line accident, and the opening / closing operation procedure of the sectional disconnector 6, which is the label, will be described. For example, assume that the distribution system 9 indicated by the distribution system information includes a main line passing through an urban area with few trees and a branch line extending from the main line to a mountainous area and surrounded by trees, and the cause of the distribution line accident is tree contact. In this case, since the cause of the distribution line accident is tree contact, it can be predicted that the accident occurrence location X is likely to be on the branch line extending to the mountainous area rather than on the main line. Therefore, for example, the opening / closing operation procedure of the sectional disconnector 6 that can preferentially check the branch line extending to the mountainous area can be used as a label.

[0032] The storage unit 13 stores the distribution system information and the cause of the distribution line accident acquired by the input data acquisition unit 11 as input data, and the opening / closing operation procedure of the sectional disconnector 6 acquired by the label acquisition unit 12 as a label. Specifically, the storage unit 13 stores the distribution system information and the cause of the distribution line accident as a set of data associated with each other. Then, the storage unit 13 stores, as a set of training data, the input data, which is a set of data including the distribution system information and the cause of the distribution line accident associated with each other, and the label, which is the opening / closing operation procedure of the sectional disconnector 6 for the input data, associated with each other. In addition, the storage unit 13 also stores the learning model constructed by the learning model construction unit 14.

[0033] The learning model construction unit 14 constructs a learning model for determining the opening / closing operation procedure of the sectional disconnector 6 in the power outage section 7 of the distribution line accident by performing supervised learning using the set of input data and the label stored in the storage unit 13 as training data. Then, the learning model construction unit 14 transmits the constructed learning model to the opening / closing operation procedure determination device 2.

[0034] The learning model construction unit 14 can be realized, for example, by using a Support Vector Machine (hereinafter also referred to as SVM).

[0035] The opening / closing operation procedure determination device 2 will be described with reference to FIG. 4. FIG. 4 is a functional block diagram of the opening / closing operation procedure determination device 2. The opening / closing operation procedure determination device 2 includes a control unit 20, a storage unit 26, a communication unit 27, and a display unit 28.

[0036] The control unit 20 is a part that controls the entire opening / closing operation procedure determination device 2. By appropriately reading various programs from a storage area such as a ROM, RAM, flash memory, or hard disk drive (HDD) and executing them, various functions in the present embodiment are realized. The control unit 20 may be a CPU. The control unit 20 includes a determination data acquisition unit 21, a prediction data acquisition unit 22, an accident cause prediction unit 23, an opening / closing operation procedure determination unit 24, and an output unit 25.

[0037] The determination data acquisition unit 21 acquires determination data for determining the opening / closing operation procedure of the sectionalizer 6 in the power outage section 7 of a new distribution line accident. The determination data includes at least the distribution system information of the power outage section 7 of the distribution line accident. The determination data acquisition unit 21 includes a distribution system information acquisition unit 211 and an accident cause acquisition unit 212.

[0038] The distribution system information acquisition unit 211 acquires the distribution system information of the power outage section 7 of a new distribution line accident separately from the distribution system information of the power outage section 7 of the training data used during machine learning. Specifically, the distribution system information acquisition unit 211 acquires the position information of the distribution line 3 and the position information of a plurality of sectionalizers 6 that divide the distribution line 3. The distribution system information acquisition unit 211 may acquire the distribution system information of the power outage section 7 of the distribution line accident from a management system of the distribution system 9 such as a distribution automation system when a power outage occurs due to a distribution line accident, for example.

[0039] The accident cause acquisition unit 212 acquires the accident cause of a new distribution line accident separately from the accident cause of the distribution line accident included in the teacher data used in machine learning. Examples of the accident cause of the distribution line accident include tree contact, damage by birds and beasts, equipment failure, etc. The accident cause acquired by the accident cause acquisition unit 212 may be, for example, the accident cause predicted by the operator and input via the input device provided in the opening / closing operation procedure determination device 2, or the accident cause predicted by the accident cause prediction unit 23.

[0040] The prediction data acquisition unit 22 acquires prediction data for the accident cause prediction unit 23 to predict the accident cause of the distribution line accident. The prediction data acquisition unit 22 includes a weather information acquisition unit 221, a time information acquisition unit 222, a zero-phase voltage information acquisition unit 223, and a history information acquisition unit 224.

[0041] The weather information acquisition unit 221 acquires the weather information in the power outage section 7 at the time of the distribution line accident as prediction data. Examples of the weather information include weather, wind speed, temperature, humidity, rainfall, snowfall, number of lightning strikes, etc. These weather information may be the announced values of the weather observation station, or the values obtained from measuring instruments that measure wind speed, temperature, humidity, etc.

[0042] The time information acquisition unit 222 acquires the information regarding the date and time of the distribution line accident as prediction data.

[0043] The zero-phase voltage information acquisition unit 223 acquires the information regarding the zero-phase voltage of the distribution line 3 in the power outage section 7 due to the distribution line accident as prediction data. The information regarding the zero-phase voltage of the distribution line 3 may be acquired from, for example, the management system of the distribution line 3.

[0044] The history information acquisition unit 224 acquires, as prediction data, history information such as felling history information, inspection history information, and nesting history information related to the power distribution system 9 in the power outage section 7 of the distribution line accident. The felling history information is information on the felling history of trees existing around the distribution line 3 or the like in the power outage section 7. More specifically, the felling history information is, for example, the number of years elapsed since felling and the distance between the tree and the facilities of the power distribution system 9 such as the distribution line 3. The inspection history information is information such as the inspection date and time, defective locations, and repair history of the power distribution system 9 in the power outage section 7. The nesting history information is information regarding nests made by birds on facilities such as the distribution line 3 of the power distribution system 9 in the power outage section 7. These history information are acquired, for example, from a management system or the like that manages the power distribution system 9.

[0045] The accident cause prediction unit 23 predicts the accident cause of the distribution line accident based on at least any one of the time zone of the distribution line accident, weather information, the value of the zero-phase voltage of the distribution line 3, and history information such as felling history information acquired by the prediction data acquisition unit 22, and the power distribution system information acquired by the power distribution system information acquisition unit 211. Specifically, the accident cause prediction unit 23 predicts the accident cause of the distribution line accident based on the relationship between the prediction data and the power distribution system information determined in advance based on past records and the accident cause of the distribution line accident. For example, when the weather information at the time of the distribution line accident indicates strong wind, snow accumulation, etc., the power distribution system information indicates that the distribution line 3 exists in a mountainous area, and the felling history information indicates that the number of years elapsed since the felling of the trees around the distribution line 3 is longer than a predetermined period, the accident cause may be predicted as tree contact. Also, for example, when the inspection history information of the distribution line 3 within the power outage section 7 indicates that defects have frequently occurred in the past, the accident cause may be predicted as equipment failure.

[0046] The opening / closing operation procedure determination unit 24 determines the opening / closing operation procedure of the sectional disconnector 6 in the power outage section 7 of the newly occurred distribution line accident based on the distribution system information and the accident cause of the distribution line accident acquired by the determination data acquisition unit 21 and the learning model constructed by the machine learning device 1. For example, when the accident cause of the distribution line accident is a device failure predicted based on the inspection history information, the opening / closing operation procedure determination unit 24 can determine the opening / closing operation procedure of the sectional disconnector 6 that can preferentially check the location predicted to have a device failure.

[0047] The output unit 25 outputs the opening / closing operation procedure of the sectional disconnector 6 in the power outage section 7 of the distribution line accident determined by the opening / closing operation procedure determination unit 24 to the display unit 28. In addition to the opening / closing operation procedure of the sectional disconnector 6, the output unit 25 may also output the accident cause of the distribution line accident predicted by the accident cause prediction unit 23 to the display unit 28.

[0048] The storage unit 26 stores the learning model acquired from the machine learning device 1. Further, the storage unit 26 may use the distribution system information of the power outage section 7 of the distribution line accident and the accident cause of the distribution line accident acquired by the determination data acquisition unit 21 as a set of input data, and store the opening / closing operation procedure of the sectional disconnector 6 determined for the input data as a label.

[0049] The communication unit 27 transmits and receives data to and from the machine learning device 1. For example, the opening / closing operation procedure determination system 100 can transmit the learning model constructed by the learning model construction unit 14 from the machine learning device 1 to the opening / closing operation procedure determination device 2 via the communication unit 27. Conversely, the opening / closing operation procedure determination system 100 can transmit the distribution system information, accident cause, determined opening / closing operation procedure, etc. of the power outage section 7 of the new distribution line accident acquired by the determination data acquisition unit 21 from the opening / closing operation procedure determination device 2 to the machine learning device 1 via the communication unit 27.

[0050] The display unit 28 is a monitor that displays the determination result of the opening / closing operation procedure of the sectional disconnector 6 in the power outage section 7 output from the output unit 25 of the control unit 20.

[0051] Next, the operation during machine learning in the opening / closing operation procedure determination system 100 according to this embodiment will be described. FIG. 5 is a flowchart showing the operation of the machine learning device 1 during this machine learning.

[0052] In step S11, the input data acquisition unit 11 of the machine learning device 1 acquires the power distribution system information of the power outage section 7 of the power distribution line accident as input data, and the cause of the power distribution line accident associated with the power distribution system information.

[0053] In step S12, the label acquisition unit 12 of the machine learning device 1 acquires the opening / closing operation procedure of the sectionalizer 6 in the power outage section 7 of the power distribution line accident as a label.

[0054] In step S13, the learning model construction unit 14 of the machine learning device 1 receives the pair of the input data and the label as teacher data.

[0055] In step S14, the learning model construction unit 14 of the machine learning device 1 executes machine learning using this teacher data.

[0056] In step S15, the learning model construction unit 14 determines whether to end the machine learning or repeat the machine learning. When the learning model construction unit 14 determines to repeat the machine learning (No in step S15), the process returns to step S11. Then, the machine learning device 1 repeats the same operation. On the other hand, when the learning model construction unit 14 determines to end the machine learning (Yes in step S15), the process proceeds to step S16. Note that the condition for ending the machine learning can be arbitrarily determined. For example, when the machine learning is repeated a predetermined number of times, the machine learning may be ended.

[0057] In step S16, the machine learning device 1 transmits the learning model constructed by the machine learning up to that point to the opening / closing operation procedure determination device 2 via a network or the like.

[0058] In addition, the storage unit 13 of the machine learning device 1 stores this learning model. As a result, when the opening / closing operation procedure determination device 2 requests the learning model, the learning model can be transmitted to the opening / closing operation procedure determination device 2. Also, when new teacher data is acquired, further machine learning can be performed on the learning model.

[0059] Next, an example of the process of determining the opening / closing operation procedure by the opening / closing operation procedure determination device 2 will be described with reference to FIG. 6. FIG. 6 is a flowchart showing the flow of the process of determining the opening / closing operation procedure of the sectionalizer 6 in the power outage section 7 by the opening / closing operation procedure determination device 2.

[0060] In step S21, when a new distribution line accident occurs, the distribution system information acquisition unit 211 of the determination data acquisition unit 21 acquires the distribution system information of the power outage section 7 of the distribution line accident.

[0061] In step S22, the prediction data acquisition unit 22 acquires prediction data such as weather information of the power outage section 7 at the time of the distribution line accident, the date and time of the occurrence of the distribution line accident, information on the zero-phase voltage of the distribution line 3, logging history information, inspection history information, and nest-building history information.

[0062] In step S23, the accident cause prediction unit 23 predicts the accident cause of the distribution line accident based on the distribution system information acquired in step S21 and the prediction data acquired in step S22.

[0063] In step S24, the accident cause acquisition unit 212 of the determination data acquisition unit 21 acquires the accident cause predicted in step S23 from the accident cause prediction unit 23.

[0064] In step S25, the opening / closing operation procedure determination unit 24 determines the opening / closing operation procedure of the sectionalizer 6 in the power outage section 7 based on the distribution system information acquired in step S21, the accident cause acquired in step S24, and the learning model transmitted from the machine learning device 1.

[0065] In step S26, the output unit 25 outputs the cause of the distribution line accident predicted in step S23 and the switching operation procedure determined in step S25 to the display unit 28. As a result, the cause of the newly occurred distribution line accident and the switching operation procedure of the sectionalizer 6 in the power outage section 7 of the distribution line accident are displayed on the display unit 28.

[0066] According to the machine learning device 1 or the switching operation procedure determination device 2 of the switching operation procedure determination system 100 according to the present embodiment described above, the following effects can be obtained.

[0067] The machine learning device 1 according to the present embodiment is a machine learning device 1 that constructs a learning model for assisting the restoration work in the event of a distribution line accident. The input data acquisition unit 11 acquires, as input data, distribution system information including the position information of the distribution line 3 in the power outage section 7 of the distribution line accident and the position information of a plurality of sectionalizers 6 that divide the distribution line 3. The label acquisition unit 12 acquires the switching operation procedures of the plurality of sectionalizers 6 as labels. The learning model construction unit 14 constructs a learning model for determining the switching operation procedures of the plurality of sectionalizers 6 in the power outage section 7 by performing supervised learning using the set of the input data and the labels as the training data.

[0068] Thereby, a learning model for more efficiently determining the switching operation procedures of the sectionalizers 6 in the power outage section 7 of the distribution line accident can be constructed. Therefore, the operator who performs the restoration work in the event of a distribution line accident can efficiently obtain the switching operation procedures considering the past performance regardless of his / her years of experience and amount of knowledge.

[0069] In the machine learning device 1 according to the present embodiment, the input data acquisition unit 11 further acquires the cause of the distribution line accident as input data.

[0070] Thereby, since the switching operation procedures of the sectionalizers 6 can be determined in consideration of the cause of the distribution line accident, a learning model for determining the switching operation procedures that can restore the distribution system 9 from the power outage state earlier can be constructed.

[0071] The switching operation procedure determination device 2 according to this embodiment is a switching operation procedure determination device that uses the learning model constructed by the machine learning device 1, and includes a determination data acquisition unit 21 that acquires determination data including at least distribution system information, and a switching operation procedure determination unit 24 that determines the switching operation procedures of a plurality of sectional switches 6 in a power outage section 7 of a new distribution line accident based on the determination data and the learning model.

[0072] Accordingly, by inputting the distribution system information in the power outage section 7 of a new distribution line accident into the learning model of this embodiment, the switching operation procedures of the sectional switches 6 are determined, so that the restoration work in the event of a distribution line accident can be advanced more efficiently.

[0073] Although the embodiments of the present invention have been described above, the present invention is not limited to the above embodiments and can be appropriately modified.

[0074] In the above embodiment, the machine learning device 1 has performed supervised learning by acquiring the distribution system information in the power outage section 7 of the distribution line accident and the cause of the accident of the distribution line accident as input data. However, a configuration may be adopted in which supervised learning is performed using only the distribution system information as input data without acquiring the cause of the accident of the distribution line accident.

[0075] In the above embodiment, the switching operation procedure determination device 2 has acquired the distribution system information in the power outage section 7 of the distribution line accident and the cause of the accident of the distribution line accident as determination data. However, the switching operation procedures of the sectional switches 6 in the power outage section 7 may be determined based on the distribution system information and the learning model without acquiring the cause of the accident of the distribution line accident.

Explanation of Reference Numerals

[0076] 1 Machine learning device 2 Switching operation procedure determination device 3 Distribution line 6 Sectional switch 7 Power outage section 11 Input data acquisition unit 12 Label acquisition unit 14 Learning model construction unit 21 Determination data acquisition unit 24 Opening / closing operation procedure determination unit

Claims

1. A machine learning device for constructing a learning model to assist in restoration work during a distribution line accident, comprising: an input data acquisition unit that acquires, as input data, distribution system information including the position information of the distribution line in the power outage section of the distribution line accident and the position information of a plurality of sectional switches that divide the distribution line, and the cause of the distribution line accident; a label acquisition unit that acquires, as a label, an opening / closing operation procedure of a sectional switch, which is a procedure for restoring from a power outage state caused by the distribution line accident and indicates which sectional switch among the plurality of sectional switches is to be opened / closed at which timing; a learning model construction unit that constructs a learning model for determining the opening / closing operation procedure of the sectional switch in the power outage section by performing supervised learning using a pair of the input data and the label as training data. A machine learning device comprising:

2. An opening / closing operation procedure determination device using the learning model constructed by the machine learning device according to Claim 1, comprising: a determination data acquisition unit that acquires determination data including the distribution system information and the cause of the distribution line accident; an opening / closing operation procedure determination unit that determines the opening / closing operation procedure of the sectional switch in the power outage section of a new distribution line accident based on the determination data and the learning model. An opening / closing operation procedure determination device comprising:

Images