Monitoring device, monitoring method, and program

Abstract

This technology provides a mechanism for monitoring power systems and outputting a predetermined value if an abnormality occurs within the power system. [Solution] The monitoring device 2 includes, as its function, an acquisition unit 231 that acquires waveform data from an oscilloscope device 3 that acquires and outputs waveform data of voltage and current of the power system; an inference unit 232 that uses the waveform data acquired by the acquisition unit 231 as input and makes inferences as to whether the operation within the power system is normal, based on a learning model 223 that uses past waveform data and operation status data at the date and time indicated by the time axis of the past waveform data as training data; a determination result output unit 233 that outputs result data indicating that an abnormality has occurred in the power system if the inference result is that an abnormality has occurred in the power system; and a related data output unit 234 that outputs predetermined related data related to the power system based on the inference result.

Description

Technical Field

[0001] The present invention relates to a monitoring device, a monitoring method, and a program that monitor a power system and perform a predetermined output when an abnormality occurs in the power system.

Background Art

[0002] In a power system that supplies power from a power plant or a substation to a general household or other consumers, when a power outage occurs due to a cause such as an accident, it will widely affect general households, business operators of various businesses, etc. Therefore, the power system is monitored to detect abnormalities.

[0003] To monitor such a power system, there is a known monitoring and control device that captures oscilloscope waveform data of the voltage and current of the power system and controls the power system using the operation information of a protection device that operates during an accident of the power system as a trigger (see, for example, Patent Document 1).

[0004] Also, there is a known method for estimating the cause of an accident on a distribution line that detects the waveforms of the voltage and current of the power system, collects data, analyzes the waveforms from multiple viewpoints, inputs the analysis values into a neural network, and identifies the fault phase (accident point) (see, for example, Patent Document 2).

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0006] When monitoring a power system using the monitoring and control device described in Patent Document 1, the person monitoring the system judged the occurrence of an accident or other incident based on oscilloscope waveform data. This judgment could vary from person to person, making it difficult to accurately monitor the occurrence of accidents or other incidents. Therefore, power systems are now monitored using the accident cause estimation method described in Patent Document 2. However, it is important to verify whether the estimation results are correct and to decide on subsequent actions based on those estimation results.

[0007] Therefore, the present invention aims to provide a monitoring device, monitoring method, and program that monitor power systems and output a predetermined output when an abnormality occurs within the power system. [Means for solving the problem]

[0008] To solve the above problems, the invention of claim 1 is a monitoring device for monitoring a power system, comprising: an acquisition unit that acquires waveform data from an oscilloscope device that acquires voltage and current data of the power system and outputs the data as waveform data; an inference unit that takes the waveform data as input and performs inference on whether the operation in the power system is normal or not, based on a learning model that uses past waveform data and operation status data indicating the operation status of relays in the power system at the date and time indicated by the time axis of the past waveform data as training data; a determination result output unit that outputs result data indicating that an abnormality has occurred in the power system if the result of the inference is that an abnormality has occurred in the power system; and a related data output unit that outputs predetermined related data related to the power system with respect to the result of the inference.

[0009] The invention of claim 2 is a monitoring device according to claim 1, wherein the inference unit is the power system This system is characterized by its ability to infer the occurrence of accidents, including short-circuit and ground-fault accidents, within the system.

[0010] The invention of claim 3 is characterized in that, in the monitoring device described in claim 2, the inference unit performs an inference as to whether the duration of the fault in the power system is appropriate, and the related data output unit outputs the result of the inference as to whether the duration of the fault is appropriate.

[0011] The invention of claim 4 is characterized in that the monitoring device described in claim 2 is provided with an operation stop unit that stops operation at the location of a fault within the power system.

[0012] The invention of claim 5 is a monitoring device according to any one of claims 1 to 4, wherein the acquisition unit acquires operating status data indicating the operating status of relays in the power system, and the inference unit takes the waveform data and the operating status data as input to infer the cause of an abnormality when an abnormality occurs in the power system.

[0013] The invention of claim 6 is characterized in that, in the monitoring device described in claim 5, the related data output unit outputs, as related data, data of contact information related to the location where the abnormality occurred in the power system.

[0014] The invention of claim 7 is characterized in that, in the monitoring device described in claim 5, the related data output unit outputs, as related data, captured image data output from an ITV device that monitors the power system and is installed at the location where the abnormality in the power system is occurring.

[0015] The invention of claim 8 is characterized in that, in the monitoring device described in claim 1, the related data output unit outputs the waveform data at the timing when an abnormality occurs in the power system as the related data.

[0016] The invention of claim 9 is a monitoring method for monitoring a power system, which is executed on a computer comprising a processor and memory, characterized in that the processor performs the following steps: acquiring waveform data from an oscilloscope that acquires voltage and current data of the power system and outputs the data as waveform data; using the waveform data as input to perform inference about whether the operation within the power system is normal, based on a learning model which uses past waveform data and operation status data indicating the operation status of relays within the power system at the date and time indicated by the time axis of the past waveform data as training data; outputting result data indicating that an abnormality has occurred within the power system if the result of the inference indicates that an abnormality has occurred within the power system; and outputting predetermined related data related to the power system with respect to the result of the inference.

[0017] The invention of claim 10 is a program for monitoring a power system, which is executed by a computer having a processor and memory, and is characterized in that the processor is made to perform the following steps: acquire waveform data from an oscilloscope that acquires voltage and current data of the power system and outputs the data as waveform data; perform inference on the waveform data as input whether the operation in the power system is normal, based on a learning model which uses past waveform data and operation status data indicating the operation status of relays in the power system at the date and time indicated by the time axis of the past waveform data as training data; output result data indicating that an abnormality has occurred in the power system if the result of the inference indicates that an abnormality has occurred in the power system; and output predetermined related data related to the power system with respect to the result of the inference. [Effects of the Invention]

[0018] According to the inventions described in claims 1, 9, and 10, based on a learning model that uses past waveform data and data on the operation status of relays in the power system at the date and time indicated by the time axis of the waveform data as training data, the system takes voltage and current waveform data of the power system as input, performs inference on whether the operation in the power system is normal or not, and outputs result data indicating that an abnormality has occurred in the power system. In addition, predetermined related data related to the power system is output based on the inference result.

[0019] In other words, based on a learning model, it outputs an inference result regarding whether the operation within the power system is normal or not, based on voltage and current waveform data of the power system, as well as predetermined related data related to the power system. Therefore, it becomes possible to determine whether the operation of the power system is normal or not without human judgment. This prevents accidents from being missed or overlooked due to differing judgments by people, and makes it possible to verify whether the inference result is correct and to understand subsequent actions in response to the inference result.

[0020] According to the invention described in claim 2, inferences are made regarding the occurrence of accidents, including short-circuit accidents and ground fault accidents, within a power system. This makes it possible to prevent the omission or oversight of specific accidents within the power system.

[0021] According to the invention described in claim 3, the system infers whether the duration of an accident within the power system is appropriate and outputs the result. This makes it possible to verify the appropriateness of the response time for a specific accident within the power system.

[0022] According to the invention described in claim 4, the power system is equipped with a stop-operation unit that stops operation at the location of a fault within the power system. This makes it possible to take the necessary action in response to a fault within the power system.

[0023] According to the invention described in claim 5, operation status data indicating the operation status of relays in the power system is acquired, and based on the waveform data and the operation status data as inputs, inferences are made about the causes of abnormalities when an abnormality occurs in the power system. As a result, more accurate inferences can be made about the occurrence of accidents in the power system and the causes of those accidents, making it possible to identify the causes.

[0024] According to the invention described in claim 6, as related data, data of contact information related to the location where an abnormality has occurred in the power system is output. This makes it possible to instruct necessary actions for accidents in the power system.

[0025] According to the invention described in claim 7, as related data, imaging image data output from an ITV device that monitors the power system and is installed at the location where an abnormality has occurred in the power system is output. This makes it possible to instruct necessary actions for accidents in the power system.

[0026] According to the invention described in claim 8, as related data, waveform data at the timing when an abnormality occurs in the power system is output. This makes it possible to verify the validity of the inference results for specific accidents in the power system.

Brief Description of the Drawings

[0027] [Figure 1] It is a block diagram showing the overall configuration of a monitoring system 1 for monitoring a power system using the monitoring device according to Embodiment 1 of the present invention. [Figure 2] It is an image diagram showing an example of waveform data displayed on the oscilloscope device 3 in FIG. 1. [Figure 3] It is a block diagram showing the functions of the monitoring device 2 in FIG. 1. [Figure 4] It is a flowchart showing the procedure of abnormality output processing by the control unit 230 in FIG. 2. [Figure 5] It is a flowchart showing the procedure of learning processing by the control unit 230 in FIG. 2. [Figure 6] Figure 2 is a block diagram showing the machine learning and inference procedures performed by the learning unit 236. [Figure 7] This diagram shows the functions of computer 700 according to Embodiment 2 of the present invention. [Modes for carrying out the invention]

[0028] The present invention will be described below based on the illustrated embodiments.

[0029] (overview) The monitoring device according to an embodiment of the present invention is a device for monitoring a power system and outputting data to notify of the occurrence of an accident or other incident. In a specific embodiment, this monitoring device holds a learning model that uses past voltage and current waveform data of the power system and operation status data showing the operation status of relays within the power system in the past as training data. Based on the learning model, this monitoring device takes the voltage and current waveform data of the power system as input and infers whether the operation within the power system is normal or not. As a result of this inference, if an abnormality has occurred in the power system, it outputs result data indicating that result. It also outputs predetermined related data related to the power system.

[0030] Furthermore, the monitoring device according to the embodiment of the present invention takes waveform data of voltage and current in the power system and operating status data indicating the operating status of relays in the power system as inputs, based on a learning model, and makes inferences as to whether the operation in the power system is normal or not.

[0031] Furthermore, the monitoring device according to the embodiment of the present invention outputs related data such as waveform data at the time the abnormality occurs, captured image data output from the ITV device that monitors the power system, and contact data related to the location where the abnormality is occurring.

[0032] This configuration makes it possible to determine whether the power system is functioning normally without relying on human judgment. This prevents accidents from being overlooked or missed due to differing judgments among individuals, and allows for verification of whether the inference results are correct and for subsequent actions to be taken in response to those results.

[0033] (Embodiment 1) <Structure> Figures 1 to 6 illustrate this embodiment, and Figure 1 is a block diagram showing the overall configuration of a monitoring system 1 for monitoring a power grid using a monitoring device according to Embodiment 1 of the present invention. This monitoring system 1 is a system composed of devices for monitoring a power grid, and is a system used when installed in an electric station, for example, a substation. The monitoring system 1 comprises a monitoring device 2, an oscilloscope device 3, and an ITV device 4. The monitoring device 2, oscilloscope device 3, and ITV device 4 are connected to a communication device via a wired or wireless network and are connected to each other so as to be able to communicate with each other. The monitoring device 2, oscilloscope device 3, and ITV device 4 are connected to a network by communicating with communication equipment such as a wireless base station compatible with communication standards such as 4G, 5G, and LTE (Long Term Evolution), and a wireless LAN router compatible with wireless LAN (Local Area Network) standards such as IEEE (Institute of Electrical and Electronics Engineers) 802.11.

[0034] The monitoring system 1 is configured to acquire operational status data indicating the operating status of relays 6 and CT / PT7 from multiple distributed control devices 5 installed in the power grid. The monitoring device 2 and the distributed control devices 5, and the distributed control devices 5, relays 6 and CT / PT7 are connected in a manner that Each device is connected to a communication device via a wired or wireless network and is interconnected in a way that allows them to communicate with each other.

[0035] Monitoring device 2 is a device that monitors the power system and outputs a signal to notify of the occurrence of accidents or other incidents. Monitoring device 2 may be composed of a computer (desktop, laptop, tablet, etc.), such as a stationary PC (personal computer) or laptop PC, or it may be composed of a server device such as a web server (including a cloud server).

[0036] Oscilloscope device 3 is a device that acquires voltage and current data from the power system and outputs it as waveform data, and is connected to the busbar or transmission / distribution line of the power system.

[0037] Figure 2 is an image illustrating an example of waveform data displayed on the oscilloscope device 3 in Figure 1. The oscilloscope device 3 outputs, for example, the current and voltage values ​​of the a-phase, b-phase, c-phase, and zero-phase of a busbar or power transmission / distribution line as waveform data showing their changes over time, as shown by curves L1 to L8 in Figure 2. The oscilloscope device 3 also has a function to transmit the waveform data to the monitoring device 2 under the control of the monitoring device 2.

[0038] The ITV (Industrial Television) device 4 is a surveillance camera installed in a specific location within a power system facility, such as a substation, to capture images of that specific location within the facility. The ITV device 4 stores, for example, the captured image data of a specific location within the substation. The ITV device 4 also has the function of transmitting the captured image data to the surveillance device 2 under the control of the surveillance device 2.

[0039] The distributed control device 5 is a device that monitors and controls the operating status of relays 6 and CT / PT7 installed in the power system and acquires operating status data. The distributed control device 5 may be composed of computers (desktops, laptops, tablets, etc.) such as stationary PCs and laptop PCs, or it may be composed of server devices such as web servers (including cloud servers).

[0040] Relay 6 is a device installed in a power system to protect various equipment or power lines within the power system. For example, to protect CT / PT7 in the power system from overload power, relay 6 cuts off power when a load exceeding a predetermined level flows through it. Relay 6 is also operated under the control of the distributed control device 5.

[0041] CT / PT7 is a current transformer (CT) installed in a power system that outputs a current signal for measurement corresponding to the current in the busbars or transmission / distribution lines of the power system, or a voltage transformer (PT) that outputs a voltage signal for measurement corresponding to the voltage in the busbars or transmission / distribution lines of the power system. CT / PT7 operates under the control of the distributed control unit 5.

[0042] Figure 3 is a block diagram showing the functions of the monitoring device 2 in Figure 1.

[0043] The monitoring device 2 comprises a communication unit 210, a storage unit 220, and a control unit 230. As shown in Figure 3, the communication unit 210, the storage unit 220, and the control unit 230 are electrically connected by a bus or the like.

[0044] The communication unit 210 is a communication interface for wired or wireless communication, and any communication protocol may be used as long as communication between the units is possible. This communication unit 210 communicates using a communication protocol such as TCP / IP (Transmission Control Protocol / Internet Protocol).

[0045] The memory unit 220 stores programs and input data for executing various control processes and functions within the control unit 230, and includes RAM (Random Access Memory) and ROM (Read Memory including (Only Memory), HDD (Hard Disk Drive), SSD (Solid State Drive), etc. It consists of storage including a drive, DRAM (Dynamic Random Access Memory), etc. The storage unit 220 also stores a waveform database 221, a power system database 222, and a learning model 223.

[0046] The waveform database 221 stores waveform data of past power system voltages and currents acquired by the acquisition unit 231, which will be described later. The waveform data is waveform data that outputs the changes in voltage and current values ​​at predetermined points in the power system over time, and is time-series data like the curves L1 to L8 shown in Figure 2. The waveform database 221 may also store image data of the waveform data, or it may store numerical data in which the voltage and current values ​​of the power system that form the basis of the waveform data are linked to time (or elapsed time from a predetermined time).

[0047] The power system database 222 stores operational status data that shows the operating status of various devices installed in the power system, such as relays. Specifically, the operational status data includes various data necessary to identify the cause of an accident, such as the relay's installation location, operating time, display information on the display panel, accident occurrence information and its cause (short circuit, ground fault, open circuit), accident type (1-phase, 2-phase, 3-phase), and accident duration.

[0048] The learning model 223 is a learning model that takes waveform data of voltage and current in the power system as input and infers whether the operation within the power system is normal or not. It is an inference algorithm that has been trained (learned) using past waveform data and operation status data showing the past operation status of relay 6 in the power system as training data. The learning model 223 can be any machine learning model, deep learning model, or artificial intelligence model that has undergone appropriate training. The learning model 223 is, for example, a learning model that has been trained by extracting features from past waveform data of voltage and current in the power system and operation status data showing the past operation status of relays in the power system, and recognizing patterns such as correlations between them. The learning model 223 may also be a learning model that takes waveform data and operation status data as input and infers the cause of an anomaly when an anomaly occurs in the power system. The learning model 223 may be generated by learning performed as a function of the learning unit 236 described later, or it may be a learning model generated as a result of learning performed by another device. The learning algorithm for generating the learning model will be described later.

[0049] The control unit 230 controls the overall operation of the monitoring device 2 by executing a program stored in the memory unit 220, and is composed of devices including a CPU (Central Processing Unit), MPU (Micro Processing Unit), GPU (Graphics Processing Unit), microprocessor, processor core, multiprocessor, ASIC (Application-Specific Integrated Circuit), and FPGA (Field Programmable Gate Array). The functions of the control unit 230 include an acquisition unit 231, an inference unit 232, a judgment result output unit 233, a related data output unit 234, an operation stop unit 235, and a learning unit 236. The acquisition unit 231, inference unit 232, judgment result output unit 233, related data output unit 234, operation stop unit 235, and learning unit 236 are activated by a program stored in the memory unit 220 and executed by the monitoring device 2.

[0050] The acquisition unit 231 controls the process of acquiring waveform data from the oscilloscope device 3, which acquires and outputs waveform data of voltage and current in the power system. The acquisition unit 231 communicates with the oscilloscope device 3. The unit 210 acquires time-series waveform data such as the curves L1 to L8 shown in Figure 2. The acquisition unit 231 may acquire image data of the waveform data from the oscilloscope 3, or it may acquire numerical data in which the voltage and current values ​​of the power system that form the basis of the waveform data are linked to time (or elapsed time from a predetermined time).

[0051] The acquisition unit 231 may acquire waveform data image data or numerical data from the oscilloscope 3 by accessing the oscilloscope 3 at a predetermined time, by having the oscilloscope 3 transmit data at a predetermined time, or by maintaining a constant connection with the oscilloscope 3. The acquired waveform data may be used for training to generate or update the learning model 223 in the learning unit 236, which will be described later. The acquisition unit 231 also stores the acquired waveform data in, for example, the waveform database 221.

[0052] In certain situations, the acquisition unit 231 may acquire operating status data from the distributed control device 5 that indicates the operating status of relays within the power system.

[0053] The inference unit 232 controls the process of inferring whether the operation within the power system is normal or not, using the waveform data acquired by the acquisition unit 231 as input, based on the learning model 223. The inference unit 232 performs inference based on the inference algorithm of the learning model 223, which uses past waveform data and operation status data indicating the operation status of relays within the power system at the date and time indicated by the time axis of the past waveform data as training data.

[0054] In certain situations, the inference unit 232 may, based on the learning model 223, infer the occurrence of specific faults, including short-circuit faults and ground fault faults, within the power system. The inference unit 232 may, based on the learning model 223, infer whether the duration of a fault within the power system is appropriate or not. In certain situations, the inference unit 232 may, using the waveform data and operating status data acquired by the acquisition unit 231 as input, infer the cause of an abnormality in the power system when an abnormality occurs, based on the learning model 223.

[0055] The inference unit 232 performs the inference regarding whether the operation within the power system is normal or not, based on the learning model 223, as follows: The inference unit 232 recognizes patterns such as correlations between past waveform data and past operation status data of relays 6, etc., stored in the learning model 223, from the waveform data acquired by the acquisition unit 231. At this time, the inference unit 232 extracts features from the waveform data and operation status data and recognizes patterns such as correlations between them. The inference unit 232 then determines whether the recognized waveform data pattern indicates that the operation within the power system is normal or not. Alternatively, the inference unit 232 may extract features from the waveform data and operation status data respectively, recognize patterns such as correlations, and perform inference about the cause of an abnormality in the power system if an abnormality has occurred, based on the learning model 223.

[0056] The determination result output unit 233 controls the process of outputting result data indicating that an abnormality has occurred in the power system if the inference result from the inference unit 232 indicates that an abnormality has occurred in the power system. Specifically, if the operation in the power system determined by the inference unit 232 is abnormal, the determination result output unit 233 outputs a message such as a warning indicating that the operation in the power system is abnormal to an output device (not shown) such as a display provided by the monitoring device 2 as result data indicating that the operation in the power system is abnormal.

[0057] In certain situations, if the inference result from the inference unit 232 indicates that an abnormality has occurred in the power system, the judgment result output unit 233 may output the cause of the abnormality.

[0058] The related data output unit 234 controls the process of outputting predetermined related data related to the power system based on the inference results of the inference unit 232. As a specific example of related data, the related data output unit 234 may output waveform data acquired by the acquisition unit 231 from the oscilloscope 3 at the time when an abnormality occurs in the power system. The related data output unit 234 may also output contact information related to the location of the cause of the abnormality in the power system based on the inference results of the inference unit 232. The related data output unit 234 may also output captured image data output from the ITV device 4 installed at the location of the cause of the abnormality in the power system based on the inference results of the inference unit 232. Furthermore, the related data output unit 234 may output data indicating whether the duration of the accident in the power system when an abnormality occurs in the power system is appropriate or not, based on the inference results of the inference unit 232.

[0059] The operation stop unit 235 controls the process of stopping operation at the location of the fault if the inference result from the inference unit 232 indicates that the operation within the power system is abnormal and a fault has occurred within the power system. The acquisition unit 231 may output waveform data acquired from the oscilloscope 3. Specifically, the operation stop unit 235 transmits a signal to the distributed control device 5 at the location where the fault has occurred to stop the power supply.

[0060] The learning unit 236 uses past waveform data acquired by the acquisition unit 231 and operation status data indicating the past operation status of relay 6 in the power system as training data to perform learning, and controls the process of generating or updating the learning model 223 based on the learning results. At this time, the learning unit 236 may update the existing learning model 223 by retraining, or it may generate a new learning model.

[0061] The learning unit 236 trains (learns) using past voltage and current waveform data of the power system and operation status data showing the operation status of relays within the power system in the past as training data, and records patterns such as correlations between past waveform data and operation status data. At this time, the operation status data may be labeled to indicate whether the operation status was normal or abnormal. Then, the learning unit 236 takes the waveform data as input and generates an inference algorithm through learning that performs inference about whether the operation within the power system is normal or not.

[0062] Furthermore, the learning unit 236 may use waveform data and operating status data as input to generate a learning model 23 that also infers the cause of an anomaly when one occurs in the power system.

[0063] Learning in the learning unit 236 may be performed using any machine learning algorithm, deep learning algorithm, etc., with past waveform data and operating status data as training data. In addition, in the learning unit 236, data processing (annotation) may be performed on the training data, which consists of past waveform data and operating status data, by adding tag information such as the interpretation of what each data represents, for example, whether the operating status was normal or abnormal. Furthermore, when updating the learning model 223, the learning unit 236 may update it by retraining, or it may be configured to update only when the performance of the learning model 223 has improved after a performance evaluation.

[0064] <Processing flow> Referring to Figures 4 to 6, an example of the flow of control processing (monitoring method) by the control unit 230 of the monitoring device 2 will be explained. Figure 4 is a flowchart showing the procedure for abnormal output processing by the control unit 230 in Figure 2.

[0065] As part of the process in step S101, the acquisition unit 231 of the control unit 230 acquires waveform data from the oscilloscope device 3, which acquires and outputs waveform data of voltage and current of the power system. The waveform data acquired in step S101 is waveform data such as curves L1 to L8, which show the time-series changes of voltage and current of the power system, as shown in the example in Figure 2.

[0066] In step S102, the inference unit 232 of the control unit 230 uses the waveform data acquired in step S101 as input to perform inference on whether the operation within the power system is normal or not, based on the learning model 223. Specifically, in step S102, inference is performed based on the inference algorithm of the learning model 223, which uses past waveform data and operation status data indicating the operating status of relays within the power system at the date and time indicated by the time axis of the past waveform data as training data.

[0067] As part of the processing in step S103, the determination result output unit 233 of the control unit 230 outputs result data indicating that an abnormality has occurred in the power system if the inference result in step S102 indicates that an abnormality has occurred in the power system. Specifically in step S103, if the operation in the power system determined by the inference unit 232 is abnormal, a message such as a warning indicating that the operation in the power system is abnormal is output as result data to an output device (not shown) such as a display of the monitoring device 2.

[0068] As part of the processing in step S104, the related data output unit 234 of the control unit 230 outputs predetermined related data related to the power system based on the inference results in step S102. Specifically in step S104, the acquisition unit 231 may output waveform data acquired from the oscilloscope 3 at the time when an abnormality occurs in the power system, or it may output contact data related to the location causing the abnormality in the power system, or it may output captured image data output from the ITV device 4.

[0069] Figure 5 is a flowchart showing the learning process performed by the control unit 230 in Figure 2. The learning process shown in Figure 5 may be performed automatically at predetermined intervals (e.g., quarterly, annually), or it may be performed at any time by user instruction.

[0070] As part of the process in step S201, the acquisition unit 231 of the control unit 230 acquires waveform data such as curves L1 to L8 shown in Figure 2 from the oscilloscope device 3, which acquires and outputs voltage and current waveform data of the power system, similar to the process in step S101.

[0071] As part of step S202, the control unit 230 acquires operation status data indicating the past operation status of the relay 6 in the power system.

[0072] As part of the processing in step S203, the learning unit 236 of the control unit 230 uses the waveform data acquired in step S201 as training data to perform learning, and generates or updates the learning model 223 based on the learning results. Specifically in step S203, the learning unit 236 performs training (learning) using waveform data of past voltage and current of the power system and operation status data showing the operation status of relays in the power system in the past as training data, and records patterns such as correlations between past waveform data and operation status data. Then, the learning unit 236 takes the waveform data as input and generates an inference algorithm through learning that performs inference about whether the operation in the power system is normal or not.

[0073] As part of the process in step S204, the learning unit 236 of the control unit 230 updates the learning model 223 based on the learning results performed in step S203. If a different new learning model is generated, the process in step S204 does not need to be performed. This updates the learning model 223 to the latest learning state.

[0074] The learning and inference procedure by the learning unit 236 will be explained with reference to Figure 6. Figure 6 is a block diagram showing the learning and inference procedure by the learning unit 236 in Figure 3.

[0075] As part of the process in step S301, the acquisition unit 231 of the control unit 230 collects training data for learning. Examples of training data include waveform data of voltage and current in the past power system and operation status data showing the operation status of relays in the power system in the past. The operation status data may also include data showing labeling results indicating whether the operation status was normal or abnormal.

[0076] As part of step S302, the learning unit 236 of the control unit 230 performs learning using an arbitrary machine learning algorithm, deep learning algorithm, etc., with the training data collected in step S301, and generates or updates the learning model 223.

[0077] As part of the process in step S303, the acquisition unit 231 of the control unit 230 acquires waveform data, which is input data for inference.

[0078] As part of the process in step S304, the inference unit 232 of the control unit 230 takes the waveform data acquired in step S303 as input and performs inference on whether or not the operation within the power system is normal.

[0079] As part of the processing in step S305, the determination result output unit 233 of the control unit 230 outputs the inference result from step S304.

[0080] <Effects> According to the monitoring device 2 and monitoring method of Embodiment 1, based on a learning model 223 that uses voltage and current waveform data of the power system as input to infer whether the operation within the power system is normal, waveform data acquired from the oscilloscope device 3 is used as input to infer whether the operation within the power system is normal. If the inference result indicates that an abnormality has occurred within the power system, result data indicating that an abnormality has occurred within the power system is output. In addition, predetermined related data related to the power system is output based on the inference result. Therefore, it becomes possible to understand whether the operation of the power system is normal without relying on human judgment. This prevents accidents from being missed or overlooked due to differing judgments by people, and makes it possible to verify whether the inference result is correct and to understand subsequent actions in response to the inference result.

[0081] Furthermore, the monitoring device 2 and monitoring method perform inference using waveform data acquired from the oscilloscope device 3 and operating status data as input. If the inference result indicates that an abnormality has occurred in the power system, the cause of the abnormality is also output. Therefore, it becomes possible to perform more accurate inferences about the occurrence of accidents in the power system and their causes, and to understand the causes.

[0082] Furthermore, according to the monitoring device 2 and monitoring method, the acquisition unit 231 outputs waveform data acquired from the oscilloscope device 3, contact data related to the location causing the abnormality in the power system, or captured image data output from the ITV device 4 installed at the location causing the abnormality in the power system, at the time an abnormality occurs in the power system. This makes it possible to instruct necessary actions in response to an accident in the power system.

[0083] (Embodiment 2 (Program)) Figure 7 is a functional block diagram showing an example of the configuration of a computer (electronic computer) 700 according to Embodiment 2 of the present invention. The computer 700 includes a CPU 701, a main memory 702, an auxiliary memory 703, and an interface 704.

[0084] Here, we will describe in detail the programs for realizing each function that constitutes the acquisition unit 231, inference unit 232, determination result output unit 233, related data output unit 234, operation stop unit 235, and learning unit 236 according to Embodiment 1. These functional blocks are implemented in the computer 700. The operation of each of these components is stored in auxiliary storage device 703 in the form of a program. The CPU 701 reads the program from auxiliary storage device 703, expands it in main memory device 702, and executes the above processing according to the program. The CPU 701 also allocates memory areas in main memory device 702 corresponding to the above-mentioned memory units according to the program.

[0085] Specifically, the program is a program that uses a computer 700 to implement the following steps: acquiring waveform data from an oscilloscope that acquires voltage and current data of a power system and outputs it as waveform data; using a learning model that uses past waveform data and operation status data indicating the operation status of relays in the power system at the date and time indicated by the time axis of the past waveform data as training data, it takes the waveform data as input and makes inferences as to whether the operation in the power system is normal; if the inference result indicates that an abnormality has occurred in the power system, it outputs result data indicating that an abnormality has occurred in the power system; and outputting predetermined related data related to the power system in relation to the inference result.

[0086] The CPU 701 is hardware for executing the instruction set described in the program, and consists of an arithmetic unit, registers, peripheral circuits, etc. The CPU 701 consists of at least one processor, which is typically a microprocessor, but may also be other types of processors, including MPUs (Micro Processing Units), GPUs (Graphics Processing Units), microprocessors, processor cores, and multiprocessors. At least one processor may be single-core or multi-core. In addition, at least one processor may be a broader type of processor, such as a hardware circuit that performs some or all of the processing (e.g., an FPGA (Field-Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit)).

[0087] The main memory 702 is for temporarily storing programs and data processed by programs, etc., and is a memory such as DRAM (Dynamic Random Access Memory). The main memory 702 consists of one or more memory devices, which are typically composed of main memory devices, and at least one of the memory devices may be volatile memory or non-volatile memory.

[0088] The auxiliary storage device 703 is an example of a non-temporary tangible medium. Other examples of non-temporary tangible media include magnetic disks, magneto-optical disks, CD-ROMs, DVD-ROMs, and semiconductor memory connected via the interface 704. Furthermore, if this program is distributed to computer 700 via a network, the receiving computer 700 may expand the program into main memory 702 and execute the above processing.

[0089] Furthermore, the program may be intended to implement some of the functions described above. Furthermore, the program may be a so-called differential file (differential program) that implements the above-mentioned functions in combination with other programs already stored in the auxiliary storage device 703.

[0090] Although embodiments of this invention have been described in detail above, the specific configuration is not limited to these embodiments, and any design changes, etc., that do not depart from the spirit of this invention are also included. [Explanation of Symbols]

[0091] 1: Monitoring system 2: Monitoring device 3: Oscilloscope 4:ITV equipment 5: Distributed control system 6: Relay 210: Communications Department 220: Storage section 221: Waveform Database 222: Power System Database 223: Learning Model 230: Control Unit 231: Acquisition Department 232: Reasoning part 233: Judgment result output unit 234: Related data output section 235: Operation stop section 236: Learning Department 700: Computer 701: CPU 702: Main memory 703 :Auxiliary storage device 704: Interface

Claims

1. A monitoring device for monitoring power systems, An acquisition unit that acquires waveform data from an oscilloscope device that acquires voltage and current data of the power system and outputs it as waveform data, An inference unit that uses the waveform data as input and performs inference on whether the operation within the power system is normal or not, based on a learning model that uses the past waveform data and operation status data indicating the operation status of relays within the power system at the date and time indicated by the time axis of the past waveform data as training data, If the result of the inference indicates that an abnormality has occurred in the power system, the determination result output unit outputs result data indicating that an abnormality has occurred in the power system. The system includes a related data output unit that outputs predetermined related data related to the power system based on the results of the aforementioned inference, A monitoring device characterized by the following features.

2. The inference unit performs inferences regarding the occurrence of accidents, including short-circuit and ground fault accidents, within the power system. The monitoring device according to feature 1.

3. The inference unit performs an inference as to whether the duration of the fault within the power system is appropriate. The aforementioned related data output unit outputs the result of the inference regarding whether the duration of the accident is appropriate or not. The monitoring device according to feature 2.

4. The monitoring device according to claim 2, further comprising a stop-operation unit for stopping operation at the location of a fault within the power system.

5. The acquisition unit acquires operating status data indicating the operating status of relays within the power system. The inference unit takes the waveform data and the operating status data as input and performs inference about the cause of an abnormality when an abnormality occurs in the power system. A monitoring device according to any one of claims 1 to 4.

6. The related data output unit outputs, as related data, data of contact information related to the location where the abnormality occurred within the power system. The monitoring device according to feature 5.

7. The related data output unit outputs, as related data, captured image data output from an ITV device that monitors the power system and is installed at the location where the abnormality occurred within the power system. The monitoring device according to feature 5.

8. The related data output unit outputs the waveform data at the timing when an abnormality occurs in the power system as related data. The monitoring device according to feature 1.

9. A monitoring method for monitoring a power system, which is executed on a computer having a processor and memory, The aforementioned processor, An oscilloscope device that acquires voltage and current data from the aforementioned power system and outputs it as waveform data. The steps include acquiring the waveform data and A learning model that uses past waveform data and operation status data indicating the operation status of relays within the power system at the date and time indicated by the time axis of the past waveform data as training data, takes the waveform data as input and performs inference as to whether the operation within the power system is normal or not. If the result of the inference indicates that an abnormality has occurred in the power system, the step is to output result data indicating that an abnormality has occurred in the power system. The following steps are performed regarding the results of the inference: outputting predetermined related data related to the power system; A monitoring method characterized by the following features.

10. A program to be executed on a computer equipped with a processor and memory, for monitoring power systems, The aforementioned processor, The step of acquiring waveform data from an oscilloscope that acquires voltage and current data of the power system and outputs it as waveform data, A learning model that uses past waveform data and operation status data indicating the operation status of relays within the power system at the date and time indicated by the time axis of the past waveform data as training data, takes the waveform data as input and performs inference as to whether the operation within the power system is normal or not. If the result of the inference indicates that an abnormality has occurred in the power system, the step is to output result data indicating that an abnormality has occurred in the power system. The procedure involves performing the steps of outputting predetermined related data related to the power system based on the results of the aforementioned inference. A program characterized by the following features.

Images