Power equipment monitoring device, power equipment monitoring method, and power equipment monitoring program

Abstract

To eliminate noise even when the environment changes.SOLUTION: According to an embodiment, the power equipment monitoring device includes a sensor and a noise removal unit. The sensor is attached to power equipment and measures a signal related to partial discharge. The noise removal unit generates a denoising signal by inputting the measured signal into a denoising model. A denoising model is a model learned so that a signal including noise is used as an input and a denoising signal obtained by removing noise is used as an output. The noise removal unit generates the denoising signal using a first denoising model and a second denoising model. The first denoising model is a denoising model learned using a learning data set generated using a signal measured by the sensor. The second denoising model is a denoising model learned using a learning data set generated using signals measured from other power equipment.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] Embodiments of the present invention relate to a power equipment monitoring device, a power equipment monitoring method, and a power equipment monitoring program.

Background Art

[0002] Due to aging, the insulation performance of the insulator on the surface or inside of the power equipment deteriorates. When the insulation performance deteriorates, partial discharge occurs at the deteriorated location. Furthermore, if the deterioration of the insulation performance progresses, insulation breakdown occurs. Therefore, in the maintenance work of power equipment, it is required to determine the presence or absence of partial discharge.

[0003] The waveform of partial discharge can be obtained by measuring it in a laboratory or the like in advance. On the other hand, since noise exists in the environment where the power equipment is installed, in order to determine partial discharge, it is required to remove noise from the measured signal. The noise may change when the environment where the power equipment is installed changes. In this case, it may not be possible to appropriately remove the noise after the change in the environment, and there is a possibility of misjudging the partial discharge.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] The problem to be solved by the present invention is to provide a power equipment monitoring device, a power equipment monitoring method, and a power equipment monitoring program that can remove noise even when the environment changes.

Means for Solving the Problems

[0006] The power equipment monitoring device of this embodiment includes a sensor and a noise reduction unit. The sensor is attached to the power equipment and measures signals related to partial discharge. The noise reduction unit generates a denoised signal by inputting the measured signal to a denoising model. A denoising model is a model that has been trained to take a signal containing noise as input and output a denoised signal from which the noise has been removed. The noise reduction unit generates the denoised signal using a first denoising model and a second denoising model. The first denoising model is a denoising model trained using a training dataset generated using signals measured by the sensor. The second denoising model is a denoising model trained using a training dataset generated using signals measured from other power equipment. [Brief explanation of the drawing]

[0007] [Figure 1] A schematic diagram showing the configuration of the detection system according to the first embodiment. [Figure 2] A schematic block diagram showing the configuration of a power equipment monitoring device according to the first embodiment. [Figure 3] A flowchart illustrating the operation of a power equipment monitoring device according to the first embodiment. [Figure 4] A figure showing an example of recording the determination result according to the first embodiment. [Figure 5] A schematic block diagram showing the configuration of a power equipment monitoring device according to the second embodiment. [Figure 6] A flowchart illustrating the operation of the power equipment monitoring device according to the second embodiment. [Figure 7] A figure showing an example of recording the determination result according to the second embodiment. [Figure 8] A schematic block diagram showing the configuration of a power equipment monitoring device according to the third embodiment. [Figure 9] A flowchart illustrating the operation of a power equipment monitoring device according to the third embodiment. [Figure 10] A figure showing an example of recording the determination result according to the third embodiment. [Figure 11]A flowchart illustrating the operation of a power equipment monitoring device according to the fourth embodiment. [Figure 12] A schematic block diagram showing the configuration of a power equipment monitoring device according to the fifth embodiment. [Figure 13] A flowchart illustrating the model retraining process by a power equipment monitoring device according to the fifth embodiment. [Figure 14] A schematic block diagram showing the configuration of a power equipment monitoring device according to the sixth embodiment. [Figure 15] An example of a results display screen from the power equipment monitoring device of the sixth embodiment. [Figure 16] A figure showing an example of noise source information according to the seventh embodiment. [Figure 17] An example of a results display screen from the power equipment monitoring device of the seventh embodiment. [Modes for carrying out the invention]

[0008] The power equipment monitoring device, power equipment monitoring method, and power equipment monitoring program of the embodiment will be described below with reference to the drawings.

[0009] (First Embodiment) Figure 1 is a schematic diagram showing the configuration of the detection system 1 according to the first embodiment. Detection system 1 detects partial discharges of power equipment M installed at multiple sites S, such as substations. Detects abnormalities. Examples of power equipment M include circuit breakers, disconnectors, current transformers, transformers, gas-insulated switches, generators, motors, reactors, etc. Power equipment M receives high voltage and high current from an external source via a power cable. Power equipment M has a function to cut off the power supply in the event of an abnormality. A grounding electrode is connected to the bottom of the enclosure housing the power equipment M.

[0010] The detection system 1 comprises power equipment monitoring devices 10 installed at multiple sites S. The multiple power equipment monitoring devices 10 are connected to each other via a wide-area network N such as the Internet.

[0011] The power equipment monitoring device 10 includes a sensor 11 and an arithmetic unit 12 installed for each power equipment M. The power equipment monitoring device 10 is provided near the power equipment M. The sensor 11 of the power equipment monitoring device 10 measures fluctuations generated by the operation of the power equipment M, such as electromagnetic waves, sound waves, transient ground currents, and currents on the wall surface derived from electromagnetic waves. Examples of the sensor 11 include a CT (Current Transformer) sensor, a TEV (Transient Earth Voltage) sensor, an AE (Acoustic Emission) sensor, etc. As the sensor 11, one type of sensor may be used, or a combination of multiple types may be used.

[0012] The arithmetic unit 12 detects partial discharges generated from the power equipment M based on the measurement data of the sensor 11. The measurement data represents the waveform of the fluctuation. The arithmetic unit 12 performs preprocessing of the measurement data, noise removal, and determination of the presence or absence of partial discharges according to predetermined parameters. Examples of the parameters include the amplification gain of the measurement data, the detection threshold, the matching threshold, the phase resolution, the band-pass frequency band, etc. For example, the arithmetic unit 12 amplifies the measurement data according to the amplification gain and samples the measurement data according to the phase resolution and the band-pass frequency band. The arithmetic unit 12 removes noise from the preprocessed measurement data and determines the presence or absence of partial discharges based on the detection threshold and the matching threshold.

[0013] FIG. 2 is a schematic block diagram showing the configuration of the power equipment monitoring device 10 according to the first embodiment. The arithmetic unit 12 is an information processing device such as a PC, a smartphone, or a tablet computer. The arithmetic unit 12 determines the presence or absence of the occurrence of partial discharges based on the electrical signals output from the sensor 11. The arithmetic unit 12 includes a processor 121, a main memory 122, a storage 123, and an interface 124.

[0014] Examples of processors 121 include CPUs (Central Processing Units), GPUs (Graphic Processing Units), and microprocessors. In other embodiments, the arithmetic unit 12 may include, in addition to or instead of the above configuration, a custom LSI (Large Scale Integrated Circuit) such as a PLD (Programmable Logic Device). Examples of PLDs include PALs (Programmable Array Logic), GALs (Generic Array Logic), CPLDs (Complex Programmable Logic Devices), and FPGAs (Field Programmable Gate Arrays). In this case, some or all of the functions realized by processor 121, as described later, may be realized by the integrated circuit. Such integrated circuits are also included as examples of processors.

[0015] Examples of storage 123 include magnetic disks, magneto-optical disks, optical disks, and semiconductor memory. Storage 123 may be an internal medium directly connected to the bus of the arithmetic unit 12, or it may be an external medium connected to the arithmetic unit 12 via interface 124 or a communication line. Storage 123 records denoising models for each of the multiple sites S.

[0016] A denoising model is a pre-trained model whose parameters have been learned to take a noisy signal as input and output a denoised signal (hereinafter referred to as a denoising signal). A denoising model is composed of machine learning models such as a neural network model or a decision tree model. The denoising model is trained using a training dataset generated based on signals acquired from sensor 11 during a predetermined training period immediately after the start of operation of the power equipment M. The training period is set to be sufficiently shorter than the period during which partial discharge due to the deterioration of the power equipment M begins to occur. In other words, the probability that signals acquired from sensor 11 during the training period include signals related to partial discharge is low. The training dataset includes, for example, sets of input samples where a signal acquired from sensor 11 is combined with a known partial discharge signal and the resulting partial discharge signal is used as the output sample, and sets of input samples where a signal acquired from sensor 11 is used and a signal with zero intensity is used as the output sample. By using such a training dataset, the parameters of the denoising model are learned to output a denoised signal with the noise removed when a noisy signal is input. Each sample in the training dataset may be, for example, data representing the waveform of an electrical signal, or data representing an electrical signal as a ΦQN pattern. Storage 123 stores not only the denoising model for site S where the power equipment M detected by the power equipment monitoring device 10 is installed, but also denoising models for other sites S. The denoising models for other sites S are received from power equipment monitoring devices 10 at other sites S.

[0017] Interface 124 is connected to multiple sensors 11, a communication device 125, an input device 126, and a display device 127. The communication device 125 communicates with an external communication device via a wide-area network N. The input device 126 accepts user input. Examples of the input device 126 include a touch panel, mouse, and keyboard. The display device 127 displays the calculation results of the arithmetic unit 12 to the user. Examples of the display device 127 include a CRT (Cathode Ray Tube) display, a liquid crystal display, and an organic EL (Electro Luminescence) display.

[0018] When the processor 121 executes the power equipment monitoring program stored in the storage 123, the processor 121 functions as a signal acquisition unit 131, a noise reduction unit 132, and a partial discharge determination unit 133. The signal acquisition unit 131 acquires the signal measured by the sensor 11. The noise reduction unit 132 inputs the signal acquired by the signal acquisition unit 131 into a denoising model stored in the storage 123 to obtain a denoised signal. The partial discharge determination unit 133 analyzes the denoised signal and determines whether or not there is a partial discharge. For example, the partial discharge determination unit 133 determines that there is a partial discharge if the denoised signal contains signal components with an intensity above a predetermined threshold. The partial discharge determination unit 133 may record the partial discharge determination result in the storage 123.

[0019] Figure 3 is a flowchart showing the operation of the power equipment monitoring device 10 according to the first embodiment. The power equipment monitoring device 10 performs the partial discharge determination process shown below at predetermined intervals (for example, every hour).

[0020] The signal acquisition unit 131 acquires the signal measured by the sensor 11 (step S1). Next, the power equipment monitoring device 10 selects one by one from the multiple denoising models stored in the storage 123 (step S2), and performs the following steps S3 and S4 for the selected denoising model.

[0021] The noise reduction unit 132 inputs the signal acquired in step S1 to the denoising model selected in step S2 to obtain a denoised signal (step S3). The partial discharge determination unit 133 analyzes the denoised signal and determines whether or not a partial discharge has occurred (step S4).

[0022] When the presence or absence of partial discharge is determined using each denoising model, the partial discharge determination unit 133 records the determination result in the storage 123 (step S5). Figure 4 is a diagram showing an example of recording of the determination result according to the first embodiment. As shown in Figure 4, the storage 123 records the determination results of the presence or absence of partial discharge using denoising models for multiple sites S for signals measured at the same time on the same day at one site S (sensor 11).

[0023] Thus, the power equipment monitoring device 10 according to the first embodiment can be expected to improve the accuracy of determining the presence or absence of partial discharge by performing noise reduction using denoising models related to multiple sites S. For example, if only a denoising model generated from noise data of a certain site S is used, it may become impossible to make a correct determination if the environmental noise of that site S changes. In this case, if the cause of the change in environmental noise is the additional installation of equipment that acts as a noise source (e.g., inverters or relays), and equivalent equipment has already been installed at other sites S, the noise signals of such equipment are considered to be similar, and it may be possible to correctly remove the noise by using the denoising model of those other sites S.

[0024] While solutions from pre-trained models generated through machine learning are not always correct, using multiple models in an ensemble can be expected to improve the accuracy of the decisions. It is conceivable to construct a single denoising model using noise signals from multiple sites S, but this presents the following challenges. • Training a denoising model from noise signals from multiple sites S requires transmitting the noise signals over a wide-area network N, which incurs a communication load. • It requires processing a large amount of noise signals with large data sizes, resulting in a high computational cost. • Each time a site S is added, the common denoising model needs to be updated for all sites S. Depending on the terms of the contract with Site S, it may be necessary to rebuild the denoising model after removing the noise signals from Site S, which has had its contract terminated. Therefore, as in the power equipment monitoring device 10 according to the first embodiment, by sharing a denoising model related to multiple sites S, a highly accurate denoising signal can be obtained with a small load.

[0025] (Second embodiment) Figure 5 is a schematic block diagram showing the configuration of the power equipment monitoring device 10 according to the second embodiment. In addition to the configuration of the first embodiment, the power equipment monitoring device 10 according to the second embodiment further includes an integrated determination unit 134 as a function of the processor 121. The integrated determination unit 134 makes an integrated determination of whether or not there is partial discharge based on the partial discharge determination results for each of the multiple denoising models recorded in the storage 123. For example, the integrated determination unit 134 determines that there is partial discharge if all of the multiple determination results are determined to be present, and determines that there is no partial discharge if even one of them is determined to be absent. This is because, if the input signal is only noise, if there is even one denoising model that has learned the noise, the data after noise removal using that denoising model will have the signal removed, and the result of the partial discharge determination unit 133 is expected to be determined to be absent.

[0026] Figure 6 is a flowchart showing the operation of the power equipment monitoring device 10 according to the second embodiment. The power equipment monitoring device 10 performs the partial discharge determination process shown below at predetermined intervals (e.g., every hour). The processes from step S1 to step S4 are the same as in the first embodiment. Once determination results are obtained for all denoising models, the comprehensive determination unit 134 performs a comprehensive determination based on the recorded determination results (step S11) and records the result in the storage 123 (step S12). Figure 7 is a diagram showing an example of recording of determination results according to the second embodiment. As shown in Figure 7, the storage 123 records the determination results of whether or not there is a partial discharge using denoising models for multiple sites S for signals measured at the same time on the same day at one site S (sensor 11), and also records the comprehensive determination result for the partial discharge at the site S at the same time on the same day.

[0027] According to the second embodiment, the power equipment monitoring device 10 can assign an overall determination result to the determination result of whether or not there is partial discharge for each denoise signal. This allows the user to easily recognize whether or not there is partial discharge at site S.

[0028] (Third embodiment) Figure 8 is a schematic block diagram showing the configuration of the power equipment monitoring device 10 according to the third embodiment. In addition to the configuration of the second embodiment, the power equipment monitoring device 10 according to the third embodiment further includes an unlearned determination unit 135 as a function of the processor 121. The unlearned determination unit 135 determines whether the signal to be determined is unlearned by the denoising model of the site S. Specifically, the unlearned determination unit 135 determines that a signal is unlearned if, for a given signal, it is determined that there is partial discharge in the denoising signal generated from the denoising model related to the site S, while at the same time, it is determined that there is no partial discharge in the overall determination.

[0029] Figure 9 is a flowchart showing the operation of the power equipment monitoring device 10 according to the third embodiment. The power equipment monitoring device 10 performs the following partial discharge determination process at predetermined intervals (e.g., every hour). The processing from steps S1 to S11 is the same as in the second embodiment. When the comprehensive determination unit 134 performs a comprehensive determination in step S11, the unlearned determination unit 135 determines whether the determination result of partial discharge for the denoising signal generated from the denoising model related to the site S matches the comprehensive determination result (step S21). If the determination result for the denoising signal related to the site S matches the comprehensive determination result (step S21: YES), it is determined that the noise contained in the measured signal has been learned (step S22). On the other hand, if the determination result for the denoising signal related to the site S does not match the comprehensive determination result (step S21: NO), it is determined that the noise contained in the measured signal has not been learned (step S23). The unlearned determination unit 135 then records the determination result in the storage 123 (step S24).

[0030] Figure 10 shows an example of recording the determination result according to the third embodiment. As shown in Figure 10, the storage 123 records the determination result of whether or not there is partial discharge using a denoising model of multiple sites S, the overall determination result, and whether or not the signal has been learned, for signals measured at the same time on the same day at one site S (sensor 11).

[0031] Thus, according to the third embodiment, the determination result of whether the noise generated at site S is unlearned or not is stored in storage 123. The user can recognize the increasing trend of unlearned noise in the determination result of whether or not it is unlearned stored in storage 123. If the amount of unlearned noise is increasing, the user can determine that the denoising model for site S is no longer able to remove the actual noise. As a result, the user can expect to maintain and restore the noise removal accuracy of the denoising model by recollecting noise from site S and retraining the denoising model.

[0032] (Fourth embodiment) In the third embodiment, the power equipment monitoring device 10 records in the storage 123 whether or not the noise is unlearned. In contrast, in the fourth embodiment, the power equipment monitoring device 10 records in the storage 123 any signal that is determined to contain unlearned noise. The configuration of the power equipment monitoring device 10 in the fourth embodiment is the same as in the third embodiment.

[0033] Figure 11 is a flowchart showing the operation of the power equipment monitoring device 10 according to the fourth embodiment. The power equipment monitoring device 10 performs the partial discharge determination process shown below at predetermined intervals (for example, every hour). The processes from step S1 to step S23 are the same as in the third embodiment. If the unlearned determination unit 135 determines in step S23 that the device is unlearned, it records the signal acquired in step S1 in the storage 123 (step S31). Then, in step S24, the unlearned determination unit 135 records the determination result in the storage 123.

[0034] Thus, according to the fourth embodiment, the power equipment monitoring device 10 records the unlearned noise in the denoising model of site S in the storage 123. As a result, when the power equipment monitoring device 10 retrains the denoising model, it can efficiently retrain the denoising model by using the unlearned noise recorded in the storage 123.

[0035] (Fifth embodiment) Figure 12 is a schematic block diagram showing the configuration of the power equipment monitoring device 10 according to the fifth embodiment. In addition to the configuration according to the fourth embodiment, the power equipment monitoring device 10 according to the fifth embodiment further includes a learning unit 136 as a function of the processor 121. The learning unit 136 retrains the denoising model using unlearned noise recorded in the storage 123.

[0036] The power equipment monitoring device 10 according to the fifth embodiment performs the partial discharge determination process shown in Figure 11 at predetermined intervals (e.g., every hour), similar to the fourth embodiment. As a result, unlearned noise is accumulated in the storage 123. Figure 13 is a flowchart showing the model retraining process by the power equipment monitoring device 10 according to the fifth embodiment. The power equipment monitoring device 10 according to the fifth embodiment performs the model retraining process in parallel with the partial discharge determination process shown in Figure 11.

[0037] The learning unit 136 determines whether the conditions for updating the denoising model are met (step S41). The conditions for updating the denoising model may be, for example, that the number of unlearned noise samples recorded in the storage 123 has reached a certain amount, or that a certain period of time has elapsed since the last retraining of the denoising model. If the conditions for updating are not met (step S41: NO), the learning unit 136 does not retrain the denoising model.

[0038] On the other hand, if the update conditions are met (step S41: YES), the learning unit 136 generates a training dataset using the unlearned noise recorded in the storage 123 (step S42). The learning unit 136 generates training datasets such as sets in which the input sample is the unlearned noise with a known partial discharge signal added to it, and the output sample is the unlearned noise, or sets in which the input sample is the unlearned noise and the output sample is a signal with zero intensity. The signals recorded in the storage 123 as unlearned noise are those that have been determined to have no partial discharge in the overall judgment, so there is a high probability that they will contain only noise. Furthermore, the unlearned noise used in the training dataset may be accumulated since the previous training of the denoising model, or it may be a predetermined number of the most recent ones.

[0039] The learning unit 136 replicates the current denoising model for site S stored in storage 123, and uses the parameters of the replicated denoising model as initial values ​​to train the parameters of the denoising model using the generated training dataset (step S43). During this training process, the power equipment monitoring device 10 can perform the partial discharge determination process shown in step S13 using the original denoising model. Once parameter training is complete, the learning unit 136 updates the current denoising model stored in storage 123 with the retrained denoising model (step S44).

[0040] Thus, according to the power equipment monitoring device 10 of the fifth embodiment, the denoising model can be kept up-to-date and capable of responding to changes in noise in the vicinity of the power equipment M by automatically retraining the denoising model.

[0041] (Sixth embodiment) Figure 14 is a schematic block diagram showing the configuration of the power equipment monitoring device 10 according to the sixth embodiment. The power equipment monitoring device 10 according to the sixth embodiment includes a display control unit 137 in addition to the configuration of the power equipment monitoring device 10 according to the fifth embodiment. The display control unit 137 displays at least one of the waveforms (time waveform or frequency waveform) of the denoising signal generated by the noise removal unit 132 and the determination result by the partial discharge determination unit 133 on the display device 127 in association with the corresponding denoising model.

[0042] Figure 15 shows an example of the results display screen of the power equipment monitoring device 10 according to the sixth embodiment. The results display screen shown in Figure 15 includes the denoised signal and partial discharge judgment result for each denoising model, as well as the overall judgment result. In the example shown in Figure 15, it can be seen that the denoising model, which was learned using past noise at site 001, was unable to remove the noise from the signal detected at site 001. On the other hand, it can be seen that the denoising model, which was learned using past noise at site 002, was able to remove the noise from the signal at site 001.

[0043] Thus, according to the sixth embodiment, the power equipment monitoring device 10 displays the noise reduction results and partial discharge judgment results for each applied denoising model, allowing the user to identify the denoising model that is performing well and to consider the reasons for its success.

[0044] (Seventh Embodiment) The power equipment monitoring device 10 according to the seventh embodiment has the same configuration as the sixth embodiment. On the other hand, the storage 123 of the seventh embodiment stores noise source information, which is information about the noise sources of each site S. Figure 16 is a diagram showing an example of noise source information according to the seventh embodiment. The noise source information stores the type of noise source (e.g., compressor, inverter, relay, etc.) installed at site S, associated with the site S.

[0045] Figure 17 shows an example of the results display screen of the power equipment monitoring device 10 according to the seventh embodiment. In the seventh embodiment, the display control unit 137 displays on the display device 127, in association with the denoising model, noise source information related to the site S where the noise used for training the denoising model was measured, in addition to the denoising signal and the determination result of partial discharge.

[0046] Thus, the power equipment monitoring device 10 according to the seventh embodiment displays noise source information for each applied denoising model, allowing the user to consider the noise source of the noise contained in the signal being diagnosed.

[0047] According to at least one embodiment described above, the power equipment monitoring device 10 includes a sensor 11 and a noise reduction unit 132. The sensor 11 is attached to the power equipment M and measures signals related to partial discharge. The noise reduction unit 132 generates a denoised signal by inputting the measured signal to a denoising model. A denoising model is a model that has been trained to take a signal containing noise as input and output a denoised signal from which the noise has been removed. The noise reduction unit 132 generates a denoised signal using a first denoising model and a second denoising model. The first denoising model is a denoising model trained using a training dataset generated using signals measured by the sensor 11. The second denoising model is a denoising model trained using a training dataset generated using signals measured from other power equipment M. As a result, the power equipment monitoring device 10 can remove noise from the measured signal even if the environment of the power equipment M changes.

[0048] While several embodiments of the present invention have been described, these embodiments are presented as examples only and are not intended to limit the scope of the invention. These embodiments can be carried out in a variety of other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents.

[0049] For example, in other embodiments, a power equipment monitoring device 10 is provided for each site S, but is not limited to this. For example, in other embodiments, some or all of the functions of the power equipment monitoring device 10 may be implemented by an information processing device connected to a wide area network N. For example, the information processing device may centrally manage the denoising models of multiple sites S, and each power equipment monitoring device 10 may download and use the denoising models of other sites S from the information processing device. In this case, if the power equipment monitoring device 10 updates the denoising model, it uploads the updated denoising model to the information processing device.

[0050] In the embodiments described above, the partial discharge determination unit 133 determines the presence or absence of partial discharge by comparing the intensity of the signal component in the denoised signal with a threshold, but is not limited to this. For example, the partial discharge determination unit 133 in other embodiments may determine the presence or absence of partial discharge using a trained machine learning model (partial discharge determination model). The training dataset used to train the partial discharge determination model consists of a denoised signal as the input sample and a bit value indicating the presence or absence of partial discharge as the output sample. The training dataset may include, for example, a set in which a known partial discharge signal is used as the input sample and a bit value "1" indicating the presence of partial discharge is used as the output sample, or a set in which a signal denoised by a denoising model is used as the input sample and a bit value indicating the presence or absence of partial discharge in that signal is used as the output sample. By using such a training dataset, the parameters of the partial discharge determination model are trained to output the probability that partial discharge exists when a denoised signal is input. [Explanation of Symbols]

[0051] 1...Detection system 10...Power equipment monitoring device 11...Sensor 12...Calculation unit 121...Processor 122...Main memory 123...Storage 124...Interface 125...Communication device 126...Input device 127...Display device 131...Signal acquisition unit 132...Noise reduction unit 133...Partial discharge determination unit 134...Comprehensive determination unit 135...Unlearned determination unit 136...Learning unit 137...Display control unit M...Power equipment N...Wide area network S...Site

Claims

1. A sensor attached to a power equipment to be monitored, which measures a signal related to partial discharge, A noise reduction unit generates a denoised signal by inputting the signal measured by the sensor attached to the power equipment being monitored to a denoising model that has been trained to take a noisy signal as input and output a denoised signal from which the noise has been removed. Equipped with, The noise reduction unit is The denoised signal is generated using a first denoising model, which is a denoising model trained using a training dataset generated using signals measured from the power equipment being monitored, and a second denoising model, which is a denoising model trained using a training dataset generated using signals measured from other power equipment. Power equipment monitoring equipment.

2. The system includes a storage unit that stores multiple denoising models, including the first denoising model and the second denoising model, which are trained using training datasets for each power device generated using signals measured from multiple power devices. The noise reduction unit generates the denoised signal using two or more denoising models from among the multiple denoising models stored in the storage unit. The power equipment monitoring device according to claim 1.

3. Partial discharge determination unit that determines whether or not a partial discharge has occurred from the denoise signal. A power equipment monitoring device according to claim 1 or claim 2, comprising:

4. A comprehensive determination unit determines the presence or absence of partial discharge based on the determination results of whether or not partial discharge occurs for each of the multiple denoising signals generated from different denoising models. The power equipment monitoring device according to claim 3, comprising:

5. If the partial discharge determination unit determines that there is a partial discharge in the denoise signal generated from the first denoise model, and the overall determination unit determines that there is no partial discharge, the unlearned noise determination unit determines that the measured signal is noise unknown to the first denoise model. The power equipment monitoring device according to claim 4, comprising:

6. The system includes a signal storage unit that stores the signal measured by the aforementioned sensor, The unlearned noise determination unit records the signal determined to be an unknown noise in the signal storage unit. The power equipment monitoring device according to claim 5.

7. The system includes a learning unit that learns the parameters of the first denoising model using the aforementioned training dataset. The parameters of the first denoising model are retrained using the training dataset generated using the signal determined to be unknown noise. The power equipment monitoring device according to claim 5 or claim 6.

8. A display control unit that displays a display screen associated with the denoising model used to generate a plurality of denoising signals, each of which is generated by inputting a single signal measured by the sensor attached to the power equipment to be monitored into a different denoising model. A power equipment monitoring device according to any one of claims 1 to 7, comprising:

9. A display control unit that displays a display screen associated with the denoise model used to generate the denoise signal, which determines whether or not a partial discharge occurs for each of a plurality of denoise signals generated by inputting a single signal measured by the sensor attached to the power equipment to be monitored into each of the denoise models. A power equipment monitoring device according to any one of claims 3 to 7, comprising:

10. It is equipped with a noise source memory unit that stores noise source information for each power device, The display control unit displays the noise source information on the display screen in association with the corresponding denoising model. The power equipment monitoring device according to claim 8 or claim 9.

11. A step of acquiring a signal relating to partial discharge measured by a sensor attached to the power equipment to be monitored, The steps include: generating a first denoising signal by inputting the signal measured by the sensor attached to the power equipment being monitored into a first denoising model that has been trained to take a signal containing noise as input and output a denoising signal from which the noise has been removed, using a training dataset generated using the signal measured from the power equipment being monitored; The steps include: generating a second denoising signal by inputting the signal measured by the sensor attached to the power equipment being monitored into a second denoising model, which has been trained to take a noisy signal as input and output a denoised signal from which the noise has been removed, using a training dataset generated using signals measured by sensors attached to other power equipment; A method for monitoring power equipment.

12. A power equipment monitoring device comprising a sensor attached to the power equipment to be monitored for measuring signals related to partial discharge, and a computing device, wherein the computing device is A denoising model, which has been trained to take a noise-containing signal as input and output a denoised signal from which the noise has been removed, is made to function as a noise reduction unit that generates a denoised signal by inputting the signal measured by the sensor attached to the power equipment to be monitored. The noise reduction unit generates the denoised signal using a first denoising model, which is a denoising model trained using a training dataset generated using signals measured from the power equipment being monitored, and a second denoising model, which is a denoising model trained using a training dataset generated using signals measured from other power equipment. program.

Images