Learning device, insulation diagnosis system and program

Abstract

A learning device according to the present disclosure is a learning device (1) that obtains a first trained model that infers non-destructive electrical characteristics related to the dielectric strength of a main insulating layer of a stator coil of a rotating electric machine using operating information related to the operating history of the rotating electric machine, a learning data acquisition unit (11) that acquires learning data including operating information and nondestructive electrical characteristics obtained when the device is operated under conditions indicated in the operating information; and a model generation unit (12) that generates a first trained model for inferring non-destructive electrical characteristics from operating information using the training data.

Description

[Technical Field]

[0001] The present disclosure relates to a learning device, an insulation diagnosis system, and a program. [Background technology]

[0002] Coil insulation in rotating electrical machines and other devices is subject to various types of degradation, including electrical, thermal, mechanical, and environmental degradation, depending on the operating history of the machine. This causes numerous small voids to form throughout the insulation layer, and larger voids may also form locally. When a high voltage is applied to the voids in the insulation layer, discharge occurs in the voids, which become conductors, shortening the insulation distance and lowering the breakdown voltage. Non-destructive tests such as dielectric loss tangent tests, AC tests, and partial discharge tests are commonly performed to determine the state of insulation degradation.

[0003] A conventional method for estimating the deterioration of coil insulation involves using operational history data and the non-destructive test data as input, converting this into reference data, determining a composite degree of deterioration using fuzzy inference, estimating the remaining dielectric strength rate from the determined composite degree of deterioration, and learning the actual remaining dielectric strength rate obtained from a destructive test to correct the estimated remaining dielectric strength rate, thereby estimating the remaining life. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 5-264645 Summary of the Invention [Problem to be solved by the invention]

[0005] The method for estimating the deterioration of insulators disclosed in the above-mentioned Patent Document 1 requires learning the actual dielectric strength remaining rate obtained from the breakdown voltage, so that, for example, when the insulator is renewed, an actual destructive test must be performed and the actual dielectric strength remaining rate must be obtained from the results. This requires that the rotating electric machine be stopped during operation, which poses a problem of the time and effort required to measure the dielectric strength remaining rate. The present disclosure has been made to solve the above-mentioned problems, and aims to provide a learning device, an insulation diagnosis system, and a program that can diagnose the insulation performance of a rotating electric machine while reducing the effort required to measure the electrical characteristics of the rotating electric machine, such as the dielectric strength remaining rate. [Means for solving the problem]

[0006] The learning device according to the present disclosure includes a first trained model that uses operating information related to the operating history of a rotating electric machine to infer non-destructive electrical characteristics related to the dielectric strength of a main insulating layer of a stator coil of the rotating electric machine. The second trained model uses the above to estimate the residual dielectric strength of the main insulation layer. A learning device for obtaining a data acquisition unit that acquires driving information; and a first inference unit that outputs nondestructive electrical characteristics obtained when the driving information is input using a first trained model; Driving information and non-destructive electrical characteristics obtained by the first inference unit when the conditions indicated in the operation information are input; When the rotating electric machine is operated under the conditions indicated by the operation information Dielectric strength remaining rate a learning data acquisition unit that acquires learning data including the above; and a learning data acquisition unit that acquires learning data including the above from driving information using the learning data. Dielectric strength remaining rate The first step to infer 2 and a model generation unit that generates a trained model. [Effects of the Invention]

[0007] The learning device according to the present disclosure includes: Since non-destructive electrical characteristics can be inferred using the first trained model, it is possible to infer the residual dielectric strength rate using operating information and the inferred values ​​of non-destructive electrical characteristics, eliminating the need to actually measure them. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a block diagram showing a configuration of a learning device according to a first embodiment. [Figure 2] FIG. 2 is a diagram illustrating an example of a neural network of the learning device according to the first embodiment. [Figure 3]4 is a flowchart illustrating processing by the learning device according to the first embodiment. [Figure 4] 1 is a block diagram showing a configuration of an insulation diagnosis system according to a first embodiment. [Figure 5] 4 is a flowchart for explaining the processing of the insulation diagnosis system according to the first embodiment. [Figure 6] FIG. 4 is a block diagram showing another configuration of the insulation diagnosis system in the first embodiment. [Figure 7] FIG. 10 is a block diagram showing the configuration of a learning device that is an element constituting the learning device according to the second embodiment. [Figure 8] 8 is a flowchart for explaining the processing of the learning device shown in FIG. 7. [Figure 9] FIG. 8 is a block diagram showing the configuration of an insulation diagnosis system that utilizes a trained model generated by the learning device shown in FIG. [Figure 10] 10 is a flowchart for explaining the processing of the insulation diagnosis system shown in FIG. [Figure 11] FIG. 8 is a block diagram showing another configuration of an insulation diagnosis system that utilizes a trained model generated by the learning device shown in FIG. [Figure 12] FIG. 10 is a block diagram showing the configuration of a learning device according to a second embodiment. [Figure 13] 10 is a flowchart illustrating processing by a learning device according to the second embodiment. [Figure 14] FIG. 10 is a block diagram showing the configuration of an insulation diagnosis system according to a second embodiment. [Figure 15] 10 is a flowchart for explaining the processing of the insulation diagnosis system according to the second embodiment. [Figure 16] FIG. 1 is a diagram illustrating an example of hardware of a learning device, a learned model storage unit, an inference device, a diagnosis unit, and an insulation diagnosis system. DETAILED DESCRIPTION OF THE INVENTION

[0009] The following describes embodiments in detail with reference to the accompanying drawings. Note that the embodiments described below are merely examples. The embodiments can be implemented in appropriate combination.

[0010] Embodiment 1 <Learning Phase> 1 is a block diagram showing the configuration of a learning device 1 according to embodiment 1. The learning device 1 obtains a first trained model for inferring non-destructive electrical characteristics of the main insulation layer of a stator coil of a rotating electric machine in an insulation diagnosis system for the rotating electric machine. The learning device 1 includes a training data acquisition unit 11 and a model generation unit 12.

[0011] The learning data acquisition unit 11 acquires learning data that is a combination of first input data and second input data. Here, the first input data is operation information related to the operation history of the rotating electric machine. The operation information of the rotating electric machine may include, for example, at least one of the number of starts and stops of the rotating electric machine and the operation time of the rotating electric machine.

[0012] The second input data is non-destructive electrical characteristics related to the dielectric strength of the main insulation layer of the stator coil in the rotating electric machine operated under the conditions indicated by the first input data. The non-destructive electrical characteristics of the main insulation layer may include at least one of an increment in dielectric tangent of the main insulation layer, a first current surge voltage of the main insulation layer, and a maximum partial discharge charge amount of the main insulation layer.

[0013] As insulation degradation progresses and voids increase in the insulation layer, partial discharges occur, causing an increase in the dielectric loss tangent, which serves as an indicator of insulation degradation. Furthermore, in AC current testing, when voltage is applied to a good insulator, the voltage-current curve is linear. However, if the insulation absorbs moisture, increasing its dielectric constant, or if partial discharges occur, the current increase rate increases sharply. When AC current-voltage characteristics are measured, two sharp points in the current increase rate are observed; here, we focus on the first current increase voltage. Furthermore, when measuring the relationship between partial discharge charge and applied voltage, the partial discharge charge increases as the applied voltage increases. However, the maximum partial discharge charge at normal voltage (E / √3, E: rated voltage) indicates localized insulation degradation. As mentioned above, all of the nondestructive electrical characteristics used here are related to the dielectric strength of the main insulation layer. However, other characteristics related to the dielectric strength of the main insulation layer may also be used.

[0014] The learning data acquisition unit 11 may acquire input data from a data input device or database (not shown). The data input device is, for example, a keyboard, mouse, keypad, or touch panel, which is operated by a user. The database is, for example, a storage device, which stores design and operation information of the rotating electric machine.

[0015] The model generation unit 12 learns the nondestructive electrical characteristics of the main insulation layer of the stator coil of a rotating electric machine operated under the conditions indicated by the first input data, based on the learning data, which is a combination of the first input data and the second input data output from the learning data acquisition unit 11. That is, the model generation unit 12 generates a first trained model that infers the nondestructive electrical characteristics of the main insulation layer of the stator coil of a rotating electric machine operated under the conditions indicated by the first input data, from the learning data, which is a combination of the first input data and the second input data. Here, the learning data is data that associates the first input data and the second input data with each other.

[0016] The learning algorithm used by the model generation unit 12 may be any known algorithm, such as supervised learning, unsupervised learning, reinforcement learning, or semi-supervised learning. An example will be described in which a neural network is used. The model generation unit 12 learns the nondestructive electrical characteristics of the main insulating layer, for example, by so-called supervised learning, according to a neural network model. Here, supervised learning refers to a technique in which pairs of input and result (label) data are provided to the learning device 1, and the learning device learns the features of the learning data and infers the result from the input.

[0017] A neural network is composed of an input layer consisting of multiple neurons, an intermediate layer (hidden layer) consisting of multiple neurons, and an output layer consisting of multiple neurons. The intermediate layer may be one layer or two or more layers. Figure 2 is a diagram showing an example of a neural network of the learning device 1 according to the first embodiment. For example, in a three-layer neural network as shown in Figure 2, when multiple inputs are input to the input layer (X1 to X3), the values ​​are multiplied by a weight w1 (w11 to w16) set according to the transmission path and input to the intermediate layer (Y1 to Y2), and the result is further multiplied by a weight w2 (w21 to w26) set according to the transmission path and output from the output layer (Z1 to Z3). This output result varies depending on the values ​​of the weights w1 and w2.

[0018] In the learning device 1 according to the first embodiment, the neural network uses learning data created based on a combination of first input data and second input data acquired by the learning data acquisition unit 11 to learn, through so-called supervised learning, the nondestructive electrical characteristics of the main insulation layer of a stator coil of a rotating electric machine operated under conditions related to the operating information of the rotating electric machine indicated by the first input data. That is, the neural network learns by inputting the first input data, which is the operating information of the rotating electric machine, into the input layer and adjusting the weights w1 and w2 so that the results output from the output layer approach the nondestructive electrical characteristics of the main insulation layer of the stator coil of a rotating electric machine operated under the conditions indicated by the first input data. The model generation unit 12 generates and outputs a first trained model by executing the above-described learning. The trained model storage unit 2 stores the first trained model output from the model generation unit 12.

[0019] Next, we will explain the learning process of the learning device 1. Figure 3 is a flowchart for explaining the process of the learning device 1 and the trained model storage unit 2 according to embodiment 1. Step S01 is a learning data acquisition step, step S02 is a learning process step, and step S03 is a trained model storage step.

[0020] In step S01, the training data acquisition unit 11 acquires training data and outputs it to the model generation unit 12, and then the process proceeds to step S02. Note that the training data acquisition unit 11 may acquire the first input data and the second input data included in the training data simultaneously, or may acquire the first input data and the second input data at different times. In short, it is sufficient if the first input data and the second input data can be acquired in association with each other.

[0021] In step S02, the model generation unit 12 uses learning data created based on a combination of the first input data and the second input data acquired by the learning data acquisition unit 11 to learn the nondestructive electrical characteristics of the main insulation layer of the stator coil in a rotating electric machine operated under the conditions indicated by the first input data by so-called supervised learning. The model generation unit 12 then generates a first trained model and outputs the generated first trained model to the trained model storage unit 2, and the process proceeds to step S03. In step S03, the trained model storage unit 2 stores the first trained model acquired from the model generation unit 12, and the process ends.

[0022] <Utilization phase> Next, an insulation diagnosis system for a rotating electric machine according to the first embodiment will be described. FIG. 4 is a block diagram showing the configuration of an insulation diagnosis system 100 for a rotating electric machine according to the first embodiment. The insulation diagnosis system 100 for a rotating electric machine diagnoses the insulation performance of the rotating electric machine. The insulation diagnosis system 100 for a rotating electric machine includes a trained model storage unit 2, an inference device 3, and a diagnosis unit 4. The inference device 3 is composed of a data acquisition unit 31 and a first inference unit 32. The trained model storage unit 2, the inference device 3, and the diagnosis unit 4 may be configured on the same hardware, or may be separate devices connected via a network. The trained model storage unit 2, the inference device 3, and the diagnosis unit 4 may reside on a cloud server.

[0023] The data acquisition unit 31 of the inference device 3 acquires, as input data, operating information of the rotating electric machine when diagnosing the insulation performance of the rotating electric machine. The operating information of the rotating electric machine may include, for example, at least one of the number of starts and stops of the rotating electric machine and the operating time of the rotating electric machine. The data acquisition unit 31 may acquire the input data from a data input device or database (not shown). The data input device is, for example, a keyboard, mouse, keypad, touch panel, etc., which is operated by a user. The database is, for example, a storage device, etc., which stores operating information of the rotating electric machine.

[0024] The first inference unit 32 infers the nondestructive electrical characteristics of the main insulation layer under the conditions indicated by the input data, using the first trained model stored in the trained model storage unit 2. That is, by inputting the input data acquired by the data acquisition unit 31 into this first trained model, the first inference unit 32 can infer the nondestructive electrical characteristics of the main insulation layer of the stator coil in a rotating electric machine operated under the conditions indicated by the input data, specifically infer at least one of the increase in the dielectric tangent of the main insulation layer, the first current sudden increase voltage of the main insulation layer, and the maximum partial discharge charge amount of the main insulation layer, and output the inference result.

[0025] In this embodiment, the first inference unit 32 has been described as outputting the non-destructive electrical characteristics of the main insulation layer using the first trained model trained by the model generation unit 12, but it may also be configured to acquire the first trained model from an external source, such as another learning device, and output the non-destructive electrical characteristics of the main insulation layer as output based on this first trained model.

[0026] The diagnosis unit 4 receives as input at least one of the increment in the dielectric tangent of the main insulation layer, the first current-surge voltage of the main insulation layer, and the maximum partial discharge charge amount of the main insulation layer, which are outputs of the first inference unit 32. It determines whether the nondestructive electrical characteristics of the main insulation layer satisfy a predetermined standard and outputs the determination result. The method for determining the standard value depends on the type of insulating material, but for example, it is possible to determine that an increment in the dielectric tangent of the main insulation layer exceeds 6.5% as an abnormality. The first current-surge voltage and the maximum partial discharge charge amount can also be determined quantitatively in a similar manner.

[0027] 5 is a flowchart for explaining the processing of the insulation diagnosis system 100 according to embodiment 1. Step S11 is a data acquisition step, step S12 is an inference processing step, and step S13 is a diagnosis processing step.

[0028] In step S11, the data acquisition unit 31 acquires operating information of the rotating electric machine as input data when diagnosing the insulation performance of the rotating electric machine and outputs it to the first inference unit 32, and the process proceeds to step S12. In step S12, the first inference unit 32 inputs the input data acquired from the data acquisition unit 31 into the first learned model stored in the learned model storage unit 2, acquires nondestructive electrical characteristics of the main insulation layer of the stator coil of the rotating electric machine operated under the conditions specified by the input data, and outputs the acquired nondestructive electrical characteristics to the diagnosis unit 4, and the process proceeds to step S13. In step S13, the diagnosis unit 4 determines whether the nondestructive electrical characteristics of the main insulation layer, which are the output data acquired from the first inference unit 32, satisfy a predetermined standard, outputs the determination result, and ends the process. Nondestructive electrical characteristic data can be acquired without stopping the operation of the rotating electric machine. Therefore, there is no need to stop the rotating electric machine even at the stage of acquiring nondestructive electrical characteristics and generating the first learned model in the learning phase. Therefore, the insulation performance of the rotating electric machine can be diagnosed without stopping the rotating electric machine during operation, including the learning phase and the utilization phase.

[0029] Fig. 6 is a block diagram showing another configuration of the insulation diagnosis system 100 for a rotating electric machine according to the first embodiment. The insulation diagnosis system 100 for a rotating electric machine shown in Fig. 6 includes a learning device 1, a trained model storage unit 2, an inference device 3, and a diagnosis unit 4. In the insulation diagnosis system 100 for a rotating electric machine shown in Fig. 6, the operations of the learning device 1, the trained model storage unit 2, the inference device 3, and the diagnosis unit 4 are the same as those of the learning device 1, the trained model storage unit 2, the inference device 3, and the diagnosis unit 4 shown in Figs. 1 and 4.

[0030] The learning device 1, the trained model storage unit 2, the inference device 3, and the diagnosis unit 4 may be configured on the same hardware, or may be separate devices connected via a network. The learning device 1, the trained model storage unit 2, the inference device 3, and the diagnosis unit 4 may also exist on a cloud server. The learning device 1 and the inference device 3 may be realized by a single device, in which case the learning data acquisition unit 11 may also function as the data acquisition unit 31.

[0031] The model generation unit 12 may learn the nondestructive electrical characteristics of the main insulation layer in accordance with learning data created for insulation diagnosis systems for multiple rotating electric machines. Note that the model generation unit 12 may acquire learning data from insulation diagnosis systems for multiple rotating electric machines used in the same area, or may learn the nondestructive electrical characteristics of the main insulation layer using learning data collected from insulation diagnosis systems for multiple rotating electric machines that operate independently in different areas.

[0032] It is also possible to add or remove insulation diagnosis systems for rotating electric machines that collect learning data during the process. Furthermore, a learning device that has learned the nondestructive electrical characteristics of the main insulation layer for an insulation diagnosis system for a rotating electric machine may be applied to an insulation diagnosis system for another rotating electric machine, and the nondestructive electrical characteristics of the main insulation layer for the insulation diagnosis system for that other rotating electric machine may be re-learned and updated.

[0033] As described above, the learning device 1 of this embodiment is a learning device 1 that uses operating information related to the operating history of a rotating electric machine to obtain a first learned model that infers non-destructive electrical characteristics related to the dielectric strength of the main insulation layer of the stator coil of the rotating electric machine, and is equipped with a learning data acquisition unit 11 that acquires learning data including the operating information and the non-destructive electrical characteristics when the rotating electric machine is operated under the conditions indicated in the operating information, and a model generation unit 12 that uses the learning data to generate a first learned model for inferring non-destructive electrical characteristics from the operating information.

[0034] The learning device 1 according to this embodiment acquires learning data including operating information and nondestructive electrical characteristics obtained when the rotating electric machine is operated under the conditions indicated in the operating information, and uses the learning data to generate a first trained model for inferring nondestructive electrical characteristics from the operating information. Therefore, it is only necessary to use the operating information and data on nondestructive electrical characteristics that can be acquired even when the rotating electric machine is operating. Therefore, compared to acquiring data on the dielectric strength remaining rate, it is possible to reduce the effort of having to stop the rotating electric machine during operation to acquire data.

[0035] Furthermore, the learning data acquisition unit 11 may acquire the learning data including design information related to the dielectric strength of the main insulation layer of the stator coil in addition to the operating information and the nondestructive electrical characteristics obtained when the rotating electrical machine is operated under the conditions indicated in the operating information, and the model generation unit 12 may use the learning data to generate a third trained model for inferring the nondestructive electrical characteristics from the operating information and the design information. Here, the design information related to the dielectric strength of the main insulation layer of the stator coil may include, for example, at least one of the rated voltage of the rotating electrical machine and the thickness of the main insulation layer.

[0036] Furthermore, the insulation diagnosis system 100 can be configured to include a data acquisition unit 31 that acquires operating information and design information, a first inference unit 32 that outputs non-destructive electrical characteristics obtained when operating information and design information are input using the third trained model, and a diagnosis unit 4 that determines whether the non-destructive electrical characteristics satisfy predetermined standards and outputs the determination results.

[0037] By adding design information to the training data in this way, the same insulation diagnosis system can be applied to multiple models with different design conditions, reducing training costs. Another advantage is that insulation diagnosis systems trained using previously acquired training data can be reused for new models.

[0038] Embodiment 2 <Explanation of the Learning Device 1a and Insulation Diagnosis System 100a as Components> In explaining the learning device and insulation diagnosis system according to the second embodiment, we will first explain the configurations of a learning device 1a and an insulation diagnosis system 100a, which are elements that make up the learning device and insulation diagnosis system. Fig. 7 is a block diagram showing the configuration of a learning device 1a, which is one element that makes up the learning device according to the second embodiment. The learning device 1a obtains a trained model that infers the dielectric strength remaining rate of the main insulation layer of the stator coil of a rotating electric machine in the insulation diagnosis system 100a for a rotating electric machine. The learning device 1a includes a learning data acquisition unit 11a and a model generation unit 12a.

[0039] The learning data acquisition unit 11a acquires learning data that is a combination of first input data and second input data. Here, the first input data includes operation information of the rotating electric machine. The operation information of the rotating electric machine may include, for example, at least one of the number of starts and stops of the rotating electric machine or the operating time of the rotating electric machine. The second input data is the dielectric strength remaining rate of the main insulation layer of the stator coil of the rotating electric machine operated under the conditions indicated by the first input data. The learning data acquisition unit 11a may acquire the input data from a data input device or a database (not shown). The data input device is, for example, a keyboard, mouse, keypad, touch panel, etc., which is operated by a user. The database is, for example, a storage device, etc., which stores data including operation information of the rotating electric machine.

[0040] The model generation unit 12a learns the dielectric strength remaining rate of the main insulation layer of the stator coil in a rotating electric machine operated under the conditions indicated by the first input data, based on the learning data which is a combination of the first input data and the second input data output from the learning data acquisition unit 11a. That is, the model generation unit 12a generates a trained model that infers the dielectric strength remaining rate of the main insulation layer of the stator coil in a rotating electric machine operated under the conditions indicated by the first input data, from the learning data which is a combination of the first input data and the second input data. Here, the learning data is data that associates the first input data and the second input data with each other.

[0041] The learning algorithm used by the model generation unit 12a may be a known algorithm such as supervised learning, unsupervised learning, reinforcement learning, or semi-supervised learning. An example will be described in which a neural network is used. The model generation unit 12a learns the dielectric strength remaining rate of the main insulation layer, for example, by so-called supervised learning in accordance with a neural network model. Here, supervised learning refers to a technique in which a set of input and result (label) data is provided to the learning device 1a, whereby the device learns the features of the learning data and infers the result from the input.

[0042] In this learning device 1a, the neural network learns the dielectric strength remaining ratio of the main insulation layer of the stator coil of a rotating electric machine operated under the conditions indicated by the first input data through so-called supervised learning in accordance with learning data created based on a combination of first input data and second input data acquired by the learning data acquisition unit 11a. That is, the neural network learns by inputting the first input data including operating information of the rotating electric machine to the input layer and adjusting the weights so that the result output from the output layer approaches the dielectric strength remaining ratio of the main insulation layer of the stator coil of the rotating electric machine operated under the conditions indicated by the first input data. The model generation unit 12a generates and outputs a trained model by executing the above-described learning. The trained model storage unit 2a stores the trained model output from the model generation unit 12a.

[0043] Next, the learning process of the learning device 1a will be described. Figure 8 is a flowchart for explaining the process of the learning device 1a and the trained model storage unit 2a. Step S21 is a learning data acquisition step, step S22 is a learning process step, and step S23 is a trained model storage step.

[0044] In step S21, the learning data acquisition unit 11a acquires learning data and outputs it to the model generation unit 12a, and the process proceeds to step S22. Note that the learning data acquisition unit 11a may simultaneously acquire the first input data and the second input data included in the learning data, or may acquire the first input data and the second input data at different times as long as the first input data and the second input data are acquired in association with each other.

[0045] In step S22, the model generation unit 12a learns the dielectric strength remaining rate of the main insulation layer of the stator coil in the rotating electric machine operated under the conditions indicated by the first input data by so-called supervised learning in accordance with the learning data created based on the combination of the first input data and the second input data acquired by the learning data acquisition unit 11a, generates a trained model, outputs the generated trained model to the trained model storage unit 2a, and proceeds to step S23. In step S23, the trained model storage unit 2a stores the trained model acquired from the model generation unit 12a, and the process ends.

[0046] <Utilization phase> Next, an insulation diagnosis system for a rotating electric machine that utilizes a trained model generated by the learning device 1a will be described. FIG. 9 is a block diagram showing the configuration of this insulation diagnosis system 100a for a rotating electric machine. The insulation diagnosis system 100a for a rotating electric machine diagnoses the insulation performance of the rotating electric machine. The insulation diagnosis system 100a for a rotating electric machine includes a trained model storage unit 2a, an inference device 3a, and a diagnosis unit 4a. The inference device 3a includes a data acquisition unit 31a and an inference unit 32a. The trained model storage unit 2a, the inference device 3a, and the diagnosis unit 4a may be configured on the same hardware, or may be separate devices connected via a network. The trained model storage unit 2a, the inference device 3a, and the diagnosis unit 4a may reside on a cloud server.

[0047] The data acquisition unit 31a of the inference device 3a acquires data including operating information of the rotating electric machine as input data when diagnosing the insulation performance of the rotating electric machine. The operating information of the rotating electric machine may include, for example, at least one of the number of starts and stops of the rotating electric machine and the operating time of the rotating electric machine. The data acquisition unit 31a may acquire the input data from a data input device or database (not shown). The data input device is, for example, a keyboard, mouse, keypad, touch panel, etc., and is operated by a user. The database is, for example, a storage device, etc., and stores data including design and operating information of the rotating electric machine.

[0048] The inference unit 32a uses the trained model stored in the trained model storage unit 2a to infer the dielectric strength remaining rate of the main insulation layer under the conditions indicated by the input data. That is, by inputting the input data acquired by the data acquisition unit 31a into this trained model, it is possible to infer the dielectric strength remaining rate of the main insulation layer of the stator coil in a rotating electric machine operated under the conditions indicated by the input data and output the inference result.

[0049] In the second embodiment, the inference unit 32a has been described as outputting the dielectric strength remaining ratio of the main insulation layer using the trained model trained by the model generation unit 12a, but the trained model may be acquired from an external device such as another learning device, and the output dielectric strength remaining ratio of the main insulation layer may be output based on this trained model. The diagnosis unit 4a determines whether the dielectric strength remaining ratio of the main insulation layer, which is the output of the inference unit 32a, satisfies a predetermined standard, and outputs the determination result.

[0050] 10 is a flowchart for explaining the processing of the inference device 3a and the diagnosis unit 4a in embodiment 2. Step S31 is a data acquisition step, step S32 is an inference processing step, and step S33 is a diagnosis processing step.

[0051] In step S31, the data acquisition unit 31a acquires data including operating information of the rotating electric machine as input data when diagnosing the insulation performance of the rotating electric machine, outputs the data to the inference unit 32a, and proceeds to step S32. In step S32, the inference unit 32a inputs the input data acquired from the data acquisition unit 31a to the learned model stored in the learned model storage unit 2a, acquires the dielectric strength remaining rate of the main insulation layer of the stator coil in the rotating electric machine operated under the conditions indicated by the input data, and outputs the acquired data to the diagnosis unit 4a, and proceeds to step S33. In step S33, the diagnosis unit 4a determines whether the dielectric strength remaining rate of the main insulation layer, which is the output data acquired from the inference unit 32a, satisfies a predetermined standard, outputs the determination result, and ends the process.

[0052] 11 is a block diagram showing another configuration of an insulation diagnosis system 100a for a rotating electric machine that utilizes a trained model generated by a learning device 1a. The insulation diagnosis system 100a for a rotating electric machine shown in FIG. 11 includes a learning device 1a, a trained model storage unit 2a, an inference device 3a, and a diagnosis unit 4a. In the insulation diagnosis system 100a for a rotating electric machine shown in FIG. 11, the operations of the learning device 1a, the trained model storage unit 2a, the inference device 3a, and the diagnosis unit 4a are the same as those of the learning device 1a, the trained model storage unit 2a, the inference device 3a, and the diagnosis unit 4a shown in FIGS. 7 and 9.

[0053] The learning device 1a, the trained model storage unit 2a, the inference device 3a, and the diagnosis unit 4a may be configured on the same hardware, or may be separate devices connected via a network. The learning device 1a, the trained model storage unit 2a, the inference device 3a, and the diagnosis unit 4a may reside on a cloud server. When the learning device 1a and the inference device 3a are implemented by a single device, the learning data acquisition unit 11a may also function as the data acquisition unit 31a.

[0054] The model generating unit 12a may learn the dielectric strength remaining rate of the main insulation layer in accordance with learning data created for insulation diagnosis systems for multiple rotating electric machines. Note that the model generating unit 12a may acquire learning data from insulation diagnosis systems for multiple rotating electric machines used in the same area, or may learn the dielectric strength remaining rate of the main insulation layer using learning data collected from insulation diagnosis systems for multiple rotating electric machines that operate independently in different areas.

[0055] It is also possible to add or remove insulation diagnosis systems for rotating electric machines that collect learning data from the targets during the process. Furthermore, a learning device that has learned the dielectric strength remaining rate of the main insulation layer for an insulation diagnosis system for a rotating electric machine may be applied to an insulation diagnosis system for another rotating electric machine, and the remaining dielectric strength rate of the main insulation layer for the insulation diagnosis system for the other rotating electric machine may be re-learned and updated.

[0056] As described above, the configuration and operation of the learning device 1a and insulation diagnosis system 100a, which are elements constituting the learning device and insulation diagnosis system according to this embodiment, have been described. By combining these with the inference device 3 shown in embodiment 1, it is possible to configure the learning device 1b and insulation diagnosis system 100b according to embodiment 2. The details are described below.

[0057] <Learning Phase> FIG. 12 is a block diagram showing the configuration of a learning device 1b according to the second embodiment. Compared to the learning device 1a shown in FIG. 7, the learning device 1b according to the present embodiment shown in FIG. 12 additionally includes a trained model storage unit 2 and an inference device 3. The other components of the learning device 1b according to the second embodiment, namely, the training data acquisition unit 11a and the model generation unit 12a, have the same functions as those of the learning device 1a described above. Furthermore, the trained model storage unit 2 and the inference device 3 have the same functions as those of the trained model storage unit 2 and the inference device 3 according to the first embodiment.

[0058] The learning data acquiring unit 11a acquires learning data, which is a combination of first input data and second input data. Here, the first input data includes operating information related to the operating history of the rotating electric machine, as in the first embodiment, but may further include nondestructive electrical characteristics of the main insulation layer of the stator coil of the rotating electric machine, inferred from the operating information of the rotating electric machine. Here, the nondestructive electrical characteristics are obtained by performing inference from the operating information of the rotating electric machine using the first trained model stored in the trained model storage unit 2. The operating information of the rotating electric machine may include, for example, at least one of the number of starts and stops of the rotating electric machine and the operating time of the rotating electric machine. Furthermore, the nondestructive electrical characteristics of the main insulation layer may include at least one of the increase in the dielectric tangent of the main insulation layer, the first current surge voltage of the main insulation layer, and the maximum partial discharge charge amount of the main insulation layer. When the inference device 3 and the learning device 1a are implemented by a single device, the data acquiring unit 31 may also function as the learning data acquiring unit 11a.

[0059] On the other hand, the second input data is the dielectric strength remaining ratio of the main insulation layer of the stator coil of a rotating electric machine operated under the conditions indicated by the first input data. Based on the learning data, which is a combination of the first input data and the second input data output from the learning data acquisition unit 11a, the model generation unit 12a learns the dielectric strength remaining ratio of the main insulation layer of the stator coil of a rotating electric machine operated under the conditions indicated by the rotating electric machine operation information of the first input data. That is, a second trained model is generated from the learning data, which is a combination of the first input data and the second input data, to infer the dielectric strength remaining ratio of the main insulation layer of the stator coil of a rotating electric machine operated under the conditions indicated by the first input data. Here, the learning data is data in which the first input data and the second input data are associated with each other. The trained model storage unit 2a stores the second trained model output from the model generation unit 12a.

[0060] Next, the learning process of the learning device 1b will be described. Fig. 13 is a flowchart for explaining the process of the learning device 1b according to the second embodiment. In step S41, the data acquisition unit 31 acquires, as input data, operating information of the rotating electric machine when diagnosing the insulation performance of the rotating electric machine, and outputs the information to the first inference unit 32, and the process proceeds to step S42. In step S42, the first inference unit 32 inputs the input data acquired from the data acquisition unit 31 into the first learned model stored in the learned model storage unit 2, and infers and acquires the nondestructive electrical characteristics of the main insulation layer of the stator coil in the rotating electric machine operated under the conditions indicated by the input data.

[0061] In step S43, the learning data acquisition unit 11a acquires a combination of the operating information of the rotating electric machine, the dielectric strength survival rate of the main insulation layer of the stator coil in the rotating electric machine operated under the conditions indicated by this operating information, and the non-destructive electrical characteristics acquired in step S42 as learning data, outputs this to the model generation unit 12, and proceeds to step S44. Note that the learning data acquisition unit 11a may acquire the various data included in the learning data simultaneously, or may acquire each data at a different timing. In short, it is sufficient if the operating information, dielectric strength survival rate, and non-destructive electrical characteristics can be acquired in an associated manner.

[0062] In step S44, the model generation unit 12a uses the learning data acquired by the learning data acquisition unit 11a to learn the dielectric strength remaining rate in association with the operating information and the nondestructive electrical characteristics through so-called supervised learning. Furthermore, the model generation unit 12a generates a second trained model, outputs the generated second trained model to the trained model storage unit 2a, and proceeds to step S45. In step S45, the trained model storage unit 2a stores the second trained model acquired from the model generation unit 12a, and the process ends.

[0063] <Utilization phase> Next, an insulation diagnosis system for a rotating electric machine according to a second embodiment will be described. FIG. 14 is a block diagram showing the configuration of an insulation diagnosis system 100b for a rotating electric machine according to a third embodiment. When comparing the insulation diagnosis system 100b for a rotating electric machine according to the second embodiment shown in FIG. 14 with the insulation diagnosis system 100a for a rotating electric machine shown in FIG. 9, a learned model storage unit 2 and an inference device 3 are added. Here, the inference device 3 has the same functions as that shown in FIG. 12 and is composed of a data acquisition unit 31 and a first inference unit 32. Moreover, the inference device 3 has the same functions as that shown in FIG. 9 and is composed of a data acquisition unit 31a and a second inference unit 32b. Moreover, the learned model storage unit 2 stores the first learned model, and the learned model storage unit 2a stores the second learned model.

[0064] The trained model storage unit 2, the inference device 3, the trained model storage unit 2a, the inference device 3a, and the diagnosis unit 4a may be configured on the same hardware or may be separate devices connected via a network. The trained model storage unit 2, the inference device 3, the trained model storage unit 2a, the inference device 3a, and the diagnosis unit 4a may reside on a cloud server. When the inference device 3 and the inference device 3a are implemented by a single device, the data acquisition unit 31 may also function as the data acquisition unit 31a.

[0065] The data acquisition unit 31 of the inference device 3 acquires data including operation information of the rotating electric machine as input data. Based on this operation information, the first inference unit 32 infers the nondestructive electrical characteristics of the rotating electric machine using the first learned model stored in the learned model storage unit 2.

[0066] Meanwhile, the data acquisition unit 31a of the inference device 3a acquires, as input data, data including operating information of the rotating electric machine when diagnosing the insulation performance of the rotating electric machine. Here, the data including operating information of the rotating electric machine may further include non-destructive electrical characteristics of the main insulation layer of the stator coil of the rotating electric machine inferred from the operating information of the rotating electric machine. The non-destructive electrical characteristics of the main insulation layer may include at least one of an increment in the dielectric tangent of the main insulation layer, a first current surge voltage of the main insulation layer, or a maximum partial discharge charge amount of the main insulation layer.

[0067] The second inference unit 32b uses the second trained model stored in the trained model storage unit 2a to infer the dielectric strength remaining rate of the main insulation layer under the conditions indicated by the operating information of the rotating electric machine acquired by the data acquisition unit 31a. That is, by inputting the operating information of the rotating electric machine acquired by the data acquisition unit 31a into this second trained model, it is possible to infer the dielectric strength remaining rate of the main insulation layer of the stator coil in the rotating electric machine operated under the conditions indicated by the operating information and output the inference result.

[0068] In the second embodiment, the second inference unit 32b has been described as outputting the dielectric strength remaining ratio of the main insulation layer using the second trained model trained by the model generation unit 12a, but the second trained model may be acquired from an external device such as another learning device, and the output dielectric strength remaining ratio of the main insulation layer may be output based on this second trained model. The diagnosis unit 4a determines whether the dielectric strength remaining ratio of the main insulation layer, which is the output of the second inference unit 32b, satisfies a predetermined standard and outputs the determination result.

[0069] 15 is a flowchart for explaining the processing of the insulation diagnosis system 100b according to embodiment 2. In step S51, the data acquisition unit 31 acquires, as input data, operating information of the rotating electric machine when diagnosing the insulation performance of the rotating electric machine, and outputs the information to the first inference unit 32, and the process proceeds to step S52. In step S52, the first inference unit 32 inputs the input data acquired from the data acquisition unit 31 into the first trained model stored in the trained model storage unit 2, and infers (first) and acquires the nondestructive electrical characteristics of the main insulation layer of the stator coil in the rotating electric machine operated under the conditions indicated by the input data.

[0070] In step S53, the data acquisition unit 31a acquires the operating information and the nondestructive electrical characteristics acquired in step S52 as input data, outputs them to the second inference unit 32b, and proceeds to step S54. In step S54, the second inference unit 32b inputs the input data acquired from the data acquisition unit 31a to the second learned model stored in the learned model storage unit 2, infers (second) and acquires the dielectric strength remaining rate of the main insulation layer of the stator coil in the rotating electric machine operated under the conditions indicated by the input data, outputs this to the diagnosis unit 4a, and proceeds to step S55. In step S55, the diagnosis unit 4a determines whether the dielectric strength remaining rate of the main insulation layer, which is output data acquired from the second inference unit 32b, satisfies a predetermined standard, outputs the determination result, and ends the process.

[0071] As described above, the learning device 1b of embodiment 2 is a learning device 1b that uses the first learned model generated by the learning device 1 to obtain a second learned model that infers the dielectric strength remaining rate of the main insulation layer, and is equipped with a data acquisition unit 31 that acquires operating information, a first inference unit 32 that outputs non-destructive electrical characteristics obtained when operating information is input using the first learned model, a learning data acquisition unit 11a that acquires learning data including the operating information, the non-destructive electrical characteristics obtained in the first inference unit 32 when conditions indicated in the operating information are input, and the dielectric strength remaining rate when the rotating electric machine is operated under the conditions indicated by the operating information, and a model generation unit 12a that uses the learning data to generate a second learned model for inferring the dielectric strength remaining rate from the operating information.

[0072] Furthermore, the insulation diagnosis system 100b of embodiment 2 is an insulation diagnosis system that diagnoses the insulation status of the main insulation layer using a first learned model generated by the learning device 1 and a second learned model generated by the learning device 1b, and is equipped with a data acquisition unit 31 that acquires operating information, a first inference unit 32 that outputs non-destructive electrical characteristics obtained when operating information is input using the first learned model, a second inference unit 32b that outputs the dielectric strength remaining rate of the main insulation layer obtained when operating information and non-destructive electrical characteristics are input using the second learned model, and a diagnosis unit 4a that determines whether the dielectric strength remaining rate of the main insulation layer satisfies a predetermined standard and outputs the determination result.

[0073] According to the insulation diagnosis system 100b of this embodiment, the inference device 3 can infer nondestructive electrical characteristics from operating information. According to conventional techniques, the nondestructive electrical characteristics are actually measured and used to predict the dielectric strength remaining ratio. Therefore, in order to make this prediction, it is necessary to measure the nondestructive electrical characteristics in addition to the dielectric strength remaining ratio and then construct a second trained model, which requires the effort of measuring the nondestructive electrical characteristics. In contrast, according to the insulation diagnosis system 100b of this embodiment, the nondestructive electrical characteristics can be inferred using the first trained model. This eliminates the effort of actually measuring the nondestructive electrical characteristics, and enables the dielectric strength remaining ratio to be inferred using the operating information and the inferred values ​​of the nondestructive electrical characteristics. Furthermore, because the dielectric strength remaining ratio is predicted using nondestructive electrical characteristics that are highly correlated with the dielectric strength remaining ratio as input data, this system has the advantage of being able to predict the dielectric strength remaining ratio more accurately than when the dielectric strength remaining ratio is predicted using only the operating information of the rotating electric machine without using the nondestructive electrical characteristics.

[0074] Furthermore, as in the first embodiment, it is also possible to add design information of the rotating electric machine. That is, the learning data acquisition unit 11a may acquire, as learning data, the nondestructive electrical characteristics obtained by inference using a third trained model based on the operating information, design information, and conditions indicated in the operating information and design information, and the dielectric strength remaining rate when the rotating electric machine is operated under the conditions indicated by the operating information and design information. The model generation unit 12a may use this learning data to generate a fourth trained model for inferring the dielectric strength remaining rate from the operating information, design information, and nondestructive electrical characteristics. Here, the design information related to the dielectric strength of the main insulation layer of the stator coil may include, for example, at least one of the rated voltage of the rotating electric machine and the thickness of the main insulation layer.

[0075] Furthermore, the insulation diagnosis system 100b of embodiment 2 is an insulation diagnosis system that diagnoses the insulation status of the main insulation layer using a third trained model generated by the learning device 1 and a fourth trained model generated by the learning device 1b, and is equipped with a data acquisition unit 31 that acquires operating information and design information, the first inference unit 32 that outputs non-destructive electrical characteristics obtained when the operating information and design information are input using the third trained model, a second inference unit 32b that outputs the dielectric strength remaining rate of the main insulation layer obtained when the operating information, design information, and non-destructive electrical characteristics are input using the fourth trained model, and a diagnosis unit 4a that determines whether the dielectric strength remaining rate of the main insulation layer satisfies a predetermined standard and outputs the determination result.

[0076] As described above, by adding design information in addition to operational information to the learning data, the same insulation diagnosis system can be applied to multiple models with different design conditions, reducing learning costs. Another advantage is that insulation diagnosis systems trained using previously acquired learning data can be reused for new models.

[0077] The learning devices 1, 1a, and 1b, the trained model storage units 2 and 2a, the inference devices 3 and 3a, the diagnostic units 4 and 4a, and the insulation diagnostic systems 100, 100a, and 100b described in the first and second embodiments are configured, as an example of hardware shown in FIG. 14, with a processor 200 and a storage device 201 provided inside a computer. Although the storage device is not shown, it includes a volatile storage device such as a random access memory and a non-volatile auxiliary storage device such as a flash memory. Alternatively, a hard disk auxiliary storage device may be provided instead of the flash memory. The processor 200 executes a program input from the storage device 201. In this case, the program is input from the auxiliary storage device to the processor 200 via the volatile storage device. The processor 200 may output data such as calculation results to the volatile storage device of the storage device 201, or may store the data in the auxiliary storage device via the volatile storage device.

[0078] Although the present disclosure describes various exemplary embodiments and examples, the various features, aspects, and functions described in one or more embodiments are not limited to application to a particular embodiment, but may be applied to the embodiments alone or in various combinations. Therefore, countless variations not exemplified are conceivable within the scope of the technology disclosed in this specification, including, for example, cases where at least one component is modified, added, or omitted, and cases where at least one component is extracted and combined with components of another embodiment. [Explanation of symbols]

[0079] 1, 1a, 1b Learning device, 11, 11a Learning data acquisition unit, 12, 12a Model generation unit, 2, 2a Learned model storage unit, 3, 3a Inference device, 31, 31a Data acquisition unit, 32 First inference unit, 32a Inference unit, 32b Second inference unit, 4, 4a Diagnosis unit, 100, 100a, 100b Insulation diagnosis system, 200 Processor, 201 Storage device

Claims

1. A learning device that uses operating information related to the operating history of a rotating electric machine to obtain a second trained model that infers a dielectric strength remaining rate of a main insulation layer using a first trained model that infers nondestructive electrical characteristics related to a dielectric strength of a main insulation layer of a stator coil of the rotating electric machine, the second trained model comprising: a data acquisition unit that acquires the driving information; a first inference unit that outputs the nondestructive electrical characteristics obtained when the driving information is input using the first trained model; a learning data acquisition unit that acquires learning data including the operating information, the nondestructive electrical characteristics obtained by the first inference unit when the conditions indicated in the operating information are input, and the dielectric strength remaining rate when the rotating electric machine is operated under the conditions indicated by the operating information; a model generation unit that generates the second trained model for inferring the dielectric strength remaining rate from the operating information using the learning data.

2. An insulation diagnosis system for diagnosing the insulation condition of a main insulation layer using a first trained model that infers non-destructive electrical characteristics related to the dielectric strength of a main insulation layer of a stator coil of a rotating electric machine using operating information related to the operating history of the rotating electric machine, and a second trained model that infers the dielectric strength remaining rate of the main insulation layer using the operating information and the non-destructive electrical characteristics, a data acquisition unit that acquires the driving information; a first inference unit that outputs the nondestructive electrical characteristics obtained when the driving information is input using the first trained model; a second inference unit that outputs the dielectric strength remaining rate obtained when the operating information and the nondestructive electrical characteristics are input using the second trained model; a diagnostic unit that determines whether the residual dielectric strength rate satisfies a predetermined standard and outputs the determination result.

3. A learning device that obtains a third trained model that infers nondestructive electrical characteristics related to dielectric strength using operation information related to an operation history of a rotating electric machine and design information related to dielectric strength of a main insulating layer of a stator coil of the rotating electric machine, a learning data acquisition unit that acquires learning data including the operation information, the nondestructive electrical characteristics when the rotating electric machine is operated under conditions indicated in the operation information, and the design information; A learning device comprising: a model generation unit that uses the learning data to generate a third trained model for inferring the nondestructive electrical characteristics from the operating information and the design information.

4. An insulation diagnosis system that diagnoses the insulation condition of a main insulation layer by using a third trained model that infers non-destructive electrical characteristics related to the dielectric strength using operating information related to the operating history of a rotating electric machine and design information related to the dielectric strength of the main insulation layer of a stator coil of the rotating electric machine, a data acquisition unit that acquires the operation information and the design information; a first inference unit that outputs the nondestructive electrical characteristics obtained when the operating information and the design information are input using the third trained model; and a diagnostic unit that determines whether the nondestructive electrical characteristics satisfy a predetermined standard and outputs the determination result.

5. A learning device that uses operating information related to the operating history of a rotating electric machine and design information related to the dielectric strength of a main insulating layer of a stator coil of the rotating electric machine to obtain a fourth trained model that infers a non-destructive electrical characteristic related to the dielectric strength of the main insulating layer, the fourth trained model inferring a residual dielectric strength rate of the main insulating layer, a data acquisition unit that acquires the operation information and the design information; a first inference unit that outputs the nondestructive electrical characteristics obtained when the operating information and the design information are input using the third trained model; and a learning data acquisition unit that acquires learning data including the nondestructive electrical characteristics obtained by the first inference unit when the operating information, the design information, and the conditions indicated in the operating information and the design information are input, and the dielectric strength remaining rate when the rotating electric machine is operated under the conditions indicated by the operating information and the design information; a model generation unit that uses the learning data to generate the fourth trained model for inferring the dielectric strength remaining rate from the operating information and the design information.

6. An insulation diagnostic system that diagnoses the insulation condition of a main insulation layer using a third trained model that infers non-destructive electrical characteristics related to dielectric strength using operating information related to the operating history of a rotating electric machine and design information related to the dielectric strength of a main insulation layer of a stator coil of the rotating electric machine, and a fourth trained model that infers the residual dielectric strength rate of the main insulation layer using the operating information, the design information, and the non-destructive electrical characteristics, a data acquisition unit that acquires the operation information and the design information; a first inference unit that outputs the nondestructive electrical characteristics obtained when the operating information and the design information are input using the third trained model; and a second inference unit that outputs the dielectric strength remaining rate obtained when the operation information, the design information, and the nondestructive electrical characteristics are input using the fourth trained model; and a diagnostic unit that determines whether the residual dielectric strength rate satisfies a predetermined standard and outputs the determination result.

7. 6. The learning device according to claim 1, wherein the operation information includes at least one of the number of starts and stops of the rotating electric machine and the operation time of the rotating electric machine.

8. 6. A learning device according to claim 1, wherein the non-destructive electrical characteristics include at least one of an increase in the dielectric tangent of the main insulation layer, a first current surge voltage of the main insulation layer, and a maximum partial discharge charge amount of the main insulation layer.

9. 6. The learning device according to claim 3, wherein the design information includes at least one of a rated voltage of the rotating electrical machine and a thickness of the main insulating layer.

10. 7. The insulation diagnosis system according to claim 2, wherein the operation information includes at least one of the number of starts and stops of the rotating electric machine and the operating time of the rotating electric machine.

11. 7. The insulation diagnosis system according to claim 2, wherein the nondestructive electrical characteristics include at least one of an increase in dielectric tangent of the main insulation layer, a first current surge voltage of the main insulation layer, and a maximum partial discharge charge amount of the main insulation layer.

12. 7. The insulation diagnosis system according to claim 4, wherein the design information includes at least one of a rated voltage of the rotating electrical machine and a thickness of the main insulation layer.

13. A program for causing a computer to function as the learning device according to any one of claims 1, 3 and 5.

14. A program for causing a computer to function as the insulation diagnosis system according to any one of claims 2, 4 and 6.

Images