Abnormality determination device and abnormality determination system for motor

Abstract

Provided is an abnormality determination device for a motor, which is capable of determining the level of a mechanical change amount and an electrical change amount inside the motor when an abnormality occurs. An abnormality determination device (1) for an electric motor is provided with: a recorder (11) that collects time-series operation data including data of a current flowing through an electric motor (20); a feature amount calculation unit (12) that performs frequency analysis on the operation data collected by the recorder and calculates an amplitude value for each frequency as a feature amount of the operation data; and an abnormality recognition unit (13) that recognizes, on the basis of the feature amount calculated by the feature amount calculation unit, the presence or absence of an abnormality, the type of the abnormality, and the level of the abnormality. The abnormality level is the level of a mechanical change amount and an electrical change amount in the motor corresponding to the abnormality type.

Description

Technical Field

[0001] This invention relates to an abnormality detection device and an abnormality detection system for electric motors. Background Technology

[0002] In the maintenance and management of workshop equipment with electric motors, from the perspective of cost reduction, it is preferable to develop a technology for detecting abnormalities in the electric motors. As an abnormality detection system for electric motors, a system is disclosed that determines the type and location of the abnormality based on data such as vibration, temperature, and pressure of the motor during rotation (for example, see Patent Document 1). Furthermore, as another abnormality detection system, a system is disclosed that measures the power consumption of the motor during rotation and detects the presence or absence of abnormalities based on the time from the start of operation until the power consumption reaches its peak (for example, see Patent Document 2). Moreover, as yet another abnormality detection system, a system is disclosed that detects the presence or absence of abnormalities based on the vibration sound of the motor during rotation (for example, see Patent Document 3).

[0003] Patent Document 1: Japanese Patent Application Publication No. 7-174617

[0004] Patent Document 2: Japanese Patent No. 3525736

[0005] Patent Document 3: Japanese Patent Application Publication No. 11-241945 Summary of the Invention

[0006] Existing motor anomaly detection systems can detect whether a motor has an anomaly and where the anomaly is located. However, existing motor anomaly detection systems have the following problem: when an anomaly occurs, they cannot determine the level of mechanical and electrical changes within the motor.

[0007] The present invention was proposed to solve the above-mentioned problems. Its purpose is to provide an abnormality determination device for an electric motor that can determine the level of mechanical and electrical changes inside the electric motor when an abnormality occurs.

[0008] The electric motor anomaly determination device of the present invention comprises: a recorder that collects time-series operating data, including data on the current flowing in the electric motor; a feature quantity calculation unit that performs frequency analysis on the time-series operating data collected by the recorder and calculates the amplitude value for each frequency as a feature quantity of the operating data; and an anomaly identification unit that identifies whether an anomaly has occurred, the type of the anomaly, and the anomaly level based on the feature quantity calculated by the feature quantity calculation unit. Furthermore, the anomaly level is the level of mechanical and electrical changes within the electric motor corresponding to the type of anomaly.

[0009] The effects of the invention

[0010] In the abnormality determination device for an electric motor of the present invention, the abnormality identification unit is able to determine the level of mechanical and electrical changes inside the electric motor corresponding to the type of abnormality, i.e., the abnormality level. Attached Figure Description

[0011] Figure 1 This is a structural diagram of the abnormality determination system for the electric motor according to Implementation Method 1.

[0012] Figure 2 This is a flowchart illustrating the investigation of abnormal content in the abnormality determination system of the electric motor according to Embodiment 1.

[0013] Figure 3 This is a flowchart illustrating the operation of the anomaly determination device involved in Implementation 1.

[0014] Figure 4 This diagram illustrates the operation of the anomaly detection device according to Embodiment 1.

[0015] Figure 5 This is a structural diagram of the abnormality determination system for the electric motor involved in Implementation Method 2.

[0016] Figure 6 This is a structural diagram of the abnormality determination system for the electric motor involved in Implementation Method 3.

[0017] Figure 7 This is a structural diagram of the machine learning unit in the abnormal determination system of the electric motor according to Embodiment 3.

[0018] Figure 8 This is a structural diagram of the abnormality identification unit in the abnormality determination system of the electric motor according to Embodiment 3.

[0019] Figure 9 This is a diagram illustrating an example of the hardware structure of the anomaly determination benchmark setting unit, machine learning unit, feature quantity calculation unit, and anomaly recognition unit involved in embodiments 1 to 3. Detailed Implementation

[0020] The following description, with reference to the accompanying drawings, details the motor malfunction detection device and motor malfunction detection system according to embodiments of the present invention. Furthermore, in the drawings, the same reference numerals denote identical or equivalent parts.

[0021] Implementation method 1.

[0022] Figure 1This is a structural diagram of the motor fault determination system according to Embodiment 1. The motor fault determination system 10 of this embodiment is a system for determining faults in the motor 20. The fault determination system 10 of this embodiment consists of a fault determination device 1, a fault content input unit 2, a database unit 3, and a fault determination reference setting unit 4. The motor 20 is either a rotating machine that performs rotary motion or a linear motor that performs linear motion. Furthermore, the rotational speed when the motor is a rotating machine and the moving speed when the motor is a linear motor will hereby be combined and referred to as the motor's operating speed.

[0023] The anomaly detection device 1 includes a recorder 11, a feature quantity calculation unit 12, and an anomaly identification unit 13. The recorder 11 is connected to the motor 20 and collects time-series operating data, including current data flowing in the motor 20. The feature quantity calculation unit 12 performs frequency analysis on the time-series operating data collected by the recorder 11, calculating the amplitude value for each frequency as a feature quantity of the operating data. The anomaly identification unit 13 identifies whether an anomaly has occurred, the type of the anomaly, and the anomaly level based on the feature quantities calculated by the feature quantity calculation unit 12. Furthermore, the anomaly identification unit 13 outputs the identified anomaly, its type, and its level to an external anomaly content display unit 30. The current data collected by the recorder 11 can be the phase current of the motor 20 or the current converted to the dq axis. Additionally, the recorder 11 may also be included in an inverter that drives the motor 20.

[0024] In addition to current data, the operating data collected by the recorder 11 also includes vibration data collected by the vibration sensor and sound pressure data collected by the microphone. Vibration data, sound pressure data, etc., can be collected by vibration sensors, microphones, etc., which are installed outside the motor 20, so data collection is easy.

[0025] When the motor 20 operates at a constant speed, the feature quantity calculation unit 12 can perform frequency analysis on the time series operation data using Fourier transform. When the motor 20 operates with varying speeds (acceleration or deceleration), the feature quantity calculation unit 12 can perform frequency analysis on the time series operation data using Fourier transform or wavelet transform for each short time interval. By performing frequency analysis using Fourier transform or wavelet transform for each short time interval during acceleration and deceleration, the feature quantity calculation unit 12 can calculate feature quantities independently of the operating state of the motor 20.

[0026] The anomaly input unit 2 inputs the type of anomaly determined through the anomaly investigation described later, the amount of internal mechanical change in the motor 20 caused by the anomaly, and the amount of electrical change in the motor 20 caused by the anomaly. Examples of anomalies include increased winding resistance, broken windings, bearing damage, and magnet damage. The amount of internal mechanical change in the motor 20 caused by the anomaly includes, for example, changes in the shape of the motor 20 relative to its normal state, such as internal scratches, cracks, or wear. The amount of electrical change in the motor 20 caused by the anomaly includes, for example, changes in the impedance between the terminals of the motor 20 from a reference value.

[0027] The database unit 3 stores information related to the type of abnormality, mechanical changes, electrical changes, and characteristic quantities of the operating data calculated by the characteristic quantity calculation unit 12 when an abnormality occurs, which are input from the abnormality content input unit 2.

[0028] The anomaly determination benchmark setting unit 4 sets an anomaly determination threshold and a determination benchmark for anomaly level based on the anomaly determination threshold. This anomaly determination threshold is used to determine the anomaly level, representing the degree of the anomaly, for each type of anomaly based on information related to the type of anomaly, mechanical changes, electrical changes, and characteristic quantities of operating data stored in the database unit 3. Here, the anomaly level refers to the level of mechanical and electrical changes within the motor. Furthermore, the level can be recorded as a numerical value or marked as large, medium, or small. Alternatively, the level can be converted into the remaining lifespan of the motor or component and marked as time. As described above, the higher the anomaly level, the larger the displayed value. Conversely, the higher the anomaly level, the smaller the displayed value.

[0029] Next, we will explain the investigation of abnormal content.

[0030] Figure 2 This is a flowchart illustrating the anomaly investigation process in the motor anomaly determination system according to this embodiment. This anomaly investigation is conducted by the user or the motor manufacturer on a motor that has experienced an anomaly. The investigation pre-determines the type of anomaly and is performed on multiple motors with varying degrees of anomaly. Here, motors with varying degrees of anomaly refer to motors with different internal mechanical and electrical variations.

[0031] In step S01, the impedance of the motor is measured. In step S02, it is determined whether the measured impedance differs from a reference value. Furthermore, the reference value for impedance is the impedance value under normal motor conditions. In step S02, if the measured impedance differs from the reference value (YES), the process proceeds to step S03. In step S03, after inputting the type of abnormality and the electrical change amount into the abnormality content input unit 2, the process proceeds to step S04. Here, the electrical change amount is the difference between the measured impedance and the reference value. In step S02, if the measured impedance is the same as the reference value (NO), the process proceeds to step S04.

[0032] In step S04, the user or the motor manufacturer's investigator disassembles the motor. Furthermore, in step S05, it is determined whether a mechanical change has occurred in the motor. If, in step S05, a mechanical change has occurred in the motor (YES), the process proceeds to step S06. In step S06, the user or the motor manufacturer's investigator measures the amount of mechanical change, such as scratches or cracks, in the bearings and magnets inside the motor. Here, the amount of mechanical change refers to the size of the scratches or cracks in the bearings and magnets. In step S07, the type of anomaly and the amount of mechanical change are input to the anomaly content input unit 2. As described above, the anomaly type, the amount of mechanical change inside the motor 20 caused by the anomaly, and the amount of electrical change in the motor 20 caused by the anomaly are input to the anomaly content input unit 2. The anomaly content input unit 2 outputs the input anomaly type, amount of mechanical change, and amount of electrical change to the database unit 3.

[0033] Furthermore, if the type of anomaly and the amount of mechanical change inside the motor can be determined without disassembling the motor, then disassembling the motor is not necessary.

[0034] Next, the database unit 3 and the abnormality determination reference setting unit 4 in the electric motor abnormality determination system 10 according to this embodiment will be described.

[0035] Based on Figure 2 The flowchart shown illustrates the investigation of multiple motors and anomalies. Figure 1 The abnormality detection system for the motor shown is connected. Recorder 11 collects time-series operating data of the motor for which the type of abnormality has been predetermined. Feature quantity calculation unit 12 performs frequency analysis on the time-series operating data collected by recorder 11, calculating the amplitude value for each frequency as a feature quantity of the operating data. Feature quantity calculation unit 12 outputs the feature quantities of the operating data to database unit 3.

[0036] Database unit 3 stores the following table, which associates the type of anomaly, mechanical changes, and electrical changes input from anomaly content input unit 2 with the characteristic values ​​of the operating data input from characteristic value calculation unit 12. That is, database unit 3 stores the following table, which associates the type of anomaly, the mechanical changes and electrical changes in motors with different degrees of anomaly, and the characteristic values ​​of the operating data.

[0037] Furthermore, it is preferable that the table includes the correlation between mechanical and electrical changes in a normal motor and characteristic quantities of operating data. By including data on normal motors, it is possible to determine whether an anomaly is normal during anomaly detection.

[0038] The anomaly determination benchmark setting unit 4 sets anomaly determination benchmark based on a table stored in the database unit 3 that associates the types of anomalies, the mechanical and electrical changes in motors with different degrees of anomalies, and characteristic quantities of operating data. Specifically, for each type of anomaly, the anomaly determination benchmark setting unit 4 calculates the relational expression using, for example, a polynomial approximation using the least squares method, based on data of the mechanical and electrical changes in motors with different degrees of anomalies and characteristic quantities of the operating data of these motors, and determines multiple thresholds for the characteristic quantities based on a predetermined number of divisions. Furthermore, based on the characteristic quantities divided by multiple thresholds, anomaly level determination benchmarks are set. The degree of the polynomial can be an integer such as 1 or 2. When the degree is 1, it becomes a linear approximation, which cannot represent a nonlinear relationship, but it is less likely to cause problems such as data fluctuations or over-approximation of noise. As described above, the anomaly determination benchmark setting unit 4 determines multiple thresholds for the characteristic quantities and sets anomaly level determination benchmarks based on these thresholds. Furthermore, the anomaly determination benchmark setting unit 4 outputs the determined multiple thresholds for the characteristic quantities and the anomaly level determination benchmarks set based on these thresholds to the anomaly identification unit 13.

[0039] Next, the operation of the anomaly detection device 1 in this embodiment will be explained.

[0040] Figure 3 This is a flowchart illustrating the operation of the anomaly detection device in this embodiment. An electric motor 20, which is the object of investigation, is connected to the anomaly detection device 1. In step S11, the recorder 11 collects time-series operation data. The operation data is time-series data including, for example, operating speed, and current data. The operation data is continuously or intermittently stored in the recorder 11. Next, in step S12, the feature quantity calculation unit 12 calculates feature quantities based on the time-series operation data stored in the recorder 11.

[0041] In step S13, the anomaly identification unit 13 determines the type of anomaly occurring in the motor 20 based on feature quantities. Furthermore, in step S14, the anomaly identification unit 13 determines the level of the anomaly occurring in the motor 20 based on feature quantities. Finally, in step S15, the anomaly identification unit 13 outputs the presence or absence of an anomaly, the type of an anomaly, and the anomaly level to the anomaly content display unit 30. The anomaly content display unit 30 displays the input information.

[0042] Figure 4 This diagram illustrates an example of the operation of the anomaly detection device in this embodiment. Figure 4 In the diagram, the horizontal axis represents frequency, and the vertical axis represents amplitude. For example... Figure 4 As shown, the feature quantity calculation unit 12 performs frequency analysis on the time-series operation data collected by the recorder 11, and calculates the amplitude value for each frequency as a feature quantity of the operation data. The anomaly identification unit 13 pre-stores multiple thresholds for the feature quantities determined by the anomaly determination criterion setting unit 4 for each type of anomaly, and a determination criterion for the anomaly level set based on these thresholds. For example, the anomaly identification unit 13 stores an increase in amplitude at frequency F1 when anomaly type A occurs, and an increase in amplitude at frequency F2 when anomaly type B occurs. Furthermore, the anomaly identification unit 13 stores thresholds A1, A2, ..., A5 for the amplitude of frequency F1, which serves as the determination criterion for the anomaly level when anomaly type A occurs.

[0043] As an example, let's describe the case where a scratch has occurred on the bearing of an electric motor. Assume that the amplitude of frequency F1 increases when a scratch has occurred on the bearing. In this case, the anomaly identification unit 13 can determine the anomaly level as follows: if the amplitude is less than or equal to A1, it is determined that there is no anomaly; if the amplitude exceeds A1 but is less than or equal to A2, it is determined that a scratch of less than or equal to 0.2 mm has occurred on the bearing; if the amplitude exceeds A2 but is less than or equal to A3, it is determined that a scratch of less than or equal to 0.5 mm has occurred on the bearing, and so on.

[0044] In the motor anomaly determination device configured as described above, the anomaly level can be represented by the mechanical and electrical changes within the motor corresponding to the type of anomaly. Therefore, the motor anomaly determination device of this embodiment can determine the level of the mechanical and electrical changes within the motor when an anomaly occurs.

[0045] Furthermore, if the motor that has been identified as abnormal can be recovered, it is preferable to recover the motor and conduct a decomposition investigation. By conducting a decomposition investigation on the recovered motor, the correlation between the actual mechanical changes generated in the motor and the characteristic quantities calculated by the characteristic quantity calculation unit 12 can be confirmed. Moreover, the mechanical changes obtained through the decomposition investigation are input to the abnormality content input unit 2, thereby updating the multiple thresholds for the characteristic quantities determined by the abnormality determination benchmark setting unit 4 and the judgment benchmark for the abnormality level based on these thresholds.

[0046] In the motor anomaly detection system of this embodiment, the database unit 3 can be a remotely configured server. When the database unit 3 is a server connected to other devices via a network, a database unit that can be shared with multiple anomaly detection systems can be constructed.

[0047] Implementation method 2.

[0048] Figure 5 This is a structural diagram of the motor anomaly determination system according to Embodiment 2. The motor anomaly determination system 10 of this embodiment is a system for determining anomalies of multiple motors 20. The anomaly determination system 10 of this embodiment is composed of multiple anomaly determination devices 1, an anomaly content input unit 2, a database unit 3, and an anomaly determination reference setting unit 4.

[0049] Multiple anomaly detection devices 1 are connected to multiple electric motors 20 respectively. Each anomaly detection device 1 has a recorder 11, a feature quantity calculation unit 12, and an anomaly recognition unit 13. The recorder 11 is connected to one electric motor 20 and collects time-series operation data, including the current data flowing in the electric motor 20. The feature quantity calculation unit 12 performs frequency analysis on the time-series operation data collected by the recorder 11 and calculates the amplitude value for each frequency as a feature quantity of the operation data. The anomaly recognition unit 13 identifies whether an anomaly has occurred, the type of an anomaly, and the anomaly level based on the feature quantity calculated by the feature quantity calculation unit 12. In addition, the anomaly recognition unit 13 outputs the identified anomaly, the type of an anomaly, and the anomaly level to an external anomaly content display unit 30.

[0050] The abnormality input unit 2 receives the type of abnormality when each of the multiple motors 20 malfunctions, the amount of internal mechanical change of the motor 20 caused by the abnormality, and the amount of electrical change of the motor 20 caused by the abnormality.

[0051] The database unit 3 is connected to the feature quantity calculation unit 12 of multiple anomaly determination devices. The database unit 3 stores information that associates the type of anomaly, mechanical change, and electrical change related to each motor 20 with the feature quantities of the operating data calculated by the feature quantity calculation unit 12 when an anomaly occurs.

[0052] The anomaly determination benchmark setting unit 4 sets an anomaly determination threshold and an anomaly level determination benchmark based on the anomaly determination threshold. The anomaly determination threshold is used to determine the anomaly level representing the degree of anomaly for each type of anomaly based on information related to the type of anomaly, mechanical change, electrical change, and characteristic quantity of the operating data of each motor 20 stored in the database unit 3.

[0053] In the motor anomaly determination system configured as described above, the anomaly determination reference setting unit 4 sets an anomaly determination threshold and an anomaly level determination reference based on the anomaly determination threshold. The anomaly determination threshold is used to determine the anomaly level representing the degree of anomaly for each type of anomaly based on information that associates the type of anomaly, mechanical change, electrical change and characteristic quantity of the operating data of each motor. Therefore, data when multiple motors have anomalies can be used effectively.

[0054] Furthermore, by conducting a disassembly investigation of the recovered motors, the frequency of confirming the correlation between the actual mechanical changes occurring in the motors and the characteristic quantities calculated by the characteristic quantity calculation unit 12 can be increased. Therefore, the frequency of updating the multiple thresholds for the characteristic quantities determined by the anomaly determination benchmark setting unit 4 and the anomaly level determination benchmark based on those thresholds is increased.

[0055] Implementation method 3.

[0056] Figure 6 This is a structural diagram of the motor fault determination system according to Embodiment 3. In the fault determination system 10 of this embodiment, the fault determination system of Embodiment 1... Figure 1 The anomaly determination benchmark setting unit of the anomaly determination system shown is composed of a machine learning unit 5.

[0057] In the anomaly determination system 10 of this embodiment, the investigation of anomaly content is similar to that in embodiment 1. Figure 2 The flowchart shown is the same. Furthermore, the process continues as in Embodiment 1 until the database unit 3 saves a table relating the types of anomalies, mechanical changes, and electrical changes input from the anomaly content input unit 2 to the characteristic values ​​of the operating data input from the characteristic value calculation unit 12.

[0058] Figure 7 This is a structural diagram of the machine learning unit in the electric motor anomaly detection system according to this embodiment. Figure 7 As shown, the machine learning unit 5 includes a data acquisition unit 51 and a model generation unit 52. The data acquisition unit 51 acquires tables stored in the database unit 3 as learning data. The learning data is a table that correlates the mechanical and electrical changes in an electric motor with different types and degrees of anomalies with characteristic quantities of operating data. The model generation unit 52 learns the correlation between characteristic quantities and the types and levels of anomalies occurring in the electric motor based on the learning data acquired by the data acquisition unit 51. That is, the model generation unit 52 generates a learning model that infers the presence or absence of anomalies, the type of anomaly when it occurs, and the level of the anomaly based on the characteristic quantities of the operating electric motor calculated by the characteristic quantity calculation unit 12. The machine learning unit 5 outputs this trained model to the anomaly recognition unit 13. The learning in the machine learning unit 5 is performed each time an anomaly content is input to the anomaly content input unit 2. This updates the trained model.

[0059] Figure 8 This is a structural diagram of the anomaly identification unit in the motor anomaly determination system according to this embodiment. For example... Figure 8 As shown, the anomaly identification unit 13 includes a feature acquisition unit 131 and an inference unit 132. The feature acquisition unit 131 acquires feature quantities of the operating motor from the feature calculation unit 12. The inference unit 132 receives a trained model from the machine learning unit 5. Using the trained model, the inference unit 132 infers the presence or absence of an anomaly, the type of an anomaly when it occurs, and the anomaly level based on the feature quantities of the operating motor acquired by the feature acquisition unit 131. The inference unit 132 outputs the inferred presence or absence of an anomaly, the type of an anomaly when it occurs, and the anomaly level.

[0060] In the motor anomaly detection system configured as described above, it is not necessary to set multiple thresholds for feature quantities based on a predetermined number of segments, as is the case in the motor anomaly detection system of Embodiment 1. Furthermore, the machine learning performed by the machine learning unit 5 can be, for example, multinomial approximation or neural networks. The machine learning method performed by the machine learning unit 5 can be any other machine learning method, which takes feature quantities calculated based on motor operating data as input and generates a trained model capable of determining the presence or absence of anomalies, the type of anomaly, and the level of anomaly.

[0061] Furthermore, the anomaly determination criterion setting unit 4, machine learning unit 5, feature quantity calculation unit 12, and anomaly recognition unit 13 in embodiments 1 to 3, for example, are hardware components such as... Figure 9As shown, it consists of a processor 40 and a storage device 50. The storage device is not shown, but it includes volatile storage devices such as random access memory and non-volatile auxiliary storage devices such as flash memory. Alternatively, a hard disk can be used as an auxiliary storage device instead of flash memory. The processor 40 executes a program input from the storage device 50. In this case, the program is input to the processor 40 from the auxiliary storage device via the volatile storage device. Furthermore, the processor 40 can output data such as calculation results to the volatile storage device of the storage device 50, or it can save data in the auxiliary storage device via the volatile storage device.

[0062] This invention describes various exemplary embodiments, but the various features, methods and functions described in one or more embodiments are not limited to specific embodiments, and can also be applied to embodiments individually or in various combinations.

[0063] Therefore, within the scope of the technology disclosed in this invention, numerous variations not illustrated will be conceived. For example, there are cases involving the modification, addition, or omission of at least one structural element, and cases involving the extraction of at least one structural element and its combination with structural elements of other embodiments.

[0064] Explanation of the label

[0065] 1 Anomaly detection device, 2 Anomaly content input unit, 3 Database unit, 4 Anomaly detection benchmark setting unit, 5 Machine learning unit, 10 Anomaly detection system, 11 Recorder, 12 Feature quantity calculation unit, 13 Anomaly recognition unit, 20 Motor, 30 Anomaly content display unit, 40 Processor, 50 Storage device, 51 Data acquisition unit, 52 Model generation unit, 131 Feature quantity acquisition unit, 132 Inference unit.

Claims

1. A fault detection device for an electric motor, comprising: A recorder that collects time-series operational data, including data on the current flowing in the motor. The feature quantity calculation unit performs frequency analysis on the time series operation data collected by the recorder, and calculates the amplitude value for each frequency as a feature quantity of the operation data. as well as The anomaly identification unit identifies whether an anomaly has occurred, the type of an anomaly, and the level of the anomaly based on the feature quantities calculated by the feature quantity calculation unit. The abnormality detection device for this electric motor is characterized by, The anomaly level is the level of the mechanical and electrical changes inside the motor that corresponds to the type of anomaly.

2. A fault detection system for an electric motor, characterized in that, have: The abnormal content input unit receives abnormal content including the type of abnormality of the motor that has malfunctioned, the amount of mechanical changes and electrical changes inside the motor. An abnormality determination device for the electric motor as described in claim 1, connected to the electric motor; The database unit stores the abnormal content in association with the feature quantities calculated by the feature quantity calculation unit of the abnormality determination device of the motor; and The anomaly determination benchmark setting unit sets a benchmark for the level of mechanical and electrical changes inside the motor corresponding to the type of anomaly, based on the correlation between the anomaly content and the feature quantity stored in the database unit.

3. A fault detection system for an electric motor, characterized in that, have: The abnormal content input unit receives abnormal content including the type of abnormality of multiple motors that have malfunctioned, the amount of mechanical changes and electrical changes inside the motors; An abnormality determination device for the plurality of motors as described in claim 1, which is connected to the plurality of motors respectively; The database unit stores the abnormal content in association with the feature quantities calculated by the feature quantity calculation unit of the multiple abnormal determination devices of the motors; and The anomaly determination benchmark setting unit sets a benchmark for the level of mechanical and electrical changes inside the motor corresponding to the type of anomaly, based on the correlation between the anomaly content and the feature quantity stored in the database unit.

4. The abnormality determination system for an electric motor according to claim 2 or 3, characterized in that, The anomaly determination benchmark setting unit sets a benchmark for the level of the mechanical and electrical changes inside the motor corresponding to the type of anomaly, based on a trained model generated by using machine learning, which uses the correlation between the anomaly content and the feature quantity stored in the database unit as learning data.

5. The abnormality determination system for an electric motor according to any one of claims 2 to 4, characterized in that, The type of anomaly, the amount of mechanical change and the amount of electrical change inside the motor, which are input to the anomaly input unit, are information obtained by decomposing and investigating the motor that has malfunctioned.

Images