Learning device and learning method, and fault diagnosis device and fault diagnosis method for diagnosing the presence or absence of abnormality in equipment using a signal reconstruction model

Abstract

A learning device and a learning method, and a fault diagnosis device and a fault diagnosis method for diagnosing the presence or absence of an abnormality in equipment using a signal reconstruction model are disclosed. The learning device includes a memory for storing instructions executable by a processor, and a processor, wherein the processor acquires, via a vibration sensor, vibration frequency signals that measure vibrations generated when the equipment is operating normally, learns a signal reconstruction model that generates a reconstructed signal corresponding to the vibration frequency signal based on the acquired vibration frequency signals, acquires the reconstructed signal corresponding to the vibration frequency signal via the trained signal reconstruction model, determines a reconstruction error value indicating a difference between the vibration frequency signal and the reconstructed signal, and determines a threshold value for the reconstruction error value for each of a plurality of predefined frequency bands.

Description

[Technical Field]

[0001] The present disclosure relates to a learning device and a learning method, and a fault diagnosis device and a fault diagnosis method that diagnose the presence or absence of an abnormality in equipment using a signal reconstruction model. [Background technology]

[0002] Equipment fault diagnosis technology using equipment vibration is a technology that diagnoses whether or not there is an abnormality in the equipment by analyzing vibration frequency signals measured by the equipment. An unsupervised learning AI model can be used to diagnose equipment faults using equipment vibration. The unsupervised learning AI model can be trained using a large amount of vibration data collected during normal operation of the equipment. The equipment fault diagnosis technology using the unsupervised learning AI model can diagnose whether or not there is an abnormality in the equipment across the entire range of measured vibration frequency signals. Summary of the Invention [Means for solving the problem]

[0003] According to one embodiment, a learning device includes: a memory that stores instructions executable by a processor; and the processor, wherein the processor acquires, via a vibration sensor, vibration frequency signals that measure vibrations generated when equipment is operated normally; learns a signal reconstruction model that generates reconstructed signals (reconstructed signal data) corresponding to the vibration frequency signals based on the acquired vibration frequency signals; acquires reconstructed signals corresponding to the vibration frequency signals via the trained signal reconstruction model; determines a reconstruction error value indicating a difference between the vibration frequency signals and the reconstructed signal; and determines a threshold value for the reconstruction error value for each of a plurality of predefined frequency bands.

[0004] The signal reconstruction model may include at least one of an autoencoder, a stacked autoencoder, a long short-term memory autoencoder (LSTM), and a convolutional autoencoder model.

[0005] The processor may determine the reconstruction error value based on at least one of an average value of the difference between the vibration frequency signal and the reconstructed signal, an average value of the square of the difference, and a square root value of the average value of the square of the difference.

[0006] The processor may determine a threshold for the reconstruction error value based on the maximum value of the reconstruction error value or the 3 SIGMA rule.

[0007] According to one embodiment, a fault diagnosis device for diagnosing whether or not an anomaly exists in equipment using a signal reconstruction model includes: a memory for storing instructions executable by a processor; and the processor, wherein the processor acquires vibration frequency signals measuring vibrations occurring in equipment via a vibration sensor; inputs the vibration frequency signals to a trained signal reconstruction model to acquire reconstructed signals corresponding to the vibration frequency signals from the signal reconstruction model; determines reconstruction error values ​​indicating signal differences between the vibration frequency signals and the reconstructed signals for each predefined frequency band; and determines whether or not an anomaly exists in the equipment based on the reconstruction error values ​​determined for each frequency band and threshold values ​​determined for each frequency band.

[0008] The processor may determine that there is an abnormality in the equipment when at least one of the reconstruction error values ​​determined for each frequency band is greater than a threshold determined for each frequency band.

[0009] The processor may determine that the equipment is in a normal state if the total reconstruction error value determined for each frequency band is equal to or less than a threshold determined for each frequency band.

[0010] The processor may determine a reconstruction error value based on at least one of an average value of the difference between the input signal and the reconstructed signal, an average value of the square of the difference, and a square root value of the average value of the square of the difference.

[0011] The processor compares the reconstruction error value determined for each frequency band with a threshold value determined for each frequency band, and when a frequency band having a reconstruction error value greater than the threshold value is detected, the processor can estimate an abnormality type based on the detected frequency band.

[0012] A learning method performed by a learning device according to one embodiment may include an operation of acquiring, via a vibration sensor, a vibration frequency signal that measures vibrations generated when equipment is operated normally; an operation of training a signal reconstruction model that generates a reconstructed signal (reconstructed signal data) corresponding to the vibration frequency signal based on the acquired vibration frequency signal; an operation of acquiring a reconstructed signal corresponding to the vibration frequency signal via the trained signal reconstruction model; an operation of determining a reconstruction error value indicating a difference between the vibration frequency signal and the reconstructed signal; and an operation of determining a threshold value for the reconstruction error value for each of a plurality of predefined frequency bands.

[0013] The operation of determining the reconstruction error value may include an operation of determining a reconstruction error value based on at least one of an average value of a difference between the vibration frequency signal and the reconstructed signal, an average value of the squared difference, and a square root value of the average value of the squared difference.

[0014] The operation of determining the threshold value may determine a threshold value for the reconstruction error value based on the maximum value of the reconstruction error value or the 3 sigma rule.

[0015] According to an embodiment, a fault diagnosis method performed by a fault diagnosis device that diagnoses whether or not an abnormality exists in equipment using a signal reconstruction model may include an operation of acquiring a vibration frequency signal that measures vibrations occurring in equipment via a vibration sensor; inputting the vibration frequency signal to a trained signal reconstruction model and acquiring a reconstructed signal corresponding to the vibration frequency signal from the signal reconstruction model; determining a reconstruction error value indicating a difference between the vibration frequency signal and the reconstructed signal for each predefined frequency band; and determining whether or not an abnormality exists in the equipment based on the reconstruction error value determined for each frequency band and a threshold value determined for each frequency band.

[0016] The operation of determining whether or not there is an abnormality may include an operation of estimating that there is an abnormality in the equipment if at least one of the reconstruction error values ​​determined for each frequency band is greater than a threshold determined for each frequency band.

[0017] The fault diagnosis method may further include determining that the equipment is in a normal state if the total reconstruction error value determined for each frequency band is equal to or less than a threshold value determined for each frequency band.

[0018] The operation of determining the reconstruction error value may include an operation of determining the reconstruction error value based on at least one of an average value of the difference between the input signal and the reconstructed signal, an average value of the squared difference, and a square root value of the average value of the squared difference.

[0019] The fault diagnosis method may further include an operation of, when a frequency band having a reconstruction error value greater than a threshold value is detected as a result of comparing the reconstruction error value determined for each frequency band with a threshold value determined for each frequency band, estimating an abnormality type based on the detected frequency band. [Brief explanation of the drawings]

[0020] [Figure 1]4 is a diagram illustrating how a fault diagnosis device according to an embodiment diagnoses whether or not a fault exists in equipment for each frequency band; [Figure 2] FIG. 1 is a block diagram illustrating a configuration of a learning device according to an embodiment. [Figure 3] 1 is a block diagram for explaining the configuration of a fault diagnosis device according to an embodiment; [Figure 4] FIG. 10 is a diagram illustrating determining a reconstruction error value using a signal reconstruction model according to an embodiment. [Figure 5] 10A and 10B are diagrams illustrating determining a reconstruction error value for different frequency band ranges according to an embodiment. [Figure 6] FIG. 10 illustrates determining a reconstruction error value for a particular vibration frequency signal according to one embodiment. [Figure 7] FIG. 10 is a diagram illustrating how a learning device according to an embodiment determines a threshold value. [Figure 8] FIG. 10 is a diagram illustrating how a learning device according to an embodiment determines thresholds in multiple frequency bands. [Figure 9] 2 is a diagram illustrating how a fault diagnosis device according to an embodiment diagnoses whether or not a fault exists in equipment; FIG. [Figure 10] 1 is a diagram illustrating how a fault diagnosis device according to an embodiment estimates an abnormality type of equipment for each frequency band. [Figure 11] 10 is a flowchart illustrating the operation of a learning method performed by a learning device according to an embodiment. [Figure 12] 4 is a flowchart illustrating the operation of a fault diagnosis method performed by a fault diagnosis device according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0021] Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only and may be modified and implemented in various forms. Therefore, the actual implementation is not limited to the specific embodiments disclosed, and the scope of this specification includes modifications, equivalents, or alternatives included in the technical ideas described in the embodiments.

[0022] Although terms such as "first" or "second" may be used to describe various components, these terms should be construed only to distinguish one component from another. For example, a first component can be designated as a second component, and similarly, a second component can be designated as a first component.

[0023] When a component is described as being "coupled" to another component, it should be understood that the component may be directly coupled or connected to the other component, or there may be other components intervening therebetween.

[0024] "Singular expressions include plural expressions unless the context clearly dictates otherwise. In this document, phrases such as "A or B," "at least one of A and B," "at least one of A or B," "A, B or C," "at least one of A, B and C," and "at least one of A, B, or C" each include any and all possible combinations of the items listed together with the corresponding phrase. In this specification, terms such as "comprise" or "have" specify the presence of stated features, numbers, steps, operations, components, parts, or combinations thereof, but should be understood not to preclude the presence or possible addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0025] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art. Terms defined in commonly used dictionaries should be interpreted to have a meaning consistent with the meaning they have in the context of the relevant art, and should not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0026] The term "module" as used in this document refers to software or hardware components such as FPGAs or ASICs, and the "module" may function in either way. However, the term "module" is not limited to software or hardware. A "module" may reside on an addressable storage medium or may be configured to implement one or more processors. For example, a "module" may include components such as software components, object-oriented software components, class components, or task components, as well as processes, functions, attributes, procedures, subroutines, program code segments, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The components and functionality provided within a "module" may be combined into fewer components and modules or further separated into additional components and modules. Furthermore, a component and a "module" may be implemented to implement one or more CPUs within a device or security multimedia card. A "module" may also include one or more processors.

[0027] Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. In the description with reference to the accompanying drawings, the same components will be given the same reference numerals regardless of the reference numerals, and redundant description thereof will be omitted.

[0028] FIG. 1 is a diagram illustrating how a fault diagnosis device according to an embodiment diagnoses whether or not a fault exists in equipment for each frequency band.

[0029] Referring to FIG. 1, a fault diagnosis device (e.g., fault diagnosis device 300 of FIG. 3) can output a signal (hereinafter, referred to as a reconstructed signal) similar to an input vibration frequency signal using a signal reconstruction model (e.g., signal reconstruction model 220 of FIG. 2) that has completed learning. The fault diagnosis device can diagnose equipment faults by comparing the difference between the input frequency vibration signal and the reconstructed signal (hereinafter, referred to as a reconstruction error) with a predefined threshold. The fault diagnosis device can diagnose the presence or absence of equipment faults in specific frequency regions, such as low frequency and high frequency regions, by comparing the reconstruction error with a threshold for each frequency band. The fault diagnosis device can diagnose the presence or absence of equipment abnormalities for each frequency band. The fault diagnosis device can diagnose the presence or absence of equipment abnormalities using the signal reconstructed by the signal reconstruction model in each frequency band and the reconstruction error. The number and range of divided frequency bands can vary depending on the operating environment of the equipment, etc. The fault diagnosis device can diagnose whether or not there is an abnormality in the facility equipment for each frequency band by comparing the reconstruction error for each frequency band with a threshold value predefined for each frequency band. The fault diagnosis device can diagnose whether or not there is an abnormality for each frequency band and diagnose whether or not there is an abnormality in the facility equipment by performing an OR operation on the abnormality determination results for each frequency band. Since the fault diagnosis device determines the abnormality result for each frequency band, information on whether or not there is an abnormality in a specific frequency band can be obtained. In one embodiment, graph 100 is a graph showing a vibration frequency signal and a reconstruction signal of the facility equipment. The frequency band of graph 100 can be divided into a first frequency band 110 and a second frequency band 120. The fault diagnosis device can diagnose that there is an abnormality in the equipment related to the first frequency band 110 if the first reconstruction error value of the first frequency band 110 is greater than the threshold value of the first frequency band 110 (e.g., if "yes" in operation 130), and can diagnose that there is an abnormality in the equipment related to the second frequency band 120 if the second reconstruction error value is greater than the threshold value of the second frequency band 120 (e.g., if "yes" in operation 140).The fault diagnosis device can diagnose that the equipment is faulty when it diagnoses that there is an abnormality in the equipment in at least one frequency band among the multiple frequency bands. The fault diagnosis device can diagnose that the equipment is normal when the first reconstruction error value of the first frequency band 110 is smaller than the threshold value of the first frequency band 110 (e.g., if "No" in operation 130) and the second reconstruction error value is smaller than the threshold value of the second frequency band 120 (e.g., if "NO" in operation 140). The fault diagnosis device can determine that the equipment is normal when all of the abnormality diagnosis results in the first frequency band 110 and the second frequency band 120 are normal.

[0030] FIG. 2 is a block diagram illustrating the configuration of a learning device according to an embodiment.

[0031] Referring to FIG. 2 , the learning device 200 can train a signal reconstruction model using a vibration frequency signal. The learning device 200 may include a frequency signal acquirer 210, a signal reconstruction model 220, a reconstruction error value determiner 230, and a threshold determiner 240. According to an embodiment, the frequency signal acquirer 210 may acquire a vibration frequency signal measuring vibrations generated during normal operation of equipment via a vibration sensor. For example, the learning device 200 may acquire a vibration frequency signal during normal operation of equipment using a vibration sensor such as a micro-electromechanical system (MEMS) or integrated electronics piezoelectric (IEPE) type. The frequency band range of the acquired vibration frequency signal may vary depending on the specifications of the vibration sensor and the environment in which the signal is collected. The acquired vibration frequency signal may be represented on a graph as a function of vibration frequency. In an embodiment, the trained signal reconstruction model 220 may generate a reconstructed signal corresponding to the vibration frequency signal using the acquired vibration frequency signal. The learning device 200 may acquire the reconstructed signal from the trained signal reconstruction model 220. The signal reconstruction model 220 generating a reconstructed signal will be described in detail with reference to FIG. 4. The reconstruction error value determiner 230 according to an embodiment may determine a reconstruction error value. The reconstruction error value is a difference between a vibration frequency signal and a reconstructed signal. The reconstruction error value determiner 230 determining a reconstruction error value will be described in detail with reference to FIG. 6. The threshold determiner 240 according to an embodiment may determine a threshold for the reconstruction error value for each of a plurality of predefined frequency bands. The threshold determiner 240 determining a threshold for the error value will be described in detail with reference to FIGS. 7 and 8.

[0032] FIG. 3 is a block diagram for explaining the configuration of the fault diagnosis device according to one embodiment.

[0033] Referring to FIG. 3, the fault diagnosis device 300 may diagnose whether or not an equipment fault exists using the signal reconstruction model 220. The fault diagnosis device 300 may include a frequency signal acquirer 210, a signal reconstruction model 220, a reconstruction error value determiner 230, and an abnormality determiner 310. The frequency signal acquirer 210, the signal reconstruction model 220, and the reconstruction error value determiner 230 correspond to the frequency signal acquirer 210, the signal reconstruction model 220, and the reconstruction error value determiner 230 of FIG. 2, and a description thereof will be omitted below. According to an embodiment, the abnormality determiner 310 may determine whether or not an abnormality exists in the equipment. For example, the abnormality determiner 310 may determine whether or not an abnormality exists in the equipment based on the reconstruction error value determined for each frequency band and the threshold value determined for each frequency band. When the abnormality determiner 310 diagnoses whether or not an abnormality exists for each frequency band, it may diagnose whether or not an abnormality exists in a portion of the equipment corresponding to a specific frequency band. The frequency band may vary depending on the vibration sensor and the environment in which the equipment is operated. The determination by the abnormality determination unit 310 as to whether or not there is an abnormality in the equipment will be described in detail with reference to FIG.

[0034] FIG. 4 is a diagram illustrating determining a reconstruction error value using a signal reconstruction model according to an embodiment.

[0035] Referring to FIG. 4, the signal reconstruction model 220 may receive an input of a vibration frequency signal 410 and generate a reconstructed signal 420. The signal reconstruction model may include at least one of an autoencoder, a stacked autoencoder, an LSTM autoencoder, and a convolutional autoencoder model. Determining a reconstruction error value using the signal reconstruction model 220 may be performed by a learning device (e.g., the learning device 200 of FIG. 2) and a fault diagnosis device (e.g., the fault diagnosis device 300 of FIG. 3). The learning device according to an embodiment trains the signal reconstruction model 220 using a vibration frequency signal 410 generated when equipment is normally operated, and the trained signal reconstruction model 220 may generate a reconstructed signal 420 similar to the vibration frequency signal 410. A reconstruction error value determiner (e.g., the reconstruction error value determiner 230 of FIG. 2) included in the learning device according to an embodiment may determine the difference between the vibration frequency signal 410 and the reconstructed signal 420 as the reconstruction error 430. For example, the learning device may determine, as the reconstruction error, a difference 430 between a signal included in the vibration frequency signal 410 and a signal included in the reconstructed signal 420 for each frequency band. According to an embodiment, a reconstruction error value determiner (e.g., reconstruction error value determiner 230 of FIG. 3) included in the fault diagnosis device inputs a vibration frequency signal (not shown) to a signal reconstruction model 220, and the signal reconstruction model 220 may generate a reconstructed signal (not shown) for the vibration frequency signal (not shown) generated when the equipment is operated. The fault diagnosis device may determine, as the reconstruction error, a signal (not shown) corresponding to the difference between the vibration frequency signal (not shown) and the reconstructed signal (not shown). The reconstruction error value determiner's determination of the reconstruction error will be described in detail with reference to FIG. 6.

[0036] FIG. 5 is a diagram illustrating determining a reconstruction error value for different frequency band ranges according to an embodiment.

[0037] Referring to FIG. 5, a reconstruction error value determiner (e.g., reconstruction error value determiner 230 of FIG. 2) included in a learning device (e.g., learning device 200 of FIG. 2) or a fault diagnosis device (e.g., fault diagnosis device 300 of FIG. 3) may determine a reconstruction error value for different frequency band ranges. The frequency band ranges according to an embodiment may vary depending on the vibration sensor and the environment in which the equipment is operated. For example, graph 510 may include two frequency bands 511 and 512 having similar frequency band ranges. The reconstruction error value determiner may determine an error value for each of the vibration frequency signals included in frequency band 511 and frequency band 512. Graph 520 may include frequency band 521 and frequency band 522. Frequency band 521 includes a vibration frequency signal with the maximum intensity among the vibration frequency signals, and the range of the frequency band may be smaller than that of frequency band 522. Graph 530 may include frequency band 531 and frequency band 532. Frequency band 531 includes most of the vibration frequency signals and may have a larger range than frequency band 532 .

[0038] FIG. 6 is a diagram illustrating determining a reconstruction error value for a particular vibration frequency signal according to one embodiment.

[0039] 6, a reconstruction error value determiner (e.g., reconstruction error value determiner 230 in FIG. 2) can determine a reconstruction error value using a vibration frequency signal and a reconstructed signal. Graph 610 is a graph showing the vibration frequency signal over the entire frequency range of the measured vibration frequency signal. Graph 620 is a graph in which the vibration frequency signal with the maximum intensity is enlarged. According to an embodiment, the reconstruction error value determiner can determine the reconstruction error value based on the intensity of vibration frequency signal 621 at frequency 623 and the intensity of reconstructed signal 622. For example, the reconstruction error value determiner can determine the reconstruction error value based on at least one of the average value of the difference between the input signal and the reconstructed signal, the average value of the squared difference, and the square root value of the average value of the squared difference.

[0040] The reconstruction error value determination unit determines the reconstruction error value using the average value of the difference between the input vibration frequency signal and the reconstruction signal, which can be expressed by the following Equation 1.

[0041] [Formula 1] JPEG2025539291000002.jpg1641

[0042] The reconstruction error value determination unit determines the reconstruction error value using the average value of the square of the difference between the input vibration frequency signal and the reconstruction signal, which can be expressed by the following Equation 2.

[0043] [Formula 2] JPEG2025539291000003.jpg1643

[0044] The reconstruction error value determination unit determines the reconstruction error value using the square root of the average value of the square of the difference between the input vibration frequency signal and the reconstruction signal, which can be expressed by the following Equation 3.

[0045] [Formula 3] JPEG2025539291000004.jpg2643

[0046] JPEG2025539291000005.jpg14170

[0047] FIG. 7 is a diagram for explaining how a learning device according to an embodiment determines a threshold value.

[0048] Referring to FIG. 7, a threshold determiner (e.g., threshold determiner 240 of FIG. 2) included in a learning device (e.g., learning device 200 of FIG. 2) may determine a threshold based on the distribution of reconstruction error values. Graph 700 shows a histogram representing the strength of reconstruction error values ​​according to frequency domains. The learning device may determine a threshold using the reconstruction error distribution for a vibration frequency signal of equipment in normal operation. According to an embodiment, the threshold determiner may determine a threshold for the reconstruction error value based on a maximum value 720 of the reconstruction error values ​​or a 3-Sigma rule. For example, the threshold determiner may determine the maximum value 720 of the reconstruction error values ​​as the threshold, or may determine a value 710 that is three times the standard deviation of the reconstruction error values ​​from the average value of the reconstruction error values ​​as the threshold.

[0049] FIG. 8 is a diagram for explaining that the learning device according to one embodiment determines threshold values ​​for a plurality of frequency bands.

[0050] 8, a threshold determiner (e.g., threshold determiner 240 of FIG. 2) included in the learning device (e.g., learning device 200 of FIG. 2) may determine a threshold for each frequency band (e.g., for each of N frequency bands). For example, graph 810 shows reconstruction error values ​​in a first frequency band. The threshold determiner may determine, as a first threshold, a maximum value 812 of the reconstruction error values ​​for the first frequency band or a value 811 that is three times the standard deviation of the reconstruction error values ​​from the average value of the reconstruction error values. Graph 820 shows reconstruction error values ​​in a second frequency band. The threshold determiner may determine, as a second threshold, a maximum value 822 of the reconstruction error values ​​for the second frequency band or a value 821 that is three times the standard deviation of the reconstruction error values ​​from the average value of the reconstruction error values.

[0051] FIG. 9 is a diagram for explaining how the fault diagnosis device according to one embodiment diagnoses whether or not there is a fault in the equipment.

[0052] Referring to FIG. 9, a fault diagnosis device (e.g., fault diagnosis device 300 of FIG. 3) may diagnose the presence or absence of an abnormality in equipment for each frequency band to determine whether or not the equipment is faulty. The fault diagnosis device may determine that the equipment is in a normal state if the presence or absence of an abnormality is determined to be normal in all frequency bands. In one embodiment, the fault diagnosis device may determine that the equipment is in an abnormal state if at least one of the reconstruction error values ​​determined for each frequency band is greater than a threshold determined for each frequency band. For example, if a first reconstruction error value is greater than a threshold for the first frequency band (e.g., if “yes” in operation 910) or a second reconstruction error value is greater than a threshold for the second frequency band (e.g., if “yes” in operation 920), the fault diagnosis device may determine that the equipment is in an abnormal state. In one embodiment, the fault diagnosis device may determine that the equipment is in a normal state if all of the reconstruction error values ​​determined for each frequency band are equal to or less than the threshold determined for each frequency band. For example, if the first reconstruction error value is less than or equal to the threshold value for the first frequency band (e.g., if "NO" in operation 910) and the second reconstruction error value is less than or equal to the threshold value for the second frequency band (e.g., if "NO" in operation 920), the fault diagnosis device can determine that the equipment is in a normal state.

[0053] FIG. 10 is a diagram illustrating how the fault diagnosis device according to an embodiment estimates the type of abnormality in the equipment for each frequency band.

[0054] Referring to FIG. 10, a fault diagnosis device (e.g., fault diagnosis device 300 of FIG. 3) can compare a reconstruction error value determined for each frequency band with a threshold value determined for each frequency band. When a frequency band in which the reconstruction error value is greater than the threshold value is detected as a result of comparing the reconstruction error value with the threshold value, the fault diagnosis device can estimate an abnormality type based on the detected frequency band. When the reconstruction error value is equal to or less than the threshold value in frequency band 1011 of graph 1010 and a vibration frequency signal is not detected in other frequency bands, the fault diagnosis device can determine that the equipment is in a normal state. When the reconstruction error value is greater than the threshold value in frequency band 1021 of graph 1020, the fault diagnosis device can estimate an abnormality type (e.g., 1X type) of the equipment associated with frequency band 1021. When the reconstruction error value is greater than the threshold value in each of frequency bands 1031, 1032, and 1033 of graph 1030, the fault diagnosis device can estimate an abnormality type (e.g., 1X type, 2X type, or 3X type) of the equipment associated with each of frequency bands 1031, 1032, and 1033. If the reconstruction error values ​​corresponding to frequency band 1041 and frequency band 1042 in graph 1040 are greater than a threshold value, the fault diagnosis device can estimate the abnormality type (e.g., 1X type, 2X type, 3X type, 4X type, 5X type, 6X type, 7X type, 8X type, 9X type) of the equipment associated with each frequency band.

[0055] FIG. 11 is a flowchart illustrating operations of a learning method performed by a learning device according to an embodiment. The learning method may be performed by a learning device (e.g., the learning device 200 of FIG. 2). In OPERATION 1110, the learning device may acquire a vibration frequency signal via a vibration sensor. The learning device may acquire a vibration frequency signal measuring vibrations generated during normal operation of equipment via the vibration sensor. For example, the learning device may acquire the vibration frequency signal from the equipment using a vibration sensor such as a MEMS or IEPE type. In OPERATION 1120, the learning device may train a signal reconstruction model using the vibration frequency signal. The learning device may train the signal reconstruction model to generate a reconstructed signal corresponding to the vibration frequency signal based on the acquired vibration frequency signal. The signal reconstruction model may include at least one of an autoencoder, a stacked autoencoder, an LSTM autoencoder, and a convolutional autoencoder model. In OPERATION 1130, the learning device may acquire a reconstructed signal from the signal reconstruction model. The trained signal reconstruction model may generate and output a reconstructed signal similar to the vibration frequency signal based on the input vibration frequency signal. The learning device may acquire a reconstructed signal corresponding to the frequency signal. In Operation 1140, the learning device may determine a reconstruction error value based on the reconstructed signal. The learning device may determine a reconstruction error value indicating the difference between the vibration frequency signal and the reconstructed signal. For example, the learning device may determine the reconstruction error value based on any of the average value of the differences between the vibration frequency signal and the reconstructed signal, the average value of the squared differences, and the square root value of the average value of the squared differences. In Operation 1150, the learning device may determine a threshold value for the reconstruction error value. The learning device may determine the threshold value for the reconstruction error value for each of a plurality of predefined frequency bands. For example, the learning device may determine the maximum value of the reconstruction error value as the threshold value, or may use the 3-Sigma rule to determine the threshold value as a value that is three times the standard deviation of the reconstruction error value from the average value of the reconstruction error value.

[0056] FIG. 12 is a flowchart illustrating operations of a fault diagnosis method performed by a fault diagnosis apparatus according to an embodiment. The operations of the fault diagnosis method may be performed by a fault diagnosis apparatus (e.g., the fault diagnosis apparatus 300 of FIG. 3) that diagnoses the presence or absence of an abnormality in equipment using a signal reconstruction model. In OPERATION 1210, the fault diagnosis apparatus may acquire a frequency signal via a vibration sensor. For example, the fault diagnosis apparatus may acquire a vibration frequency signal measuring vibrations occurring during equipment operation via a vibration sensor such as a MEMS or IEPE type. In OPERATION 1220, the fault diagnosis apparatus may acquire a reconstructed signal from a signal reconstruction model. The fault diagnosis apparatus may input the vibration frequency signal into a trained signal reconstruction model to acquire a reconstructed signal corresponding to the vibration frequency signal. For example, the fault diagnosis apparatus may acquire the reconstructed signal from a signal reconstruction model including at least one of an autoencoder, a stacked autoencoder, a LSTM autoencoder, and a convolutional autoencoder model. In OPERATION 1230, the fault diagnosis apparatus may determine a reconstruction error value based on the vibration frequency signal and the reconstructed signal. The fault diagnosis device may determine a reconstruction error value indicating a signal difference between the vibration frequency signal and the reconstructed signal for each predetermined frequency band. For example, the fault diagnosis device may determine the reconstruction error value for each frequency band based on the average value of the difference between the input signal and the reconstructed signal, the average value of the squared difference, or the square root value of the average value of the squared difference. In one embodiment, if at least one of the reconstruction error values ​​determined for each frequency band is not greater than a threshold value determined for each frequency band (e.g., if “NO” in operation 1240), the fault diagnosis device may determine (1250) that the equipment is in a normal state. For example, the fault diagnosis device may determine that the equipment is in a normal state if all of the reconstruction error values ​​determined for each frequency band are equal to or less than the threshold value determined for each frequency band.In one embodiment, if at least one of the reconstruction error values ​​determined for each frequency band is greater than the threshold determined for that frequency band (e.g., if operation 1240 returns "yes"), the fault diagnosis device may determine (1250) that the equipment is in an abnormal state. In one embodiment, the fault diagnosis device may determine whether the equipment is abnormal based on the reconstruction error value determined for each frequency band and the threshold determined for that frequency band. For example, if the fault diagnosis device compares the reconstruction error value determined for each frequency band with the threshold determined for that frequency band and detects a frequency band in which the reconstruction error value is greater than the threshold, the fault diagnosis device may estimate the type of abnormality based on the detected frequency band. The fault diagnosis device may detect equipment abnormalities such as equipment imbalance, misalignment, bearing faults, and looseness based on the detected frequency band. The fault diagnosis device may estimate the type of abnormality by detecting an equipment abnormality for each frequency band using the vibration frequency signal. The vibration frequency signal generated in the equipment exhibits high amplitude in a specific frequency range depending on the type of defect, and therefore the fault diagnosis device can estimate the type of abnormality in the equipment using the vibration frequency signal.

[0057] The above-described embodiments may be implemented using hardware components, software components, and / or a combination of hardware and software components. For example, the devices, methods, and components described in the embodiments may be implemented using a general-purpose computer or a special-purpose computer, such as a processor, controller, arithmetic logic unit (ALU), digital signal processor, microcomputer, field programmable gate array (FPGA), programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and software applications that run on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to the execution of software. For ease of understanding, a processing device may be described as being a single device, but those skilled in the art will recognize that a processing device may include multiple processing elements and / or multiple types of processing elements. For example, a processing device may include multiple processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

[0058] Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing device or, independently or collectively, instruct the processing device to operate as desired. The software and / or data may be permanently or temporarily embodied in some type of machine, component, physical device, virtual device, computer storage medium or device, or transmitted signal wave to be interpreted by the processing device or to provide instructions or data to the processing device. The software may also be distributed across computer systems connected by a network, stored or executed in a distributed manner. The software and data may be stored on a computer-readable recording medium.

[0059] Methods according to the embodiments may be embodied in the form of program instructions that can be executed by various computer means and may be recorded on a computer-readable medium. The computer-readable medium may store program instructions, data files, data structures, and the like, alone or in combination. The program instructions stored on the medium may be those specially designed and constructed for the embodiments, or may be those well known and available to those skilled in the art of computer software. Examples of computer-readable storage media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include not only machine language code produced by a compiler, but also high-level language code that can be executed by a computer using an interpreter, for example.

[0060] The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

[0061] Although the embodiments have been described above using limited drawings, those skilled in the art can apply various technical modifications and variations thereto. For example, the described techniques may be performed in a different order than described, and / or the components of the described systems, structures, devices, circuits, etc. may be combined or combined in a different manner than described, or replaced with other components or equivalents, and still achieve suitable results.

[0062] Accordingly, other implementations, other embodiments, and equivalents of the claims are within the scope of the following claims.

Claims

1. A learning device, comprising: a memory storing instructions executable by a processor; the processor; The processor: Vibration sensors are used to measure vibrations that occur when the equipment is operating normally, and vibration frequency signals are acquired. Based on the acquired vibration frequency signal, a signal reconstruction model is trained to generate a reconstructed signal corresponding to the vibration frequency signal; obtaining a reconstructed signal corresponding to the vibration frequency signal through the learned signal reconstruction model; determining a reconstruction error value indicative of the difference between the vibration frequency signal and the reconstructed signal; A learning device that determines a threshold value for the reconstruction error value for each of a plurality of predefined frequency bands.

2. The signal reconstruction model is 2. The learning device of claim 1, comprising at least one of an autoencoder, a stacked autoencoder, a long short-term memory autoencoder (LSTM) autoencoder, and a convolutional autoencoder model.

3. The processor: The learning device of claim 1 , wherein the reconstruction error value is determined based on at least one of an average value of a difference between the vibration frequency signal and the reconstructed signal, an average value of the square of the difference, and a square root value of the average value of the square of the difference.

4. The processor: The learning device according to claim 1 , wherein the threshold value for the reconstruction error value is determined based on a maximum value of the reconstruction error value or a 3 SIGMA rule.

5. A fault diagnosis device that diagnoses the presence or absence of an abnormality in equipment using a signal reconstruction model, a memory storing instructions executable by a processor; the processor; The processor: Vibration sensors measure vibrations occurring in the equipment and acquire vibration frequency signals. inputting the vibration frequency signal into a trained signal reconstruction model to obtain a reconstructed signal corresponding to the vibration frequency signal from the signal reconstruction model; determining a reconstruction error value indicating a signal difference between the vibration frequency signal and a reconstruction signal for each predefined frequency band; A fault diagnosis device that determines whether or not there is an abnormality in the equipment based on the reconstruction error value determined for each frequency band and the threshold value determined for each frequency band.

6. The processor: The fault diagnosis device of claim 5 , wherein if at least one of the reconstruction error values ​​determined for each frequency band is greater than a threshold determined for each frequency band, it is determined that there is an abnormality in the equipment.

7. The processor: The fault diagnosis device according to claim 5 , wherein the device determines that the equipment is in a normal state when the total reconstruction error value determined for each frequency band is equal to or less than a threshold value determined for each frequency band.

8. The signal reconstruction model is The fault diagnosis device of claim 5 , comprising at least one of an autoencoder, a stacked autoencoder, an LSTM autoencoder, and a convolutional autoencoder model.

9. The processor: The fault diagnosis device according to claim 5 , wherein the reconstruction error value is determined based on at least one of an average value of the difference between the input signal and the reconstructed signal, an average value of the square of the difference, and a square root value of the average value of the square of the difference.

10. The processor:

6. The fault diagnosis device according to claim 5, wherein, when a frequency band having a reconstruction error value greater than the threshold value is detected as a result of comparing the reconstruction error value determined for each frequency band with the threshold value determined for each frequency band, an abnormality type is estimated based on the detected frequency band.

11. A learning method performed by a learning device, comprising: An operation of acquiring a vibration frequency signal that measures vibrations generated when the equipment is operating normally through a vibration sensor; an operation of training a signal reconstruction model that generates a reconstructed signal corresponding to the vibration frequency signal based on the acquired vibration frequency signal; obtaining a reconstructed signal corresponding to the vibration frequency signal via the learned signal reconstruction model; determining a reconstruction error value indicative of a difference between the vibration frequency signal and the reconstructed signal; and A learning method including an operation of determining a threshold value for the reconstruction error value for each of a plurality of predefined frequency bands.

12. The signal reconstruction model is 12. The method of claim 11, comprising at least one of an autoencoder, a stacked autoencoder, an LSTM autoencoder, and a convolutional autoencoder model.

13. The operation of determining the reconstruction error value comprises:

12. The training method of claim 11, comprising determining a reconstruction error value based on at least one of an average value of the difference between the vibration frequency signal and the reconstructed signal, an average value of the squared difference, and a square root value of the average value of the squared difference.

14. The act of determining the threshold value includes: The learning method according to claim 11 , wherein the threshold value for the reconstruction error value is determined based on a maximum value of the reconstruction error value or a 3 SIGMA rule.

15. A fault diagnosis method performed by a fault diagnosis device that diagnoses the presence or absence of an abnormality in equipment using a signal reconstruction model, comprising: An operation of acquiring a vibration frequency signal that measures vibrations occurring in the equipment via a vibration sensor; an operation of inputting the vibration frequency signal into a trained signal reconstruction model and obtaining a reconstructed signal corresponding to the vibration frequency signal from the signal reconstruction model; determining a reconstruction error value indicating a difference between the vibration frequency signal and a reconstructed signal for each predefined frequency band; and The fault diagnosis method includes determining whether or not there is an abnormality in the equipment based on the reconstruction error value determined for each frequency band and a threshold value determined for each frequency band.

16. The operation of determining whether an abnormality exists includes: The fault diagnosis method according to claim 15, further comprising: estimating that there is an abnormality in the equipment when at least one of the reconstruction errors determined for each frequency band is greater than a threshold determined for each frequency band.

17. The fault diagnosis method includes: The fault diagnosis method of claim 15, further comprising determining that the equipment is in a normal state if the total reconstruction error value determined for each frequency band is equal to or less than a threshold value determined for each frequency band.

18. The signal reconstruction model is 16. The fault diagnosis method of claim 15, comprising at least one of an autoencoder, a stacked autoencoder, an LSTM autoencoder, and a convolutional autoencoder model.

19. The operation of determining the reconstruction error value comprises: The fault diagnosis method of claim 15 , comprising determining a reconstruction error value based on an average value of the difference between the input signal and the reconstructed signal, an average value for the square of the difference, and a square root value of the average value for the square of the difference.

20. The fault diagnosis method includes:

16. The fault diagnosis method of claim 15, further comprising: when a frequency band in which a reconstruction error value is greater than a threshold value is detected as a result of comparing the reconstruction error value determined for each frequency band with a threshold value determined for each frequency band, estimating an abnormality type based on the detected frequency band.

Images