Vibration signal-based device and method for diagnosing gear box

Abstract

According to an embodiment of the present invention, a device for diagnosing a gear box may include: a preprocessor configured to preprocess a vibration signal obtained from the gear box; an estimator configured to estimate a driving speed of the gear box on the basis of the vibration signal and generate a first energy distribution with respect to a failure estimation signal on the basis of the driving speed; an extractor configured to extract a mono-component signal from the preprocessed vibration signal and perform amplitude demodulation on the mono-component signal to generate a second energy distribution with respect to the mono-component signal; and a diagnosis device configured to diagnose the gear box on the basis of similarity between the first energy distribution and the second energy distribution.

Description

Gearbox diagnostic device and method based on vibration signals

[0001] The present invention relates to a gearbox diagnostic device and method, and more particularly, to a device and method for performing fault diagnosis of a gearbox based on a vibration signal obtained from the gearbox.

[0002] Recently, extensive research has been conducted on data-driven fault diagnosis methods to improve the reliability and stability of rotating systems used in various mechanical equipment. In particular, machine learning-based fault diagnosis techniques, such as deep learning and neural network models, which can learn characteristic factors extracted from data, are being actively developed. However, to enhance the robustness of deep learning and / or neural network models, it is crucial to ensure that the characteristic factors used for learning reflect the physical characteristics of the rotating body. However, existing fault diagnosis methods simply utilize data-driven machine learning methods and do not consider the physical characteristics of the rotating body. Therefore, accurate fault diagnosis is difficult for machines and / or rotating bodies operating under diverse operating conditions due to noise caused by road conditions, transmission conditions, and other factors. Therefore, there is a need for methods that can more accurately diagnose faults in machines and / or rotating bodies operating under diverse operating conditions.

[0003] The technical problem of the present invention is to provide a device and method capable of more accurately diagnosing whether a machine and / or a rotating body is broken under various operating conditions based on a vibration signal obtained from a gear box.

[0004] Another technical problem of the present invention is to provide a device and method capable of diagnosing whether a gear box is faulty by considering the physical characteristics of the gear box.

[0005] The problems to be solved by the present invention are not limited to those mentioned above, and other problems to be solved that are not mentioned will be clearly understood by those skilled in the art to which the present invention pertains from the description below.

[0006] According to one embodiment of the present invention, a device for diagnosing a gear box may include a preprocessor configured to preprocess a vibration signal obtained from the gear box, an estimator configured to estimate an operating speed of the gear box based on the vibration signal and generate a first energy distribution for a fault estimation signal based on the operating speed, an extractor configured to extract a mono-component signal from the preprocessed vibration signal and perform amplitude demodulation on the mono-component signal to generate a second energy distribution for the mono-component signal, and a diagnostic device configured to diagnose the gear box based on a similarity between the first energy distribution and the second energy distribution.

[0007] Here, the preprocessor can convert the vibration signal into a signal in the time-frequency domain, calculate the average energy for each frequency in the acceleration section of the converted signal, and then divide the energy for each time period of each frequency by the average energy.

[0008] Additionally, the estimator can estimate a sprocket speed using harmonic signals for a GMF (Gear Mesh Frequency) of the gear box obtained from the vibration signal, and generate the first energy distribution based on a speed distribution for the sprocket speed.

[0009] At this time, the estimator can determine the frequency bands of harmonics for the GMF based on CAN (Control Area Network) data of the machine including the gear box, and estimate the sprocket speed by multiplying the harmonic signals obtained in each frequency band.

[0010] Additionally, the estimator can estimate the speed distribution by marginalizing the sprocket speed, and multiply the speed distribution by a reduction gear ratio of the gear box to generate the first energy distribution.

[0011] Meanwhile, the extractor can extract the single component signal by applying a band-pass filter to the preprocessed vibration signal, perform amplitude demodulation on the single component signal to extract an envelope of the single component signal, and then generate the second energy distribution based on the energy of the envelope.

[0012] Additionally, the diagnostic device can diagnose the gear box using a pre-learned diagnostic model based on normalized feature values ​​based on the similarity between the first energy distribution and the second energy distribution.

[0013] Additionally, the diagnostic device can determine that a fault exists in the gear box based on the similarity between the first energy distribution and the second energy distribution.

[0014] According to another embodiment of the present invention, a method for diagnosing a gearbox by a gearbox diagnosis device may include a step of preprocessing a vibration signal obtained from the gearbox, a step of generating a first energy distribution for a fault estimation signal based on an operating speed of the gearbox estimated based on the vibration signal, a step of performing amplitude demodulation on a single component signal extracted from the preprocessed vibration signal to generate a second energy distribution for the single component signal, and a step of diagnosing the gearbox based on a similarity between the first energy distribution and the second energy distribution.

[0015] According to another embodiment of the present invention, a computer program stored in a computer-readable recording medium may include instructions for causing the processor to perform a gearbox diagnosis method, the method comprising the steps of: preprocessing a vibration signal obtained from a gearbox; generating a first energy distribution for a fault estimation signal based on an operating speed of the gearbox estimated based on the vibration signal; performing amplitude demodulation on a single component signal extracted from the preprocessed vibration signal to generate a second energy distribution for the single component signal; and diagnosing the gearbox based on a similarity between the first energy distribution and the second energy distribution, when executed by a processor.

[0016] According to one embodiment of the present invention, since the similarity between the speed distribution estimated from the vibration signal and the distribution of the fault signal is used as a characteristic factor, the diagnostic model for diagnosing the gear box can learn the physical characteristics of the gear box, and thus robustness can be improved.

[0017] According to one embodiment of the present invention, since the gearbox is diagnosed using a diagnostic model learned based on characteristic factors reflecting the physical characteristics of the gearbox, the fault diagnosis performance for a machine used in a harsh environment and operating conditions and / or a gearbox included therein can be improved.

[0018] FIG. 1 is a drawing showing a gear box diagnostic device according to one embodiment of the present invention.

[0019] FIG. 2 is a drawing showing a process of preprocessing a vibration signal by a gear box diagnostic device according to one embodiment of the present invention.

[0020] FIG. 3 is a diagram showing a process of removing noise from a vibration signal by a gear box diagnostic device according to one embodiment of the present invention.

[0021] FIG. 4 is a diagram showing a process in which a gear box diagnostic device according to one embodiment of the present invention extracts speed information from a vibration signal.

[0022] FIG. 5 is a diagram showing a process in which a gearbox diagnostic device according to one embodiment of the present invention generates an energy distribution for a failure estimation signal based on speed information.

[0023] FIG. 6 is a diagram showing a process in which a gearbox diagnostic device according to one embodiment of the present invention generates an energy distribution for a fault estimation signal based on amplitude demodulation.

[0024] FIG. 7 is a diagram showing a process in which a speed estimation device according to one embodiment of the present invention diagnoses a gear box.

[0025] FIG. 8 is a drawing showing a gear box diagnosis method according to one embodiment of the present invention.

[0026] The advantages and features of the present invention, and methods for achieving them, will become clearer with reference to the embodiments described below together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below and may be implemented in various different forms. These embodiments are provided only to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. Accordingly, in some embodiments, well-known processing steps, well-known device structures, and well-known techniques are not specifically described to avoid ambiguity in the interpretation of the present invention.

[0027] The terms used in this specification have been selected from among widely used, current terms, taking into account the functions of the present invention. However, these terms may vary depending on the intentions of those skilled in the art, precedents, the emergence of new technologies, etc. Furthermore, in certain cases, terms may be arbitrarily selected by the applicant, and in such cases, their meanings will be described in detail in the relevant description of the invention. Therefore, the terms used in this specification should be interpreted based on their inherent meanings and the overall content of the present invention, rather than simply their names.

[0028] When it is said throughout this specification that a part "includes" a component, this does not mean that it excludes other components, but rather that it may include other components, unless otherwise stated.

[0029] In addition, the components within the gearbox diagnostic device in this specification may mean software or hardware components such as FPGA (Field Programmable Gate Array) or ASIC (Application Specific Integrated Circuit), and perform at least one function or operation. However, the components are not limited to software or hardware. The components may be configured to be on an addressable storage medium, and may be configured to reproduce one or more processors. Accordingly, as an example, the components include components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, properties, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The functions provided by the components of the present invention may be combined into a smaller number of components or further separated into additional components.

[0030] Additionally, while terms such as first, second, and third may be used herein to describe various components, these components are not limited by these terms. These terms are used to distinguish one component from another.

[0031] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings so that those skilled in the art can easily implement the present invention. In order to clearly illustrate the present invention, portions irrelevant to the description in the drawings are omitted. Like reference numerals designate like elements throughout the specification.

[0032] FIG. 1 is a drawing showing a gear box diagnostic device according to one embodiment of the present invention.

[0033] Referring to FIG. 1, a gear box diagnostic device (100) according to one embodiment of the present invention may include a preprocessor (110), an estimator (120), an extractor (130), and a diagnostic device (140).

[0034] The preprocessor (110) can preprocess a vibration signal (or vibration data) obtained from a gear box to remove noise from the vibration signal. Here, the gear box is a device including a plurality of gears for power transmission, and may include a planetary gear box equipped in a driving part of a construction machine such as an excavator. The gear box may be used interchangeably with terms such as gear case and gear set.

[0035] For example, the preprocessor (110) converts a vibration signal obtained from a gear box into a signal in the time-frequency domain, calculates the average energy for each frequency in the acceleration section of the signal converted into the time-frequency domain, and then divides the energy for each frequency by the average energy to remove noise from the vibration signal.

[0036] The estimator (120) may be configured to estimate the operating speed of the gearbox based on a vibration signal obtained from the gearbox, and generate an energy distribution for a signal estimated to be a failure based on the operating speed (hereinafter, referred to as a failure estimation signal). Hereinafter, the energy distribution generated based on the speed information estimated by the estimator (120) is referred to as a first energy distribution.

[0037] For example, the estimator (120) can obtain harmonic signals for the GMF (Gear Mesh Frequency) of the gearbox from a vibration signal, and estimate the sprocket speed of the corresponding gearbox using the harmonic signals. In addition, the estimator (120) can generate a first energy distribution based on the speed distribution for the sprocket speed. To this end, for example, the estimator (120) can determine the frequency bands of the GMF harmonics based on CAN (Control Area Network) data of the machine including the gearbox, and estimate the sprocket speed by multiplying the harmonic signals obtained in each frequency band. In addition, the estimator (120) can marginalize the sprocket speed to estimate or generate a speed distribution (probability distribution for speed), and multiply the speed distribution by a reduction gear ratio of the corresponding gearbox to generate the first energy distribution.

[0038] The extractor (130) may be configured to extract a mono-component signal from a vibration signal preprocessed by the preprocessor (110), i.e., a vibration signal from which noise has been removed, and perform amplitude demodulation on the mono-component signal to generate an energy distribution for the mono-component signal. Hereinafter, the energy distribution generated based on amplitude demodulation in the extractor (130) is referred to as a second energy distribution.

[0039] For example, the extractor (130) may apply a band-pass filter to the preprocessed vibration signal to extract one component signal from among multiple harmonic signals included in the vibration signal, and perform amplitude demodulation on the extracted component signal to extract the envelope of the signal. In addition, the extractor (130) may generate a second energy distribution based on the energy of the envelope.

[0040] The diagnostic device (140) may be configured to diagnose the gearbox based on the similarity between the first energy distribution generated by the estimator (120) and the second energy distribution generated by the extractor (130). For example, the diagnostic device (140) may diagnose the gearbox using a pre-learned diagnostic model based on normalized feature values ​​based on the similarity between the first energy distribution and the second energy distribution. In this case, the diagnostic device (140) may determine that there is a fault in the gearbox if the first energy distribution and the second energy distribution are similar to each other, that is, if the similarity between the first energy distribution and the second energy distribution is high, and may determine that the gearbox is normal if the first energy distribution and the second energy distribution are not similar to each other, that is, if the similarity between the first energy distribution and the second energy distribution is low.

[0041] Hereinafter, a process of diagnosing a gear box by a gear box diagnostic device according to one embodiment of the present invention will be described in more detail.

[0042] FIG. 2 is a drawing showing a process in which a gear box diagnostic device according to one embodiment of the present invention preprocesses a vibration signal, and FIG. 3 is a drawing showing a process in which a gear box diagnostic device according to one embodiment of the present invention removes noise from a vibration signal.

[0043] According to one embodiment of the present invention, a gearbox diagnostic device can determine whether a gearbox is faulty by using, as a characteristic factor, the similarity between the first energy distribution derived based on speed information estimated from a vibration signal and the second energy distribution of a single component signal extracted from the vibration signal. Therefore, in order to accurately estimate speed and / or extract a fault signal, it is necessary to first remove noise components of a certain frequency existing in the vibration signal due to resonance, electrical noise, etc. caused by a harsh environment. Referring to FIG. 2, a frequency band (210) where noise exists exhibits high energy even in an acceleration section (220). Therefore, the gearbox diagnostic device can convert a one-dimensional vibration signal into a two-dimensional image using a time-frequency representation technique such as STFT (Short Time Fourier Transform), and calculate an average energy value for each frequency in the acceleration section (220) using the following mathematical expression 1.

[0044]

[0045] In addition, the gearbox diagnostic device can remove noise from the vibration signal by normalizing the energy value of each frequency in the entire time-frequency domain by dividing it by the average energy value of the corresponding frequency in the acceleration section derived based on mathematical expression 1 using the following mathematical expression 2.

[0046]

[0047] For example, as illustrated in FIG. 3, if there is noise indicating an energy value of '5' at a specific frequency (310) in the time-frequency representation, dividing the time-domain energy value of the frequency (310) by '5', which is the average energy value of the specific frequency (310) in the acceleration section, results in 5 / 5 = 1, confirming that the noise is removed. In the case of a frequency in which the time-domain energy value of the acceleration section represents (1, 1, 1, 7), the average energy value of the frequency is '2.5', and if the time-domain energy value of each time-domain of the frequency is divided by '2.5', which is the average energy value of the frequency, the time-domain energy value of the frequency becomes (0.4, 0.4, 0.4, 2.8). Therefore, if noise of a certain frequency is removed from a vibration signal using mathematical expressions 1 and 2, signal distortion is prevented, enabling accurate speed estimation and / or fault signal extraction from the signal.

[0048] FIG. 4 is a diagram showing a process in which a gear box diagnostic device according to one embodiment of the present invention extracts speed information from a vibration signal, and FIG. 5 is a diagram showing a process in which a gear box diagnostic device according to one embodiment of the present invention generates an energy distribution for a fault estimation signal based on speed information.

[0049] A gearbox diagnostic device according to one embodiment of the present invention can directly estimate the operating speed of a corresponding gearbox based on a vibration signal converted into the time-frequency domain. In this case, multiple harmonics, rather than a single harmonic, can be used for robustness.

[0050] For example, referring to FIG. 4, the gearbox diagnostic device sets the range of the sprocket speed (or operating speed) based on the CAN (Control Area Network) data of the machine including the gearbox, and calculates the frequency band corresponding to each harmonic of the GMF with the gear ratio of the corresponding gearbox, thereby separating the frequency band corresponding to each harmonic of the GMF from the image (or data) generated through STFT. In addition, the gearbox diagnostic device can derive the probability density function for the sprocket frequency by considering the harmonic signal of each frequency band as a probability density function and multiplying (overlapping) each probability density function equally. At this time, the gearbox diagnostic device can utilize three harmonics (1st harmonic, 2nd harmonic, and 3rd harmonic) for the GMF when estimating the sprocket speed as illustrated in FIG. 4, and for this purpose, for example, the gearbox diagnostic device can use the Multi-Order Probabilistic Approach (MOPA).

[0051] The operating speed of the gearbox is proportional to the frequency of the fault signal. Therefore, when a probability density function for the sprocket speed is derived through the above-described process, the gearbox diagnostic device can estimate the sprocket speed based on the probability density function as illustrated in Fig. 5, and estimate or generate a speed distribution for the sprocket speed by marginalizing the estimated sprocket speed. In addition, the gearbox diagnostic device can generate a first energy distribution for the fault estimation signal related to the sprocket speed by multiplying the speed distribution by the reduction gear ratio or gear ratio of the gearbox.

[0052] FIG. 6 is a diagram showing a process in which a gearbox diagnostic device according to one embodiment of the present invention generates an energy distribution for a fault estimation signal based on amplitude demodulation.

[0053] Referring to FIG. 6, the gearbox diagnostic device can extract a single component signal (600) from a preprocessed vibration signal by applying a band-pass filter to the preprocessed vibration signal. Here, the preprocessed vibration signal refers to a vibration signal from which noise has been removed from a raw signal obtained from a gearbox, and may refer to, for example, a vibration signal obtained by performing an ISTFT (Inverse STFT) on an image (or data) from which noise of a certain frequency has been removed based on the mathematical expressions 1 and 2. In addition, the single component signal (600) may refer to a signal for one harmonic among a plurality of harmonics for GMF.

[0054] If a fault occurs in the gearbox, a vibration signal obtained from the gearbox has sideband components (610) around the center frequency (GMF or its harmonics) as illustrated in FIG. 6. The gap between the center frequency and the sidebands indicates a modulation related to the fault. If there are no sideband components, an energy change (620) of the center frequency may indicate a modulation related to the fault. Therefore, the gearbox diagnostic device can perform amplitude demodulation based on a Hilbert transform or the like on a single component signal extracted through a bandpass filter to extract an envelope of the signal, and then generate a second energy distribution based on the energy of the envelope. If a fault occurs in the gearbox, the envelope of the fault signal (modulated signal) can be extracted by amplitude demodulation, and if the gearbox is normal, the envelope of the normal signal (GMF or its harmonic signal) can be extracted by amplitude demodulation.

[0055] FIG. 7 is a diagram showing a process in which a speed estimation device according to one embodiment of the present invention diagnoses a gear box.

[0056] Referring to Fig. 7, for example, an energy distribution generated based on the MOPA result, an energy distribution generated based on the envelope of a fault signal, and an energy distribution generated based on the envelope of a settlement signal are illustrated. The envelope energy of a normal signal has a constant distribution in all frequency bands, as illustrated in Fig. 7. On the other hand, the envelope energy of a fault signal has a distribution similar to the energy distribution generated from the operating speed. Therefore, when the energy distribution generated based on the MOPA result (first energy distribution) and the envelope energy distribution generated based on the amplitude demodulation result (second energy distribution) are similar to each other, it can be determined that a fault exists in the corresponding gear box, and when the first energy distribution and the second energy distribution are not similar to each other, the corresponding gear box can be determined to be normal.

[0057] For this purpose, as an example, the gearbox diagnostic device can quantify the similarity between the first energy distribution and the second energy distribution through KLD (Kullback-Leibler divergence), and derive a normalized feature value using the following mathematical expression 3 based on the similarity between the first energy distribution and the second energy distribution.

[0058]

[0059] The gearbox diagnostic device can input the feature values ​​derived as in mathematical expression 3 into a pre-learned diagnostic model and diagnose the gearbox based on the classification results of the diagnostic model. The diagnostic model can be implemented as, for example, a binary linear classification model, and its cost function can be determined by considering the optimization of hyperparameters and data balance.

[0060] FIG. 8 is a drawing showing a gear box diagnosis method according to one embodiment of the present invention.

[0061] Referring to FIG. 8, a gearbox diagnostic device according to an embodiment of the present invention can preprocess a vibration signal obtained from a gearbox to remove noise of a certain frequency from the vibration signal (S800). For example, the gearbox diagnostic device performs STFT on a vibration signal obtained from a gearbox to generate an image expressing the vibration signal in the time-frequency domain, calculates an average energy for each frequency in an acceleration section of the vibration signal based on the image, and then divides the energy for the entire time period of each frequency by the average energy, thereby removing noise from the vibration signal. Here, the image may be called a spectrogram.

[0062] Meanwhile, the gearbox diagnostic device can generate a first energy distribution for a fault estimation signal based on the operating speed of the gearbox estimated based on the vibration signal (S810). For example, the gearbox diagnostic device can obtain harmonic signals for the GMF (Gear Mesh Frequency) of the gearbox from the image (spectrogram) converted through the STFT, and estimate the sprocket speed of the gearbox by multiplying the harmonic signals. In addition, the gearbox diagnostic device can generate the first energy distribution based on the speed distribution for the sprocket speed. To this end, the gearbox diagnostic device determines a range of sprocket speeds based on CAN data of a machine including the gearbox, calculates a frequency band corresponding to each harmonic of GMF by multiplying the sprocket speed by a gear ratio of the gearbox, thereby dividing the image into frequency bands corresponding to each GMF harmonic, and multiplying harmonic signals (probability density functions) obtained in each frequency band based on MOPA to estimate the sprocket speed (probability density function for the sprocket frequency). Thereafter, the gearbox diagnostic device marginalizes the sprocket speed to estimate a speed distribution, and multiplies the speed distribution by a reduction ratio of the gearbox to generate a first energy distribution.

[0063] In addition, the gearbox diagnostic device can perform amplitude demodulation on a single component signal extracted from the preprocessed vibration signal, i.e., a noise-removed vibration signal, to generate a second energy distribution (S820). For example, the gearbox diagnostic device can bandpass filter the preprocessed vibration signal to extract one component signal from among a plurality of harmonic signals included in the signal, perform Hilbert transform-based amplitude demodulation on the extracted component signal to extract an envelope of the signal, and then generate a second energy distribution based on the energy of the envelope.

[0064] When the first energy distribution and the second energy distribution are generated, the gearbox diagnostic device can diagnose the corresponding gearbox based on the similarity between the first energy distribution and the second energy distribution (S830). For example, the gearbox diagnostic device can quantify the similarity between the first energy distribution and the second energy distribution to derive a feature value, normalize it, and input it into a pre-learned diagnostic model. Then, based on the classification result of the diagnostic model, it can determine whether the corresponding gearbox is faulty. If the feature value is greater than a threshold value (e.g., 0), that is, if the similarity between the first energy distribution and the second energy distribution is high, the gearbox diagnostic device can determine that the corresponding gearbox is faulty. However, if the feature value is less than the threshold value, that is, if the similarity between the first energy distribution and the second energy distribution is low, the gearbox diagnostic device can determine that the corresponding gearbox is normal.

[0065] Meanwhile, each step included in the gearbox diagnosis method performed by the gearbox diagnosis device according to the above-described embodiment can be implemented as a computer program including commands for causing a processor to perform these steps.

[0066] In addition, each step included in the gearbox diagnosis method performed by the gearbox diagnosis device according to the above-described embodiment can be implemented in a computer-readable recording medium having recorded thereon a computer program including instructions for causing a processor to perform these steps.

[0067] The combination of each step of the flowchart attached to the present invention may be performed by computer program instructions. These computer program instructions may be installed in a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing equipment, so that the instructions executed by the processor of the computer or other programmable data processing equipment create a means for performing the functions described in each step of the flowchart. These computer program instructions may also be stored in a computer-usable or computer-readable recording medium that can direct a computer or other programmable data processing equipment to implement the functions in a specific manner, so that the instructions stored in the computer-usable or computer-readable recording medium can also produce a manufactured article that includes an instruction means for performing the functions described in each step of the flowchart. Since the computer program instructions can also be installed on a computer or other programmable data processing device, a series of operational steps can be performed on the computer or other programmable data processing device to create a computer-executable process, and the instructions that cause the computer or other programmable data processing device to perform can also provide steps for performing the functions described in each step of the flowchart.

[0068] Additionally, each step may represent a module, segment, or portion of code that includes one or more executable instructions for performing a specific logical function(s). Furthermore, in some alternative embodiments, the functions described in the steps may occur out of order. For example, two steps depicted in succession may be performed substantially simultaneously, or the steps may sometimes be performed in reverse order, depending on the corresponding function.

[0069] The above description is merely an illustrative illustration of the technical idea of ​​the present invention, and those skilled in the art will appreciate that various modifications and variations can be made without departing from the essential quality of the present invention. Therefore, the embodiments disclosed in the present invention are intended to illustrate, rather than limit, the technical idea of ​​the present invention, and the scope of the technical idea of ​​the present invention is not limited by these embodiments. The scope of protection of the present invention should be interpreted by the following claims, and all technical ideas within a scope equivalent thereto should be interpreted as being included in the scope of the rights of the present invention.

[0070] [Explanation of symbols]

[0071] 100: Gearbox Diagnostic Device

[0072] 110: Preprocessor

[0073] 120: Estimator

[0074] 130: Extractor

[0075] 140: Diagnostic Device

Claims

1. In a device for diagnosing a gear box, A preprocessor configured to preprocess a vibration signal obtained from the gear box; An estimator configured to estimate an operating speed of the gear box based on the vibration signal and generate a first energy distribution for a fault estimation signal based on the operating speed; An extractor configured to extract a mono-component signal from the preprocessed vibration signal and perform amplitude demodulation on the mono-component signal to generate a second energy distribution for the mono-component signal; and A diagnostic device configured to diagnose the gear box based on the similarity between the first energy distribution and the second energy distribution. A device comprising:

2. In paragraph 1, The above preprocessor is, A device characterized in that it converts the vibration signal into a signal in the time-frequency domain, calculates the average energy for each frequency in the acceleration section of the converted signal, and then divides the energy for each time period of each frequency by the average energy.

3. In paragraph 1, The above estimator is, A device characterized in that it estimates a sprocket speed using harmonic signals for the GMF (Gear Mesh Frequency) of the gear box obtained from the vibration signal, and generates the first energy distribution based on the speed distribution for the sprocket speed.

4. In paragraph 3, The above estimator is, A device characterized in that the frequency bands of harmonics for the GMF are determined based on CAN (Control Area Network) data of a machine including the gear box, and the sprocket speed is estimated by multiplying the harmonic signals obtained in each frequency band.

5. In paragraph 3, The above estimator is, A device characterized in that the speed distribution is estimated by marginalizing the sprocket speed, and the first energy distribution is generated by multiplying the speed distribution by the reduction gear ratio of the gear box.

6. In paragraph 1, The above extractor, A device characterized in that a band-pass filter is applied to the preprocessed vibration signal to extract the single component signal, amplitude demodulation is performed on the single component signal to extract an envelope of the single component signal, and then the second energy distribution is generated based on the energy of the envelope.

7. In paragraph 1, The above diagnostic device, A device characterized in that the gear box is diagnosed using a pre-learned diagnostic model based on normalized feature values ​​based on the similarity between the first energy distribution and the second energy distribution.

8. In paragraph 1, The above diagnostic device, A device characterized in that it is determined that a fault exists in the gear box based on the first energy distribution and the second energy distribution being similar to each other.

9. In a method for diagnosing a gearbox by a gearbox diagnostic device, A step of preprocessing a vibration signal obtained from the gear box; A step of generating a first energy distribution for a fault estimation signal based on an operating speed of the gear box estimated based on the vibration signal; A step of performing amplitude demodulation on a mono-component signal extracted from the preprocessed vibration signal to generate a second energy distribution for the mono-component signal; and A step of diagnosing the gear box based on the similarity between the first energy distribution and the second energy distribution. A method comprising:

10. In paragraph 9, The above preprocessing step is, A step of converting the above vibration signal into a signal in the time-frequency domain; A step of calculating the average energy for each frequency in the acceleration section of the above-mentioned converted signal; and A method characterized by comprising a step of dividing the energy of each frequency by the average energy.

11. In paragraph 9, The step of generating the above first energy distribution is: A step of estimating a sprocket speed using harmonic signals for the GMF (Gear Mesh Frequency) of the gear box obtained from the vibration signal; and A method characterized by comprising a step of generating the first energy distribution based on a speed distribution for the sprocket speed.

12. In paragraph 11, The above estimating steps are: A step of determining the frequency bands of harmonics for the GMF based on CAN (Control Area Network) data of the machine including the gear box; and A method characterized by including a step of estimating the sprocket speed by multiplying harmonic signals obtained in each of the above frequency bands.

13. In paragraph 11, The above generating steps are: A step of marginalizing the sprocket speed to generate the speed distribution; and A method characterized by comprising a step of generating the first energy distribution by multiplying the speed distribution by a reduction gear ratio of the gear box.

14. In paragraph 9, The step of generating the second energy distribution is: A step of extracting the single component signal by applying a band-pass filter to the preprocessed vibration signal; A step of performing amplitude demodulation on the single component signal to extract the envelope of the single component signal; and A method characterized by comprising a step of generating the second energy distribution based on the energy of the envelope.

15. In paragraph 9, The above diagnosing steps are: A method characterized by comprising a step of diagnosing the gear box using a pre-learned diagnostic model based on normalized feature values ​​based on the similarity between the first energy distribution and the second energy distribution.

16. In paragraph 9, The above diagnosing steps are: A method characterized by including a step of determining that a fault exists in the gear box based on the first energy distribution and the second energy distribution being similar to each other.

17. A computer program stored in a computer-readable recording medium, The above computer program, when executed by a processor, A step of preprocessing the vibration signal obtained from the gear box, A step of generating a first energy distribution for a fault estimation signal based on the operating speed of the gear box estimated based on the vibration signal; A step of performing amplitude demodulation on a mono-component signal extracted from the above-mentioned preprocessed vibration signal to generate a second energy distribution for the mono-component signal; A computer program comprising instructions for causing the processor to perform a gear box diagnosis method, the method including a step of diagnosing the gear box based on a similarity between the first energy distribution and the second energy distribution.

18. In paragraph 17, The above preprocessing step is: A step of converting the above vibration signal into a signal in the time-frequency domain, A step of calculating the average energy for each frequency in the acceleration section of the above-mentioned converted signal, A computer program characterized by comprising a step of dividing the energy of each frequency by the average energy.

19. In paragraph 17, The step of generating the above first energy distribution is: A step of estimating a sprocket speed using harmonics of the GMF (Gear Mesh Frequency) of the gear box obtained from the vibration signal, A computer program characterized by comprising a step of generating the first energy distribution based on the speed distribution for the sprocket speed.

20. In paragraph 17, The step of generating the second energy distribution is: A step of extracting the single component signal by applying a band-pass filter to the preprocessed vibration signal, A step of performing amplitude demodulation on the single component signal to extract the envelope of the single component signal, A computer program characterized by comprising a step of generating the second energy distribution based on the energy of the envelope.

Images