A method and system for detecting gear wear in a geared motor

Abstract

This invention relates to the field of gear testing technology, specifically to a method and system for detecting gear wear in a geared motor. The invention first obtains all fault characteristic frequencies based on the historical vibration signals of the faulty gear. Then, it divides the vibration signal of the gear under test into several local impact segments and obtains the local spectrum of each local impact segment and the kurtosis of each frequency within it. Furthermore, it combines the power distribution at each fault characteristic frequency in the time spectrum to fit a fuzzy estimation signal of the vibration signal. Then, it modulates the vibration signal based on the motor load conditions. Finally, it analyzes the spectral signal of the modulated vibration signal to assess the early wear type of the gear under test. This invention modulates the vibration signal by analyzing the masking interference of noise and motor load on local wear impacts, and further analyzes the frequency modulation and amplitude modulation effects in the frequency domain of the modulated vibration signal, thereby accurately distinguishing early wear types and improving the gear wear detection effect.

Description

Technical Field

[0001] This invention relates to the field of gear inspection technology, and specifically to a method and system for detecting gear wear in a geared motor. Background Technology

[0002] As a core transmission component in industrial equipment, geared motors are highly susceptible to gear wear, which directly impacts equipment lifespan and operational stability. Existing gear wear detection technologies primarily include vibration analysis, noise detection, lubricant particle monitoring, visual inspection, and dimensional measurement. Furthermore, with the development of intelligent monitoring technology, novel methods have emerged, such as fault identification based on the time-frequency characteristics of rotational speed signals and joint diagnosis of impact pulses and vibration.

[0003] However, these methods still have limitations in identifying wear characteristics under continuous operation; in particular, they cannot effectively identify early wear features caused by the propagation of micro-pitting and micro-cracks on the gear surface during continuous operation of the geared motor. Because the vibration signal amplitude generated by early wear is low, typically only 5%-10% of that under normal operating conditions, and is easily masked by background noise such as bearing friction and load fluctuations, traditional vibration spectrum analysis lacks sufficient sensitivity in identifying early gear wear. This makes it impossible to accurately distinguish between micro-pitting and micro-cracks, resulting in poor gear wear detection performance. Summary of the Invention

[0004] To address the technical problem of ineffective gear wear detection, the present invention aims to provide a method and system for detecting gear wear in a geared motor. The specific technical solution adopted is as follows:

[0005] A method for detecting gear wear in a geared motor, the method comprising:

[0006] The vibration signal of the gear under test is acquired, the historical vibration signal of the faulty gear is used as the sample signal, and the fault characteristic frequency of each sample signal is acquired.

[0007] Based on the energy changes of the vibration signal, it is divided into several local impact segments; the vibration signal is subjected to a short-time Fourier transform to obtain the local spectrum of each local impact segment in the time spectrum of the vibration signal; the kurtosis of each local impact segment is used as the kurtosis of each frequency in the corresponding local spectrum; in the time spectrum, based on the power distribution and kurtosis at each fault characteristic frequency, the fuzzy estimate value at each fault characteristic frequency is determined, and combined with the temporal sequence of each fault characteristic frequency in the time domain, the fuzzy estimate signal of the vibration signal is fitted.

[0008] Based on the fuzzy estimation signal and the sideband energy of each fault characteristic frequency, combined with the motor load, the vibration signal is modulated to obtain a modulated vibration signal. In the spectrum signal of the modulated vibration signal, based on the frequency band difference on both sides of the gear rotation frequency of the gear under test and the frequency band difference on both sides of the meshing frequency of the gear under test, combined with the sideband interval of the fault characteristic frequency of the vibration signal, the early wear type of the gear under test is evaluated.

[0009] Furthermore, the method for obtaining the fault characteristic frequency includes:

[0010] The target frequency band is determined based on the meshing frequency of the faulty gear corresponding to each sample signal; the low resonance component in the sample signal is extracted using the resonance sparse decomposition algorithm, and the component sub-segment corresponding to the target frequency band in the low resonance component is obtained.

[0011] Time delay energy accumulation analysis is performed on the component segments to determine the vibration period, and the reciprocal of the vibration period is used as the fault characteristic frequency of the corresponding sample signal.

[0012] Furthermore, the method for obtaining the local impact segment and the local spectrum includes:

[0013] Starting from the origin of the vibration signal, the vibration signal is divided into several local segments of corresponding window lengths using a window. Based on the total number of local segments and the dispersion of signal energy in all local segments, a window objective function is defined.

[0014] The initial window and the iteration step size of the window length are defined, and the window is obtained iteratively. When the window objective function takes the minimum value, the length of the corresponding window is taken as the target length. The local sub-segment corresponding to the target length is taken as the local impact sub-segment.

[0015] Using the target length as the length of the short-time window, the vibration signal is subjected to a short-time Fourier transform to obtain the time spectrum, and the spectrum in the time domain corresponding to each local impact segment in the time spectrum is used as the local spectrum of the corresponding local impact segment.

[0016] Furthermore, the method for obtaining the fuzzy estimate and the method for fitting the fuzzy estimate signal include:

[0017] In the local spectrum corresponding to each local impact segment, the ratio of the power spectral density to the corresponding kurtosis at each fault characteristic frequency is used as the fuzzy estimate.

[0018] In the time spectrum, the time corresponding to each fault characteristic frequency in the time domain is determined, and the fuzzy estimate of each fault characteristic frequency is mapped to a timestamp according to the corresponding time, and the fuzzy estimate signal is fitted.

[0019] Furthermore, the method for acquiring the modulated vibration signal includes:

[0020] At each fault characteristic frequency, the sideband energy of a preset order is aggregated on both sides to obtain the enhanced energy; the load fluctuation signal and its envelope signal of the geared motor are obtained, and the envelope signal and fuzzy estimation signal are fused to obtain the characteristic signal;

[0021] The negative correlation mapping result of the largest amplitude in the envelope signal is used as the interference weight. The enhanced energy at each fault characteristic frequency is weighted using the interference weight to obtain the confidence enhanced energy. The average amplitude level of the characteristic signal is used as the characteristic information. The confidence enhanced energy and the characteristic information are fused to obtain the modulation compensation amount at the corresponding fault characteristic frequency.

[0022] The vibration amplitude at the time corresponding to the fault characteristic frequency is modulated by the modulation compensation amount to obtain the modulated vibration signal.

[0023] Furthermore, methods for assessing the early wear type of the gear under test include:

[0024] In the spectrum signal of the modulated vibration signal, the first wear characteristic parameter is obtained based on the frequency band difference on both sides of the gear rotation frequency of the gear under test and the frequency band difference on both sides of the meshing frequency of the gear under test.

[0025] The frequency modulation sideband interval is determined based on the gear rotation frequency of the gear under test, and the amplitude modulation sideband interval is determined based on the meshing frequency of the gear under test; the actual sideband interval of the fault characteristic frequency in the spectrum signal of the vibration signal is obtained; the second wear characteristic parameter is obtained based on the deviation of the actual sideband interval from the frequency modulation sideband interval and the deviation of the actual sideband interval from the amplitude modulation sideband interval.

[0026] By combining the first wear characteristic parameter and the second wear characteristic parameter, the wear characteristic coefficient of the gear under test is obtained; the early wear type of the gear under test is determined based on the wear characteristic coefficient.

[0027] Furthermore, the method for obtaining the first wear characteristic parameter includes:

[0028] In the spectral signal of the modulated vibration signal, the frequency modulation asymmetry is obtained based on the difference between the total energy of a preset number of sidebands to the left of the gear frequency and the total energy of a preset number of sidebands to the right of the gear frequency; the amplitude modulation asymmetry is obtained based on the difference between the total energy of a preset number of sidebands to the left of the meshing frequency and the total energy of a preset number of sidebands to the right of the gear frequency; the amplitude modulation asymmetry is used as the denominator, the negative correlation normalized value of the frequency modulation asymmetry is used as the numerator, and the fractional ratio is used as the first wear characteristic parameter.

[0029] Furthermore, the method for obtaining the second wear characteristic parameter includes:

[0030] The gear rotation frequency is used as the frequency modulation sideband interval, and twice the meshing frequency is used as the amplitude modulation sideband interval. The negative correlation mapping result of the absolute value of the difference between the actual sideband interval and the frequency modulation sideband interval is used as the frequency modulation matching parameter. The negative correlation mapping result of the absolute value of the difference between the actual sideband interval and the amplitude modulation sideband interval is used as the amplitude modulation matching parameter. The amplitude modulation matching parameter is used as the denominator, the frequency modulation matching parameter is used as the numerator, and the fractional ratio is used as the second wear characteristic parameter.

[0031] Furthermore, the method for determining the early wear type of the gear under test based on the wear characteristic coefficient includes:

[0032] When the wear characteristic coefficient is greater than a preset first threshold, it is determined that the gear under test has micro-pitting wear; when the wear characteristic coefficient is less than a preset second threshold, it is determined that the gear under test has micro-crack wear; wherein, the preset first threshold is greater than the preset second threshold.

[0033] A gear wear detection system for a geared motor includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the gear wear detection method for the geared motor.

[0034] The present invention has the following beneficial effects:

[0035] This invention first acquires the vibration signal of the gear under test, using the historical vibration signal of the faulty gear as a sample signal and obtaining its fault characteristic frequency to provide historical reference for subsequent analysis of the wear impact characteristics of the vibration signal. Further, the vibration signal is divided into several local impact segments, and a short-time Fourier transform is performed to obtain the local spectrum of each local impact segment. Then, the kurtosis of each local impact segment is used as the kurtosis of each frequency in the corresponding local spectrum, preparing for subsequent evaluation of the ambiguity masking of wear impact by noise. In the time spectrum, based on the power distribution and kurtosis at each fault characteristic frequency, a fuzzy estimate is determined for each fault characteristic frequency, and a fitting is made to reflect the masking of the impact characteristics of the vibration signal. The fuzzy estimation signal of the situation prepares for subsequent enhancement of the covered sparse impact; then, based on the fuzzy estimation signal and the sideband energy of each fault characteristic frequency, combined with the motor load, the vibration signal is modulated to obtain the modulated vibration signal. The modulated signal features masked by motor load or noise are prepared for more accurate capture of wear impact characteristics. Finally, in the spectrum signal of the modulated vibration signal, based on the frequency band differences on both sides of the gear rotation frequency and the meshing frequency of the gear under test, combined with the sideband interval of the fault characteristic frequency of the vibration signal, the distinguishing features of the frequency modulation effect caused by micro-pitting and the amplitude modulation effect caused by micro-cracks are analyzed, thereby assessing the early wear type of the gear under test. This invention modulates the vibration signal by analyzing the masking interference of noise and motor load on local wear impact, and further analyzes the frequency modulation and amplitude modulation effects in the frequency domain of the modulated vibration signal, thereby accurately distinguishing the early wear type and improving the gear wear detection effect. Attached Figure Description

[0036] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 A flowchart illustrating a method for detecting gear wear in a geared motor, as provided in an embodiment of the present invention.

[0038] Figure 2 This is a flowchart of a method for evaluating the early wear type of a gear under test, provided as an embodiment of the present invention. Detailed Implementation

[0039] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a gear wear detection method and system for a geared motor according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0040] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0041] The following description, in conjunction with the accompanying drawings, details the specific scheme of the gear wear detection method and system for a geared motor provided by the present invention.

[0042] It should be noted that gear wear detection is a well-known existing technology in this field. Here is a brief description of part of the wear detection process:

[0043] 1. Acoustic emission co-detection module: A wideband acoustic emission sensor (frequency range 20-400kHz) is installed on the outer surface of the housing in the gear meshing area. The sensor is coupled by a waveguide rod to reduce signal attenuation and capture the elastic wave signal generated by pitting. The threshold triggering mode is adopted. When the amplitude of the elastic wave signal exceeds the triggering threshold (e.g., 40dB), a gear wear warning is issued.

[0044] 2. Online oil monitoring module: A ferrometer (1μm resolution) is connected in series in the lubricating oil circuit to detect the concentration of abrasive particles in the lubricating oil in real time (alarm threshold >200ppm); a moisture sensor is embedded in the bottom of the oil tank to monitor the free water content (warning value >0.1%); when the abrasive particle concentration or free water content exceeds the corresponding threshold, a gear wear warning is issued.

[0045] 3. High-precision vibration monitoring module: A triaxial MEMS accelerometer (range ±50g, sampling rate 50kHz) is installed on the surface of the housing near the bearing in the gear meshing area, such as at the support bearings of the input and output shafts. It is fixed with magnetic attraction or thread to ensure high-frequency signal coupling. Vibration signals are collected in the range of 5 times the meshing frequency, with a focus on monitoring the 0.5-5kHz frequency band. The vibration signals are decomposed into several intrinsic mode functions (IMFs) through empirical mode decomposition (EMD), and the instantaneous energy entropy of the IMF components in the 3-5kHz frequency band is calculated. When the instantaneous energy entropy value in the 3-5kHz frequency band is greater than 1.8, a gear wear warning is issued.

[0046] It should be noted that the above module can only provide wear warnings when the gear has already shown clear and obvious defects, which has a lag. The present invention is aimed at accurately distinguishing between micro-pitting and micro-cracks in the early wear type of gear. When no warnings are given during the monitoring process, the vibration signal is further analyzed using the following steps to assess the early wear type.

[0047] Please see Figure 1 The diagram illustrates a method flowchart for detecting gear wear in a geared motor according to an embodiment of the present invention, specifically including:

[0048] Step S1: Obtain the vibration signal of the gear under test, take the historical vibration signal of the faulty gear as the sample signal, and obtain the fault characteristic frequency of each sample signal.

[0049] In one embodiment of the present invention, the vibration signal of the gear under test is first obtained using the high-precision vibration monitoring module described above, which will not be elaborated further; at the same time, the historical vibration signal of the monitored faulty gear is used as a sample signal to analyze the periodic wear impact characteristics of each gear, evaluate the fault characteristic frequency in each sample signal, and provide historical reference for subsequent analysis of the impact of the vibration signal of the gear under test to evaluate the early wear type.

[0050] Preferably, in one embodiment of the present invention, considering that early wear impact has transient resonance characteristics and exhibits a sparse distribution in the time-frequency domain; and considering that the gear rotates continuously, the wear impact caused by micro-pitting and micro-cracks on the tooth surface will exhibit a periodic energy correlation near the theoretical meshing cycle, while noise does not have this characteristic; therefore, the target frequency band can first be determined based on the meshing frequency of the faulty gear, and then the periodicity of vibration impact within the target frequency band can be analyzed; and time-delay energy accumulation analysis can help determine the vibration impact cycle, thereby helping to determine the fault characteristic frequency; based on this, the method for obtaining the fault characteristic frequency includes:

[0051] The target frequency band is determined based on the meshing frequency of the faulty gear corresponding to each sample signal; the low resonance component in the sample signal is extracted using the resonance sparse decomposition algorithm, and the component sub-segment corresponding to the target frequency band in the low resonance component is obtained.

[0052] Time delay energy accumulation analysis is performed on the component segments to determine the vibration period, and the reciprocal of the vibration period is used as the fault characteristic frequency of the corresponding sample signal.

[0053] As an example, taking any sample signal as an example, first calculate the meshing frequency of the faulty gear. The target frequency band is defined as [0.8]. 1.2 The implementer can also use the meshing frequency. Using [0.9] as the central reference, automatically adjust the target frequency band as follows: 1.1 Then define a frequency domain detection window and mark the component sub-segments corresponding to the target frequency band in the low resonance components;

[0054] Using the same idea as the autocorrelation function, a time-delay energy accumulation function is constructed. ;in, The delay parameter is iterable; t is the time index of the signal point in the component segment; This represents the total number of signal points in the component segment; Let be the energy value of the signal point at time t in the component segment; For the first sub-segment of the component The energy value of the signal point at a given moment; change You can obtain different The time-delay energy accumulation function value is given below. When the low-resonance component has periodic characteristics, similar energy impact characteristics appear after a certain time delay, that is, the time-delay energy accumulation function will reach its maximum value; therefore, the value corresponding to the maximum value is... The vibration period is used to obtain the fault characteristic frequency of the sample signal.

[0055] In another embodiment of the present invention, after obtaining the low resonance component, autocorrelation analysis can also be performed directly on the low resonance component to determine the vibration period and fault characteristic frequency. This is already existing technology and will not be described in detail here.

[0056] It should be noted that obtaining the meshing frequency of the faulty gear is already existing technology and will not be elaborated further.

[0057] Step S2: Based on the energy change of the vibration signal, divide it into several local impact segments; perform a short-time Fourier transform on the vibration signal to obtain the local spectrum of each local impact segment in the time spectrum of the vibration signal; use the kurtosis of each local impact segment as the kurtosis of each frequency in the corresponding local spectrum; in the time spectrum, determine the fuzzy estimate value of each fault characteristic frequency based on the power distribution and kurtosis at each fault characteristic frequency, and combine the time sequence of each fault characteristic frequency in the time domain to fit the fuzzy estimate signal of the vibration signal.

[0058] Considering that when the gear under test exhibits micro-pitting or micro-crack wear, its corresponding vibration signal will also show characteristics of local impact, but mainly manifested as transient and random features, which are easily masked by random noise impacts, making it difficult to identify weak potential early wear; therefore, this embodiment of the invention first divides the vibration signal into several local impact segments based on the energy changes of the vibration signal, thereby facilitating the more accurate capture of the weak impact characteristics caused by early wear within the local impact segments, so as to accurately distinguish the type of early wear.

[0059] Furthermore, considering that performing a short-time Fourier transform on the vibration signal can help to simultaneously observe the time-domain and frequency-domain information of the local impact segment, thereby helping to more accurately capture the periodic impact components and frequency characteristics; therefore, this embodiment of the invention further performs a short-time Fourier transform on the vibration signal to obtain the time spectrum of the vibration signal, and obtains the local spectrum of each local impact segment in the time spectrum.

[0060] Preferably, in one embodiment of the present invention, considering that the more local impact segments there are, the more finely the vibration signal is divided, which can help identify minute changes in the signal and periodic impact characteristics caused by early wear; and the more uniform the energy within all local impact segments of the vibration signal, the less energy concentration can be avoided, reducing the excessive influence of noise on a single segment; therefore, a segmentation window can be set, and a segmentation target can be defined to divide the local impact segments; at the same time, a short-time Fourier transform of the vibration signal can be performed using the segmentation window to obtain the local spectrum of the local impact segments within the time spectrum; based on this, the method for obtaining the local impact segments and the local spectrum includes:

[0061] Starting from the origin of the vibration signal, the vibration signal is divided into several local segments of corresponding window lengths using a window. Based on the total number of local segments and the dispersion of signal energy in all local segments, the objective function of the window is defined.

[0062] The initial window and the iteration step size of the window length are defined, and the window is obtained iteratively. When the window objective function takes the minimum value, the length of the corresponding window is taken as the target length. The local sub-segment corresponding to the target length is taken as the local impact sub-segment.

[0063] Using the target length as the length of the short-time window, a short-time Fourier transform is performed on the vibration signal to obtain the time spectrum. The spectrum in the time domain corresponding to each local impact segment in the time spectrum is then used as the local spectrum of the corresponding local impact segment.

[0064] As an example, first define an initial window with a length of 20, and then set the iteration step size to 5. That is, the length of the window increases iteratively with 20, 25, and 30 until it reaches a preset threshold, such as half the total number of vibration signal points. Implementers can also customize the length of the initial window, the iteration step size, and the iteration cutoff condition.

[0065] Taking any iteration as an example, starting from the beginning of the vibration signal, the vibration signal is divided into several equal-length local segments using an iterative window; the objective function of the window is set as follows. Where w is the energy value of a local segment in the vibration signal; Let N be the variance of the energy values ​​of all local segments in the vibration signal; N be the total number of local segments in the vibration signal; e be the natural constant; in the window objective function, the degree of dispersion is measured in the form of variance, through... The logic adjusts the total number of local segments by negative correlation mapping in the form of a function. The larger the total number, the smaller the exponent value. Iterate the window length. When the window objective function reaches its minimum value, the length of the corresponding window is taken as the target length, so that all local impact segments can be obtained.

[0066] Furthermore, the short-time window of the short-time Fourier transform can be determined, thereby obtaining the time spectrum of the vibration signal. Then, the spectrum in the time domain corresponding to each local impact segment in the time spectrum can be used as the local spectrum of the corresponding local impact segment.

[0067] It should be noted that the short-time Fourier transform is a well-known technique and will not be elaborated further; finding the local spectrum of a local impactor in a time-frequency diagram is an existing technique and will not be elaborated further.

[0068] Considering that kurtosis can reflect the sharpness of row data distribution, the kurtosis of local impact segments can reflect their impact characteristics in the time domain; the greater the kurtosis, the more significant the impact characteristics. Also, considering that noise components have relatively low power density in the frequency domain, while periodic impact characteristics will show high power density at certain specific frequencies in the frequency domain; since fault characteristic frequencies can be used as historical fault references, when the power density of fault characteristic frequencies in the frequency domain of a local impact segment is greater and its kurtosis in the time domain is smaller, it indicates that it is more likely to be masked by noise, and its impact characteristics in the corresponding time series are more likely to be blurred.

[0069] Therefore, in this embodiment of the invention, the kurtosis of each local impact segment is first used as the kurtosis of each frequency in the corresponding local spectrum; then, in the time spectrum, based on the power distribution and kurtosis at each fault characteristic frequency, the fuzzy estimate value at each fault characteristic frequency is determined; the fuzzy estimate value reflects the degree to which the impact characteristics at the corresponding fault characteristic frequency are masked by noise, and can then be combined with the time spectrum to analyze the temporal sequence of each fault characteristic frequency in the time domain, fit the fuzzy estimate signal of the vibration signal, and prepare for subsequent modulation of the vibration signal, accurate analysis of the modulation effect and amplitude modulation effect of the vibration signal, and differentiation of early wear types.

[0070] Preferably, in one embodiment of the present invention, considering that a larger power spectral density indicates a greater likelihood of wear impact characteristics, and a smaller kurtosis indicates that the impact characteristics are not significant and are more likely to be masked by noise, the fuzzy estimation value is also larger. Using power spectral density as the numerator and kurtosis as the denominator can reflect the corresponding characteristic relationship. Furthermore, considering that after performing a short-time Fourier transform on the vibration signal, the spectral information at each time-series position can be observed in the time domain, thereby finding the time-series position corresponding to each fault characteristic frequency, and then fitting the fuzzy estimate at each fault characteristic frequency according to the time sequence to obtain the fuzzy estimation signal of the vibration signal; therefore, the method for obtaining the fuzzy estimation value and the method for fitting the fuzzy estimation signal include:

[0071] In the local spectrum corresponding to each local impact segment, the ratio of the power spectral density to the corresponding kurtosis at each fault characteristic frequency is used as the fuzzy estimate.

[0072] In the time spectrum, the time corresponding to each fault characteristic frequency in the time domain is determined, and the fuzzy estimate of each fault characteristic frequency is mapped to a timestamp according to the corresponding time, and the fuzzy estimate signal is fitted.

[0073] It should be noted that mapping the fuzzy estimate to a timestamp to fit the fuzzy estimate signal is an existing technique and will not be elaborated further. It is a time-domain signal that shows the situation where the impact characteristics of the vibration signal are masked, which prepares for subsequent enhancement of the masked sparse impact.

[0074] Step S3: Based on the fuzzy estimation signal and the sideband energy of each fault characteristic frequency, combined with the motor load, the vibration signal is modulated to obtain the modulated vibration signal; in the spectrum signal of the modulated vibration signal, based on the frequency band difference on both sides of the gear rotation frequency of the gear under test and the frequency band difference on both sides of the meshing frequency of the gear under test, combined with the sideband interval of the fault characteristic frequency of the vibration signal, the early wear type of the gear under test is evaluated.

[0075] Considering that aggregating the sideband energy of each fault characteristic frequency can help enhance the fault-related frequency components, thereby improving the fault impact characteristics in the signal; and considering that the motor load also affects the fluctuation of the vibration signal and may also mask the true impact characteristics of the fault signal, this embodiment of the invention will modulate the vibration signal based on the fuzzy estimation signal and the sideband energy of each fault characteristic frequency, combined with the motor load, to obtain a modulated vibration signal, thereby enhancing the wear impact characteristics in the vibration signal and preparing for subsequent assessment of early wear types.

[0076] Preferably, in one embodiment of the present invention, sideband energy is first aggregated at the fault characteristic frequency to enhance the fault characteristic frequency components, preparing for subsequent modulation enhancement; then, the load fluctuation signal of the geared motor is acquired, and envelope analysis is performed to extract the main load change trend, which is then fused with the fuzzy estimation signal to comprehensively evaluate the masking of the wear and impact characteristics of the vibration signal and obtain the characteristic signal; finally, the confidence enhancement energy is obtained by combining the maximum load interference in the envelope signal, and the average level of the characteristic signal is further used as a compensation term to retain the fuzzy interference effect of load and noise on the impact signal, thereby obtaining the modulation compensation amount for modulation; based on this, the method for obtaining the modulated vibration signal includes:

[0077] At each fault characteristic frequency, the sideband energy of a preset order is aggregated on both sides to obtain the enhanced energy; the load fluctuation signal and its envelope signal of the geared motor are obtained, and the envelope signal and fuzzy estimation signal are fused to obtain the characteristic signal;

[0078] The negative correlation mapping result of the largest amplitude in the envelope signal is used as the interference weight. The enhanced energy at each fault characteristic frequency is weighted using the interference weight to obtain the confidence enhanced energy. The average amplitude level of the characteristic signal is used as the feature information. The confidence enhanced energy and the feature information are fused to obtain the modulation compensation amount at the corresponding fault characteristic frequency.

[0079] The vibration amplitude at the time corresponding to the fault characteristic frequency is modulated by the modulation compensation amount to obtain the modulated vibration signal.

[0080] As an example, the preset order is set to 3, and the third-order sideband energy is aggregated on both sides of each fault characteristic frequency to obtain enhanced energy. The implementer can also customize the preset order. Furthermore, the load fluctuation signal and its envelope signal are obtained, and the mean signal of the envelope signal and the fuzzy estimation signal is used as the characteristic signal, which reflects the interference of load and noise on the wear impact characteristics.

[0081] The maximum load in the envelope signal is further divided by the reciprocal to obtain the interference weight. The interference weight is multiplied by the enhancement energy to obtain the confidence enhancement energy at each fault characteristic frequency. The feature information is further multiplied by the confidence enhancement energy and combined to obtain the modulation compensation amount at each fault characteristic frequency.

[0082] Finally, in the time spectrum, the corresponding time of each fault characteristic frequency in the time domain is obtained, and the product of the modulation compensation amount and the vibration amplitude at the corresponding time is taken as the modulation result, thus obtaining the modulated vibration signal.

[0083] It should be noted that aggregating sideband energy at a certain frequency is an existing technology well known to those skilled in the art and will not be elaborated further; the load fluctuation signal is obtained by installing a torque sensor on the gear shaft, and its sampling frequency is the same as that of the vibration signal, thus maintaining consistency in temporal sequence; the mean signal is the signal reconstructed by averaging the signal points of the two signals at the same acquisition time, and both it and envelope analysis are existing technologies and will not be elaborated further.

[0084] Considering that micropitting is manifested as small-scale damage or wear on the tooth surface, it can cause periodic changes in the meshing stiffness of the gear, which in turn can cause periodic frequency modulation components in the vibration signal of the gear transmission. Usually, the carrier frequency of this component corresponds to the gear rotation frequency, and a frequency band will be generated in the spectrum of the vibration signal. Furthermore, a frequency modulation effect may occur near the frequency band, which is reflected in the sideband components of the vibration signal.

[0085] Microcracks can cause changes in the contact force between gears during meshing, triggering energy migration in the resonant frequency band. This means that asymmetrical sidebands will appear on both sides of the meshing frequency in the spectrum of the vibration signal, usually manifested as an amplitude modulation effect. This asymmetry can help identify the presence and propagation of cracks.

[0086] Therefore, after acquiring the modulated vibration signal, the present invention can further analyze the frequency band difference on both sides of the gear rotation frequency and the meshing frequency of the gear under test in the spectrum signal of the modulated vibration signal, and evaluate the early wear type of the gear under test by combining the sideband interval of the fault characteristic frequency of the vibration signal.

[0087] In one embodiment of the present invention, the modulated vibration signal is first subjected to Fourier transform to obtain its spectrum signal, which prepares for subsequent analysis of frequency modulation and amplitude modulation characteristics to distinguish early wear types. This is already existing technology and will not be described in detail here.

[0088] Preferably, in one embodiment of the present invention, the method for evaluating the early wear type of the gear under test includes:

[0089] Please see Figure 2 The diagram illustrates a flowchart of a method for evaluating the early wear type of a gear under test, provided by an embodiment of the present invention, specifically including:

[0090] Step S301: In the spectrum signal of the modulated vibration signal, the first wear characteristic parameter is obtained based on the frequency band difference on both sides of the gear rotation frequency of the gear under test and the frequency band difference on both sides of the meshing frequency of the gear under test.

[0091] Considering that micropitting causes the vibration signal spectrum to exhibit a frequency modulation effect at the gear rotation frequency, while microcracks cause the vibration signal spectrum to exhibit an amplitude modulation effect at the meshing frequency, the two have distinct characteristics; and considering that analyzing the sideband energy on both sides of the corresponding frequency can help analyze the frequency modulation and amplitude modulation characteristics in the spectrum, this will prepare for further differentiation of early wear types.

[0092] It should be noted that the acquisition of the meshing frequency and gear rotation frequency of the gear under test is already existing technology and will not be elaborated further.

[0093] In a preferred embodiment of the present invention, considering that the frequency modulation asymmetry and amplitude modulation asymmetry are evaluated by the frequency band difference on both sides of the frequency band, a first wear characteristic parameter can be defined; when both the frequency modulation asymmetry and amplitude modulation asymmetry are smaller, it indicates that the frequency modulation effect in the spectrum signal is more obvious, and the possibility of micro-pitting of the gear under test is relatively greater; therefore, the method for obtaining the first wear characteristic parameter includes:

[0094] In the spectrum signal of the modulated vibration signal, the frequency modulation asymmetry is obtained based on the difference between the total energy of a preset number of sidebands to the left of the gear frequency and the total energy of a preset number of sidebands to the right of the gear frequency; the amplitude modulation asymmetry is obtained based on the difference between the total energy of a preset number of sidebands to the left of the meshing frequency and the total energy of a preset number of sidebands to the right of the gear frequency; the amplitude modulation asymmetry is used as the denominator, the negative correlation normalized value of the frequency modulation asymmetry is used as the numerator, and the fractional ratio is used as the first wear characteristic parameter.

[0095] As an example, the preset quantity is set to 3, but implementers can also customize it;

[0096] FM asymmetry The calculation formula is: ;in, This represents the total energy of the third-order sideband to the right of the gear rotation frequency. This represents the total energy of the third-order sideband to the left of the gear's rotational frequency. The absolute value symbol is used; in the formula, the absolute value of the difference is used to measure the energy difference, and the sum of the energy is used as the denominator for normalization. The smaller the energy difference between the left and right sidebands, the less the energy shift occurs, the smaller the frequency modulation asymmetry, and the more obvious the frequency modulation effect characteristics.

[0097] Similarly, amplitude modulation asymmetry can be calculated. The greater the difference in sideband energy between the left and right sides of the meshing frequency, the more it indicates a shift in sideband energy, a greater amplitude asymmetry, and a more pronounced amplitude-to-frequency characteristic.

[0098] This will then be achieved through The frequency modulation asymmetry is normalized by normalizing the negative correlation, thus yielding the first wear characteristic parameter. The first wear characteristic parameter reflects the relative probability of micropitting on the gear under test. The larger the first wear characteristic parameter, the greater the probability that the early wear type is micropitting.

[0099] Step S302: Determine the frequency modulation sideband interval based on the gear rotation frequency of the gear under test, and determine the amplitude modulation sideband interval based on the meshing frequency of the gear under test; obtain the actual sideband interval of the fault characteristic frequency in the spectrum signal of the vibration signal; obtain the second wear characteristic parameter based on the deviation of the actual sideband interval relative to the frequency modulation sideband interval and the deviation of the actual sideband interval relative to the amplitude modulation sideband interval.

[0100] Considering that by comparing the sideband spacing of the fault characteristic frequency in the vibration signal with the sideband spacing produced by amplitude modulation or frequency modulation effect, it is possible to assess whether a frequency modulation effect or amplitude modulation effect is present, and thus obtain the second wear characteristic parameter.

[0101] In a preferred embodiment of the present invention, considering that the frequency modulation effect is usually caused by the periodic change in the meshing stiffness of the gear, which is mainly related to the gear rotation frequency, the theoretical sideband interval of the frequency modulation effect corresponds to the gear rotation frequency; while the amplitude modulation effect is usually caused by cracks or defects on the gear surface, leading to energy migration in the resonant frequency band, which is mainly related to the gear meshing frequency; and the energy migration caused by cracks usually involves the frequency doubling effect, so the theoretical sideband interval of the amplitude modulation effect can be set as a multiple of the meshing frequency; then the matching sideband intervals can be compared to evaluate the frequency modulation effect and the amplitude modulation effect, and obtain the second wear characteristic parameter; therefore, the method for obtaining the second wear characteristic parameter includes:

[0102] The gear rotation frequency is used as the frequency modulation sideband interval, and twice the meshing frequency is used as the amplitude modulation sideband interval. The negative correlation mapping result of the absolute value of the difference between the actual sideband interval and the frequency modulation sideband interval is used as the frequency modulation matching parameter. The negative correlation mapping result of the absolute value of the difference between the actual sideband interval and the amplitude modulation sideband interval is used as the amplitude modulation matching parameter. The amplitude modulation matching parameter is used as the denominator, the frequency modulation matching parameter is used as the numerator, and the fractional ratio is used as the second wear characteristic parameter.

[0103] In the above method, the deviation between the sideband intervals is specifically measured by the absolute value of the difference. The absolute value of the difference is used as x in the exponential function exp(-x) with the natural constant e as the base, and the negative correlation mapping is adjusted so that the smaller the absolute value of the difference, the larger the negative correlation mapping value, and the larger the corresponding amplitude modulation matching parameter or frequency modulation matching parameter. Then, the amplitude modulation matching parameter is defined as the denominator and the frequency modulation matching parameter is defined as the numerator. The larger the ratio of the fractions, the larger the second wear characteristic parameter, indicating that the frequency modulation effect is more obvious and the greater the possibility of micro-pitting of the gear under test.

[0104] Step S303: Combine the first wear characteristic parameter and the second wear characteristic parameter to obtain the wear characteristic coefficient of the gear under test; determine the early wear type of the gear under test based on the wear characteristic coefficient.

[0105] As an example, the first wear characteristic parameter and the second wear characteristic parameter are multiplied and combined to obtain the wear characteristic coefficient; in other examples, the implementer may also use methods such as addition or weighted summation to combine the two, which will not be elaborated further.

[0106] In a preferred embodiment of the present invention, the method for determining the early wear type of the gear under test includes:

[0107] When the wear characteristic coefficient is greater than the preset first threshold, the gear under test is determined to have micro-pitting wear; when the wear characteristic coefficient is less than the preset second threshold, the gear under test is determined to have micro-crack wear; wherein, the preset first threshold is greater than the preset second threshold; as an example, the preset first threshold is set to 1.2 and the preset second threshold is set to 0.8, but the implementer can also customize it.

[0108] This allows for the differentiation of early wear types in the gear under test, improving the effectiveness of gear wear detection.

[0109] One embodiment of the present invention also proposes a gear wear detection system for a geared motor. The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements all the steps of a gear wear detection method for a geared motor.

[0110] In summary, this invention first acquires the vibration signal of the gear under test, using the historical vibration signal of the faulty gear as a sample signal, and obtains its fault characteristic frequency. Further, the vibration signal is divided into several local impact segments, and the local spectrum of each local impact segment in the time spectrum of the vibration signal is obtained, thereby obtaining the kurtosis of each frequency in the corresponding local spectrum. In the time spectrum, combined with the power distribution at each fault characteristic frequency, a fuzzy estimate is determined, and the fuzzy estimate signal of the vibration signal is fitted. Then, combined with the motor load conditions, the vibration signal is modulated. Finally, in the spectrum signal of the modulated vibration signal, based on the frequency band differences on both sides of the gear rotation frequency and the frequency band differences on both sides of the meshing frequency of the gear under test, combined with the sideband spacing of the fault characteristic frequency of the vibration signal, the early wear type of the gear under test is evaluated. This invention modulates the vibration signal by analyzing the masking interference of noise and motor load on local wear impacts, and further analyzes the frequency modulation effect and amplitude modulation effect in the frequency domain of the modulated vibration signal, thereby accurately distinguishing the early wear type and improving the gear wear detection effect.

[0111] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0112] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for detecting gear wear in a geared motor, characterized in that, The method includes: The vibration signal of the gear under test is acquired, the historical vibration signal of the faulty gear is used as the sample signal, and the fault characteristic frequency of each sample signal is acquired. Based on the energy changes of the vibration signal, it is divided into several local impact segments; the vibration signal is subjected to a short-time Fourier transform to obtain the local spectrum of each local impact segment in the time spectrum of the vibration signal; the kurtosis of each local impact segment is used as the kurtosis of each frequency in the corresponding local spectrum; in the time spectrum, based on the power distribution and kurtosis at each fault characteristic frequency, the fuzzy estimate value at each fault characteristic frequency is determined, and combined with the temporal sequence of each fault characteristic frequency in the time domain, the fuzzy estimate signal of the vibration signal is fitted. Based on the fuzzy estimation signal and the sideband energy of each fault characteristic frequency, combined with the motor load, the vibration signal is modulated to obtain a modulated vibration signal. In the spectrum signal of the modulated vibration signal, based on the frequency band difference on both sides of the gear rotation frequency of the gear under test and the frequency band difference on both sides of the meshing frequency of the gear under test, combined with the sideband interval of the fault characteristic frequency of the vibration signal, the early wear type of the gear under test is evaluated.

2. The method for detecting gear wear in a geared motor according to claim 1, characterized in that, The method for obtaining the fault characteristic frequency includes: The target frequency band is determined based on the meshing frequency of the faulty gear corresponding to each sample signal; the low resonance component in the sample signal is extracted using the resonance sparse decomposition algorithm, and the component sub-segment corresponding to the target frequency band in the low resonance component is obtained. Time delay energy accumulation analysis is performed on the component segments to determine the vibration period, and the reciprocal of the vibration period is used as the fault characteristic frequency of the corresponding sample signal.

3. The method for detecting gear wear in a geared motor according to claim 1, characterized in that, The method for obtaining the local impact segment and the local spectrum includes: Starting from the origin of the vibration signal, the vibration signal is divided into several local segments of corresponding window lengths using a window. Based on the total number of local segments and the dispersion of signal energy in all local segments, a window objective function is defined. The initial window and the iteration step size of the window length are defined, and the window is obtained iteratively. When the window objective function takes the minimum value, the length of the corresponding window is taken as the target length. The local sub-segment corresponding to the target length is taken as the local impact sub-segment. Using the target length as the length of the short-time window, the vibration signal is subjected to a short-time Fourier transform to obtain the time spectrum, and the spectrum in the time domain corresponding to each local impact segment in the time spectrum is used as the local spectrum of the corresponding local impact segment.

4. The method for detecting gear wear in a geared motor according to claim 1, characterized in that, The methods for obtaining the fuzzy estimate and fitting the fuzzy estimate signal include: In the local spectrum corresponding to each local impact segment, the ratio of the power spectral density to the corresponding kurtosis at each fault characteristic frequency is used as the fuzzy estimate. In the time spectrum, the time corresponding to each fault characteristic frequency in the time domain is determined, and the fuzzy estimate of each fault characteristic frequency is mapped to a timestamp according to the corresponding time, and the fuzzy estimate signal is fitted.

5. The method for detecting gear wear in a geared motor according to claim 1, characterized in that, The method for acquiring the modulated vibration signal includes: At each fault characteristic frequency, the sideband energy of a preset order is aggregated on both sides to obtain the enhanced energy; the load fluctuation signal and its envelope signal of the geared motor are obtained, and the envelope signal and fuzzy estimation signal are fused to obtain the characteristic signal; The negative correlation mapping result of the largest amplitude in the envelope signal is used as the interference weight. The enhanced energy at each fault characteristic frequency is weighted using the interference weight to obtain the confidence enhanced energy. The average amplitude level of the characteristic signal is used as the characteristic information. The confidence enhanced energy and the characteristic information are fused to obtain the modulation compensation amount at the corresponding fault characteristic frequency. The vibration amplitude at the time corresponding to the fault characteristic frequency is modulated by the modulation compensation amount to obtain the modulated vibration signal.

6. The method for detecting gear wear in a geared motor according to claim 1, characterized in that, Methods for assessing the early wear type of a gear under test include: In the spectrum signal of the modulated vibration signal, the first wear characteristic parameter is obtained based on the frequency band difference on both sides of the gear rotation frequency of the gear under test and the frequency band difference on both sides of the meshing frequency of the gear under test. The frequency modulation sideband interval is determined based on the gear rotation frequency of the gear under test, and the amplitude modulation sideband interval is determined based on the meshing frequency of the gear under test; the actual sideband interval of the fault characteristic frequency in the spectrum signal of the vibration signal is obtained; the second wear characteristic parameter is obtained based on the deviation of the actual sideband interval from the frequency modulation sideband interval and the deviation of the actual sideband interval from the amplitude modulation sideband interval. By combining the first wear characteristic parameter and the second wear characteristic parameter, the wear characteristic coefficient of the gear under test is obtained; the early wear type of the gear under test is determined based on the wear characteristic coefficient.

7. The method for detecting gear wear in a geared motor according to claim 6, characterized in that, The method for obtaining the first wear characteristic parameter includes: In the spectral signal of the modulated vibration signal, the frequency modulation asymmetry is obtained based on the difference between the total energy of a preset number of sidebands to the left of the gear frequency and the total energy of a preset number of sidebands to the right of the gear frequency; the amplitude modulation asymmetry is obtained based on the difference between the total energy of a preset number of sidebands to the left of the meshing frequency and the total energy of a preset number of sidebands to the right of the meshing frequency; the amplitude modulation asymmetry is used as the denominator, the negative correlation normalized value of the frequency modulation asymmetry is used as the numerator, and the fractional ratio is used as the first wear characteristic parameter.

8. The method for detecting gear wear in a geared motor according to claim 6, characterized in that, The method for obtaining the second wear characteristic parameter includes: The gear rotation frequency is used as the frequency modulation sideband interval, and twice the meshing frequency is used as the amplitude modulation sideband interval. The negative correlation mapping result of the absolute value of the difference between the actual sideband interval and the frequency modulation sideband interval is used as the frequency modulation matching parameter. The negative correlation mapping result of the absolute value of the difference between the actual sideband interval and the amplitude modulation sideband interval is used as the amplitude modulation matching parameter. The amplitude modulation matching parameter is used as the denominator, the frequency modulation matching parameter is used as the numerator, and the fractional ratio is used as the second wear characteristic parameter.

9. A method for detecting gear wear in a geared motor according to claim 6, characterized in that, Methods for determining the early wear type of the gear under test based on the wear characteristic coefficient include: When the wear characteristic coefficient is greater than a preset first threshold, it is determined that the gear under test has micro-pitting wear; when the wear characteristic coefficient is less than a preset second threshold, it is determined that the gear under test has micro-crack wear; wherein, the preset first threshold is greater than the preset second threshold.

10. A gear wear detection system for a geared motor, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the gear wear detection method for a geared motor as described in any one of claims 1 to 9.

Images