A method and system for inverting gearbox structural parameters by integrating current and vibration signals

Abstract

This invention belongs to the field of gearbox parameter inversion technology, and discloses a method and system for inverting gearbox structural parameters by integrating current and vibration signals. The method includes: rearranging and compressing the time-frequency matrix of the vibration signal in the time and frequency directions through a rearranged second-order synchronous compression transform; extracting the instantaneous rotational frequency and meshing frequency ridges of each gear in the gearbox using a dynamic path optimization ridge extraction method; automatically extracting meshing frequency and instantaneous rotational frequency information at different times; and constructing a method for calculating the number of teeth from the high-speed shaft to the low-speed shaft based on the one-to-two mapping relationship between meshing frequency and rotational frequency according to the gearbox power transmission path, thereby extracting the gearbox structural parameters. This invention proposes a method for inverting the structural parameters of each stage of gears in a gearbox by integrating current and vibration signals. Under varying operating conditions of the gearbox, it automatically and intelligently inverts the number of teeth on each shaft of the gearbox, providing missing tooth count information for gearbox fault diagnosis.

Description

Technical Field

[0001] This invention belongs to the field of gearbox parameter inversion technology, and particularly relates to a method and system for inverting gearbox structural parameters by integrating current and vibration signals. Background Technology

[0002] Currently, power is transmitted through multiple stages of gears within gearboxes. Compared to other mechanical transmission methods (belt drives, chain drives), gear drives offer advantages such as constant instantaneous transmission ratio and a wide range of transmitted power, leading to their widespread application in industries such as industry and aerospace. However, many mechanical equipment have been manufactured and used for a long time, resulting in the loss of gearbox structural drawings. Without these parameters, it is difficult for workers and technicians to perform fault monitoring and diagnosis. Furthermore, gearboxes require high manufacturing and installation precision, and disassembly and reassembly would reduce this precision. Without disassembling the gearbox, it is impossible to obtain parameter information for each stage of gears within it.

[0003] Patent CN 111795819 A: Adaptive feature learning and fusion processing is performed on gearbox vibration signals and generator current signals. Based on the characteristics and complementarity of current and vibration monitoring signals, vibration is used as an auxiliary learning signal for current during the feature acquisition and learning stage. Fault diagnosis is performed using the current signal with fault information learned from the vibration signal. Patent CN 103645052 B: Vibration acceleration sensors are used to collect vibration signals of each gear shaft in the wind turbine gearbox, monitor the gearbox operating status, and perform gearbox life assessment. Patent CN 115758094 A: Modal decomposition and feature extraction are performed on current and vibration signals respectively; multiple current and vibration components are obtained; the correlation between each component is calculated; then the correlation is filtered and feature-level data fusion and dimensionality reduction are performed to obtain multiple principal components; by calculating the relationship between the principal components and their own contribution rate, a degradation index curve is obtained, realizing the life prediction of the reducer.

[0004] Current patents related to gearboxes mainly focus on fault diagnosis and life prediction of gears and bearings within gearboxes. These patents collect vibration signal data generated during gearbox operation, perform FFT and Hilbert transforms, and generate spectrum diagrams for fault characteristic analysis to determine the location of faulty bearings and gears. Existing patents do not consider using vibration and current signals to inversely derive gearbox structural parameters. When calculating gear tooth count information using rotational and meshing frequencies from vibration and current signals, obtaining the instantaneous rotational and meshing frequencies of gears under varying operating conditions is difficult and the acquisition process is complex. Therefore, a method and system for inverting gearbox structural parameters by integrating current and vibration signals is needed to deduce missing structural parameters within the gearbox.

[0005] Based on the above analysis, the problems and defects of the existing technology are as follows: the existing technology lacks a method for inverting the missing structural parameters in the gearbox; the existing technology has cumbersome methods for obtaining the instantaneous rotational frequency and meshing frequency of the gears, and the accuracy of the obtained instantaneous rotational frequency and meshing frequency information is low, which cannot be used to calculate the number of teeth of each shaft of the gearbox. Summary of the Invention

[0006] To address the problems existing in the prior art, this invention provides a method and system for inverting gearbox structural parameters by integrating current and vibration signals.

[0007] This invention is implemented as follows: a method for inverting gearbox structural parameters by integrating current and vibration signals, wherein the method includes:

[0008] Under the condition of variable speed of the gearbox, the current signal of the motor is collected during operation, and the current signal is transformed by STFT to draw a time-frequency diagram to obtain the instantaneous rotation frequency of the motor. The search range of the instantaneous rotation frequency ridge line of each gear shaft is narrowed. The vibration signal generated during the operation of the gearbox is collected and transformed by STFT to obtain the time-frequency matrix. The vibration signal time-frequency matrix is ​​rearranged and compressed in the time and frequency directions by rearranging the second-order synchronous compression transformation. Combined with the dynamic path optimization ridge line extraction method, the instantaneous rotation frequency and meshing frequency ridge line of each gear in the gearbox are extracted, and the meshing frequency and instantaneous rotation frequency information at different times are automatically extracted.

[0009] Based on the power transmission path of the gearbox and the one-to-two mapping relationship between meshing frequency and rotational frequency, a method for calculating the number of teeth from the high-speed shaft to the low-speed shaft of the gearbox is constructed. The structural parameters of the gearbox are extracted, and the number of teeth on each shaft of the gearbox is automatically and intelligently calculated under the variable working conditions of the gearbox.

[0010] Furthermore, the gearbox structural parameter inversion method that integrates current and vibration signals includes the following steps:

[0011] Step 1: Collect the current signal when the gearbox motor is running and the vibration signal generated when each gear shaft meshes, and preprocess the collected current signal and vibration signal.

[0012] Step 2: Perform STFT transformation on the preprocessed current signal and vibration signal to plot the time-frequency diagram of the AC motor current signal and the time-frequency diagram of the vibration signal; and extract the instantaneous rotational frequency of the motor.

[0013] Step 3: Based on the drawn time-frequency diagram, the dynamic path optimization ridge extraction method is used to extract the meshing frequency ridges of each gear and obtain the meshing frequency information of each gear.

[0014] Step 4: Based on the drawn time-frequency diagram, perform frequency doubling amplitude summation on the time-frequency information, determine the instantaneous frequency search range according to the instantaneous motor frequency, compress the time and frequency information of the low-frequency part, extract the instantaneous frequency ridge line within the search range, and extract the instantaneous frequency information of each shaft of the gearbox.

[0015] Step 5: Extract the instantaneous rotational frequency ridges and meshing frequency ridges of each shaft of the gearbox from the time-frequency diagram of vibration and current signals. On the ridge diagram, select the meshing frequency and instantaneous rotational frequency at the same moment to calculate the number of teeth of each gear shaft.

[0016] Furthermore, the acquisition of the current signal from the gearbox and the vibration signal from each shaft, and the preprocessing of the acquired current signal and vibration signal, includes:

[0017] First, measurement points are set at the bearing housings of each gear shaft to collect vibration signals of each shaft; a current sensor is set at the power phase line of the gearbox motor to collect motor current signals.

[0018] Secondly, the acquired current signal is subjected to mean subtraction processing:

[0019]

[0020]

[0021] in, An index representing the mean value of a current signal; Indicates the sensitivity of the current sensor; This indicates the collected current signal. ={1,2,……,N}; N represents the number of sampling points;

[0022] Then, the collected vibration signals are preprocessed:

[0023] ;

[0024] ;

[0025] in, An index representing the mean value of a current signal; This indicates the sensitivity of the vibration acceleration sensor. This indicates the collected vibration signal. ={1,2,……,N}, where N represents the number of sampling points.

[0026] Furthermore, the step of performing STFT transformation on the preprocessed current signal and vibration signal to plot the time-frequency diagrams of the AC motor current signal and vibration signal includes:

[0027] (1) Perform a short-time Fourier transform on the preprocessed current signal and take its absolute value:

[0028] ;

[0029] (2) With time t as the horizontal axis, frequency Using the estimated instantaneous localization frequency content after STFT transformation as the vertical axis, the depth of the chromaticity of the time-frequency ridge line is controlled to plot the time-frequency diagram of the AC motor current signal.

[0030] (3) Perform a short-time Fourier transform on the vibration signal and take its absolute value:

[0031] ;

[0032] (4) Plot time t as the horizontal axis and frequency Using the estimated instantaneous localized frequency content after STFT transformation as the vertical axis, the intensity of the chromaticity of the time-frequency ridge line is controlled, and the time-frequency diagram of the vibration signal acquired along each axis is plotted.

[0033] Furthermore, the instantaneous motor frequency extraction is performed as follows:

[0034] Calculate the instantaneous frequency of the motor using the following formula :

[0035]

[0036]

[0037] Where f represents the current frequency, p represents the number of pole pairs of the motor, and s represents the motor slip. This indicates the speed of the AC motor.

[0038] Furthermore, the method of extracting meshing frequency ridges using dynamic path optimization based on the drawn time-frequency graph to obtain gear meshing frequency information includes:

[0039] By comparing the color intensity of the ridges on the time-frequency graph of the AC motor current signal, the ridges with higher localized frequency content are selected as the target meshing frequency ridges for extraction. The dynamic path optimization ridge extraction method is used to extract the meshing frequency ridges and obtain the meshing frequency information of the gear.

[0040] The method of extracting meshing frequency ridges using dynamic path optimization to obtain gear meshing frequency information includes:

[0041] 1) Perform frequency-multiplication amplitude summation on the time-frequency analysis matrix after STFT transformation, and calculate the relative amplitude at the same frequency at different times by frequency slicing.

[0042] 2) Perform time-frequency analysis on the signal to obtain a time-frequency matrix with dimension N×M. Along the time axis, the amplitudes at different time-frequency points at the same moment are compared within the matrix, and the values ​​that simultaneously satisfy the following conditions are extracted. and The time and frequency points, among which, Indicates the peak point of the time frequency.

[0043] ;

[0044] in, This represents the total number of time-frequency peak points corresponding to a given time.

[0045] 3) Based on the relationship between the fundamental frequency and harmonics of the time-varying frequency component signal, an amplitude summation peak search algorithm is applied to the peak points of different time frequencies at the same moment within the matrix to determine whether the corresponding frequencies meet the following conditions:

[0046]

[0047] in, Indicates rounding down; This represents the frequency resolution of the window function in the STFT time-frequency analysis of the signal; when the conditions are met, the amplitudes of the time-frequency peak points are summed.

[0048] ;

[0049] 4) Perform the same processing on each column of time-frequency peaks within the time-frequency matrix to obtain a time-frequency peak matrix; set a threshold. The time-frequency peak points within the time-frequency peak matrix are filtered out, and when the conditions are met... If the peak value is present, retain the peak value; otherwise, set the amplitude of the peak value to zero.

[0050] 5) The ridge line is extracted by summing the amplitudes using a cost function on the time-frequency matrix;

[0051] The cost function is as follows:

[0052] .

[0053] Furthermore, step four includes:

[0054] The instantaneous rotational frequency of the gearbox input shaft is calculated using the instantaneous current frequency information of the motor. The range of the gear rotational frequency ridge line search is determined by the input shaft rotational frequency. The time-frequency information of the search range is subjected to second-order rearrangement synchronous compression transformation to determine the target instantaneous rotational frequency ridge line to be extracted. The instantaneous rotational frequency ridge line is extracted by using the dynamic path optimization ridge line extraction method to obtain the instantaneous rotational frequency information of each shaft.

[0055] The step of performing a second-order rearrangement synchronization compression transform on the time-frequency information of the search range to determine the target frequency ridge to be extracted includes:

[0056] (1) Perform STFT conversion on the vibration signal:

[0057] ;

[0058] ;

[0059] in, This represents the original vibration signal; k∈N is a positive integer. Indicates instantaneous amplitude. Represents the instantaneous phase function; Represents the short-time Fourier transform coefficients. The complex conjugate;

[0060] (2) Perform second-order local complex instantaneous time and frequency estimation:

[0061] (2.1) The second-order local complex instantaneous frequency IF is estimated using the following formula:

[0062] ;

[0063] in, This represents the instantaneous frequency IF estimated based on STFT. Im represents the imaginary part of a complex number; This represents the Taylor approximation of instantaneous time. ; and These represent the time-frequency coefficients calculated by STFT when the window function is g(t) or tg(t), respectively. , represents the second-order local complex modulation operator, ;

[0064] (2.2) Second-order local instantaneous time estimation is performed using the following formula:

[0065] ;

[0066] (3) Calculate the rearrangement time and frequency operators using the following formula:

[0067] ;

[0068] ;

[0069] in, This represents the rearrangement time operator; Represents the rearrangement frequency operator

[0070] (4) Calculate the compensation frequency distance :

[0071]

[0072] (5) The vibration signal is rearranged in the two-dimensional time-frequency direction, and then subjected to second-order compression:

[0073] Rearrange the second-order synchronous compression transformation using the following formula :

[0074] ;

[0075] in, Represents the Dirac function, This represents the time-frequency matrix of the compressed reconstruction.

[0076] Furthermore, the formula for calculating the number of teeth on each gear shaft is as follows:

[0077] ;

[0078] Where Z1 and Z2 represent the number of teeth in a pair of meshing gears; n1 and n2 represent the rotational speeds of gears Z1 and Z2, respectively. .

[0079] Another object of the present invention is to provide a gearbox structural parameter inversion system for implementing the gearbox structural parameter inversion method of fused current and vibration signal, the gearbox structural parameter inversion system comprising:

[0080] The signal acquisition and preprocessing module is used to acquire the current signal of the gearbox and the vibration signal of each shaft, and to preprocess the acquired current signal and vibration signal.

[0081] The time-frequency analysis module is used to perform STFT transformation on the preprocessed current signal and vibration signal, and to draw the time-frequency diagram of the AC motor current signal and the time-frequency diagram of the vibration signal.

[0082] The instantaneous motor frequency extraction module is used to extract the instantaneous motor frequency and determine the search range of the gearbox instantaneous frequency ridge line.

[0083] The meshing frequency extraction module is used to extract meshing frequency ridges based on the drawn time-frequency diagram using a dynamic path optimization ridge extraction method, thereby obtaining the meshing frequency information of the gears.

[0084] The instantaneous frequency extraction module is used to perform frequency doubling amplitude summation processing on the time and frequency information based on the drawn time and frequency diagram, then compress the time and frequency information of the low frequency part, extract the instantaneous frequency ridge line, and extract the instantaneous frequency information of each shaft of the gearbox.

[0085] The tooth count information calculation module is used to extract the instantaneous rotational frequency ridges and meshing frequency ridges of each shaft of the gearbox from the time-frequency diagram of vibration and current signals. On the ridge diagram, the meshing frequency and instantaneous rotational frequency at the same moment are selected to calculate the tooth count information of each gear shaft.

[0086] Another object of the present invention is to provide a computer device, the computer device including a memory and a processor, the memory storing a computer program, the computer program being executed by the processor causing the processor to perform the steps of the gearbox structure parameter inversion method of fused current and vibration signal.

[0087] Based on the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solution to be protected by this invention are as follows:

[0088] First, under the condition of variable speed of the gearbox, the present invention rearranges and compresses the time-frequency matrix of the vibration signal in the time and frequency directions by rearranging the second-order synchronous compression transformation. Combined with the dynamic path optimization ridge extraction method, the instantaneous rotation frequency and meshing frequency ridges of each gear in the gearbox are extracted. The process of automatically identifying the instantaneous rotation frequency and meshing frequency is eliminated, and the meshing frequency and instantaneous rotation frequency information at different times are automatically extracted, providing data support for the inversion of gearbox structural parameters.

[0089] Based on the power transmission path of the gearbox and considering the one-to-two mapping relationship between meshing frequency and rotational frequency, this invention establishes a method for calculating the number of teeth in the gearbox from the high-speed shaft to the low-speed shaft, thereby realizing the extraction of gearbox structural parameters. It features simple calculation steps, high calculation efficiency, and accurate calculation results.

[0090] Secondly, this invention proposes a method for inverting the structural parameters of each gear in a gearbox by integrating current and vibration signals. Under varying operating conditions of the gearbox, the instantaneous rotational frequency and meshing frequency of each gear shaft are extracted through motor current signals and gearbox vibration signals. Based on the one-to-two mapping relationship between meshing frequency and rotational frequency, a method for calculating the number of teeth in the gearbox from the high-speed shaft to the low-speed shaft is constructed. This automatically and intelligently inverts the number of teeth on each shaft of the gearbox, providing missing tooth count information for gearbox fault diagnosis.

[0091] Third, as supplementary evidence of the inventive step of the claims of this invention, it is also reflected in the following important aspects:

[0092] (1) The technical solution of the present invention is aimed at obtaining the gear tooth number information of gearboxes that are difficult to disassemble. It provides a method for inverting the structural parameters of each gear in a gearbox by integrating current and vibration signals, which solves the problem of obtaining the parameter information of gears in gearboxes that have not been disassembled, and fills the technical gap in the method of inverting the gear tooth number of each gear in a gearbox at home and abroad.

[0093] (2) The technical solution of the present invention combines the rearrangement second-order synchronous compression transformation and dynamic path optimization ridge extraction method to extract the instantaneous rotation frequency and meshing frequency ridge of each gear in the gearbox. It eliminates the need for manual identification of the instantaneous rotation frequency and meshing frequency process and automatically extracts the meshing frequency and instantaneous rotation frequency information at different times. This solves the problems of complex instantaneous rotation frequency and meshing frequency extraction steps and low accuracy of the extracted instantaneous rotation frequency and meshing frequency information. Attached Figure Description

[0094] Figure 1 This is a schematic diagram of the inversion process of gearbox parameters based on the fusion current and vibration signal provided in this embodiment of the invention;

[0095] Figure 2 This is a flowchart of the gearbox structural parameter inversion method that integrates current and vibration signals provided in this embodiment of the invention;

[0096] Figure 3 This is the dynamic path planning ridge extraction algorithm flow provided in the embodiments of the present invention;

[0097] Figure 4 This is a flowchart of the rearrangement second-order synchronous compression transformation algorithm provided in an embodiment of the present invention;

[0098] Figure 5 This is an external structural diagram of the gearbox provided in an embodiment of the present invention;

[0099] Figure 6 This is a flowchart of the gearbox structural parameter inversion provided in an embodiment of the present invention;

[0100] Figure 7 This is a time-domain diagram of the vibration signals of each bearing on the left side of the gearbox provided in an embodiment of the present invention;

[0101] Figure 8 This is a time-domain diagram of the vibration signals of each bearing on the right side of the gearbox provided in an embodiment of the present invention;

[0102] Figure 9 This is a time-domain diagram of the current signal provided in an embodiment of the present invention;

[0103] Figure 10 This is a time-frequency diagram of the motor current connected to the left input shaft of the gearbox provided in an embodiment of the present invention;

[0104] Figure 11 This is a time-frequency diagram of the vibration signal at the position of the left bearing 1 of the gearbox provided in an embodiment of the present invention;

[0105] Figure 12 This is a time-frequency diagram of the vibration signal at the position of the left bearing 2 of the gearbox provided in an embodiment of the present invention;

[0106] Figure 13This is a time-frequency diagram of the vibration signal at position 5 of the left bearing of the gearbox provided in an embodiment of the present invention;

[0107] Figure 14 This is a time-frequency diagram of the vibration signal at position 6 of the left bearing of the gearbox provided in an embodiment of the present invention;

[0108] Figure 15 This is a time-frequency diagram of the two target meshing frequency ridges extracted according to an embodiment of the present invention;

[0109] Figure 16 This is an internal structural diagram of the gear meshing relationship of the gearbox provided in an embodiment of the present invention;

[0110] Figure 17 This is a time-frequency diagram of the low-frequency information of the vibration signal of bearing 5 provided in this embodiment of the invention;

[0111] Figure 18 This is a time-frequency diagram of the low-frequency information of the vibration signal of bearing 5 provided in this embodiment of the invention;

[0112] Figure 19 This is a time-frequency diagram of the low-frequency vibration signal of bearing 6 provided in this embodiment of the invention;

[0113] Figure 20 This is a time-frequency diagram of the rotational frequency ridge line for extracting the vibration signal of bearing 1 provided in an embodiment of the present invention;

[0114] Figure 21 This is a time-frequency diagram of the rotational frequency ridge line for vibration signal extraction of bearing 5 provided in this embodiment of the invention;

[0115] Figure 22 This is a time-frequency diagram of the rotational frequency ridge line for extracting the vibration signal of bearing 6 provided in an embodiment of the present invention. Detailed Implementation

[0116] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0117] like Figure 1As shown, this invention, by analyzing the correlation between current, vibration signals, and gear parameters within a gearbox, proposes a method for extracting meshing frequency, rotational frequency, and current frequency from the time-frequency diagrams of vibration and current signals. It provides specific steps for inverting gear parameters within a gearbox that has not been disassembled, and analyzes the influence of vibration signals generated during gearbox operation and motor current signals on the calculation of the number of gear teeth. Analysis of the gearbox's working mechanism reveals that the motor transmits torque to the gearbox input shaft via a coupling. The motor speed can be calculated using the motor's current frequency. Gear meshing and gear shaft rotation both generate vibrations, with the vibration intensity gradually decreasing and being transmitted to the gearbox housing. This vibration information contains the gear meshing frequency and rotational frequency information, providing a basis for calculating the number of gear teeth. However, existing patents related to current, vibration signals, and gearboxes do not offer effective solutions to the gearbox parameter inversion problem. The gearbox parameter inversion process integrating current and vibration signals is as follows: Figure 1 As shown.

[0118] like Figure 2 As shown, the gearbox structural parameter inversion method based on the fusion of current and vibration signals provided in this embodiment of the invention includes the following steps:

[0119] S101 collects the current signal when the gearbox motor is running and the vibration signal generated when each gear shaft meshes, and preprocesses the collected current signal and vibration signal.

[0120] S102, Perform STFT transformation on the preprocessed current signal and vibration signal to plot the time-frequency diagram of AC motor current signal and vibration signal; and extract the instantaneous rotational frequency of the motor;

[0121] S103, Based on the drawn time-frequency diagram, the dynamic path optimization ridge extraction method is used to extract the meshing frequency ridge of each gear and obtain the meshing frequency information of each gear.

[0122] S104, based on the drawn time-frequency diagram, perform frequency doubling amplitude summation processing on the time-frequency information, determine the instantaneous frequency search range according to the instantaneous motor frequency, then compress the time and frequency information of the low-frequency part, extract the instantaneous frequency ridge line within the search range, and extract the instantaneous frequency information of each shaft of the gearbox.

[0123] S105 extracts the instantaneous rotational frequency ridges and meshing frequency ridges of each shaft of the gearbox from the time-frequency diagram of vibration and current signals. On the ridge diagram, the meshing frequency and instantaneous rotational frequency at the same moment are selected to calculate the number of teeth of each gear shaft.

[0124] The gearbox structural parameter inversion system that integrates current and vibration signals provided in this embodiment of the invention includes:

[0125] The signal acquisition and preprocessing module is used to acquire the current signal when the gearbox motor is running and the vibration signal generated when each gear shaft meshes, and to preprocess the acquired current signal and vibration signal.

[0126] The time-frequency analysis module is used to perform STFT transformation on the preprocessed current signal and vibration signal, and to draw the time-frequency diagram of the AC motor current signal and the time-frequency diagram of the vibration signal.

[0127] The instantaneous motor speed extraction module is used to extract the instantaneous speed of the motor.

[0128] The meshing frequency extraction module is used to extract the meshing frequency ridges of each gear based on the drawn time-frequency diagram using a dynamic path optimization ridge extraction method, and to obtain the meshing frequency information of each gear.

[0129] The instantaneous frequency extraction module is used to perform frequency multiplication amplitude summation processing on the time and frequency information based on the drawn time and frequency diagram, determine the instantaneous frequency search range according to the instantaneous frequency of the motor, compress the time and frequency information of the low frequency part, extract the instantaneous frequency ridge line within the search range, and extract the instantaneous frequency information of each shaft of the gearbox.

[0130] The tooth count information calculation module is used to extract the instantaneous rotational frequency ridges and meshing frequency ridges of each shaft of the gearbox from the time-frequency diagram of vibration and current signals. On the ridge diagram, the meshing frequency and instantaneous rotational frequency at the same moment are selected to calculate the tooth count information of each gear shaft.

[0131] The gearbox structural parameter inversion method based on the fusion of current and vibration signals provided in this embodiment of the invention specifically includes:

[0132] 1. Acquisition and preprocessing of current and vibration signals

[0133] Without disassembling the gearbox, the only way to determine the number of gear teeth is by analyzing its operating principle. During gearbox operation, gear meshing and gear shaft rotation generate vibrations. By analyzing these vibration signals, the vibration signal produced by a pair of meshing gears contains information such as the gear meshing frequency and rotational frequency. The number of gear teeth can be calculated from the gear rotational frequency and meshing frequency.

[0134] 1.1 Selection of Current and Vibration Signal Acquisition Locations

[0135] During gearbox operation, the intensity of vibration signals measured at different locations within the gearbox varies. The closer to the bearing, the higher the amplitude of the vibration signal; conversely, the farther away, the lower the amplitude. The strength of the vibration signal amplitude determines the accuracy of the matching results. Therefore, the selection of vibration measurement point locations is crucial. When collecting vibration signals from each shaft, the measurement points should be located at the bearing housing of each gear shaft. Gearbox vibration information is collected using an accelerometer. Motor current information is collected using a current sensor installed at the motor power supply phase line location.

[0136] 1.2 Preprocessing of Current and Vibration Signals

[0137] The acquired current and vibration signals contain constant data, which, when mixed in with the raw signal data, cannot be used to create graphs that reflect the amplitude of vibration changes and the trend of current signal variations. Therefore, preprocessing of the acquired current and vibration signals is necessary before performing STFT transformation. Current sensors have a fixed sampling frequency. and sensitivity The collected current signal is denoted as , ={1,2,……,N}, where N is the number of sampling points. Calculate the mean index of the current signal:

[0138]

[0139] The original current signal data is processed by subtracting the mean, and the preprocessed data is denoted as follows: :

[0140]

[0141] Vibration acceleration sensors have a fixed sampling frequency. and sensitivity The collected vibration signals are recorded as follows: , ={1,2,……,N}, where N is the number of sampling points. Calculate the mean index of the vibration signal:

[0142]

[0143] The raw vibration signal data is preprocessed, and the preprocessed data is recorded as follows: :

[0144]

[0145] 2. Time-frequency analysis of vibration and current signals

[0146] 2.1 STFT Conversion of Current Signals

[0147] Motors operate in five states: starting, accelerating, constant speed operation, deceleration, and stopping. The current frequency collected by the current sensor differs depending on the motor's operating state. Therefore, it is necessary to perform STFT transformation on the discrete signal collected by the current sensor to plot a time-frequency graph and extract the motor current frequency. Perform a short-time Fourier transform and take the absolute value:

[0148]

[0149] With time t as the horizontal axis, frequency... Using the estimated instantaneous localized frequency content after STFT transformation as the vertical axis, the chromaticity of the time-frequency ridge line is controlled to plot the time-frequency diagram of the AC motor current signal. The time-frequency ridge line on the time-frequency diagram is the instantaneous current frequency ridge line of the motor.

[0150] 2.2 STFT Transformation of Vibration Signals

[0151] Within a fixed sampling period, the accelerometer collects the gearbox vibration signal a fixed number of times. The sampling process is not continuous; the collected vibration signal consists of multiple discrete points. A time-domain plot with the collected vibration signal amplitude as the vertical axis and the sampling time as the horizontal axis is insufficient to reveal the meshing frequency and rotational frequency. Therefore, it is necessary to perform an STFT transform on the collected vibration signal and plot a time-frequency diagram for analysis. Perform a short-time Fourier transform and take the absolute value:

[0152]

[0153] With time t as the horizontal axis, frequency... As the vertical axis, the estimated instantaneous localized frequency content after STFT transformation controls the depth of the chromaticity of the time-frequency ridge line, and plots the time-frequency diagram of the vibration signal acquired on each axis to facilitate the extraction of meshing frequency and rotational frequency information.

[0154] 3. Extraction of instantaneous motor frequency

[0155] There are three methods for speed control of AC motors: adjusting the power supply current frequency, changing the number of pole pairs, and changing the slip. Changing the number of pole pairs is stepped speed control, which adjusts the speed by changing the phase sequence of the stator windings; it can only adjust a few fixed speeds. Changing the slip is stepless speed control, which adjusts the speed by changing the ratio of the stator magnetic field rotation speed to the rotor rotation speed; however, the slip cannot be too large and the speed control process is complex. Variable frequency speed control is also stepless speed control; it adjusts the speed by changing the power supply frequency, has stable mechanical characteristics, and a wide speed range, therefore it is widely used. The formula for calculating the speed of an AC motor is:

[0156]

[0157] In the formula, f is the current frequency, p is the number of pole pairs of the motor, and s is the motor slip. The formula for calculating the instantaneous frequency of the motor is:

[0158]

[0159] 4. Extraction of meshing frequency information for each shaft

[0160] 4.1 Basic Principles of Meshing Frequency Extraction

[0161] The gearbox transmits torque through a multi-stage gear transmission. During operation, it generates vibration signals of varying amplitudes. Accelerometers are installed at each bearing housing of the gearbox. The vibration signals collected by each sensor are subjected to STFT transformation, and time-frequency diagrams are plotted. The vibration signals collected by different sensors vary in intensity, resulting in differences in the color intensity of the time-frequency ridges on the plotted diagrams. Furthermore, the meshing frequency is a high-frequency frequency, and the frequency ridges exhibit high localized frequency content on the time-frequency diagrams. By comparing the color intensity of the ridges, the ridges with higher localized frequency content are selected as the target meshing frequency ridges for extraction. A dynamic path optimization ridge extraction method is used to extract the meshing frequency ridges and obtain the gear meshing frequency information.

[0162] 4.2 Dynamic Path Optimization Ridge Extraction Method

[0163] The time-frequency analysis matrix after STFT transformation is subjected to frequency-multiplying amplitude summation to increase the magnitude of the target ridge point amplitude and reduce the influence of noise components. Simultaneously, considering the significant differences in amplitude changes between constant frequency components and time-varying frequency components at the same frequency location at different times, the relative amplitude at the same frequency location at different times is calculated using frequency slicing, thereby reducing interference from constant frequency components. The dynamic path planning ridge extraction algorithm flow is as follows: Figure 3 As shown.

[0164] First, time-frequency analysis is performed on the signal. The STFT transformation result of the vibration signal is shown in Equation 9, yielding a time-frequency matrix with dimension N×M. Along the time axis, the amplitudes at different time-frequency points at the same moment are compared within the matrix, and the values ​​that simultaneously satisfy the following conditions are extracted. and The time and frequency points are called The time-frequency peak point is denoted as , and the total number of time-frequency peak points corresponding to this moment is denoted as . ;

[0165]

[0166] Secondly, based on the relationship between the fundamental frequency and harmonics of the time-varying frequency component signal, an amplitude summation peak search algorithm is applied to the peak points of different time frequencies at the same moment within the matrix to determine whether their corresponding frequencies meet the following conditions:

[0167]

[0168] In the formula Indicates rounding down; This represents the frequency resolution of the window function in the STFT time-frequency analysis of the signal. When certain conditions are met, the amplitudes of the time-frequency peak points are summed.

[0169]

[0170] To reduce computational complexity, we only consider the first four harmonics of the signal frequency components, i.e. The summed time-frequency peak points replace the amplitude values ​​at the same positions in the original matrix. This same processing is applied to each column of time-frequency peaks within the time-frequency matrix to obtain the time-frequency peak matrix. Then, a threshold is set. The time-frequency peak points within the time-frequency peak matrix are filtered out, and when the conditions are met... If the peak value is not specified, retain the peak value; otherwise, set the amplitude of the peak value to zero. The threshold value should be set according to the specific data. Setting it too high drastically reduces the number of retained peak points and leads to the loss of some valid time-frequency information; conversely, if... If the setting is too small, peak points caused by noise cannot be effectively eliminated, which will affect the next step of the analysis. The threshold size ensures that the top 1% to 5% of the larger time-frequency peaks are retained;

[0171] To eliminate interference from constant frequency components within the signal, a cost function is constructed:

[0172]

[0173] The ridge line is extracted by summing the amplitudes using a cost function on the time-frequency matrix.

[0174] 5. Extraction of instantaneous rotational frequency information for each axis

[0175] 5.1 Basic Principles of Instantaneous Frequency Extraction

[0176] The gear rotation frequency is a low-frequency component in the vibration signal spectrum, and its localized frequency content is low on the time-frequency plot of the STFT transform, resulting in an indistinct rotation frequency ridge. Therefore, it is necessary to perform frequency harmonic amplitude summation on the time-frequency information to increase the amplitude of the rotation frequency ridge position. Then, the time and frequency information should be compressed to improve the time-frequency clustering of the target rotation frequency ridge, making it more prominent and thus improving the accuracy of the extracted instantaneous rotation frequency. In the gearbox, a pair of meshing gears have a driven gear rotating at a lower speed than the driving gear. The gearbox input shaft rotation frequency is calculated using the motor current and speed information. Generally, the position with the highest gearbox rotation frequency is the input shaft. The range for searching the gear rotation frequency ridge is determined by the input shaft rotation frequency. A second-order rearrangement synchronous compression transform is performed on the time-frequency information within this range to determine the target rotation frequency ridge to be extracted. A dynamic path optimization ridge extraction method is used to extract the rotation frequency ridge, obtaining the instantaneous rotation frequency information of each shaft.

[0177] 5.2 Rearrangement of Second-Order Synchronous Compression Transformation

[0178] The original vibration signal can be expressed by a typical multi-component amplitude-frequency modulation (AM-FM) signal, defined as:

[0179]

[0180] Where k∈N is a positive integer, It is the instantaneous amplitude. This is the instantaneous phase function. For the original signal... The STFT can be represented as:

[0181]

[0182] in These are the short-time Fourier transform coefficients. yes .

[0183] The rearrangement second-order synchronous compression transform rearranges the spectrum in the two-dimensional time-frequency direction, placing it back into the time-frequency trajectory line, thus improving the time-frequency clustering of the ridges on the time-frequency plot. The algorithm flow is as follows: Figure 4 As shown.

[0184] The formula for estimating the instantaneous frequency IF based on STFT is:

[0185]

[0186] Where Im represents the imaginary part of the complex number, and its Taylor approximation over instantaneous time is:

[0187]

[0188] In the formula, and These correspond to the time-frequency coefficients calculated by STFT when the window function is g(t) or tg(t), respectively. Second-order local complex modulation operator. Represented as:

[0189]

[0190] The second-order local complex instantaneous frequency IF is estimated as follows:

[0191]

[0192] The second-order local instantaneous time estimate is:

[0193]

[0194] The rearrangement time operator is represented as The rearrangement frequency operator is expressed as The specific expression is:

[0195]

[0196]

[0197] The performance of frequency direction rearrangement is insufficient, so frequency compensation is needed to compensate for the frequency distance. The calculation expression is:

[0198]

[0199] The vibration signal's time-frequency spectrum is rearranged in a two-dimensional time-frequency direction and then compressed in a second-order manner to improve the convergence of time-frequency ridges. Specifically, this involves a rearrangement and second-order synchronous compression transform. The expression is:

[0200]

[0201] In the above formula, For the Dirac function, To compress and reconstruct the time-frequency matrix, a second-order synchronous compression transform is used to rearrange the vibration signal's time-frequency matrix in the time-frequency direction. This compresses ridges with poor time-frequency aggregation in both the time and frequency directions. From the time-frequency matrix after STFT transformation, information on ridges with poor time-frequency aggregation and low resolution is obtained. A dynamic path optimization ridge extraction method is then used to extract instantaneous frequency-shifting ridges.

[0202] 6. Gearbox tooth count calculation

[0203] According to the power transmission path of the gearbox, the vibration signal during the meshing transmission of a pair of gears includes a meshing frequency generated by the meshing of the gear pair and two rotational frequencies generated by the rotation of the two gears. Due to the change in the number of teeth in the gearbox, the rotational speed gradually decreases and the number of meshing times gradually decreases, and both the meshing frequency and rotational frequency show a decreasing trend. The instantaneous rotational frequency ridge and meshing frequency ridge of each shaft of the gearbox are extracted from the vibration and current signal time-frequency diagram. On the ridge diagram, the meshing frequency and instantaneous rotational frequency at the same moment are selected to calculate the number of teeth of each gear shaft. The specific calculation is shown in formula (24):

[0204]

[0205] In the formula, Z1 and Z2 are the number of teeth of a pair of meshing gears, and n1 and n2 are the rotational speeds of gears Z1 and Z2. The number of teeth on each gear shaft can be calculated by extracting the meshing frequency and instantaneous rotation frequency of each shaft.

[0206] This technology takes the overhead crane gearbox in a steel plant as the research object and uses the proposed technical solution to invert the missing gearbox structural parameters. First, an accelerometer is installed at the gear shaft bearing to collect vibration information generated during gearbox operation. A current sensor is installed on the motor power supply phase line to collect current information during motor operation. The current and vibration signals are subjected to STFT transformation to extract the instantaneous current frequency and calculate the gearbox input shaft rotation frequency, determining the instantaneous rotation frequency search range. Second, the vibration signal time-frequency matrix is ​​rearranged and compressed in the time and frequency directions through a rearranged second-order synchronous compression transform. Combined with a dynamic path optimization ridge extraction method, the instantaneous rotation frequency and meshing frequency ridges of each gear in the gearbox are extracted. Finally, considering the one-to-two mapping relationship between meshing frequency and rotation frequency, a method for calculating the number of teeth in the gearbox from the high-speed shaft to the low-speed shaft is established. Instantaneous rotation frequency and meshing frequency at the same moment are selected to calculate the number of teeth on each gear shaft. The results show that the present invention realizes the tooth count inversion of gearboxes with missing tooth count parameters, providing data support for gearbox fault diagnosis. This technical solution has certain practical engineering significance and high technical value.

[0207] It should be noted that embodiments of the present invention can be implemented in hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the above-described devices and methods can be implemented using computer-executable instructions and / or included in processor control code, for example, such code provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuitry such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field-programmable gate arrays, programmable logic devices, etc., or by software executed by various types of processors, or by a combination of the above-described hardware circuitry and software, such as firmware.

[0208] The embodiments of the present invention have achieved some positive results during the research and development or use process, and have indeed great advantages compared with the prior art. The following content describes them in conjunction with the data, charts and other information of the experimental process.

[0209] Taking a steel plant overhead crane gearbox as an example, the gearbox is installed on the steel plant overhead crane and has a large working load. To prevent safety hazards caused by the load falling in the event of a crane malfunction, a constant-speed dual-motor system is used to drive the gearbox. The two output shafts of the gearbox run at the same speed. When one side of the gearbox fails, the other side can continue to work normally to complete the current production operation. The external structure of the gearbox is as follows: Figure 5 As shown, the gearbox includes two input shafts and two output shafts. The input shafts are connected to the motor via couplings. Vibration acceleration sensors are installed at each bearing position of the gearbox to collect vibration signals generated by each gear shaft during operation. The gearbox parameter inversion system uses a data acquisition card with a sampling frequency of 1kHz, a 24-bit high-precision vibration acceleration sensor, and a current sensor to collect vibration signals generated during gearbox operation and current signals generated during motor operation.

[0210] Operating condition analysis of the gearbox reveals that the input shaft motors on both sides rotate at the same speed, and the output shafts rotate at the same speed. When analyzing the current and vibration signals, the left side is analyzed first to determine the number of teeth on the left gearbox. The right side can then be analyzed using the same steps to obtain the complete number of teeth on the gearbox. The gearbox structural parameter inversion steps are as follows: Figure 6 As shown.

[0211] The two input shafts of the gearbox are connected to two identical three-phase AC asynchronous motors, model YZR450L2-8. The specific parameters of the motors are shown in Table 1.

[0212] Table 1. Motor Parameter Information Table

[0213]

[0214] Figure 7 , Figure 8 The images show the time-domain plots of the raw vibration signal data collected from the bearing positions on the left and right sides of the gearbox. Figure 9 This is a time-domain diagram of the original motor current signal data for the two input shafts of the gearbox.

[0215] Since the collected signals are current and vibration signals under varying operating conditions, the instantaneous current frequency, instantaneous rotational frequency, and meshing frequency at the same moment should be selected when extracting the frequency to calculate the number of teeth on the gear. First, the current signal on the left side of the gearbox is analyzed to extract the instantaneous current frequency under varying operating conditions. The current time-domain signal is preprocessed, and an STFT transformation is performed to draw a time-frequency diagram to extract the comfort current frequency ridge. Figure 10 To obtain the time-frequency diagram of the motor current connected to the left input shaft of the gearbox, select the instantaneous frequency at 7.8 seconds from the ridge position on the time-frequency diagram to calculate the instantaneous rotational frequency of the motor at that moment. The instantaneous current frequency of the motor connected to the left input shaft at 7.8 seconds is... 50.13Hz.

[0216] Calculate the instantaneous speed of the motor connected to the left side of the gearbox at 7.8 s:

[0217]

[0218] The instantaneous frequency of the motor at that moment is calculated by measuring the instantaneous speed of the motor connected to the left side of the reducer. :

[0219]

[0220] The vibration signals of the bearings on the left side of the gearbox were analyzed. First, the vibration signals were preprocessed, and then STFT transformation was performed on the preprocessed signals to generate time-frequency diagrams. The time-frequency diagrams of the vibration signals at each bearing position on the left side of the gearbox are shown below. Figure 11 , Figure 12 , Figure 13 , Figure 14 As shown, the intensity of vibration signals measured at different locations varies. Generally, the lighter the color of the meshing frequency ridge line on the time-frequency graph, the better. By comparing the time-frequency information on each time-frequency graph, we can find the lighter-colored ridge lines with more obvious color changes on the time-frequency graphs of vibration signals measured at different locations. Through comparison, we found that the two ridge lines marked in the figure are lighter in color and have obvious color changes. These two ridge lines are identified as the target meshing frequency ridge lines. Therefore, the left side of the gearbox contains two meshing frequencies, and the number of meshing frequencies can be used to determine that there are two pairs of meshing gears on the left side of the gearbox.

[0221] In the plotted time-frequency graph, the meshing frequency ridge is affected by noise, leading to errors in the extracted meshing frequency. Two target ridges are clearly visible in the time-frequency graph of the vibration signal at bearing position 6. A dynamic path optimization ridge extraction method is used to extract the information from these two ridges. The extracted ridges are shown below. Figure 15 As shown.

[0222] Figure 15 The frequencies at 7.8 s along the two meshing frequency ridges are 259.1 Hz and 109.1 Hz, respectively. The presence of two pairs of meshing teeth indicates that the gearbox contains three rotating shafts; therefore, the low-frequency information of the vibration signal should include three rotational frequencies. Since the input and output shafts of the gearbox have the same rotational speed and the same number of intermediate transmission shafts, the meshing relationship of the gear sets on the left and right sides of the gearbox is consistent, and the positions of the gears on the left and right sides are symmetrically distributed within the gearbox. Based on the number of meshing frequencies, the positions of the input shaft, the output shaft, and the intermediate transmission shaft, a diagram of the internal structure of the gearbox's gear meshing relationship is drawn. The specific structure is as follows: Figure 16 As shown.

[0223] By calculating the instantaneous rotational frequency of the motor, the highest rotational frequency within the gearbox was determined to be 12.36Hz, thus reducing the search range for the instantaneous rotational frequency ridge. Since the ridge information below 50Hz is not obvious on the time-frequency diagram of the STFT transform, and the rotational speeds of each shaft are inconsistent, it is necessary to rearrange the time-frequency spectrum of the low-frequency components of the vibration signals of each shaft in the time-frequency direction. Based on the gear meshing structure diagram and the external structure diagram of the gearbox, it is known that frequency information acquisition requires extracting the input shaft 1 rotational frequency information from the vibration signal at bearing 1, extracting the 2nd shaft rotational frequency information from the vibration signal at bearing 2 or bearing 6, and extracting the output shaft 1 rotational frequency information from the vibration signal at bearing 5. Figure 17 , Figure 18 , Figure 19 The time-frequency charts indicate the low-frequency information bands in the frequency spectrum at the positions of bearing 1, bearing 5, and bearing 6, respectively.

[0224] The low-frequency information of the vibration signal time-frequency diagrams of bearings 1, 5, and 6 was rearranged using a second-order synchronous compression transform. The time-frequency spectrum was rearranged in both time and frequency directions. A dynamic path optimization ridge extraction method was then used to search for the low-frequency rotational ridges in the rearranged time-frequency matrix. The dynamic path optimization ridge extraction method extracted the rotational ridges for each shaft as follows: Figure 20 , Figure 21 , Figure 22 As shown.

[0225] The time-frequency diagrams of the rotational frequency ridges extracted from the vibration signals of bearings 1, 5, and 6 show that the instantaneous rotational frequencies at 7.8 s are 12.3 Hz, 3.4 Hz, and 2.2 Hz, respectively. The rotational frequency of bearing 1, extracted from the vibration signal, is 12.3 Hz, which is not significantly different from the 12.36 Hz calculated from the current frequency of the input shaft motor at 7.8 s. The meshing frequencies on the left side of the gearbox are 259.1 Hz and 109.1 Hz, respectively. Since this gearbox is a reduction gearbox, the meshing frequency from the input shaft to the output shaft gradually decreases, and the rotational speeds of each shaft within the gearbox also show a decreasing trend. Therefore, the meshing frequency of gears Z1 and Z2... The frequency is 259.1Hz, Z1 switching frequency. 12.3Hz, the meshing frequency of gears Z3 and Z4 The frequency is 109Hz. Gears Z2 and Z3 are on the same gear shaft, and the rotational frequencies of the two gears are equal. 3.4Hz, the rotational frequency of gear Z4 on the output shaft is 2.2Hz.

[0226] Calculate the number of teeth on the gear:

[0227]

[0228]

[0229]

[0230]

[0231] In the above formula Input shaft 1 speed, For 2-axis rotation speed, The output shaft speed is 1.

[0232] The steps for inverting the number of teeth on the right side of the gearbox are the same as those for the left side. During this tooth inversion process, vibration signals from each bearing in the gearbox and the input shaft motor current signal are collected, preprocessed, and subjected to STFT transformation to generate a time-frequency diagram. The instantaneous current frequency of the motor at 7.8 s is extracted from the current time-frequency diagram, and the instantaneous rotational speed of the gearbox input shaft is calculated. On the vibration signal time-frequency diagrams at each bearing position, the meshing frequency ridge and instantaneous rotational frequency ridge are intelligently identified from high-frequency and low-frequency information. The meshing frequency and instantaneous rotational frequency information between each gear at the 7.8 s position are extracted, completing the calculation of the number of teeth on each gear shaft. This allows for the inversion of the number of teeth on each gear within the gearbox without disassembling the gearbox.

[0233] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A gear box structure parameter inversion method of fusing current, vibration signal, characterized in that, The gear box structure parameter inversion method of the fusion current and vibration signal comprises the following steps: Step one, collect the current signal of the gear box motor and the vibration signal generated when each gear shaft meshes, and pretreat the collected current signal and vibration signal; Step two, perform STFT transformation on the pretreated current signal and vibration signal, draw the time-frequency diagram of the alternating current motor current signal and the vibration signal, and extract the motor instantaneous rotation frequency; Step three, based on the drawn time-frequency diagram, extract the meshing frequency ridge line of each gear by using the dynamic path optimization ridge line extraction method, and obtain the meshing frequency information of each gear; Step four, based on the drawn time-frequency diagram, perform frequency amplitude addition processing on the time-frequency information, determine the instantaneous rotation frequency search range according to the motor instantaneous rotation frequency, compress the time and frequency information of the low frequency part, extract the instantaneous rotation frequency ridge line, and extract the instantaneous rotation frequency information of each shaft of the gear box; Step five, select the meshing frequency and instantaneous rotation frequency at the same time on the ridge line atlas according to the instantaneous rotation frequency ridge line and the meshing frequency ridge line of each shaft of the gear box extracted from the vibration and current signal time-frequency diagram, and calculate the gear number information of each gear shaft.

2. The method according to claim 1, wherein the fusion current, vibration signal and gearbox structure parameter inversion method is characterized in that, The current signal of the gear box and the vibration signal of each shaft are collected, and the collected current signal and vibration signal are pretreated, which comprises: Firstly, measure points are arranged at the bearing seat positions of each gear shaft to collect the vibration signal of each shaft, and a current sensor is arranged at the phase line position of the gear box motor power supply to collect the motor current signal; Secondly, the collected current signal is subjected to mean value reduction processing: wherein, denotes a mean value indicator of the current signal; denotes a sensitivity of the current sensor; denotes the acquired current signal, {1, 2, …, N}; N denotes the number of sampling points; Then, the collected vibration signal is pretreated: ； ； wherein, represents a mean value index of the current signal; represents a sensitivity of the vibration acceleration sensor, represents a collected vibration signal, ={1, 2, …, N}, N represents a number of sampling points.

3. The method according to claim 1, wherein the fusion current, vibration signal and gearbox structure parameter inversion method is characterized in that, The STFT transformation is performed on the pretreated current signal and vibration signal to draw the time-frequency diagram of the alternating current motor current signal and the vibration signal, which comprises: (1) performing short-time Fourier transformation on the pretreated current signal and taking the absolute value: ； (2) take time t as the horizontal axis, frequency The STFT transformed instantaneous localized frequency content estimate is taken as the vertical axis to control the depth of the time-frequency ridge line color, and a time-frequency diagram of the AC motor current signal is drawn. (3) performing short-time Fourier transformation on the vibration signal and taking the absolute value: ； (4) Take time t as the horizontal axis, frequency As the vertical axis, the STFT transformed instantaneous localized frequency content estimate controls the lightness and darkness of the time-frequency ridge line color, and the time-frequency graph of the vibration signal collected on each axis is drawn.

4. The method of claim 1, wherein the fusion current, vibration signal and gearbox structure parameter inversion method is characterized by, The motor instantaneous rotation frequency extraction is as follows: The motor instantaneous rotational frequency is calculated using the following equation : where f represents the current frequency, p represents the number of motor pole pairs, and s represents the motor slip rate; represents the rotational speed of the alternating current motor.

5. The method of claim 1, wherein the fusion current, vibration signal and gearbox structure parameter inversion method is characterized in that, Based on the drawn time-frequency diagram, the meshing frequency ridge line is extracted by using the dynamic path optimization ridge line extraction method, and the meshing frequency information of the gear is obtained, which comprises: By comparing the color depth of the ridge line on the time-frequency diagram of the alternating current motor current signal, the ridge line with higher localized frequency content is selected as the target meshing frequency ridge line, the dynamic path optimization ridge line extraction method is used to extract the meshing frequency ridge line, and the meshing frequency information of the gear is obtained; The dynamic path optimization ridge line extraction method is used to extract the meshing frequency ridge line, and the meshing frequency information of the gear is obtained, which comprises: 1) the STFT transformed time-frequency analysis matrix is subjected to frequency amplitude addition processing, and the relative amplitude size at the same frequency at different times is calculated in the form of frequency slice; 2) Time-frequency analysis of the signal, obtaining a time-frequency matrix with dimension N x M Comparing the amplitudes of different time-frequency points at the same time in the matrix along the time axis, extracting time-frequency points that satisfy and at the same time, wherein, represents the time-frequency peak point, ； wherein, represents the total number of time-frequency peak points corresponding to the time point. 3) based on the relationship between the base frequency and the frequency multiplication of the time-varying frequency component signal, the amplitude addition peak value search algorithm is performed on the different time-frequency peak points in the matrix at the same time, and it is judged whether the corresponding frequency meets the following conditions: wherein denotes rounding down; denotes the frequency resolution of the window function in the signal STFT time-frequency analysis; when the condition is met, the time-frequency peak point amplitude is added up: ； 4) same processing is performed on each column of time-frequency peak value in the time-frequency matrix to obtain a time-frequency peak value matrix; a threshold is set , the time-frequency peak value point in the time-frequency peak value matrix is screened, when , the time-frequency peak value point is reserved, otherwise the amplitude of the peak value point is set to zero; 5) the ridge line is extracted from the amplitude-added time-frequency matrix by using the cost function; The cost function is as follows: 。 6. The method of claim 1, wherein the fusion current, vibration signal and gearbox structure parameter inversion method is characterized by, Step four comprises: The gear box input shaft instantaneous rotation frequency is calculated through the motor instantaneous current frequency information, the rotation frequency ridge line search range of the gear is determined through the input shaft rotation frequency, and the time-frequency information of the search range is subjected to second-order rearrangement synchronous compression transformation to determine the target instantaneous rotation frequency ridge line to be extracted. The second-order rearrangement synchronous compression transformation of the time-frequency information of the search range to determine the target rotation frequency ridge line includes: (1) STFT conversion is performed on the vibration signal: ； ； wherein denotes the original vibration signal; k e N is a positive integer, denotes the instantaneous amplitude, denotes the instantaneous phase function; denotes the short-time Fourier transform coefficient, the complex conjugate of (2) second-order local complex instantaneous time and frequency estimation is performed: (2.1) second-order local complex instantaneous frequency IF estimation is performed using the following formula: ； wherein denotes the instantaneous frequency IF estimated based on the STFT, ; Im denotes the imaginary part of a complex number; denotes a Taylor approximation representation of the instantaneous time, ; and denote the time-frequency coefficients computed from the STFT with a window function g(t) or tg(t), respectively; denotes a second-order local complex modulation operator, ; (2.2) second-order local instantaneous time estimation is performed using the following formula: ； (3) the rearrangement time and frequency operator is calculated using the following formula: ； ； wherein denotes a rearrangement time operator; denotes a rearrangement frequency operator (4) Calculate the compensation frequency distance : (5) the time-frequency spectrum of the vibration signal is rearranged from the two-dimensional time-frequency direction, and second-order compression is performed: The second-order synchronous compression transform is rearranged using the following formula : ； wherein, denotes the Dirac function, denotes the compressed reconstructed time-frequency matrix.

7. The method of claim 1, wherein the fusion current, vibration signal and gearbox structure parameter inversion method is characterized by, The gear number information calculation formula of each gear shaft is as follows: ； wherein Z1, Z2 represent the number of teeth of a pair of meshing gears; n1, n2 represent the rotational speed of the gears Z1, Z2, rotational speed .

8. A fusion current, vibration signal of a gear box structure parameter inversion system for implementing the fusion current, vibration signal of a gear box structure parameter inversion method according to any one of claims 1-7, characterized in that, The gear box structure parameter inversion system fusing current and vibration signals includes: A signal acquisition and preprocessing module is configured to acquire current signals and vibration signals of each shaft of the gear box, and to preprocess the acquired current signals and vibration signals; A time-frequency analysis module is configured to perform STFT conversion on the preprocessed current signals and vibration signals, and to draw time-frequency diagrams of the alternating current motor current signals and the vibration signals; A motor instantaneous rotation frequency extraction module is configured to extract motor instantaneous rotation frequency, and to determine a gear box instantaneous rotation frequency ridge line search range; An engagement frequency extraction module is configured to extract an engagement frequency ridge line based on the drawn time-frequency diagrams using a dynamic path optimization ridge line extraction method, and to acquire engagement frequency information of the gear; An instantaneous rotation frequency extraction module is configured to perform frequency amplitude addition processing on the time-frequency information based on the drawn time-frequency diagrams, to compress the time and frequency information of the low frequency part, to extract an instantaneous rotation frequency ridge line, and to extract instantaneous rotation frequency information of each shaft of the gear box; A gear number information calculation module is configured to select engagement frequency and instantaneous rotation frequency at the same time on the ridge line atlas based on the extracted instantaneous rotation frequency ridge line and engagement frequency ridge line of each shaft of the gear box from the vibration and current signal time-frequency diagrams, and to calculate gear number information of each gear shaft.

9. A computer device, comprising: The computer device includes a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the gear box structure parameter inversion method fusing current and vibration signals according to any one of claims 1-7.

Images