A method and system for optimizing vibration and noise in an electric drive assembly.

Abstract

This invention relates to the field of vibration and noise optimization technology, and particularly to a method and system for optimizing the vibration and noise of an electric drive assembly. The method includes the following steps: collecting vibration and noise data of the transmission motor in the electric drive assembly under multiple operating conditions and performing frequency domain conversion to generate vibration and noise frequency domain data; based on this data, performing gear meshing force off-center load limit calculation, simulating the cyclic extensibility of tooth surface pitting, and identifying the bearing clearance expansion increment; optimizing the bearing structure based on the bearing clearance expansion increment data to obtain bearing structure optimization data; performing tooth profile bulging modification analysis based on the tooth surface pitting data to obtain modification data; finally, combining the bearing optimization and tooth profile modification data to perform vibration-resistant strengthening design, generating structural vibration-resistant strengthening design data, and sending it to a terminal to execute structural optimization and reduce the impact of vibration and noise. This invention improves vibration and noise optimization technology through optimization processing.

Description

Technical Field

[0001] This invention relates to the field of vibration and noise optimization technology, and in particular to a method and system for optimizing the vibration and noise of an electric drive assembly. Background Technology

[0002] Electric drive assemblies typically consist of multiple components such as motors, transmissions, and bearings. Among these, the interaction between the motor and transmission, and the vibration and noise issues within the electric drive assembly under various operating conditions, are particularly prominent. This study combines vibration and noise data analysis with mechanical structure optimization. By performing frequency domain transformation on the vibration and noise data, characteristic frequency information is extracted, and the impact of vibration and noise on gear meshing forces and bearing clearance is analyzed to propose structural optimization schemes. However, traditional vibration and noise optimization methods for electric drive assemblies suffer from inaccurate analysis of the impact of vibration and noise on gears and bearings, resulting in large errors in the optimization of the mechanical structure's vibration and noise resistance. Summary of the Invention

[0003] Therefore, it is necessary to provide a vibration and noise optimization method and system for electric drive assemblies to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a method for optimizing the vibration and noise of an electric drive assembly is provided, the method comprising the following steps:

[0005] Step S1: Collect vibration and noise under multiple operating conditions of the transmission motor in the electric drive assembly under experimental conditions; perform frequency domain conversion processing on the vibration and noise under multiple operating conditions to generate vibration and noise frequency domain data;

[0006] Step S2: Perform the gear meshing force off-center load limit calculation based on vibration and noise frequency domain data, and then perform tooth surface pitting cyclic ductility simulation to obtain tooth surface pitting cyclic ductility data; identify bearing clearance expansion increment based on vibration and noise frequency domain data to obtain bearing clearance expansion increment data.

[0007] Step S3: Based on the bearing clearance expansion increment data, perform bearing structure optimization processing to obtain bearing structure optimization data; perform tooth-direction drum shape modification analysis based on tooth surface pitting cycle extension data to obtain tooth-direction drum shape modification data; perform vibration-resistant strengthening design based on bearing structure optimization data and tooth-direction drum shape modification data to generate structural vibration-resistant strengthening design data; send the structural vibration-resistant strengthening design data to the terminal to perform structural optimization affected by vibration and noise.

[0008] Preferably, step S1 includes the following steps:

[0009] Step S11: Collect vibration and noise data under various operating conditions of the transmission motor in the electric drive assembly during the experimental process;

[0010] Step S12: Perform data cleaning on the multi-condition vibration and noise to generate multi-condition vibration and noise cleaning data;

[0011] Step S13: Perform detail enhancement processing on the multi-condition vibration and noise cleaning data to obtain vibration and noise detail enhancement data;

[0012] Step S14: Perform frequency domain transformation on the vibration noise detail enhancement data to generate vibration noise frequency domain data.

[0013] Preferably, step S2 includes the following steps:

[0014] Step S21: Analyze the vibration energy transfer intensity diffusion based on the vibration noise frequency domain data to obtain vibration energy transfer intensity diffusion data;

[0015] Step S22: Calculate the gear meshing force eccentric load limit based on the vibration energy transfer intensity diffusion data to obtain the gear meshing force eccentric load limit data;

[0016] Step S23: Simulate the cyclic ductility of tooth surface pitting based on the gear meshing force eccentric load limit data to obtain the cyclic ductility data of tooth surface pitting.

[0017] Step S24: Identify the bearing clearance expansion increment based on the vibration energy transfer intensity diffusion data, thereby obtaining the bearing clearance expansion increment data.

[0018] Preferably, step S22 includes the following steps:

[0019] Step S221: Obtain the gear meshing vibration frequency and basic characteristics of the gear material under the current working condition; perform diffusion dispersion superposition frequency doubling quantization on the vibration energy transfer intensity diffusion data to obtain energy dispersion superposition frequency doubling data;

[0020] Step S222: Analyze the energy density evolution ratio of the gear meshing vibration frequency based on the energy dispersion superposition overtone data, and generate the energy density evolution ratio of the resonance peak;

[0021] Step S223: Based on the energy density evolution ratio of the resonance peak, perform incremental simulation of the progressive gradient of fatigue deformation of the gear shaft to estimate the basic properties of the gear material, thereby obtaining incremental data of the progressive gradient of fatigue deformation.

[0022] Step S224: Dynamically simulate and calculate the parallelism / perpendicularity deviation data of gear meshing using fatigue deformation progressive gradient incremental data;

[0023] Step S225: Calculate the gear meshing force off-center load limit based on the parallelism / perpendicularity deviation data to obtain the gear meshing force off-center load limit data.

[0024] Preferably, step S225 includes the following steps:

[0025] Dynamic integration of meshing line offset is performed based on parallelism / perpendicularity deviation data to obtain dynamic meshing line offset data;

[0026] The axial force and radial force imbalance index is calculated based on the dynamic offset data of the meshing line to obtain the axial force / radial force imbalance index.

[0027] The imbalance index of axial force / radial force was analyzed by proportional variation of the imbalance index at adjacent time points to obtain the proportional variation data of the imbalance index at adjacent time points;

[0028] Based on the proportional change data of the imbalance index and the dynamic offset data of the meshing line, the rotational force misalignment deviation of the gear helix angle is quantified to obtain the rotational force misalignment deviation data.

[0029] Based on the rotational force misalignment deviation data, the proportional change data of the imbalance index, and the dynamic offset data of the meshing line, the gear meshing force off-center load limit is calculated, thereby obtaining the gear meshing force off-center load limit data.

[0030] Preferably, step S23 includes the following steps:

[0031] Step S231: Analyze the load distribution along the tooth width skew of the gear meshing force limit data to obtain the load tooth width skew distribution data;

[0032] Step S232: Based on the gear meshing force eccentric load limit data, the load tooth width skew distribution data is subjected to tooth width edge contact stress geometric growth quantification to obtain tooth width edge stress geometric growth data.

[0033] Step S233: Iteratively couple the stress crack propagation degree to the geometric growth data of tooth width edge stress to obtain iterative stress crack propagation degree data;

[0034] Step S234: Based on the geometric growth data of tooth width edge stress and the iterative stress crack propagation data, simulate the cyclic extension of tooth surface pitting to obtain the cyclic extension data of tooth surface pitting.

[0035] Preferably, step S24 includes the following steps:

[0036] Step S241: Perform high-frequency impact vector analysis on the vibration energy transfer intensity diffusion data between the bearings to obtain the high-frequency impact vector;

[0037] Step S242: Calculate the radial load distribution between bearings based on the high-frequency vibration impact vector to obtain bearing radial load distribution data;

[0038] Step S243: Based on the high-frequency vibration impact vector, perform Hertzian contact stress iterative simulation and deduction on the radial load distribution data of the bearing to generate Hertzian contact stress iterative data;

[0039] Step S244: Solve the nonlinear variation of the bearing contact angle based on the Hertzian contact stress iteration data and the bearing radial load distribution data to obtain the bearing contact angle variation data;

[0040] Step S245: Identify the bearing clearance increase based on the bearing contact angle change data, thereby obtaining the bearing clearance increase data.

[0041] Preferably, step S3 includes the following steps:

[0042] Step S31: Perform logical learning on the bearing clearance increase increment data to generate bearing clearance increase increment learning data;

[0043] Step S32: Based on the learning data of bearing clearance expansion increment, perform bearing structure optimization processing to obtain bearing structure optimization data;

[0044] Step S33: Perform tooth-direction bulging modification analysis based on the tooth surface pitting cycle extension data to obtain tooth-direction bulging modification data;

[0045] Step S34: Based on the bearing structure optimization data and tooth profile drum shape modification data, conduct vibration-resistant strengthening design to mitigate the impact of vibration and generate structural vibration-resistant strengthening design data;

[0046] Step S35: Send the structural vibration strengthening design data to the terminal to perform structural optimization for the impact of vibration and noise.

[0047] Preferably, step S32 includes the following steps:

[0048] Step S321: Measure the bearing radial stiffness strengthening ratio based on the bearing clearance expansion increment learning data to obtain the bearing radial stiffness strengthening ratio;

[0049] Step S322: Based on the learning data of bearing radial stiffness enhancement ratio and bearing clearance increase increment, couple the contact angle to increase by an equal amount to generate contact angle increase data;

[0050] Step S323: Perform adaptive design matching of the runout angle by increasing the bearing radial stiffness enhancement ratio and contact angle by the same amount to obtain the runout angle matching data;

[0051] Step S324: Based on the bearing radial stiffness enhancement ratio, contact angle increase data, and yaw angle matching data, the bearing structure is optimized to obtain bearing structure optimization data.

[0052] Preferably, the present invention also provides a vibration and noise optimization system for an electric drive assembly, used to execute the vibration and noise optimization method for the electric drive assembly as described above, the vibration and noise optimization system for the electric drive assembly comprising:

[0053] The frequency domain conversion module is used to collect vibration and noise under various operating conditions of the transmission motor in the electric drive assembly during experimental conditions; and to perform frequency domain conversion processing on the vibration and noise under various operating conditions to generate vibration and noise frequency domain data.

[0054] The vibration and noise impact analysis module is used to perform gear meshing force off-center load limit calculation based on vibration and noise frequency domain data, and then perform tooth surface pitting cyclic ductility simulation to obtain tooth surface pitting cyclic ductility data; and to identify bearing clearance expansion increment based on vibration and noise frequency domain data to obtain bearing clearance expansion increment data.

[0055] The vibration-strengthening design module is used to optimize the bearing structure based on the bearing clearance expansion increment data to obtain bearing structure optimization data; to perform tooth-direction drum shape modification analysis based on tooth surface pitting cycle extension data to obtain tooth-direction drum shape modification data; to perform vibration-strengthening design based on the bearing structure optimization data and tooth-direction drum shape modification data to generate structural vibration-strengthening design data; and to send the structural vibration-strengthening design data to the terminal to perform structural optimization affected by vibration and noise.

[0056] The beneficial effects of this invention are as follows: by collecting multi-condition vibration and noise data of the transmission motor in the electric drive assembly under experimental conditions and performing frequency domain transformation on these noises, the frequency domain characteristics of the vibration and noise can be extracted, accurately identifying the noise source and its frequency distribution. The generated vibration and noise frequency domain data provides accurate basic data for subsequent analysis and optimization, ensuring clear identification and location of the noise source, providing a scientific basis for further optimization, and improving the accuracy of noise control. Using the vibration and noise frequency domain data to perform gear meshing force off-center load limit calculations, the stress situation of the gear under different operating conditions can be analyzed, thereby identifying the risk of tooth surface pitting caused by vibration and noise. Then, the cyclic extension data of tooth surface pitting is simulated to obtain the cyclic extension data of tooth surface pitting, enabling the prediction of tooth surface damage trends during the design stage. Simultaneously, the increase in bearing clearance can be identified based on the vibration and noise frequency domain data, thereby accurately grasping the changes in bearing clearance and ensuring that the impact of noise and vibration is effectively reduced through structural optimization design. By optimizing the bearing structure through incremental data of bearing clearance expansion, the operating condition of the bearing can be effectively improved, reducing the negative impact of vibration and noise. The optimized bearing structure can improve its stability and reliability, further reducing the impact of noise. Simultaneously, tooth profile modification analysis based on tooth surface pitting cycle extension data can effectively adjust the gear meshing accuracy, reduce tooth surface damage, and improve gear meshing smoothness through tooth profile modification optimization. Therefore, this invention is an improvement on a traditional vibration and noise optimization method for electric drive assemblies. It solves the problem that traditional methods for optimizing vibration and noise in electric drive assemblies suffer from inaccurate analysis of the impact of vibration and noise on gears and bearings, resulting in large errors in the vibration and noise resistance optimization of the mechanical structure. This invention improves the accuracy of the analysis of the impact of vibration and noise on gears and bearings, thereby reducing the error in the vibration and noise resistance optimization of the mechanical structure. Attached Figure Description

[0057] Figure 1 This is a schematic diagram of the steps involved in a vibration and noise optimization method for an electric drive assembly.

[0058] Figure 2 for Figure 1 A detailed flowchart illustrating the implementation steps of step S2.

[0059] Figure 3 for Figure 1 A detailed flowchart illustrating the implementation steps of step S3. Detailed Implementation

[0060] Please see Figures 1 to 3 A method for optimizing the vibration and noise of an electric drive assembly, the method comprising the following steps:

[0061] Step S1: Collect vibration and noise under multiple operating conditions of the transmission motor in the electric drive assembly under experimental conditions; perform frequency domain conversion processing on the vibration and noise under multiple operating conditions to generate vibration and noise frequency domain data;

[0062] Step S2: Perform the gear meshing force off-center load limit calculation based on vibration and noise frequency domain data, and then perform tooth surface pitting cyclic ductility simulation to obtain tooth surface pitting cyclic ductility data; identify bearing clearance expansion increment based on vibration and noise frequency domain data to obtain bearing clearance expansion increment data.

[0063] Step S3: Based on the bearing clearance expansion increment data, perform bearing structure optimization processing to obtain bearing structure optimization data; perform tooth-direction drum shape modification analysis based on tooth surface pitting cycle extension data to obtain tooth-direction drum shape modification data; perform vibration-resistant strengthening design based on bearing structure optimization data and tooth-direction drum shape modification data to generate structural vibration-resistant strengthening design data; send the structural vibration-resistant strengthening design data to the terminal to perform structural optimization affected by vibration and noise.

[0064] In this embodiment of the invention, reference Figure 1 The above is a schematic flowchart of the vibration and noise optimization method for an electric drive assembly according to the present invention. In this example, the vibration and noise optimization method for the electric drive assembly includes the following steps:

[0065] Step S1: Collect vibration and noise under multiple operating conditions of the transmission motor in the electric drive assembly under experimental conditions; perform frequency domain conversion processing on the vibration and noise under multiple operating conditions to generate vibration and noise frequency domain data;

[0066] In this embodiment of the invention, a high-sensitivity triaxial accelerometer is used to test the transmission motor in the electric drive assembly in an NVH laboratory environment. The test conditions cover different speed ranges such as 1000 rpm, 3000 rpm, and 5000 rpm, and different load conditions such as 0 Nm, 50 Nm, and 100 Nm torque are superimposed in each speed range. The sampling frequency is set to 25600 Hz to ensure complete acquisition of the high-frequency characteristics of the vibration signal. The obtained multi-condition vibration and noise raw data are processed by time-domain window function to correct the endpoints. The Hanning window function is selected to alleviate the truncation effect of the data boundary. After the window function processing is completed, the fast Fourier transform (FFT) algorithm is used to convert the time-domain vibration signal to the frequency domain. The frequency domain resolution is set to 0.5 Hz to meet the accurate analysis of gear meshing order frequencies such as gear mesh frequency and its harmonics. The final generated vibration and noise frequency domain data is expressed in dB / Hz as the spectral amplitude density unit.

[0067] In another embodiment, when collecting vibration and noise data from the transmission motor in the electric drive assembly under various operating conditions during testing, a PCB356A45 triaxial accelerometer was first installed at five key locations on the surface of the transmission housing. The sensor sensitivity was set to 10 mV / g, the sampling frequency was fixed at 25600 Hz, and the acquisition time was set to 180 seconds. Simultaneously, the motor speed range was recorded from 800 rpm to 6000 rpm, and the torque load from 50 Nm to 350 Nm. Vibration signals were collected at 20 different operating points using LMS SCADAS. The mobile data acquisition system has 16 acquisition channels, uses IEPE power supply mode for signal conditioning, and sets the preamplifier gain to 10dB. The acquired vibration signals are then preprocessed using a Hanning window function with a window length of 4096 points and an overlap rate of 50%. A Fast Fourier Transform (FFT) algorithm is then used to convert the time-domain signal, with the analysis frequency range set to 0-10000Hz and the frequency resolution controlled at 2Hz. The resulting vibration noise frequency domain data includes amplitude and phase spectra, with vibration amplitude units uniformly in g-values ​​and phase units in degrees. The frequency domain data pays special attention to the gear meshing frequency (GMF) and its harmonic components. Bearing characteristic frequencies include the rolling element passing through the inner ring frequency (BPFI), rolling element passing through the outer ring frequency (BPFO), rolling element rotation frequency (BSF), and cage frequency (FTF). Resonance peaks in the spectrum are marked, ultimately forming a frequency domain data matrix containing the spectrum data of all operating conditions. The matrix dimension is 20×8192, with each row representing an operating condition point and each column representing the vibration amplitude of a frequency point.

[0068] Step S2: Perform the gear meshing force off-center load limit calculation based on vibration and noise frequency domain data, and then perform tooth surface pitting cyclic ductility simulation to obtain tooth surface pitting cyclic ductility data; identify bearing clearance expansion increment based on vibration and noise frequency domain data to obtain bearing clearance expansion increment data.

[0069] In this embodiment of the invention, vibration energy distribution analysis is performed on the vibration noise frequency domain data obtained in step S1. The transform point method is used to identify the meshing frequency segments where the main energy is concentrated. The multibody dynamics analysis module in the finite element method is used to extract the local stress time history of the gear meshing contact area. Based on the maximum principal stress amplitude within the time history, statistics are performed on different meshing positions. The Goodman correction criterion is used to perform the gear meshing force eccentric load limit calculation. By comparing the limit calculation results with the theoretical contact stress at different meshing positions, the gear force distribution limit curve is obtained. Subsequently, the limit curve is input into the gear fatigue life prediction module, and the Miner linear cumulative damage theory combined with the SN curve is used to analyze the pitting corrosion of the gear material. The number of crack initiation cycles was calculated, and then the pitting crack depth propagation stage was iteratively simulated using crack propagation rate test data. The step size was set to 0.05 mm to improve the fitting accuracy of the pitting propagation trajectory, and the pitting cycle extension data of the tooth surface was obtained. In addition, the characteristic segments of bearing clearance change were extracted based on vibration and noise frequency domain data. Short-time Fourier transform was used to perform time segmentation processing on the local high-frequency band signal. The amplitude change of the harmonic order of the clearance change segment was accumulated and integrated by combining the bearing clearance characteristic frequency formula. Then, the incremental change of clearance was identified by differential change rate analysis to obtain the bearing clearance expansion increment data. All data outputs were saved in CSV format and input into the subsequent optimization process.

[0070] In another embodiment, when performing the limit calculation of gear meshing load imbalance based on vibration noise frequency domain data, the gear meshing frequency and its 1st to 8th order harmonic vibration energy distribution are first extracted from the frequency domain data. The amplitude ratio of each harmonic is calculated to determine the increase or decrease law of the harmonics. The benchmark harmonic ratio is set to 1.35. When the measured harmonic ratio exceeds 1.85, it is judged as an abnormal meshing load imbalance. Then, a gear meshing stiffness fluctuation model is constructed, with the stiffness fluctuation coefficient ranging from 0.15 to 0.28, increasing in 5% increments. The transfer matrix method is used to calculate the dynamic load at the meshing point. Load distribution was described using a piecewise continuous function along the tooth width, with a load concentration factor of 1.62 and an edge stress concentration factor of 1.78. Then, the AGMA standard contact stress calculation formula was applied, substituting a gear module of 7mm, a pressure angle of 20 degrees, a helix angle of 15 degrees, a tooth width of 65mm, and 2.05 meshing teeth to calculate the maximum contact stress on the tooth surface. When the maximum contact stress exceeded 85% of the material's yield strength of 1450MPa, it was determined to be a critical state of eccentric loading. Finally, the Palmgern-Miner cumulative fatigue loss was calculated. Based on the damage theory, the fatigue accumulation process of tooth surface contact was simulated with a cycle size of 10,000 cycles. The fatigue damage coefficient was set to 0.85, and the cumulative cycle count was set to 5 million cycles. Pitting failure was considered to have occurred when the damage value reached 1. The pitting initiation position, depth, and propagation rate were recorded to obtain the tooth surface pitting cycle extension data. Simultaneously, the bearing characteristic frequency and its sidebands were extracted from the vibration and noise frequency domain data. The deviation between the sideband interval and the theoretical value was calculated, and the deviation threshold was set to 3 Hz. Based on the bearing characteristic frequency modulation depth analysis, the clearance change was analyzed, and envelope demodulation technology was used to extract the shaft clearance. The bearing is subjected to an impact pulse signal with a pulse amplitude threshold of 0.15g and a pulse interval period error controlled within 5%. Combined with the parameters of a bearing contact angle of 15 degrees, 12 rolling elements, a rolling element diameter of 12mm, and a pitch circle diameter of 85mm, the radial clearance increment of the bearing is calculated through the nonlinear stiffness-displacement relationship. The standard clearance value is 0.025mm. When the clearance expansion exceeds 0.045mm, it is judged as abnormal. The clearance expansion rate is calculated to be 80%, and the spatial distribution non-uniformity coefficient of clearance expansion is taken as 1.32. The bearing clearance expansion increment data are obtained.

[0071] Step S3: Based on the bearing clearance expansion increment data, perform bearing structure optimization processing to obtain bearing structure optimization data; perform tooth-direction drum shape modification analysis based on tooth surface pitting cycle extension data to obtain tooth-direction drum shape modification data; perform vibration-resistant strengthening design based on bearing structure optimization data and tooth-direction drum shape modification data to generate structural vibration-resistant strengthening design data; send the structural vibration-resistant strengthening design data to the terminal to perform structural optimization affected by vibration and noise.

[0072] In this embodiment of the invention, multivariate regression analysis is performed on the bearing clearance increase data obtained in step S2. The relationship between the clearance change and the contact stress between the inner and outer rings of the bearing is selected as the dependent variable input. The least squares method is used to fit a linear model of the influence of bearing clearance on contact stiffness. Then, based on the fitted curve, the bearing inner ring thickness and roller length parameters are recalculated. Finite element contact simulation software such as ANSYS is used to model the bearing structure. Axisymmetric elements are selected for mesh generation, and the element size is controlled within 0.1 mm to ensure the analytical accuracy of stress gradient changes. The radial stiffness distribution change of the optimized bearing structure under working conditions is simulated by applying actual load conditions to form bearing structure optimization data. At the same time, tooth surface pitting cyclic extension data obtained in step S2 are analyzed for tooth bulging modification. Multi-order B-spline curves are used to fit the original tooth shape, and a cubic B-spline function is selected as the modification base curve. The length is set to 1% of the tooth length. By corresponding the pitting extension area with the tooth width position coordinates, the offset of the drum-shaped curve is calculated, and the tooth-direction drum-shaped modification data is finally generated. Then, the bearing structure optimization data and the tooth-direction drum-shaped modification data are jointly input. The vibration transmission path analysis method VTPA is used to quantify the natural frequency offset of the overall structure of the electric drive assembly at each meshing order frequency. The transfer function is optimized and iterated using the structural impedance matrix method. The local thickness and support stiffness of key structural units are adjusted. The TOPSIS multi-objective decision optimization method is used to select the best stiffness and mass distribution schemes for the connection area of ​​the shell and support seat. Finally, the structural vibration-resistant strengthening design data is generated. All optimized design parameters, including bearing preload, gear modification offset, and shell stiffener distribution, are finally sent to the electric drive assembly control terminal via CAN bus protocol to guide the subsequent structural optimization and manufacturing process.

[0073] In another embodiment, when optimizing the bearing structure based on the bearing clearance increase data, the clearance value in the measured bearing clearance increase data is first adjusted from 0.045mm to 0.035mm. This is achieved by increasing the bearing preload from 1500N to 2200N, while simultaneously increasing the bearing contact angle from 15 degrees to 18 degrees. The ratio of the inner and outer ring groove curvature coefficients is maintained at 1.08. The inner ring groove curvature radius is increased from 5.46mm to 5.52mm, and the outer ring groove curvature radius is increased from 5.85mm to 5.92mm. The consistency fit between the rolling elements and raceways is controlled between 0.98 and 1.02. The internal radial clearance of the bearing is then modified to... Within the range of 0.015mm to 0.020mm, the bearing cage structure is made of stamped brass with a window-type cage design. The cage guide clearance is reduced to 0.18mm, the circumferential clearance between the rolling elements and the cage is controlled at 0.25mm, the rolling element ball diameter tolerance grade is adjusted to G5, and the diameter deviation is controlled within ±0.0005mm. The cylindricity accuracy of the bearing housing bore is improved to 0.005mm. Bearing structure optimization data is obtained. Based on the tooth surface pitting cycle extension data, tooth camber modification analysis is performed. The tooth camber amount is gradually varied from both ends to the middle, with the maximum camber amount set at 0.025mm. The camber curve adopts a quadratic parabola form, and the curve coefficient is set to... The initial position of the drum-shaped profile is set at 7.5mm from the edge of the tooth width, with the length of the straight section in the middle being 35% of the tooth width. The length of the contour transition zone is 25% of the tooth width. The tooth profile contouring uses a head contouring amount of 0.018mm, with the contouring starting position located 2.5mm below the tooth tip circle. The contouring length is 22% of the tooth height. The contouring transition curve adopts an exponential function form with an exponential coefficient of 1.75. Based on the bearing structure optimization data and the tooth contour contouring data, vibration-resistant reinforcement design is carried out to address the impact of vibration. The thickness of the transmission housing ribs is increased from 4mm to 6mm, and the rib spacing is reduced from 120mm to 90mm. The housing material is made of aluminum alloy. The AC4B model has been changed to AC4CH-T6. The housing wall thickness has been increased from 5mm to 7mm. Radial reinforcing rings with a thickness of 8mm and a width of 35mm have been added at key support points. The thickness of the flange connecting the gearbox and motor has been increased from 12mm to 15mm. The number of flange connection bolts has been increased from 8 to 12, and the bolt preload has been increased from 25kN to 32kN. Epoxy resin vibration damping material with a density of 1.85g / cm³, a Shore D75 hardness, and a loss factor of 0.08 has been filled between the motor housing and the stator. A 0.12mm thick damping gasket made of nitrile rubber and copper powder composite material with a damping ratio of 0 has been added between the bearing housing and the housing contact surface.15. Four hydraulic vibration damping supports are added at the connection between the gearbox housing and the bracket. The support stiffness is 65 N / mm, and the damping coefficient is 1200 N·s / m. Structural vibration strengthening design data is generated and sent to the Siemens S7-1500 PLC controller via the OPC UA communication protocol. The communication rate is set to 100 Mbps, the data packet size is 512 KB, and the transmission delay is controlled within 15 ms. The receiving end uses the CRC32 algorithm for data verification with a verification accuracy of 99.99%. After successful data transmission, an execution command is triggered, and the CNC machining center is started to execute the machining program. The machining accuracy is controlled within 0.01 mm, the surface roughness requirement is Ra1.6, and the heat treatment hardness requirement is HRC45-50, in order to perform structural optimization to mitigate the impact of vibration and noise.

[0074] Step S1 includes the following steps:

[0075] Step S11: Collect vibration and noise data under various operating conditions of the transmission motor in the electric drive assembly during the experimental process;

[0076] Step S12: Perform data cleaning on the multi-condition vibration and noise to generate multi-condition vibration and noise cleaning data;

[0077] Step S13: Perform detail enhancement processing on the multi-condition vibration and noise cleaning data to obtain vibration and noise detail enhancement data;

[0078] Step S14: Perform frequency domain transformation on the vibration noise detail enhancement data to generate vibration noise frequency domain data.

[0079] In this embodiment of the invention, a triaxial accelerometer is used to collect vibration and noise data of a sample electric drive assembly. The sensor is fixed at the top of the gearbox housing of the electric drive assembly. The data collection is conducted in a constant temperature laboratory environment at 25°C. The drive motor is powered by an AVL variable frequency power supply system, with set speed ranges of 1000rpm, 3000rpm, 5000rpm, and 7000rpm. Three output torques of 0Nm, 50Nm, and 100Nm are applied in each speed range. Noise testing is performed simultaneously. A GRAS 40AE free-field electroacoustic sensor is used for noise acquisition. The microphone is positioned 500mm away from the housing of the electric drive assembly. The sampling frequency is fixed at 25600Hz, and the data acquisition duration is set to 120 seconds per condition. The time-domain vibration signal and sound pressure signal are recorded in real time through the LMS Test.Lab data acquisition platform. The original multi-condition vibration and noise data obtained in step S11 are cleaned. First, a self-written Matlab script is used to remove outliers from the time-domain vibration signal. The sliding window method is used to perform mean square error analysis on the data. The window length is set to 1024 sampling points and the sliding step size is 256 sampling points. Outlier segments with a mean square error significantly greater than 3 times the standard deviation of the sample mean are removed. Linear interpolation is used to fill the removed segments. Then, baseline drift correction is performed. The least squares linear fitting method is used to fit the baseline of the original signal in multiple segments. The segment length is fixed at 5% of the total sampling time. The fitted curve is differentially removed to eliminate the influence of low-frequency drift. Then, repeated segments are identified and removed. The correlation coefficient method is used to perform multi-window similarity analysis on the time-series signal. The window length is 2048 points and the similarity threshold is set to 0.98. Highly repeated segments are removed to avoid redundant interference in subsequent feature extraction. Finally, the cleaned multi-condition vibration and noise data is output.

[0080] The multi-scale time-frequency domain enhancement method is used to enhance the details of the multi-condition vibration and noise cleaning data obtained in step S12. First, the cleaning data is processed by wavelet packet decomposition, and Daubechies is selected. An 8th-order wavelet function was used as the mother wavelet, with a decomposition layer of 6. Energy discrimination was performed on the sub-band signals of each layer, and the energy proportion coefficient of each sub-band was calculated. The sub-band signals with the top 20% energy proportion were selected for signal reconstruction to enhance the energy distribution of the main vibration characteristics. Subsequently, Hilbert instantaneous energy analysis was performed on the reconstructed signal, and the signal envelope was normalized. The instantaneous energy curve was locally smoothed using the sliding root mean square method, with a sliding window length of 512 points and a step size of 256 points. At the same time, detail stretching was performed on the signal in the frequency domain. Short-time Fourier transform (STFT) was used to perform time-frequency domain distribution analysis on the signal, with a window function length of 1024 points and an overlap length of 512 points. Amplitude nonlinear compression was performed on the low-amplitude range, and cubic spline interpolation was used for amplitude range mapping to ensure that the characteristic amplification factor of the low-amplitude energy segment was not less than 3 times. The vibration noise detail enhancement data obtained in step S13 is processed by frequency domain transformation. The Fast Fourier Transform (FFT) algorithm is used for frequency domain analysis. First, the time-domain signal is weighted using a window function. A Hanning window is selected to reduce leakage effects, with a window length of 4096 points and an overlap length of 2048 points. Then, zero-padding is performed, with the zero-padding length being twice the original signal length to improve the frequency domain resolution to 0.1Hz. After the FFT transformation, the amplitude of the obtained complex spectrum is calculated, and logarithmic normalization is used, with the unit set to dB re. To match the subsequent gear meshing order feature recognition requirements, the amplitude-frequency characteristic curve was bandpass filtered. The passband range of the filter was set to 20Hz to 10000Hz, and the filter type was a fourth-order Butterworth filter. Multi-order harmonic features were extracted from the filtered spectrum. The feature interval was divided into meshing frequency interval, harmonic interval and high-frequency random interval. Finally, the vibration noise frequency domain data was output and saved as a three-dimensional matrix format. Each row corresponds to a frequency point, each column corresponds to different test conditions, and each layer corresponds to different sensor channels. The matrix was stored in MATLAB .mat file format.

[0081] Step S2 includes the following steps:

[0082] Step S21: Analyze the vibration energy transfer intensity diffusion based on the vibration noise frequency domain data to obtain vibration energy transfer intensity diffusion data;

[0083] Step S22: Calculate the gear meshing force eccentric load limit based on the vibration energy transfer intensity diffusion data to obtain the gear meshing force eccentric load limit data;

[0084] Step S23: Simulate the cyclic ductility of tooth surface pitting based on the gear meshing force eccentric load limit data to obtain the cyclic ductility data of tooth surface pitting.

[0085] Step S24: Identify the bearing clearance expansion increment based on the vibration energy transfer intensity diffusion data, thereby obtaining the bearing clearance expansion increment data.

[0086] As an example of the present invention, reference is made to Figure 2 As shown, in this example, step S2 includes:

[0087] Step S21: Analyze the vibration energy transfer intensity diffusion based on the vibration noise frequency domain data to obtain vibration energy transfer intensity diffusion data;

[0088] In this embodiment of the invention, the vibration noise frequency domain data obtained in step S14 is subjected to vibration energy transfer intensity diffusion analysis. First, the frequency domain data is reconstructed by time segmentation using short-time Fourier transform, with a time window length of 1024 points and an overlap length of 512 points. The energy change trend of different meshing order frequency intervals is identified using the time-frequency energy distribution matrix. The weighted energy center method is used to calculate the weight coefficients for the energy concentration areas. The weight coefficient calculation range covers the frequency interval from 20Hz to 10000Hz. The vibration energy density in different frequency bands is normalized based on the weight coefficients. The normalized spectrum signal is then processed using the multi-resolution wavelet packet decomposition method. Subband energy distribution extraction was performed, with the db8 wavelet selected as the mother wavelet and a decomposition layer of 6 layers. The top 15% of subbands with the highest energy proportions were selected for energy diffusion trend feature extraction. The gradient field of the time-frequency two-dimensional energy distribution was calculated using the energy gradient change method. The energy gradient vectors in the frequency and time directions were calculated using the five-point difference method. The obtained gradient vectors were processed by nonlinear amplitude squared processing to obtain the vibration energy transfer intensity distribution matrix. The spatial-frequency joint weighting method was further used to enhance the diffusion characteristics of the distribution matrix, with a diffusion weight coefficient of 0.7. The enhanced vibration energy transfer intensity diffusion data was used in the subsequent gear meshing force off-center load limit analysis step.

[0089] In another embodiment, when analyzing the intensity diffusion of vibration energy transfer based on vibration noise frequency domain data, the transfer function method is first used to analyze the transmission characteristics of vibration energy in the structure. By measuring the vibration response at different locations on the gearbox housing and combining it with the acquired frequency domain data, a frequency response function matrix is ​​established between each measuring point. The frequency response function measurement adopts a single-point excitation and multi-point response method. The excitation point is set at the connection between the gearbox housing and the bearing seat, and the response points are distributed at 15 key locations on the housing surface. The consistency index of the frequency response function is controlled above 0.95. The measurement frequency range is 20Hz to 10000Hz, and the frequency resolution is set to 1.25Hz. Then, the acoustic transmission path analysis (TPA) technique is used to construct a vibration transmission path model. By setting the impedance matrix of the housing structure, the impedance matrix dimension is 15×15, and the impedance value range is 10^5 to 10^7. The contribution of each transmission path was calculated using a matrix inversion algorithm, with the contribution expressed in decibels. After normalization, the values ​​ranged from 0 to 10. Next, a frequency domain decomposition method was applied to extract the energy distribution of each vibration source. The frequency domain decomposition used a singular value decomposition algorithm, with the singular value threshold set at 85% of the original signal energy and a decomposition order of 12. This yielded the spatial distribution of the key vibration source frequency components. The vibration source intensity was expressed as acceleration power spectral density, with units of (g^2) / Hz. The vibration source frequencies were concentrated at gear meshing frequencies of 456Hz, 912Hz, and 1368Hz, and the bearing characteristic frequency of 87Hz. At 152Hz and 278Hz, the acoustic and vibration energy diffusion path was then simulated. The structural transmission characteristics were discretized using the finite volume method with a mesh size of 15mm and a total of 235,478 meshes. The calculation time step was 0.00005 seconds and the physical time was 0.5 seconds. The distribution law of energy in space and frequency was obtained by solving the energy diffusion equation, and finally the vibration energy transmission intensity diffusion data was formed. This data contains an energy distribution matrix of 15 measurement points and 72 frequency bands. The energy transmission coefficient matrix has a dimension of 15×15×72 and records the energy attenuation rate on each transmission path. The energy attenuation rate ranges from 0.05 to 0.65 dB / mm.

[0090] Step S22: Calculate the gear meshing force eccentric load limit based on the vibration energy transfer intensity diffusion data to obtain the gear meshing force eccentric load limit data;

[0091] In this embodiment of the invention, the gear meshing force off-center load limit calculation is performed based on the vibration energy transfer intensity diffusion data obtained in step S21. First, the diffusion data is processed by order extraction. The order synchronous averaging method is used to synchronously extract the characteristic orders related to the fundamental frequency, gear meshing frequency and its harmonics. The reference rotational speed signal is acquired in real time by a photoelectric encoder with a resolution of 4096 pulses per revolution. The order signal is normalized by envelope demodulation technology with a normalization range of 0 to 1. Then, multi-point contact fatigue stress calculation is performed. Based on the gear material properties, the Young's stress of 16MnCr5 steel is input into the module. With a modulus of 206 GPa and a Poisson's ratio of 0.3, the gear contact area was discretized into a mesh with a meshing step size of 0.2 mm. The Gaussian integral method was used to integrate the stress at the contact point, and the average principal stress method was used to perform equivalent conversion of the contact stress at different meshing positions. Then, the load non-uniformity coefficient of the off-center load section was calculated using a piecewise linear fitting method. After determining the off-center load limit point, the peak stress of the limit section was amplified, and the off-center load limit curve was continuously corrected using a cubic interpolation method. Finally, the off-center load limit data of gear meshing was formed for subsequent tooth surface pitting simulation steps.

[0092] In another embodiment, when calculating the gear meshing load limit based on vibration energy transfer intensity diffusion data, the energy distribution at the gear meshing frequency is first extracted from the diffusion data. The gear meshing frequency is 456Hz at a motor speed of 2500rpm and its third harmonic 1368Hz, with extracted energy values ​​of 0.42(g^2) / Hz and 0.28(g^2) / Hz, respectively. Then, the sideband distribution of the meshing frequency is determined using cepstral analysis. The sideband spacing is 41.7Hz, the sideband amplitude ratio is 1.85, and the sideband symmetry deviation is 15%. Next, a gear system dynamic model is established. In the model, the gear parameters are: module 6mm, number of teeth 35 / 42, pressure angle 20 degrees, helix angle 15 degrees, tooth width 60mm, gear material 20CrMnTi, hardness 58HRC, and elastic modulus 210GPa. The meshing stiffness under ideal meshing conditions is calculated to be 18.5×10^7 through the model. The meshing damping ratio was 0.06. Then, based on the vibration energy distribution, the actual meshing stiffness distribution of the gear was calculated. Using a segmented analysis method, the tooth width direction was divided into 12 equal segments, each 5 mm wide. The actual meshing stiffness value of each segment was calculated, ranging from 12.8 × 10^7 to 19.2 × 10^7 N / m, with a maximum stiffness deviation of 31%. The stiffness variation curve exhibited a non-linear distribution. Next, LTCA (Load Track Contact Analysis) technology was used to analyze the tooth surface load distribution. A total of 60 calculation points were set along the tooth width direction on the meshing line, with a spacing of 1 mm. The contact stress value at each calculation point ranged from 850 MPa to 1680 MPa, and the load distribution non-uniformity coefficient was 1.95, significantly higher than the normal value of 1.35. The load was concentrated within a 15 mm range on the left side of the tooth width, with insufficient contact on the right side. Then, combining the phase information from the vibration energy transmission intensity diffusion data, the actual meshing line inclination angle of the gear was calculated. The least squares method was used to fit the meshing... The meshing line shape was fitted with an accuracy of 0.01mm, resulting in a meshing line inclination angle of 0.082 degrees, deviating from the design value by 0.025 degrees. Next, the meshing force off-center load limit was calculated. Through comprehensive calculations of load distribution, meshing line inclination angle, and vibration energy, a weighted average method was used to determine the off-center load coefficient. The weight allocation was 55% for load distribution, 30% for meshing line inclination, and 15% for vibration energy. The final determined gear meshing force off-center load coefficient was 1.92, exceeding the safety threshold of 1.65, reaching the limit state. The generated gear meshing force off-center load limit data includes the off-center load coefficient, stress distribution, meshing line shape, and dangerous area distribution information for each measuring point.

[0093] Step S23: Simulate the cyclic ductility of tooth surface pitting based on the gear meshing force eccentric load limit data to obtain the cyclic ductility data of tooth surface pitting.

[0094] In this embodiment of the invention, the gear meshing force eccentric load limit data obtained in step S22 is used to simulate the cyclic ductility of pitting on the gear tooth surface. First, the gear tooth meshing line is discretized based on the eccentric load limit data. A meshing strategy with a step size of 0.25 mm in the tooth width direction and 0.5 mm in the tooth length direction is adopted. Contact fatigue damage accumulation is calculated for each mesh element. The Palmgren-Miner linear damage superposition theory is used to calculate the cumulative number of fatigue cycles. The fatigue limit at the SN inflection point of the material SN curve is set to 1800 MPa, and the fatigue strength coefficient is selected as 12. With a fatigue index of -0.12 and a pressure of 0.00 MPa, pitting initiation was determined for grid points with cumulative fatigue damage greater than 1. Crack propagation simulation was performed on the initiation area using the Paris crack propagation model. The crack growth step size was 0.01 mm, and the iteration cycle was 1000 cycles. The crack propagation direction was determined by the maximum tangential stress criterion. The stress intensity factor at the crack tip was calculated in real time using the finite element method. The crack length was updated based on the crack propagation rate curve. The iteration was repeated until the crack depth reached 30% of the tooth surface thickness, and finally, the tooth surface pitting cycle propagation data was generated.

[0095] In another embodiment, when simulating the cyclic ductility of pitting on the gear surface using the off-center load limit data of gear meshing, the contact fatigue characteristic parameters of the gear material 20CrMnTi were first determined. The contact fatigue limit was 1450 MPa, the slope of the SN curve was -0.075, and the fatigue life distribution coefficient was 1.6. Then, the contact stress distribution on the gear surface was extracted based on the off-center load limit data. The maximum contact stress point was located 7 mm from the tooth end on the left side of the tooth width, with a maximum contact stress value of 1680 MPa, exceeding the material fatigue limit value by 15.9%. Next, a pitting initiation model was established, and the Lundberg-Palmgren theory was used to calculate the subsurface... The maximum shear stress, which occurs subsurfacely at 0.27 mm below the surface, is 665 MPa. The fatigue damage accumulation theory is then used to predict the pitting corrosion initiation time. Cyclic loads are accumulated in increments of 1 million cycles, and the damage amount per cycle is calculated using Miner's linear accumulation rule. The damage threshold is set to 1.0, resulting in an initial pitting corrosion initiation cycle count of 3.2 million cycles. A pitting corrosion propagation model is then constructed, with an initial pit size of 0.15 mm × 0.12 mm and a depth of 0.08 mm. The Paris propagation law is used to describe the pitting corrosion propagation process, with the Paris equation parameter C set to 5. The stress intensity factor (ΔK) was set to 0.8 × 10⁻⁸, with the exponent m set to 3.25. A semi-elliptical crack model was used to calculate the stress intensity factor range ΔK, with an initial stress intensity factor of 12.8 MPa·m⁰.⁵. The expansion of pitting size after 1 million cycles was calculated iteratively, with an expansion rate set from 0.06 mm / 1 million cycles to 0.15 mm / 1 million cycles. The expansion rate varied non-linearly with stress level. The influence of tooth surface lubrication on the expansion rate was also considered, with a lubrication film thickness of 1.2 μm and a viscosity of 80 cSt. Under lubrication conditions, the expansion rate reduction coefficient was 0.85. Then, a multi-pitting interaction analysis was performed, considering the distance between the centers of two pitting points. When the thickness is less than 4 mm, the stress field superposition effect is considered, and the stress field superposition coefficient is 1.25. The distribution pattern of multiple pitting corrosions on the tooth surface is simulated. The number of pitting corrosions increases from the initial 3 to 15, and the pitting corrosion coverage area increases from 0.8% of the tooth surface area to 5.2%. Finally, pitting corrosion cycle extension data is generated, including pitting corrosion location coordinates, size, depth, expansion rate and predicted future expansion trend. The pitting corrosion expansion curve shows a piecewise linear characteristic within 10 million cycles. The expansion rate is 0.06 mm / 1 million cycles from 0 to 5 million cycles, and the expansion rate increases to 0.12 mm / 1 million cycles from 5 million to 10 million cycles.

[0096] Step S24: Identify the bearing clearance expansion increment based on the vibration energy transfer intensity diffusion data, thereby obtaining the bearing clearance expansion increment data.

[0097] In this embodiment of the invention, the bearing clearance expansion increment is identified based on the vibration energy transfer intensity diffusion data obtained in step S21. First, the diffusion data is separated into bearing characteristic frequency segments. Based on the target electric drive assembly bearing model, the inner ring passing frequency, outer ring passing frequency, rolling element passing frequency, and cage passing frequency are selected as characteristic order segments. Short-time Fourier transform is used to extract the instantaneous energy density of each order segment. The window length is set to 512 points, and the overlap length is 256 points. The energy peak values ​​of each segment are time-series normalized. The sliding interval peak comparison method is used to track peak drift. The interval length was set to 100ms, the step size to 20ms, and then the bearing stiffness variation was simulated. Hertzian contact theory was used to perform interval inversion on the contact stress between the rolling element and the raceway. The radius of the rolling element was set to 5mm, and the radius of curvature of the raceway was set to 20mm. The difference between the actual observed frequency amplitude and the theoretical contact stiffness variation interval was compared. The radial clearance variation was derived by using the least squares fitting method. The clearance variation at multiple working conditions was fitted with quadratic spline to calculate the clearance increment variation curve under different speeds and loads. The slope of the linear segment of the clearance variation trend was extracted, and finally the bearing clearance expansion increment data was output.

[0098] In another embodiment, when identifying the bearing clearance increase based on vibration energy transfer intensity diffusion data, the vibration spectrum of the measuring point at the bearing location is first extracted from the diffusion data. The bearing is a 7208C type angular contact ball bearing with 15 rolling elements, a rolling element diameter of 12.5 mm, a pitch circle diameter of 65 mm, a contact angle of 15 degrees, and an initial radial clearance of 0.022 mm. Then, envelope analysis technology is applied to process the vibration signal. The modulation signal is extracted using the Hilbert transform method, the sampling frequency is 25600 Hz, the analysis frequency band is 1000 Hz to 5000 Hz, the bandpass filter order is 8, and the filter type is Butterworth. Subsequently, envelope spectrum analysis is performed, and the envelope spectrum resolution is... The frequency was set to 0.5Hz, and the Hanning window was used as the window function to extract the bearing characteristic frequencies. The rolling element passing through the outer ring (BPFO) frequency was 87.5Hz, the rolling element passing through the inner ring (BPFI) frequency was 128.6Hz, the rolling element rotation frequency (BSF) was 57.2Hz, and the cage frequency (FTF) was 5.8Hz. The characteristic frequencies and their harmonic amplitudes were analyzed. The characteristic frequency harmonic amplitude ratio decreased from 0.82 to 0.58, and the sideband energy proportion increased from 8% to 22%. Then, the bearing fault type and severity were determined by spectrum analysis. BPFO and its harmonics were significantly enhanced, with amplitudes 3.7 times higher than the standard state. The BPFO sideband was obvious, and the sideband interval was 41 Hz of the shaft frequency. At 7Hz, indicating wear on the outer ring, a model was established to model the relationship between bearing clearance and vibration characteristics. The model input parameters included the BPFO amplitude change rate, sideband strength, high-frequency modulation index, and impact characteristic factor. The model was trained using 65 sets of historical data, achieving a validation accuracy of 92%. The clearance increment was calculated based on the model. The initial bearing clearance was 0.022mm, and after 500 hours of operation, the measured clearance was 0.047mm, an increase of 0.025mm, representing an increment of 113.6%. Further analysis of the spatial distribution of the clearance increase was conducted by measuring the vibration response at different angles. The bearing circumference was divided into 24 equally spaced regions, each with an angle of 15 degrees, and the vibration intensity in each region was measured. The study found that the clearance increase was most significant in the 45° to 135° range, with a clearance increment of 0.033 mm, 32% higher than the average. Combined with the bearing load direction analysis, this region was aligned with the main load direction. Further analysis of the causes of clearance expansion revealed that the average abrasive particle size was 5 μm, with Fe and Cr as the main components. Temperature monitoring data showed that the highest operating temperature reached 95°C, and thermal expansion contributed 0.008 mm to the clearance change. The final bearing clearance expansion increment data included an absolute clearance value of 0.047 mm, a clearance increment of 0.025 mm, a clearance spatial distribution non-uniformity coefficient of 1.32, and a clearance expansion rate of 0.5 μm / 10 hours.

[0099] Step S22 includes the following steps:

[0100] Step S221: Obtain the gear meshing vibration frequency and basic characteristics of the gear material under the current working condition; perform diffusion dispersion superposition frequency doubling quantization on the vibration energy transfer intensity diffusion data to obtain energy dispersion superposition frequency doubling data;

[0101] Step S222: Analyze the energy density evolution ratio of the gear meshing vibration frequency based on the energy dispersion superposition overtone data, and generate the energy density evolution ratio of the resonance peak;

[0102] Step S223: Based on the energy density evolution ratio of the resonance peak, perform incremental simulation of the progressive gradient of fatigue deformation of the gear shaft to estimate the basic properties of the gear material, thereby obtaining incremental data of the progressive gradient of fatigue deformation.

[0103] Step S224: Dynamically simulate and calculate the parallelism / perpendicularity deviation data of gear meshing using fatigue deformation progressive gradient incremental data;

[0104] Step S225: Calculate the gear meshing force off-center load limit based on the parallelism / perpendicularity deviation data to obtain the gear meshing force off-center load limit data.

[0105] In this embodiment of the invention, the real-time speed signal acquired during the operation of the electric drive assembly test bench is processed by a multi-pulse encoder. The encoder pulse count is set to 4096 pulses per revolution. The current operating speed is calculated through the relationship between speed and time series. The gear meshing vibration frequency is then accurately calculated by combining the gear tooth count. The input gear tooth count is 36 teeth on the input shaft gear and 72 teeth on the output shaft gear. The calculated vibration frequency covers the fundamental frequency, gear meshing frequency, and its 2nd to 10th harmonic range. The basic properties of the gear material include Young's modulus of 206 GPa, Poisson's ratio of 0.3, and yield strength of 900 MPa. The density is 7.85 g / cm³. Subsequently, the vibration energy transfer intensity diffusion data obtained in step S21 is subjected to diffusion dispersion superposition and frequency doubling quantization processing. First, the original diffusion data is divided into meshing frequency region, frequency doubling region and high harmonic region according to frequency order interval. The interval division is based on the calculation results of rotational speed and gear meshing characteristic frequency. For each segment, the vibration energy density is integrally calculated. The trapezoidal integral method is used to accumulate the energy of each frequency doubling region. The energy density result of the frequency doubling region is subjected to quadratic polynomial fitting processing. The fitting coefficient is solved by the least squares method. The fitted energy dispersion superposition frequency doubling data is obtained. Based on the energy dispersion superposition frequency data obtained in step S221, the energy density evolution ratio of the resonant peaks of gear meshing vibration frequency is analyzed. First, an automatic extreme point extraction algorithm is used to identify local peaks in the energy dispersion superposition frequency data. Peak extraction adopts the sliding window extreme value method, with a window width of 5Hz and a step size of 1Hz. The peak energy density of each extracted resonant peak segment is calculated using the contour area integration method. The peak energy density changes under different working conditions are standardized, and the maximum value normalization method is used to normalize the energy peaks under different working conditions to the interval 0 to 1. The logarithmic rate of change of the normalized peak energy density is calculated according to the time series, with a calculation interval length of 200ms and a step size of 50ms. The energy change rate within the time series is processed by moving average, with a sliding window width of 5 sampling points. Finally, the ratio of the normalized energy change rate to the original peak energy is analyzed using the energy evolution ratio formula to generate the energy density evolution ratio of the resonant peaks.

[0106] Using the obtained resonance peak energy density evolution ratio data, a progressive gradient incremental simulation estimation of gear shaft fatigue deformation was performed to assess the fundamental properties of the gear material. First, based on the gear shaft's geometric dimensions, the shaft diameter was determined to be 35 mm and the shaft length 210 mm. The gear shaft was then meshed using the finite element method (FEM), with one-dimensional beam elements selected and an element length set to 5 mm. The applied load was converted from the energy density in the resonance peak region to the equivalent torque. The torque conversion employed a linear mapping method from energy density to torque, with the mapping coefficient determined to be 8 based on previous dynamic stiffness calibration tests. The equivalent torque after mapping was subjected to time-history loading at Nm / dB, with the loading period consistent with the sampling period. Subsequently, the radial deformation of the gear shaft was simulated in segments over time. The Euler-Borneau beam theory was used for numerical integration of the deflection curve with a step size of 0.1 mm. Tangential stress was calculated for the deformation increment at different time steps. The material elastic modulus was used to convert the stress and strain into a linear relationship. The fatigue deformation was spatially integrated in layers based on the deformation gradient distribution curve. The length of the integration segment was fixed at 2% of the shaft length. Finally, the asymptotic gradient increment data of fatigue deformation was obtained. The obtained fatigue deformation progressive gradient incremental data were used to dynamically simulate and calculate the parallelism and perpendicularity deviations of gear meshing. First, the radial displacement of the nodes at different cross-sections of the gear shaft was integrated based on the fatigue gradient data. The integration method was trapezoidal integration with a segment step size of 0.1 mm. The straightness of the axis was fitted using the least squares method based on the displacement change curve. The fitting residual is the parallelism deviation. The perpendicularity deviation was solved by derivative based on the displacement change rate of the nodes at different cross-sections in the vertical direction. The length of the derivative interval was set to 1 mm. The calculated displacement change rate was smoothed using the midpoint method with a smoothing window width of 3 sampling points. Statistical analysis was performed on the parallelism and perpendicularity deviation curves. The maximum deviation peak value was selected as the final deviation parameter output. The deviation was quantified in micrometers. Based on the obtained parallelism and perpendicularity deviation data, the gear meshing force eccentric load limit calculation is performed. First, the deviation data is linearly fitted, and the slope of the deviation change is obtained using the least squares method. The contact area of ​​the meshing line is discretized into a grid with a grid size of 0.2mm × 0.2mm. The displacement offset method is used to reposition the meshing line grid node displacements. Based on the repositioned node coordinates, the contact stress is recalculated using Hertzian contact stress theory. The input parameters include a gear module of 2.5, a pressure angle of 20 degrees, and a tooth width of 25mm. The peak contact stress values ​​of different nodes within the meshing line are sorted, and the section where the peak stress is located is selected as the stress concentration area. The ratio of the maximum stress in this section to the overall average contact stress is calculated to form the force eccentric load coefficient. The force eccentric load coefficient is compared with the material yield limit to determine the eccentric load limit point. Finally, the stress of each node in the eccentric load limit section is integrated, and the total stress load of the section is calculated using the Gaussian integration method, ultimately obtaining the gear meshing force eccentric load limit data.

[0107] Step S225 includes the following steps:

[0108] Dynamic integration of meshing line offset is performed based on parallelism / perpendicularity deviation data to obtain dynamic meshing line offset data;

[0109] The axial force and radial force imbalance index is calculated based on the dynamic offset data of the meshing line to obtain the axial force / radial force imbalance index.

[0110] The imbalance index of axial force / radial force was analyzed by proportional variation of the imbalance index at adjacent time points to obtain the proportional variation data of the imbalance index at adjacent time points;

[0111] Based on the proportional change data of the imbalance index and the dynamic offset data of the meshing line, the rotational force misalignment deviation of the gear helix angle is quantified to obtain the rotational force misalignment deviation data.

[0112] Based on the rotational force misalignment deviation data, the proportional change data of the imbalance index, and the dynamic offset data of the meshing line, the gear meshing force off-center load limit is calculated, thereby obtaining the gear meshing force off-center load limit data.

[0113] In this embodiment of the invention, specifically in the embodiment of dynamic integration of meshing line offset based on parallelism and perpendicularity deviation data, the parallelism deviation data and perpendicularity deviation data obtained in step S224 are first mapped according to discrete segments of the meshing line. The segment length is set to 0.2 mm. The displacement change of each discrete point of the meshing line during the meshing cycle is gradually integrated using the Euler forward integration method. The integration step size is 0.05 mm, and the integration interval covers the entire gear meshing process, from the meshing start point to the meshing end point. During the integration process, the five-point weighted moving average method is used to smooth the noise fluctuations. The smoothing coefficient is set to 0.8. The integration result is the dynamic offset data of the meshing line. This data reflects the cumulative displacement change trend at different meshing positions, providing a basic input for subsequent imbalance force analysis. In a specific embodiment of calculating the axial and radial force imbalance index based on the dynamic offset data of the meshing line, the tooth width is first determined to be 30mm and the tooth length resolution is 0.2mm based on the gear geometry parameters. The dynamic offset data of the meshing line obtained in step S225 is used to reposition the contact position of the gear meshing point. The static mechanical equilibrium equation is then used to calculate the axial and radial components of each discrete point. Input parameters include a unit contact point normal stiffness set to 1.2 × 10⁻⁶. 8With a contact angle of 20 degrees and a local axial and radial forces at each meshing position, the weighted summation of the local axial and radial forces is performed. The total axial and radial forces in the entire meshing section are normalized using a weighted average method. Then, the imbalance index of axial and radial forces is calculated using the imbalance force coefficient method. The imbalance index is defined as the ratio of the local force offset to the average force value of the entire section. The force offset of a section is calculated by the difference between the average values ​​of a single section and the average values ​​of the entire section. Finally, the axial and radial force imbalance index is output.

[0114] In a specific embodiment of the geometrical change analysis of the imbalance index between adjacent time points for axial and radial force imbalance index, the imbalance index sequence is first arranged in the order of time sampling points, with a sampling time interval set to 5ms. The geometrical change rate of the imbalance index between two adjacent time points is calculated, using the ratio of the current imbalance index to the imbalance index at the previous time point. The geometrical change rate data of the entire time series is smoothed and filtered using a weighted moving average method, with a window length set to 5 sampling points. The smoothing weight distribution is set according to a Gaussian weight function, with the weight peak located at the center of the window. The filtered data is then subjected to a second difference processing to identify the acceleration characteristics of the imbalance change. The second difference step size is set to 1 sampling point, finally obtaining the geometrical change data of the imbalance index between adjacent time points. In a specific embodiment of quantifying the helical force misalignment deviation of gear helix angle based on the proportional change data of the imbalance index and the dynamic offset data of the meshing line, the initial helix angle is first set to 25 degrees according to the actual design parameters of the gear helix angle. The dynamic offset data of the meshing line is spatially remapped. The tooth direction component, radial component and axial component conversion method in the three-dimensional coordinate system is used to map the dynamic offset data to the axial direction of the gear. The periodic fluctuation amplitude of the proportional change data of the imbalance index is analyzed. The frequency component of the proportional change sequence is extracted by fast Fourier transform. The frequency component that matches the gear rotation frequency and its harmonics is selected for inverse transformation reconstruction to obtain the proportional change fluctuation envelope curve. The tooth direction component offset and the envelope curve are combined for dot product superposition to generate the helical force misalignment deviation time series. The maximum deviation of the series is extracted and the largest value is selected as the gear helix angle helical force misalignment deviation data. In a specific embodiment of calculating the gear meshing force deviation limit based on helical force misalignment deviation data, imbalance index proportional change data, and meshing line dynamic offset data, the three sets of input data are first processed for interval synchronization with a time alignment accuracy of 1ms. Then, the meshing line dynamic offset data are divided into segments with a segment length set to 1 / 10 of the tooth width. The dynamic force fluctuation of each segment is integrated using the imbalance index proportional change data, employing a trapezoidal integration method with an integration step size of 0.1mm. Finally, the helical force misalignment deviation data is mapped to the tooth direction distribution using a linear mapping method. Interpolation is used to establish a one-to-one correspondence between the misalignment of different sections and the section length. Hertzian contact stress is back-calculated for the comprehensive force offset in each section using a dual-contact Hertzian analytical formula. Input parameters include the material elastic modulus of 206 GPa, Poisson's ratio of 0.3, and a contact radius of 5 mm. The local contact stress obtained by back-calculation is compared with the material yield limit, which is set to 900 MPa. The ultimate load is integrated for sections exceeding the yield limit, and the stress is integrated in the local area using the Gaussian integration method to finally obtain the gear meshing force off-center load limit data.

[0115] Step S23 includes the following steps:

[0116] Step S231: Analyze the load distribution along the tooth width skew of the gear meshing force limit data to obtain the load tooth width skew distribution data;

[0117] Step S232: Based on the gear meshing force eccentric load limit data, the load tooth width skew distribution data is subjected to tooth width edge contact stress geometric growth quantification to obtain tooth width edge stress geometric growth data.

[0118] Step S233: Iteratively couple the stress crack propagation degree to the geometric growth data of tooth width edge stress to obtain iterative stress crack propagation degree data;

[0119] Step S234: Based on the geometric growth data of tooth width edge stress and the iterative stress crack propagation data, simulate the cyclic extension of tooth surface pitting to obtain the cyclic extension data of tooth surface pitting.

[0120] In this embodiment of the invention, specifically in the analysis of the load skew distribution along the tooth width of the gear meshing force eccentricity limit data, the gear meshing force eccentricity limit data obtained in step S225 is first spatially discretized. The length of the discretized segment is set to 1 / 20 of the tooth width. The force value is integrated at each segment point in the tooth width direction. The trapezoidal integral method is used to accumulate the local contact stress in the contact line. Then, the accumulated load is fitted with a spatial slope. The fitting method is the least squares method. The segment load curve is expressed in the form of a linear function. The slope of the linear fitting is the gradient of the load skew distribution along the tooth width direction. The positive and negative directions of the slope are defined according to the position of the load offset centroid. At the same time, the load density of each segment is normalized. The normalization method is the ratio of the segment load density to the total load density of the tooth width. Finally, the load tooth width skew distribution data is formed. In a specific embodiment of quantifying the geometric growth of edge contact stress in gear tooth width based on the gear meshing force eccentricity limit data, the load tooth width eccentricity distribution data obtained in step S231 is first synchronized with the force eccentricity limit data in sections, with the time synchronization accuracy set to 1ms. The load density data of the edge sections on both sides of the tooth width are locally magnified and multiplied. The multiplication factor is determined by the scaling ratio of the actual meshing contact surface width, with a value range of 1.2 to 1.8. The stress gradient of the edge section is spatially differentiated according to different working conditions. The differentiation method adopts the first-order central difference method. The stress change rate obtained after differentiation is geometrically multiplied and fitted. The fitting function is a quadratic polynomial form, and the fitting parameters are calculated by the least squares method. The length of the fitting section is set to 5% of the tooth width. Finally, the geometric growth data of edge stress in the tooth width is output.

[0121] In a specific embodiment of iteratively coupling stress crack propagation degree with the geometric growth data of tooth width edge stress, the crack initiation point is first identified in the geometric growth data of tooth width edge stress obtained in step S232. The crack initiation point is selected with the peak position of stress gradient as the center, and an area with a length of 0.5 mm is selected for the initial crack length setting. The initial crack length is 0.05 mm. The stress intensity factor at the crack tip is calculated using the principle of linear elastic fracture mechanics. The stress intensity factor calculation adopts the Y-factor correction method. The crack propagation simulation process adopts the incremental iteration method, and the step size of each iteration is set to 0.01 mm. The stress intensity factor within each iteration step is compared with the material threshold fracture toughness. The material threshold fracture toughness is taken as 30 MPa√m. The crack segment exceeding the threshold value is extended in length. The crack propagation direction is adjusted according to the direction of maximum shear stress, and the angle adjustment step is 2 degrees. The crack propagation iteration process continues until the crack length reaches 1 mm. Finally, the iterative stress crack propagation degree data is output. In a specific embodiment of simulating the cyclic ductility of pitting on tooth surfaces based on the geometric growth data of edge stress and the iterative stress crack propagation data, the two sets of data obtained in steps S232 and S233 are first coupled in sections, with the length of the coupled section set to 0.2 mm. The stress growth rate and crack propagation rate within each section are matched point-to-point using a linear weighted superposition method. The weighting coefficients are dynamically adjusted according to the ratio of stress gradient to crack propagation rate, with initial weighting coefficients set to 0.7 and 0.3. Fatigue damage accumulation analysis is then performed on the coupled section data using a linear cumulative damage method. The fatigue damage factor for each section is integrated over time, with an integration step size set to 1 ms. The integration results are normalized to the number of cycles, with a baseline number of cycles set to 10. 6 Next, the pitting initiation zone was determined for the section where the accumulated fatigue damage exceeded 1. Then, the radial depth progression simulation of the pitting initiation zone was performed. The hyperbola growth fitting method was used to predict the trend of pitting depth change. The fitting parameters were obtained by the least squares method. Finally, the pitting cycle extension data of the tooth surface was output.

[0122] Step S24 includes the following steps:

[0123] Step S241: Perform high-frequency impact vector analysis on the vibration energy transfer intensity diffusion data between the bearings to obtain the high-frequency impact vector;

[0124] Step S242: Calculate the radial load distribution between bearings based on the high-frequency vibration impact vector to obtain bearing radial load distribution data;

[0125] Step S243: Based on the high-frequency vibration impact vector, perform Hertzian contact stress iterative simulation and deduction on the radial load distribution data of the bearing to generate Hertzian contact stress iterative data;

[0126] Step S244: Solve the nonlinear variation of the bearing contact angle based on the Hertzian contact stress iteration data and the bearing radial load distribution data to obtain the bearing contact angle variation data;

[0127] Step S245: Identify the bearing clearance increase based on the bearing contact angle change data, thereby obtaining the bearing clearance increase data.

[0128] In this embodiment of the invention, specifically in the analysis of high-frequency impact vector of vibration transmission between bearings using vibration energy transfer intensity diffusion data, the vibration energy transfer intensity diffusion data obtained in step S21 is first subjected to high-frequency bandpass filtering. A fifth-order Butterworth filter is used, with the passband frequency range set to 8kHz to 30kHz and the filtering sampling frequency set to 100kHz. A short-time Fourier transform is performed on the filtered signal, with the time window length set to 2ms and the window function type selected as a Hamming window, to obtain a high-frequency time-frequency distribution matrix. Peak detection is performed on the amplitude changes of frequency components in each time period, and the amplitude fluctuation rate is calculated using the second-order difference method. Transient change points greater than the set threshold of 0.05g are marked. Subsequently, these points are vectorized, with the vector components being triaxial acceleration vectors. The vector length is normalized using the square root sum method, and finally, the high-frequency impact vector of vibration is obtained. In a specific embodiment of calculating the radial load distribution between bearings based on high-frequency vibration impact vectors, firstly, according to the bearing layout dimensions, the bearing center distance is set to 120mm, and the radial distribution section is divided into 10 equal-length sections. The high-frequency vibration impact vectors obtained in step S241 are mapped to these sections. A spatial relationship matrix is ​​established based on the spatiotemporal propagation direction of each vector and the center point of each bearing section. The least mean square error method is used to fit the reverse trajectory of the vector propagation path to obtain the impact energy distribution ratio of each section. The impact energy ratio is normalized, and the normalization coefficient is taken as the sum of the energy of each section. Then, combined with the bearing preload setting, the preload is set to 500N. The radial load of each section is calculated by multiplying the energy distribution ratio and the preload, and finally the bearing radial load distribution data is obtained. In a specific embodiment of the Hertzian contact stress iterative simulation and deduction based on the bearing radial load distribution data using the high-frequency vibration impact vector, the bearing radial load distribution data obtained in step S242 and the high-frequency vibration impact vector are first processed for segmental synchronization, with the synchronization accuracy set to 0.5ms. The initial contact stress of each radial load segment is calculated using Hertzian contact mechanics theory, with the material elastic modulus set to 210GPa, Poisson's ratio to 0.3, and the initial contact angle to 15 degrees. The contact radius and contact pressure are updated cyclically using an iterative correction method, with the iteration stopping condition being that the contact pressure difference between two iterations is less than 0.01MPa. The maximum number of iterations is set to 50. The contact area and contact depth are dynamically updated at each step during the iteration process. The contact depth is calculated using a displacement control method, with the displacement step size set to 1μm. Finally, the Hertzian contact stress iterative data are obtained.In a specific embodiment of solving the nonlinear variation of bearing contact angle based on Hertzian contact stress iteration data and bearing radial load distribution data, the two sets of data obtained in steps S243 and S242 are first aligned into segments. The Hertzian contact stress data corresponding to each radial load segment is fitted with a nonlinear relationship. A quadratic polynomial regression method is used to establish a nonlinear functional relationship between the contact angle, contact stress, and radial load. The regression equation parameters are iteratively corrected using the least squares method. The regression residual threshold is set to 0.02 degrees. The contact angle variation calculated for all segments is interpolated and smoothed. The interpolation method is spline interpolation with a smoothing coefficient of 0.7. Finally, the bearing contact angle variation data is output. In a specific embodiment of identifying bearing clearance expansion increment based on bearing contact angle change data, the bearing contact angle change data obtained in step S244 is first resampled in time series with a sampling interval of 2ms. The first derivative of the contact angle change rate is calculated using the five-point central difference method. The contact angle change rate is linearly coupled with the elastic deformation modulus of the bearing material, with the material elastic modulus set to 210GPa. The clearance change rate is obtained by multiplying the contact angle change rate by the deformation modulus. The clearance change rate is then integrated over time with an integration step size of 1ms. The integration results are accumulated in segments with a segment length of 10ms. Finally, the accumulated segment results are compared with the initial bearing clearance reference value, which is 0.02mm. The difference results are then nonlinearly amplified with an amplification factor set to 1.5 based on empirical experiments. Finally, the bearing clearance expansion increment data is obtained.

[0129] Step S3 includes the following steps:

[0130] Step S31: Perform logical learning on the bearing clearance increase increment data to generate bearing clearance increase increment learning data;

[0131] Step S32: Based on the learning data of bearing clearance expansion increment, perform bearing structure optimization processing to obtain bearing structure optimization data;

[0132] Step S33: Perform tooth-direction bulging modification analysis based on the tooth surface pitting cycle extension data to obtain tooth-direction bulging modification data;

[0133] Step S34: Based on the bearing structure optimization data and tooth profile drum shape modification data, conduct vibration-resistant strengthening design to mitigate the impact of vibration and generate structural vibration-resistant strengthening design data;

[0134] Step S35: Send the structural vibration strengthening design data to the terminal to perform structural optimization for the impact of vibration and noise.

[0135] As an example of the present invention, reference is made to Figure 3 As shown, step S3 in this example includes:

[0136] Step S31: Perform logical learning on the bearing clearance increase increment data to generate bearing clearance increase increment learning data;

[0137] In this embodiment of the invention, the bearing clearance expansion increment data obtained in step S245 is subjected to time series expansion processing with a time series step size of 1ms. The expanded data is then normalized by segmentation using the maximum-minimum value normalization method with a sampling segment length of 50ms. The normalized data sequence is then subjected to autoregressive moving average analysis with an order of 3. The moving average sequence is then subjected to first-order differencing to highlight trend change characteristics. Then, a logic rule mining algorithm based on Bayesian decision trees is used to divide the differencing sequence by feature thresholds. The threshold is set to a clearance increment rate threshold of 0.002mm / ms based on historical experimental data. The logic learning process uses a five-round incremental iteration method, with the feature weights adjusted by an increment of 0.1 steps in each iteration. The iteration terminates when the error converges to less than 0.001, ultimately generating bearing clearance expansion increment learning data.

[0138] Step S32: Based on the learning data of bearing clearance expansion increment, perform bearing structure optimization processing to obtain bearing structure optimization data;

[0139] In this embodiment of the invention, the bearing clearance expansion increment learning data obtained in step S31 is subjected to feature vector decomposition. The feature components include three parameters: clearance change rate, contact angle change, and radial load non-uniformity coefficient. The clearance change rate range is set to 0 to 0.05 mm / ms, the contact angle change range is set to 0 to 5 degrees, and the radial load non-uniformity coefficient ranges to 1 to 5. A multi-objective weighted method is used to construct the optimization objective function for the three features. The objective function is set to minimize the clearance change rate and maximize the contact angle uniformity. The weight coefficients are set to 0.6 and 0.4, respectively. The objective function is optimized by nonlinear constraints. The constraints include that the maximum contact stress does not exceed 1800 MPa and the contact length of the rolling element edge does not exceed 85% of the raceway width. The optimization algorithm adopts the Newton-Raphson iteration method with a step size of 0.05 mm. After each iteration, the bearing geometric parameters are updated, including the rolling element diameter, the raceway curvature radius, and the raceway depth. The optimization termination condition is set to the objective function change being less than 0.0001. Finally, the bearing structure optimization data is obtained.

[0140] Step S33: Perform tooth-direction bulging modification analysis based on the tooth surface pitting cycle extension data to obtain tooth-direction bulging modification data;

[0141] In this embodiment of the invention, the tooth surface pitting cyclic extension data obtained in step S234 is first discretized in the tooth width direction, with the spacing between discrete points set to 0.2 mm. The pitting extension depth of each discrete point is then subjected to two-dimensional smoothing filtering, with the filter kernel size set to 3×3. The filtered pitting extension depth data is then subjected to gradient direction projection processing, with the projection direction being a biaxial projection along the tooth length and tooth width directions, and the projection step size set to 0.1 mm. The projection result is then fitted using the least squares method, with the fitting order set to a quadratic polynomial. The curvature of the fitted curve is then used to calculate the tooth-direction drum-shaped modification deviation, with the deviation being the difference between the maximum pitting extension depth and the ideal tooth surface modification curvature. The calculation result serves as a preliminary reference for setting the drum-shaped modification amount. The preliminary drum-shaped modification amount is then corrected for fatigue life constraints, including a fatigue limit of not less than 1200 MPa and a contact line length of not less than 80% of the total tooth length. Finally, the tooth-direction drum-shaped modification data is output.

[0142] Step S34: Based on the bearing structure optimization data and tooth profile drum shape modification data, conduct vibration-resistant strengthening design to mitigate the impact of vibration and generate structural vibration-resistant strengthening design data;

[0143] In this embodiment of the invention, a specific embodiment of the vibration-resistant strengthening design based on bearing structure optimization data and tooth profile drum shape modification data is as follows: First, the bearing structure optimization data and tooth profile drum shape modification data obtained in steps S32 and S33 are subjected to multi-objective parameter comprehensive processing. The parameter inputs include the change in bearing stiffness characteristics, clearance control increment, drum shape modification deviation curvature and pitting extension length. A vibration response sensitivity matrix is ​​established using a nonlinear multi-objective optimization algorithm. The sensitivity matrix is ​​constructed using finite element modal sensitivity analysis. The element division uses four-node solid elements with a size of 1 mm. The sensitivity matrix is ​​decomposed into eigenvectors, and the first three eigenvectors are extracted as the main control vectors. The control vectors are then subjected to vibration response constraint optimization. The constraint objectives are to increase the first-order mode frequency by no less than 10 Hz and reduce the peak axial vibration velocity by no less than 20%. The optimization iteration uses a sequential quadratic programming method with an iteration step size of 0.1. Finally, the structural vibration-resistant strengthening design data is obtained.

[0144] Step S35: Send the structural vibration strengthening design data to the terminal to perform structural optimization for the impact of vibration and noise.

[0145] In this embodiment of the invention, in a specific embodiment of sending structural vibration-strengthening design data to the terminal to perform structural optimization for vibration and noise effects, the structural vibration-strengthening design data obtained in step S34 is first forwarded via a CAN bus. The CAN bus baud rate is set to 500kbps, and the data transmission protocol adopts the ISO-TP segmented transmission protocol. The data content is subjected to CRC16 cyclic redundancy check, and the check polynomial is set to 0x8005. The data that passes the check is mapped to the terminal receiving buffer. The starting position of the buffer address space is set to 0x2000H, and the data length is set to 512 bytes. The received data is written to the hardware control register in real time. The register address range is from 0x3000H to 0x3200H, and the writing period is controlled within 5ms. After the writing is completed, the terminal is triggered to execute the structural parameter update instruction. The instruction trigger signal level is high, and the level width is set to 10ms.

[0146] Step S32 includes the following steps:

[0147] Step S321: Measure the bearing radial stiffness strengthening ratio based on the bearing clearance expansion increment learning data to obtain the bearing radial stiffness strengthening ratio;

[0148] Step S322: Based on the learning data of bearing radial stiffness enhancement ratio and bearing clearance increase increment, couple the contact angle to increase by an equal amount to generate contact angle increase data;

[0149] Step S323: Perform adaptive design matching of the runout angle by increasing the bearing radial stiffness enhancement ratio and contact angle by the same amount to obtain the runout angle matching data;

[0150] Step S324: Based on the bearing radial stiffness enhancement ratio, contact angle increase data, and yaw angle matching data, the bearing structure is optimized to obtain bearing structure optimization data.

[0151] In this embodiment of the invention, specifically in the embodiment of measuring the radial stiffness enhancement ratio of a bearing based on the learning data of bearing clearance expansion increment, the learning data of bearing clearance expansion increment obtained in step S31 is first divided into numerical intervals with an interval step size of 0.002 mm and the number of interval increment segments is set to 50. Static stiffness analysis is performed on the load deformation response corresponding to each interval of clearance increment. The static stiffness analysis method adopts the finite element element stiffness matrix iterative solution. In the finite element model, the element type is selected as a three-dimensional solid element, the element size is set to 0.5 mm, and the material property parameters are set as Young's modulus of steel 210 GPa and Poisson's ratio 0.3. Five levels of loading conditions with equal intervals from 0 to 5000 N are applied to each interval with a loading step size of 1000 N. The radial deformation during loading is collected, and the corresponding stiffness value is calculated based on the load deformation slope. The obtained stiffness data is compared with the standard design stiffness curve. The reference standard stiffness curve is given by the ISO281 standard bearing stiffness characteristic curve. Finally, the radial stiffness enhancement ratio data of the bearing is obtained.

[0152] In a specific embodiment of coupling the bearing radial stiffness enhancement ratio and bearing clearance expansion increment learning data to increase the contact angle by an equal amount, firstly, a feature coupling matrix is ​​constructed between the bearing radial stiffness enhancement ratio data obtained in step S321 and the bearing clearance expansion increment learning data obtained in step S31. In the feature matrix, the row vectors are the clearance increment gradients, and the column vectors are the stiffness enhancement ratio gradients. The clearance increment gradient range is 0 to 0.1 mm, and the stiffness enhancement ratio gradient range is 1 to 2 times. The contact angle change is predicted for each data point in the feature matrix. The prediction process uses Hertzian contact theory and the nonlinear relationship of contact angle change for iterative calculation. The input parameters include the bearing rolling element radius of 7.5 mm and the channel curvature radius of 10 mm. During the iteration process, the initial contact angle value is updated according to each level of clearance increment and stiffness change. The initial contact angle is set to 15 degrees, the iteration step size is set to 0.1 degrees, and the iteration termination condition is that the contact angle change converges to less than 0.01 degrees. Finally, the contact angle increase data is obtained.

[0153] In a specific embodiment of adaptive design matching of the yaw angle using bearing radial stiffness enhancement ratio and contact angle increase data, the bearing radial stiffness enhancement ratio data and contact angle increase data obtained in steps S321 and S322 are first normalized. The normalization interval is set to 0 to 1, and the linear interval transformation method is used for normalization. Multivariate regression analysis is performed on the normalized data. The regression dependent variable is the change in the target yaw angle, and the independent variables are the stiffness enhancement ratio and the change in the contact angle. The least squares method is used for regression analysis, and the regression goodness of fit threshold is set to 0.9. If the goodness of fit is insufficient, the sample data is smoothed by a sliding window with a window width of 5 points. Regression analysis is performed again on the smoothed data. After the regression model is established, the yaw angle is predicted step by step. The initial yaw angle is set to 0.5 degrees, and the step size is set to 0.05 degrees. After each iteration, the predicted angle is adjusted according to the fitting equation. The iteration continues until the change in the target yaw angle is less than the preset deviation limit of 0.02 degrees. Finally, the yaw angle matching data is output. In a specific embodiment of bearing structure optimization based on bearing radial stiffness enhancement ratio, contact angle increase data, and runout angle matching data, the bearing radial stiffness enhancement ratio data, contact angle increase data, and runout angle matching data obtained in steps S321, S322, and S323 are first processed by merging feature vectors. The feature vector dimensions include four parameters: radial stiffness change coefficient, contact angle increase, runout angle matching increment, and rolling element diameter change. The initial range for the rolling element diameter change is set to 0 to 0.5 mm. The feature vectors are then input into a multi-objective nonlinear constraint optimization algorithm, with the optimization objective being... The axial displacement and radial stiffness are minimized and maximized with weighting coefficients of 0.7 and 0.3, respectively. The constraints during the optimization process include a maximum contact stress not exceeding 1500 MPa, a raceway contact length not less than 80% of the groove length, and a pressure gradient at the rolling element edge contact point not exceeding 300 MPa / mm. The optimization iteration uses the Lagrange multiplier method with an iteration step size of 0.05 mm. After each iteration, the bearing structural parameters are updated, including the number of rolling elements, the raceway radius of curvature, and the inner ring wall thickness. The optimization terminates when the rate of change of the objective function is less than 0.0005. The final output is the optimized bearing structural data.

[0154] The present invention also provides a vibration and noise optimization system for an electric drive assembly, used to execute the vibration and noise optimization method for the electric drive assembly as described above, the vibration and noise optimization system for the electric drive assembly comprising:

[0155] The frequency domain conversion module is used to collect vibration and noise under various operating conditions of the transmission motor in the electric drive assembly during experimental conditions; and to perform frequency domain conversion processing on the vibration and noise under various operating conditions to generate vibration and noise frequency domain data.

[0156] The vibration and noise impact analysis module is used to perform gear meshing force off-center load limit calculation based on vibration and noise frequency domain data, and then perform tooth surface pitting cyclic ductility simulation to obtain tooth surface pitting cyclic ductility data; and to identify bearing clearance expansion increment based on vibration and noise frequency domain data to obtain bearing clearance expansion increment data.

[0157] The vibration-strengthening design module is used to optimize the bearing structure based on the bearing clearance expansion increment data to obtain bearing structure optimization data; to perform tooth-direction drum shape modification analysis based on tooth surface pitting cycle extension data to obtain tooth-direction drum shape modification data; to perform vibration-strengthening design based on the bearing structure optimization data and tooth-direction drum shape modification data to generate structural vibration-strengthening design data; and to send the structural vibration-strengthening design data to the terminal to perform structural optimization affected by vibration and noise.

[0158] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for optimizing the vibration and noise of an electric drive assembly, characterized in that, The method comprises the following steps: Step S1: collecting multi-working condition vibration noises of a transmission motor in an electric drive assembly under an experimental state; performing frequency domain conversion processing on the multi-working condition vibration noises, thereby generating vibration noise frequency domain data; Step S2: performing gear meshing stress bias load limit calculation based on the vibration noise frequency domain data, and then performing gear surface pitting cycle propagation simulation to obtain gear surface pitting cycle propagation data; identifying bearing gap expansion increment based on the vibration noise frequency domain data, thereby obtaining bearing gap expansion increment data; Step S3: performing bearing structure optimization processing based on the bearing gap expansion increment data, thereby obtaining bearing structure optimization data; performing tooth direction drum shape modification analysis based on the gear surface pitting cycle propagation data, thereby obtaining tooth direction drum shape modification data; and performing anti-vibration strengthening design of vibration influence based on the bearing structure optimization data and the tooth direction drum shape modification data, thereby generating structure anti-vibration strengthening design data; sending the structure anti-vibration strengthening design data to a terminal to perform structure optimization of vibration noise influence; Step S3 comprises the following steps: Step S31: performing logical learning on the bearing gap expansion increment data, thereby generating bearing gap expansion increment learning data; Step S32: performing bearing structure optimization processing based on the bearing gap expansion increment learning data, thereby obtaining bearing structure optimization data; Step S33: performing tooth direction drum shape modification analysis based on the gear surface pitting cycle propagation data, thereby obtaining tooth direction drum shape modification data; Step S34: performing anti-vibration strengthening design of vibration influence based on the bearing structure optimization data and the tooth direction drum shape modification data, thereby generating structure anti-vibration strengthening design data; Step S35: sending the structure anti-vibration strengthening design data to a terminal to perform structure optimization of vibration noise influence; Step S32 comprises the following steps: Step S321: performing bearing radial stiffness strengthening ratio measurement based on the bearing gap expansion increment learning data, thereby obtaining a bearing radial stiffness strengthening ratio; Step S322: performing contact angle equivalent increment coupling based on the bearing radial stiffness strengthening ratio and the bearing gap expansion increment learning data, thereby generating contact angle equivalent increment data; Step S323: performing partial swing angle adaptive design matching based on the bearing radial stiffness strengthening ratio and the contact angle equivalent increment data, thereby obtaining partial swing angle matching data; Step S324: performing bearing structure optimization processing based on the bearing radial stiffness strengthening ratio, the contact angle equivalent increment data, and the partial swing angle matching data, thereby obtaining bearing structure optimization data.

2. The method of claim 1, wherein, Step S1 comprises the following steps: Step S11: collecting multi-working condition vibration noises of a transmission motor in an electric drive assembly under an experimental state; Step S12: performing data cleaning on the multi-working condition vibration noises, thereby generating multi-working condition vibration noise cleaning data; Step S13: performing detail enhancement processing on the multi-working condition vibration noise cleaning data, thereby obtaining vibration noise detail enhancement data; Step S14: performing frequency domain conversion processing on the vibration noise detail enhancement data, thereby generating vibration noise frequency domain data.

3. The method of claim 1, wherein, Step S2 comprises the following steps: Step S21: performing vibration energy transmission intensity diffusion analysis based on the vibration noise frequency domain data, thereby obtaining vibration energy transmission intensity diffusion data; Step S22: Gear meshing force bias limit calculation is performed according to the vibration energy transmission intensity diffusion data, so as to obtain gear meshing force bias limit data; Step S23: Gear surface pitting cycle ductility simulation is performed on the gear meshing force bias limit data, so as to obtain gear surface pitting cycle ductility data; Step S24: Bearing clearance expansion increment identification is performed according to the vibration energy transmission intensity diffusion data, so as to obtain bearing clearance expansion increment data.

4. The method of claim 3, wherein, Step S22 includes the following steps: Step S221: Obtain the gear meshing vibration frequency and gear material basic characteristics under the current working condition; spread frequency dispersion superposition frequency quantization is performed on the vibration energy transmission intensity diffusion data to obtain energy frequency dispersion superposition frequency data; Step S222: Resonance peak energy density evolution ratio analysis is performed on the gear meshing vibration frequency according to the energy frequency dispersion superposition frequency data, to generate resonance peak energy density evolution ratio; Step S223: Gear shaft fatigue deformation gradual gradient increment simulation estimation is performed on the gear material basic characteristics based on the resonance peak energy density evolution ratio, so as to obtain fatigue deformation gradual gradient increment data; Step S224: Parallelism / verticality deviation data of gear meshing is dynamically simulated and calculated by using the fatigue deformation gradual gradient increment data; Step S225: Gear meshing force bias limit calculation is performed based on the parallelism / verticality deviation data, so as to obtain gear meshing force bias limit data.

5. The method of claim 4, wherein, Step S225 includes the following steps: Parallelism / verticality deviation data of gear meshing is dynamically integrated based on the parallelism / verticality deviation data, to obtain engagement line dynamic offset data; Axial force and radial force imbalance index calculation is performed according to the engagement line dynamic offset data, to obtain axial force / radial force imbalance index; Imbalance index isochronous change analysis is performed on adjacent time points, to obtain imbalance index isochronous change data; Based on the imbalance index isochronous change data and the engagement line dynamic offset data, rotational direction force misalignment deviation quantization of gear helix angle is performed, to obtain rotational direction force misalignment deviation data; Gear meshing force bias limit calculation is performed according to the rotational direction force misalignment deviation data, the imbalance index isochronous change data and the engagement line dynamic offset data, so as to obtain gear meshing force bias limit data.

6. The method of claim 4, wherein, Step S23 includes the following steps: Step S231: Load along tooth width skew distribution analysis is performed on the gear meshing force bias limit data, to obtain load tooth width skew distribution data; Step S232: Tooth width edge contact stress geometric growth quantization is performed on the load tooth width skew distribution data according to the gear meshing force bias limit data, to obtain tooth width edge stress geometric growth data; Step S233: Cycle iteration stress crack propagation degree coupling is performed on the tooth width edge stress geometric growth data, so as to obtain iteration stress crack propagation degree data; Step S234: Gear surface pitting cycle ductility simulation is performed according to the tooth width edge stress geometric growth data and the iteration stress crack propagation degree data, to obtain gear surface pitting cycle ductility data.

7. The method of claim 3, wherein, Step S24 includes the following steps: Step S241: bearing inter-conduction vibration high-frequency impact vector analysis is performed on the vibration energy transmission intensity diffusion data to obtain a vibration high-frequency impact vector; Step S242: bearing inter-radial load distribution calculation is performed based on the vibration high-frequency impact vector to obtain bearing radial load distribution data; Step S243: Hertz contact stress iteration simulation deduction is performed on the bearing radial load distribution data according to the vibration high-frequency impact vector to generate Hertz contact stress iteration data; Step S244: bearing contact angle nonlinear change solving is performed according to the Hertz contact stress iteration data and the bearing radial load distribution data to obtain bearing contact angle change data; Step S245: bearing clearance expansion increment identification is performed based on the bearing contact angle change data, thereby obtaining bearing clearance expansion increment data.

8. A system for optimizing vibration noise of an electric drive assembly, comprising: The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for an electric drive assembly. The application discloses a vibration noise optimization method for

Images