A multi-drive cooperative control method for high-precision electric drive assembly equipment

Abstract

The application discloses a kind of multi-drive collaborative control methods for high-precision electric drive assembly equipment, it is related to intelligent manufacturing control technical field, for solving the problem of insufficient collaborative control precision of multi-axis system under parameter mismatch and disturbance.By constructing multi-source data fusion and dynamic coupling analysis mechanism, the system collects the motion state and control parameter data of multi-drive shaft, generates multi-axis collaborative dynamic response sequence, calculates the dynamic coupling degree between shafts to identify system uncertainty mode;Combined with temperature field and vibration signal characteristic positioning performance degradation shaft section, generate system stability evaluation coefficient;Establish spatial mapping relationship to construct dynamic compensation factor matrix, generate control parameter adjustment amount that meets collaborative matching condition through fuzzy rule reasoning.The application avoids the control instability caused by traditional global adjustment, realizes the precise collaborative control of multi-axis system, significantly improves the control precision, operation stability and reliability of equipment under dynamic working condition.

Description

Technical Field

[0001] This invention relates to the field of high-precision electromechanical equipment control technology, and more specifically, to a multi-drive collaborative control method for high-precision electric drive assembly equipment. Background Technology

[0002] In the multi-axis collaborative control process of the electric motor assembly line of new energy vehicles, the system is caused by multi-axis dynamic coupling and state response mismatch due to sudden load changes, time-varying parameters and disturbance excitation. Traditional collaborative control methods are mostly based on the assumption of idealized system models, simplifying multi-axis collaboration to ideal tracking control under fixed parameters, and using linear control strategies to handle the independent dynamic response of each axis. They do not quantify the dynamic mapping mechanism between the uncertainty mode of the system and the accuracy of multi-axis collaboration.

[0003] In the existing technology, the lack of coupling of the real-time interaction between the system parameter mismatch, the probability distribution of disturbance amplitude and the multi-axis dynamic coupling effect in the cooperative control model leads to the distortion of system stability assessment under low current density conditions and the inaccuracy of dynamic compensation factor matrix generation. This causes the cooperative control parameters of electric drive assembly equipment with strong disturbances and parameter uncertainties to deviate from the actual operating conditions, exacerbating the tracking error and synchronous oscillation risk. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a multi-drive cooperative control method and system for high-precision electric drive assembly equipment to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A multi-drive cooperative control method for high-precision electric drive assembly equipment includes the following steps:

[0007] S1. Real-time acquisition of motion status data and real-time data of drive shaft control parameters of multiple drive shafts in the new energy vehicle motor assembly line, time-series synchronization calibration of each shaft data, and generation of multi-axis collaborative dynamic response sequence;

[0008] S2. Extract the tracking error characteristics and synchronization error characteristics of each drive axis from the multi-axis cooperative dynamic response sequence, and calculate the dynamic coupling degree between axes to identify the uncertainty mode of the system.

[0009] S3. Based on the frequency domain characteristics of the system uncertainty mode, determine the quantized value of the system model parameter mismatch and the probability distribution of the disturbance amplitude;

[0010] S4. Collect and analyze the spatial distribution characteristics of temperature field and vibration signal of each drive shaft, locate the performance degradation shaft segment according to the high frequency abnormal area, and generate system stability evaluation coefficient by combining the dynamic coupling degree between shafts.

[0011] S5. Establish the spatial mapping relationship between the performance degradation shaft segment and the drive shaft control parameters, and construct the dynamic compensation factor matrix;

[0012] S6. Based on the changing trends of the system stability evaluation coefficient and the dynamic compensation factor matrix, adjust the control parameters of each drive shaft so that the adjustment amount meets the quantification value of the system model parameter mismatch degree and the coordinated matching condition of the dynamic coupling degree between shafts.

[0013] In a preferred embodiment, real-time motion state data and drive shaft control parameter data of multiple drive shafts in the new energy vehicle motor assembly line are collected, and time-series synchronization calibration of each shaft data is performed to generate a multi-axis collaborative dynamic response sequence. Specific steps include:

[0014] High-precision encoders are deployed on each drive shaft of the new energy vehicle motor assembly line to collect position and speed data in real time as motion state data, and servo drivers collect current and torque command data as real-time data for drive shaft control parameters.

[0015] Torque command data refers to the control signals generated by the multi-axis motion controller and sent to the servo driver.

[0016] Based on the system control cycle of the multi-axis motion controller, a fixed duration window is set. Under time synchronization calibration, the motion state data of each drive shaft of the new energy vehicle motor assembly line and the real-time data of the drive shaft control parameters are combined into a multi-axis collaborative dynamic response sequence with a unified time coordinate and containing dynamic response data according to the fixed duration window.

[0017] In a preferred embodiment, the tracking error characteristics and synchronization error characteristics of each drive axis are extracted from the multi-axis cooperative dynamic response sequence, and the dynamic coupling degree between axes is calculated to identify system uncertainty modes. Specific steps include:

[0018] The actual position data and command position data of each drive axis are obtained from the multi-axis collaborative dynamic response sequence. The tracking error data of each drive axis is calculated, and the tracking statistical features of the tracking error data are extracted as tracking error features.

[0019] The relative positional relationship data between each drive axis is obtained from the multi-axis coordinated dynamic response sequence. The synchronization error data between each drive axis is calculated, and the synchronization statistical feature of the synchronization error data is extracted as the synchronization error feature.

[0020] Command position data refers to the target position information of each drive axis generated by the multi-axis motion controller according to the preset motion trajectory planning. Command position data is represented in the form of digital signals to represent the theoretical position coordinates that each drive axis is expected to reach in each control cycle.

[0021] Based on the tracking error characteristics and synchronization error characteristics of each drive shaft, the dynamic coupling degree between shafts is calculated, and the dynamic change characteristics of the dynamic coupling degree between shafts are extracted.

[0022] Based on the comparison between the inter-axis dynamic coupling value and the preset coupling value threshold, the current uncertainty mode of the system is identified.

[0023] In a preferred embodiment, the dynamic coupling degree between shafts is calculated based on the tracking error characteristics and synchronization error characteristics of each drive shaft. Specific steps include:

[0024] The tracking error characteristics and synchronization error characteristics of each drive shaft are combined to form a multi-dimensional feature vector;

[0025] Calculate the covariance matrix among the components of a multidimensional eigenvector;

[0026] The covariance matrix is ​​decomposed into a combination of eigenvalues ​​and eigenvectors through mathematical transformation. Then, the eigenvalue with the largest value is identified from all the eigenvalues, and the eigenvector corresponding to the largest eigenvalue is extracted as the principal eigenvector.

[0027] Normalize each component in the main feature vector to obtain the weight coefficient of each component, then calculate the sum of squares of each weight coefficient, then perform a square root operation on the sum of squares, and finally multiply the square root result by the system calibration coefficient to obtain the final value of the inter-axis dynamic coupling degree of the inter-axis coupling strength.

[0028] The dynamic coupling degree between axes is used as a quantitative basis for the uncertainty pattern recognition of the system.

[0029] In a preferred embodiment, based on the frequency domain characteristics of the system's uncertainty mode, the quantized value of the system model parameter mismatch and the probability distribution of the disturbance amplitude are determined. Specific steps include:

[0030] Perform a Fast Fourier Transform on the dynamic response data corresponding to the system's uncertainty mode to obtain its frequency domain characteristics;

[0031] The power spectral density distribution of a specific frequency band is extracted from the frequency domain features. The total energy value of the entire frequency band is obtained by integrating the power spectral density distribution over the entire frequency band. The ratio of the feature energy value to the total energy value of the entire frequency band is calculated. The ratio is multiplied by the system calibration coefficient to finally determine the quantification value of the system model parameter mismatch.

[0032] The statistical distribution characteristics of the power spectral density distribution amplitude were analyzed, and a probability distribution model of the perturbation amplitude was obtained by fitting a Gaussian mixture model.

[0033] Based on the fitted probability distribution model, the probability of the disturbance amplitude occurring in different intensity ranges is determined, and the probability distribution of the disturbance amplitude is generated.

[0034] The frequency domain characteristics of the system uncertainty mode are formed by combining the quantized value of the system model parameter mismatch with the probability distribution of the disturbance amplitude.

[0035] In a preferred embodiment, the spatial distribution characteristics of temperature field and vibration signal of each drive shaft are collected and analyzed. Degraded shaft segments are located based on high-frequency abnormal regions. System stability evaluation coefficients are generated by combining the dynamic coupling degree between shafts. Specific steps include:

[0036] Temperature field distribution data of each drive shaft is collected by a distributed temperature sensor array, and vibration signal data of each drive shaft is collected by a vibration acceleration sensor.

[0037] Spatial gradient calculation is performed on temperature field distribution data to identify regions of temperature abrupt change; spectral analysis is performed on vibration signal data to extract high-frequency resonance features.

[0038] Based on the spatial overlap between the temperature abrupt change region and the high-frequency resonance characteristics, the high-frequency abnormal region is located and the spatial location and degree of degradation of the performance degradation shaft segment are determined.

[0039] Quantify the degree of degradation of the performance degradation axis segment and generate a quantified value of the degree of degradation of the performance degradation axis segment;

[0040] Extract the dynamic change characteristics of the inter-axis dynamic coupling degree value, and calculate the fluctuation amplitude and change frequency;

[0041] Based on a fuzzy logic system, the quantified value of the degradation degree of the performance degradation axis segment is used as the first input variable, the fluctuation range of the dynamic coupling degree between axes is used as the second input variable, and the change frequency of the dynamic coupling degree between axes is used as the third input variable. The system stability evaluation coefficient is generated by reasoning and calculation through a preset fuzzy rule base.

[0042] In a preferred embodiment, a spatial mapping relationship between the performance degradation shaft segment and the drive shaft control parameters is established, and a dynamic compensation factor matrix is ​​constructed. Specific steps include:

[0043] Obtain the spatial location and quantified value of the degree of degradation of the performance degradation shaft segment;

[0044] Based on the spatial location and degradation degree quantification of the performance degradation shaft segment, the compensation priority and compensation intensity parameters of the corresponding drive shaft control parameters are determined.

[0045] Based on the quantified value of the degradation degree of the performance degradation shaft segment and the compensation strength parameter of the drive shaft control parameter, the compensation coefficient of the drive shaft control parameter is generated.

[0046] Establish a spatial mapping relationship between the spatial position of the performance degradation shaft segment and the compensation coefficients of the drive shaft control parameters;

[0047] Based on the topology of multi-axis cooperative motion, the compensation coefficients of the control parameters of each drive axis are arranged in a matrix according to the spatial mapping relationship to form a preliminary compensation coefficient matrix.

[0048] The values ​​of each element in the preliminary compensation coefficient matrix are dynamically updated based on the real-time operating status of the system to generate a dynamic compensation factor matrix with adaptive capabilities.

[0049] In a preferred embodiment, the compensation priority of the drive shaft control parameters is obtained by multiplying the reciprocal of the Euclidean distance from the spatial location of the performance degradation shaft segment to the key nodes of the mechanical structure of each drive shaft with the quantified value of the degradation degree of the degradation shaft segment to obtain a spatial correlation factor. Then, the spatial correlation factor values ​​are sorted from largest to smallest, and the sorting number is used as the compensation priority order of the corresponding drive shaft.

[0050] The compensation strength parameter of the drive shaft control parameter is obtained by multiplying the reciprocal of the Euclidean distance from the spatial position of the performance degradation shaft segment to the key nodes of the mechanical structure of each drive shaft with the degradation degree quantification value of the degradation shaft segment. Then, the spatial correlation factor value is input into the preset strength mapping function, and the compensation strength of the drive shaft control parameter is determined according to the output value of the mapping function.

[0051] The compensation coefficient of the drive shaft control parameters is generated by multiplying the quantified value of the degradation degree of the performance degradation shaft segment by the compensation intensity parameter of the drive shaft control parameters, and then multiplying by the system calibration coefficient.

[0052] In a preferred embodiment, the drive shaft control parameters of each drive shaft are adjusted according to the changing trends of the system stability evaluation coefficient and the dynamic compensation factor matrix, so that the adjustment amount satisfies the quantified value of the system model parameter mismatch degree and the cooperative matching condition of the dynamic coupling degree between shafts. Specific steps include:

[0053] Real-time monitoring of the numerical trend of the stability evaluation coefficient and the element update rules of the dynamic compensation factor matrix;

[0054] Based on the rate of decrease of the system stability evaluation coefficient and the direction of gradient change of the dynamic compensation factor matrix, the adjustment direction and adjustment range of the drive shaft control parameters are determined.

[0055] Based on the quantified value of the degree of mismatch of system model parameters, and combined with the adjustment direction and adjustment range of drive shaft control parameters, the basic compensation amount of drive shaft control parameters is calculated.

[0056] Determine the coupling compensation weights for multi-axis cooperative motion based on the inter-axis dynamic coupling degree values;

[0057] The fuzzy rule reasoning system uses the basic compensation amount of the drive shaft control parameters as the antecedent input and the coupling compensation weight of the multi-axis cooperative motion as the consequent adjustment factor. Based on the preset cooperative matching rule base, the system performs reasoning calculations to generate the adjustment amount of the drive shaft control parameters that meet the cooperative matching conditions.

[0058] Verify whether the adjustment amount simultaneously meets the compensation requirements for the quantification value of the system model parameter mismatch degree and the optimization requirements for the dynamic coupling degree between axes, and realize the determination of the cooperative matching conditions;

[0059] When the matching conditions are met, the final adjustment amount is applied to the control system of the corresponding drive shaft to achieve adaptive adjustment of multi-axis coordinated motion;

[0060] When the collaborative matching condition is not met, the coupling compensation weights are readjusted, and the fuzzy rule reasoning and verification steps are iteratively executed until the collaborative matching condition is met.

[0061] Compared with the prior art, the present invention has the following beneficial effects:

[0062] 1. By analyzing the multi-axis dynamic coupling characteristics and temperature field gradient, the system achieves precise capture and cross-parameter collaborative control of the drive shaft performance degradation state. Based on the construction of a multi-axis collaborative dynamic response sequence with time-series synchronous calibration, it effectively identifies the interaction mode between system model parameter mismatch and inter-axis dynamic coupling. Combined with the spatial positioning of high-frequency abnormal regions, the system uses joint analysis of the inter-axis dynamic coupling degree and degradation degree quantification values ​​to physically correlate microscopic performance degradation (such as bearing wear and transmission clearance) with macroscopic motion accuracy fluctuations, forming a multi-scale system stability assessment model. Compared with existing technologies, this system can perceive the coupling state changes of multi-axis collaborative motion in real time under high dynamic conditions. Through the spatial mapping relationship of the dynamic compensation factor matrix, it significantly improves the multi-axis collaborative control accuracy and suppresses the decrease in system stability caused by parameter mismatch and coupling oscillation.

[0063] 2. By dynamically adapting the spatial mapping relationship with the drive shaft control parameters, the problem of parameter mismatch and the superposition and amplification of inter-axis coupling effects in traditional control methods is solved. Based on the correspondence between the spatial position of the performance degradation shaft segment and the drive shaft topology, the constructed dynamic compensation factor matrix can match the shaft degradation state and multi-axis control characteristics in real time, realizing the coordinated adjustment of the compensation intensity of each drive shaft control parameter and the system stability requirements. Through the coordinated matching condition of the system model parameter mismatch degree and the dynamic coupling degree between shafts, it is ensured that the adjustment amount always matches the actual degradation degree and the system stability requirements, avoiding control instability caused by overcompensation or undercompensation. While maintaining the synchronization accuracy of multi-axis, it significantly reduces the risk of system oscillation caused by the accumulation of performance degradation, ensuring the long-term operational reliability of high-precision electric drive assembly equipment under dynamic conditions. Attached Figure Description

[0064] Figure 1 This is a schematic diagram of the structure of a multi-drive collaborative control method for high-precision electric drive assembly equipment according to the present invention. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] Example 1: Figure 1 A schematic diagram of a multi-drive cooperative control method for high-precision electric drive assembly equipment according to the present invention is provided, which includes the following steps:

[0067] S1. Real-time acquisition of motion status data and real-time data of drive shaft control parameters of multiple drive shafts in the new energy vehicle motor assembly line, time-series synchronization calibration of each shaft data, and generation of multi-axis collaborative dynamic response sequence;

[0068] S2. Extract the tracking error characteristics and synchronization error characteristics of each drive axis from the multi-axis cooperative dynamic response sequence, and calculate the dynamic coupling degree between axes to identify the uncertainty mode of the system.

[0069] S3. Based on the frequency domain characteristics of the system uncertainty mode, determine the quantized value of the system model parameter mismatch and the probability distribution of the disturbance amplitude;

[0070] S4. Collect and analyze the spatial distribution characteristics of temperature field and vibration signal of each drive shaft, locate the performance degradation shaft segment according to the high frequency abnormal area, and generate system stability evaluation coefficient by combining the dynamic coupling degree between shafts.

[0071] S5. Establish the spatial mapping relationship between the performance degradation shaft segment and the drive shaft control parameters, and construct the dynamic compensation factor matrix;

[0072] S6. Based on the changing trends of the system stability evaluation coefficient and the dynamic compensation factor matrix, adjust the control parameters of each drive shaft so that the adjustment amount meets the quantification value of the system model parameter mismatch degree and the coordinated matching condition of the dynamic coupling degree between shafts.

[0073] Real-time acquisition of motion status data and real-time control parameter data of multiple drive shafts in the new energy vehicle motor assembly line; timing synchronization calibration of data for each shaft; generation of multi-axis collaborative dynamic response sequence; specific implementation as follows:

[0074] A multi-axis motion control system consists of a multi-axis motion controller, servo drives, high-precision encoders, and data acquisition modules interconnected via a fieldbus network. In specific implementation, the multi-axis motion control system is configured as follows: First, a networked hardware architecture is established, with the multi-axis motion controller as the core processing unit. This controller achieves high-speed interconnection with all servo drives and data acquisition modules via an EtherCAT industrial Ethernet fieldbus network, forming a unified control network. Second, data acquisition terminals are deployed. High-precision encoders are installed on key mechanical structural nodes of each drive shaft in the new energy vehicle motor assembly line. These encoders are absolute encoders with micron-level resolution. Each encoder acts as an independent slave node, communicating with the upper-level multi-axis motion controller and data acquisition module via the EtherCAT bus. Simultaneously, drive execution units are integrated. Each drive shaft's servo drive is also connected to the EtherCAT network as a slave node, receiving command data from the multi-axis motion controller. Finally, the information is integrated through the data acquisition module. The drive shaft position and speed data collected in real time by all encoders (forming a motion state dataset), as well as the current command data (representing the target current value of the motor winding) and torque command data (representing the theoretical output torque value of the drive shaft) issued by the multi-axis motion controller and collected in real time by the servo drives, are all transmitted to the data acquisition module for centralized processing via the bus. These two types of data together constitute the real-time data of the drive shaft control parameters, thus physically and logically interconnecting to form a complete multi-axis motion control system with centralized data management.

[0075] In practice, high-precision encoders are installed at key positions on each drive shaft of the new energy vehicle motor assembly line. These encoders are absolute encoders with micron-level resolution and are connected to the host controller via EtherCAT industrial Ethernet. The encoders collect position and speed data of the drive shafts in real time at a fixed sampling frequency (e.g., 1kHz). These data together constitute a motion state dataset. Simultaneously, the servo drives associated with each drive shaft collect current and torque command data issued by the multi-axis motion controller in real time. The current command data represents the target current value of the motor windings, and the torque command data represents the theoretical torque value that the drive shaft needs to output.

[0076] To achieve multi-axis data acquisition, high-precision encoders are installed on each drive shaft of the new energy vehicle motor assembly line. These encoders are absolute encoders, and each encoder is connected to the host controller via an EtherCAT bus. The encoders acquire the position data of the drive shaft in real time with a micrometer-level resolution and the speed data of the drive shaft with a millisecond-level sampling frequency. These two types of data together constitute the motion state dataset. At the same time, the servo drivers associated with each drive shaft acquire current command data and torque command data in real time. The servo drivers use control commands issued by the multi-axis motion controller. The current command data represents the expected current value of the motor winding, and the torque command data represents the theoretical torque value that the drive shaft needs to output. These data together constitute the real-time data of the drive shaft control parameters.

[0077] Torque command data refers to the control signals generated by the multi-axis motion controller and sent to the servo driver.

[0078] Using the core processor clock signal of the multi-axis motion controller in the new energy vehicle motor assembly line as the synchronization reference, this clock signal is distributed to each data acquisition node via the IEEE 1588 Precision Time Protocol (a precision clock synchronization protocol standard for networked measurement and control systems). Hardware timestamp technology is used to embed a time stamp with nanosecond-level precision at each data sampling point. In specific implementation, FPGA (a semiconductor integrated circuit device that can be programmed to implement specific circuit functions) chips are installed in the data acquisition modules of each drive axis. When the encoder acquires position / speed data or the servo driver acquires current / torque command data, the FPGA chip immediately adds the timestamp of the current clock reference. After the timestamp marking is completed, time alignment processing is performed: first, the motion state data of each drive axis is sorted according to the timestamp; then, the real-time data of the drive axis control parameters is sorted according to the same time reference; finally, the two types of data are aligned and matched on a unified time axis to complete the timing synchronization calibration of the data of each axis.

[0079] Based on the system control cycle characteristics of the multi-axis motion controller, a fixed duration window is set. The duration of this window is usually set to an integer multiple of the system control cycle to ensure that the complete motion control cycle characteristics are included. On the basis of the completed timing synchronization calibration, motion state data and control parameter data of each drive shaft of the new energy vehicle motor assembly line are extracted according to the fixed duration window. The multi-axis data within the same time window are combined and encapsulated: each data packet contains the window start timestamp, the window end timestamp, and the position data sequence, speed data sequence, current command data sequence, and torque command data sequence of all drive shafts within that time period. Finally, a multi-axis collaborative dynamic response sequence with a unified time coordinate and containing dynamic response data is formed. This sequence completely preserves the spatiotemporal correlation characteristics of the motion state and control commands of each drive shaft.

[0080] The tracking error characteristics and synchronization error characteristics of each drive axis are extracted from the multi-axis cooperative dynamic response sequence, and the dynamic coupling degree between axes is calculated to identify the uncertainty mode of the system. The specific implementation is as follows:

[0081] The actual position data stream and commanded position data stream of each drive axis are analyzed from the multi-axis collaborative dynamic response sequence. The actual position data is acquired by a high-precision encoder, and the commanded position data is generated by the multi-axis motion controller according to the preset motion trajectory planning. The difference between the actual position and the commanded position of each drive axis at each sampling moment is calculated to obtain the tracking error data. Statistical analysis is performed on the tracking error data: first, the standard deviation of the tracking error data is calculated to characterize the dispersion of the tracking error data; second, the mean of the tracking error data is calculated to characterize the steady-state deviation of the system; finally, the peak value of the tracking error data is calculated to characterize the maximum deviation range during the motion process. These statistics are combined to form a tracking error feature vector as the tracking error feature.

[0082] The relative positional relationship data between each drive axis is analyzed from the multi-axis coordinated dynamic response sequence. The position difference between any two drive axes at the same moment is calculated to obtain the synchronization error data between each drive axis. Furthermore, the variance of the position data of all drive axes in the multi-axis system is calculated to characterize the degree of synchronization consistency of the multi-axis motion. Statistical analysis is performed on the synchronization error data: the root mean square value of the synchronization error data is calculated to characterize the synchronization accuracy; the range of the synchronization error data is calculated to characterize the maximum positional deviation between axes; and the rate of change of the synchronization error data is calculated to characterize the dynamic characteristics of the relative motion between axes. These statistical quantities are combined to form the synchronization error characteristic quantity, which is used as the synchronization error feature.

[0083] Command position data refers to the target position information of each drive axis generated by the multi-axis motion controller according to the preset motion trajectory planning. Command position data is represented in the form of digital signals to represent the theoretical position coordinates that each drive axis is expected to reach in each control cycle.

[0084] The tracking error characteristics and synchronization error characteristics of each drive shaft are combined into a multidimensional feature matrix. The covariance matrix of this multidimensional feature matrix is ​​calculated to characterize the correlation between the features. The covariance matrix is ​​decomposed into features, and the eigenvector corresponding to the largest eigenvalue is extracted as the principal eigenvector. Each component in the principal eigenvector is normalized to obtain the weight coefficient of each component. The geometric mean of these weight coefficients is calculated and then multiplied by the system calibration coefficient to finally obtain the inter-axis dynamic coupling degree value of the inter-axis coupling strength.

[0085] In a specific implementation case, the system extracts features from three drive axes (A, B, and C axes) to obtain feature vectors containing features of tracking and synchronization errors. First, these feature values ​​are combined to form a multidimensional feature matrix. Then, a covariance matrix is ​​constructed by calculating the covariance between any two feature vectors in the multidimensional feature matrix: first, the mean of each feature vector across all samples is calculated; then, the deviation of each sample value of each feature vector from its mean is multiplied; finally, the average of these products is calculated, resulting in the covariance value characterizing the degree of linear correlation between any two feature vectors. All these covariance values ​​are arranged in the order of the feature vectors to form a symmetric covariance matrix. The magnitude and sign of the off-diagonal elements of the matrix represent the strength and direction of the correlation between different eigenvectors, respectively, to comprehensively characterize the correlation between various eigenvalues. Subsequently, the covariance matrix is ​​decomposed into eigenvalues, and the eigenvector corresponding to the largest eigenvalue is extracted as the principal eigendirection. The components of the principal eigenvector are normalized to obtain the weight coefficients corresponding to each component. Then, the geometric mean of these weight coefficients is calculated, which is the sixth root of the product of all weight coefficients. Finally, the geometric mean is multiplied by the system coefficients pre-calibrated through experiments to obtain the precise quantified value of the interaxial dynamic coupling strength.

[0086] The above calculation process is repeated within a continuous time window to obtain time series data of the inter-axis dynamic coupling degree. Dynamic features are then extracted from this time series data: First, the first-order difference of the values ​​is calculated using the sliding window method to obtain the rate of change sequence of the inter-axis dynamic coupling degree, characterizing the instantaneous trend of the inter-axis coupling strength. Second, spectral analysis is performed on the time series data to extract the main fluctuation frequency components, characterizing the periodic variation characteristics of the inter-axis coupling strength. Finally, the variance and autocorrelation function of the time series data are calculated. The autocorrelation function is a mathematical tool used to quantify the correlation between time series data at different time points. By calculating the correlation coefficient between the sequence and its lagged version, the dependencies and periodic patterns existing in the time series data are revealed, characterizing the stability and memory effect of the inter-axis coupling strength. These dynamic change features are combined with the inter-axis dynamic coupling degree values ​​to form a complete description of the dynamic change characteristics of the inter-axis dynamic coupling degree values.

[0087] For example, the system performed dynamic coupling analysis on the three drive shaft system of a certain type of new energy vehicle motor assembly line. The system continuously collected data with a time window of 100 milliseconds, and obtained time series data containing dynamic coupling values ​​of 500 sampling points. When extracting dynamic features from the time series data, the system first used a sliding window with a length of 50 sampling points to calculate the first-order difference, and obtained the dynamic coupling change rate sequence of the shaft. The change rate sequence shows that the coupling strength has a rapid upward trend within a specific time period, with the instantaneous change rate reaching a maximum of 0.15 units per second, which clearly characterizes the instantaneous change trend of the coupling strength of the shaft.

[0088] Secondly, spectral analysis was performed on the time series data. The main fluctuation frequency components were extracted using Fast Fourier Transform. The analysis revealed significant peaks at 0.5 Hz and 2.0 Hz, with the 0.5 Hz component dominating in amplitude. This indicates a regular fluctuation in inter-axis coupling strength with a period of 2 seconds, accurately characterizing the periodic variation of inter-axis coupling strength during system operation. Finally, the statistical properties of the time series data were calculated, yielding a variance of 0.045, indicating that the inter-axis coupling strength maintained a relatively stable fluctuation level throughout the observation period. Simultaneously, the autocorrelation function was calculated, revealing that... The data shows a significant peak at a lag of 4 time windows (400 milliseconds), with a correlation coefficient of 0.72. This reveals a strong short-term memory effect in the time series data, indicating that the change in inter-axis coupling strength has an inertial characteristic lasting about 400 milliseconds. Combining these dynamic change characteristics with the inter-axis dynamic coupling degree values ​​forms a complete description of the dynamic change characteristics: the current inter-axis dynamic coupling degree value of the system is 0.68, with an instantaneous change rate of 0.15 units per second, significant fluctuations with a period of 2 seconds, and the overall fluctuation level remains stable, exhibiting a short-term memory effect of 400 milliseconds.

[0089] Preset threshold ranges for coupling degree values ​​under different operating conditions are established: 0-0.3 for low coupling, 0.3-0.7 for medium coupling, and 0.7-1.0 for strong coupling. The calculated dynamic coupling degree values ​​between axes are compared with these thresholds to determine the current coupling state of the system. Based on the coupling state, the current uncertainty mode of the system is identified: low coupling corresponds to a stable mode with good system parameter matching; medium coupling corresponds to a transitional mode with slight parameter mismatch; and strong coupling corresponds to an unstable mode with severe parameter mismatch and disturbances.

[0090] Based on the tracking error characteristics and synchronization error characteristics of each drive shaft, the dynamic coupling degree between shafts is calculated as the coupling strength between shafts. The specific implementation is as follows:

[0091] The tracking error features and synchronization error features of each drive shaft are arranged and combined in a predetermined order to form a multidimensional feature vector. The tracking error features include the standard deviation, mean error, and peak error of the tracking error of each drive shaft; the synchronization error features include the root mean square value, range, and rate of change of the synchronization error between shafts. These feature vectors are arranged in the order of drive shaft numbering to finally form a multidimensional feature vector containing feature information of all drive shafts.

[0092] To calculate the covariance matrix among the components of a multidimensional eigenvector, first calculate the mean of each eigenvector, then multiply the deviations of each eigenvector from the mean, and finally calculate the average of these products to obtain the covariance matrix among the components.

[0093] The covariance matrix is ​​mathematically transformed to decompose it into a combination of eigenvalues ​​and eigenvectors. All eigenvalues ​​are obtained by solving the characteristic equation. The eigenvalue with the largest value is identified, and the eigenvector corresponding to the largest eigenvalue is extracted as the principal eigenvector. The principal eigenvector represents the direction of the most significant change among the multidimensional eigenvectors, and the magnitude of each component reflects the degree of contribution of the corresponding eigenvector to the coupling strength of the system.

[0094] The components in the principal eigenvector are normalized, and each component value is divided by the vector's magnitude to obtain the weight coefficient of each component. The sum of squares of these weight coefficients is calculated, and then the square root of the sum of squares is taken to obtain the basic value of the coupling strength. Finally, the basic value of the coupling strength is multiplied by the system calibration coefficients determined through system calibration experiments to obtain the final value of the inter-axis dynamic coupling degree. The system calibration coefficients are scaling factors obtained by testing and calibrating standard equipment.

[0095] The calculated inter-axis dynamic coupling degree is used as the main quantitative basis for system uncertainty pattern recognition. This value directly reflects the coupling strength within the multi-axis system. The larger the value, the stronger the inter-axis coupling effect and the higher the uncertainty of the system. By comparing this value with the preset threshold range, the current operating mode of the system can be accurately identified.

[0096] In a specific implementation case, the system contains four drive axes. The multidimensional feature vector consists of 12 feature vectors (3 tracking error features and 3 synchronization error features for each drive axis). The calculated covariance matrix is ​​a 12×12 square matrix with a maximum eigenvalue of 3.45. The corresponding principal eigenvector contains 12 components. After normalization, the sum of squares of the weight coefficients is 2.17, and the square root is 1.47. The system calibration coefficient was experimentally determined to be 1.25. The final calculated inter-axis dynamic coupling degree is 1.84. Based on the preset threshold range (low coupling: 0-1.0, medium coupling: 1.0-2.0, high coupling: above 2.0), the system is identified as being in a medium coupling state.

[0097] Based on the frequency domain characteristics of the system's uncertainty mode, the quantized values ​​of the system model parameter mismatch and the probability distribution of the disturbance amplitude are determined. Specifically, the implementation is as follows:

[0098] The dynamic response data corresponding to the system's uncertainty mode is processed by a Fast Fourier Transform (FFT). This transform algorithm decomposes the time-domain signal into a series of sinusoidal components with different frequencies and amplitudes, thereby converting the dynamic response data into a frequency-domain feature representation composed of these frequency components. This frequency-domain feature contains the amplitude and phase information of each sinusoidal component, which is the basis for subsequent power spectral density analysis. In practice, a digital signal processor (DSP) or field-programmable gate array (FPGA) is used to accelerate the implementation of this transform algorithm, ensuring the real-time performance and computational accuracy of the process.

[0099] The power spectral density distribution of a specific frequency band is extracted from the frequency domain characteristics. This specific frequency band is determined based on the system's mechanical resonant frequency and control bandwidth, typically selecting a mid-to-high frequency band (e.g., 100Hz-1kHz) that reflects the system's dynamic characteristics. The characteristic energy value of this frequency band is obtained by integrating the power spectral density distribution within the selected characteristic frequency band. Simultaneously, the total energy value of the entire frequency band is obtained by integrating the power spectral density distribution across the entire frequency range. The ratio of the characteristic energy value to the total energy value of the entire frequency band is calculated, and this ratio is multiplied by the system calibration coefficients determined through system calibration experiments. Finally, a quantitative value of the system model parameter mismatch is obtained. These system calibration coefficients are obtained by comparing... The frequency response characteristics of the ideal model and the actual system are determined through multiple calibration tests. The calibration multiplier used to convert the dimensionless energy ratio into specific quantization parameters is used. For example, under standard experimental conditions, three standard test systems with known system model parameter mismatches of 20%, 50%, and 80% are selected for testing. The ratios of the three sets of characteristic energy values ​​to the total energy value of the entire frequency band are 0.05, 0.13, and 0.21, respectively. Then, a linear relationship between these ratios and the known system model parameter mismatches is established. The optimal calibration multiplier is obtained by least squares fitting, which is 3.85 with an offset of 0.01. Finally, this calibration multiplier is fixed as the system calibration coefficient.

[0100] The statistical distribution characteristics of the power spectral density amplitude are analyzed, and the distribution fit test of the power spectral density amplitude data is performed. A Gaussian mixture model is used to model the probability distribution of the disturbance amplitude. The Gaussian mixture model fits the complex probability distribution shape through a linear combination of multiple Gaussian distribution functions, which can accurately describe the multimodal distribution characteristics of the disturbance amplitude. The parameters of the Gaussian mixture model are estimated by the expectation-maximization (EM) algorithm, including the mean, variance and mixing weight of each Gaussian component. Finally, the Gaussian mixture model defined by these parameters is the probability distribution model of the obtained disturbance amplitude.

[0101] Based on the fitted probability distribution model, the probability of the disturbance amplitude occurring in different intensity intervals is calculated. The disturbance amplitude range is divided into multiple continuous intensity intervals. The integral value of the probability density function is calculated for each interval to obtain the probability value of that interval. The probability values ​​of all intensity intervals are combined to form a complete disturbance amplitude probability distribution, which comprehensively describes the statistical characteristics of the disturbance amplitude.

[0102] By combining the quantified value of the system model parameter mismatch with the probability distribution of the disturbance amplitude, a frequency domain feature of the system uncertainty mode is formed. This frequency domain feature contains both the quantified information of parameter mismatch and the statistical description of disturbance characteristics, providing a comprehensive frequency domain feature basis for subsequent control strategy formulation.

[0103] In a specific implementation case, the sampling frequency was set to 5kHz, and the analysis frequency range was 0-2.5kHz. The characteristic frequency band was selected as 200-800Hz, which contains the main mechanical resonant frequency of the system. The calculated energy value of the characteristic frequency band was 4.75J, the total energy value of the entire frequency band was 18.32J, and the ratio was 0.259. The system calibration coefficient was determined to be 3.85 through calibration. Finally, the quantification value of the system model parameter mismatch was 1.00. The Gaussian mixture model was fitted with three Gaussian components, and the goodness of fit reached above 0.95. The disturbance amplitude intensity range is divided into six intervals: 0-10%, 10-30%, 30-50%, 50-70%, 70-90%, and 90-100%. The calculated probability values ​​for each interval are 0.12, 0.28, 0.35, 0.18, 0.05, and 0.02, respectively, forming a complete disturbance amplitude probability distribution. Finally, the quantized value of the calculated system model parameter mismatch (1.00) is combined with the disturbance amplitude probability distribution (the set of probability values ​​for the six intervals) to constitute a complete frequency domain feature representation of the system uncertainty mode. This combined feature is output in the data structure form of {quantized value: 1.00, probability distribution: [0.12, 0.28, 0.35, 0.18, 0.05, 0.02]}, providing comprehensive information with both deterministic quantification results and statistical probability characteristics for subsequent control decisions.

[0104] The spatial distribution characteristics of temperature field and vibration signals of each drive shaft are collected and analyzed. Degraded shaft segments are located based on high-frequency abnormal regions. System stability evaluation coefficients are generated by combining the dynamic coupling degree between shafts. The specific implementation is as follows:

[0105] Distributed temperature sensor arrays are deployed at key locations on each drive shaft of the new energy vehicle motor assembly line. These arrays utilize infrared temperature sensors to collect temperature field distribution data of the drive shafts in a non-contact manner, with a sampling frequency of at least 100Hz and a temperature measurement accuracy of ±0.5℃. Simultaneously, a vibration acceleration sensor group is installed, employing triaxial MEMS accelerometers to collect vibration signal data from each drive shaft. The sampling frequency is set to 10kHz, and the measurement range covers ±50g to ensure the capture of high-frequency vibration characteristics.

[0106] MEMS accelerometers are miniature sensors manufactured using microelectromechanical systems technology. This technology uses microfabrication processes on silicon wafers to create micron-scale mechanical structures, electronic components, and sensor units, enabling them to convert applied acceleration (or vibration) into measurable electrical signals.

[0107] Spatial gradient calculations were performed on the collected temperature field distribution data. The Sobel operator (an edge detection algorithm widely used in image processing and computer vision) was used to detect the spatial rate of change of the temperature field, identifying temperature abrupt change regions where the temperature gradient exceeded a set temperature threshold (e.g., 10℃ / cm). Fast Fourier transform spectral analysis was performed on the collected vibration signal data, converting the vibration signal into a frequency domain spectrum. This frequency domain spectrum consists of a series of discrete frequency points and their corresponding amplitudes, fully characterizing the energy distribution of the vibration signal at each frequency component. Based on this frequency domain spectrum, the high-frequency band above 500Hz was analyzed in the frequency domain to identify resonance peaks in this band whose energy is significantly higher than the background noise. High-frequency resonance features were extracted: from the identified resonance peaks, their center frequency, peak energy, and half-power bandwidth were quantized and extracted, and these quantized parameters were combined to define the high-frequency resonance features.

[0108] Spatial overlap analysis is performed between the spatial coordinates of the temperature abrupt change region and the spatial distribution of high-frequency resonance features. In practice, a unified two-dimensional or three-dimensional spatial coordinate system is established, and the coordinates of the center point of each temperature abrupt change region are matched with the coordinates of the physical location with the highest intensity of high-frequency resonance features. The Euclidean distance between the two sets of coordinates is calculated, and the reciprocal of the Euclidean distance is the quantitative index of spatial overlap. When the distance value is less than the set spatial tolerance threshold (e.g., 5 mm), it is determined that there is spatial overlap in the region.

[0109] Establish a spatial mapping relationship between temperature field and vibration characteristics. When a region simultaneously meets the conditions of temperature gradient exceeding the limit, high-frequency resonance energy exceeding the limit, and spatial overlap exceeding the limit, the region is identified as a high-frequency abnormal region. Based on the area size, distribution density, and intensity level of the high-frequency abnormal region, determine the spatial location and degree of degradation of the performance degradation shaft segment.

[0110] Based on the high-frequency abnormal region characteristics of the performance degradation shaft segment, a degradation degree assessment index system is constructed. This system includes three core indices: temperature abnormality index, vibration abnormality index, and regional influence index. These indices are weighted and fused to generate a quantitative value of the degradation degree of the performance degradation shaft segment in the range of 0-1.

[0111] The temperature anomaly index is used to quantify the severity of performance degradation caused by overheating. The calculation process includes: first, calculating the difference between the average temperature in the high-frequency anomaly region and the ambient reference temperature to obtain the absolute temperature rise; second, calculating the average temperature gradient in the region; then, normalizing the absolute temperature rise and the average temperature gradient respectively, mapping them to the range of 0-1; finally, weighting and summing the normalized temperature rise value and the gradient value according to a predetermined weight (e.g., 6:4) to obtain the final temperature anomaly index.

[0112] The vibration anomaly index is used to quantify the severity of performance degradation caused by mechanical vibration. The calculation process includes: first, extracting the high-frequency resonance characteristics at the corresponding locations of the high-frequency anomaly region, including the center frequency and peak energy of the resonance peak; second, normalizing the peak energy; then, determining the frequency weighting factor based on the center frequency (the higher the frequency, the greater the risk of failure, and the higher the weight assigned); finally, multiplying the normalized peak energy by the frequency weighting factor to obtain the vibration anomaly index.

[0113] The regional impact index is used to quantify the degree of influence of the spatial attributes of anomaly regions on the overall performance of the system. The calculation process includes: first, calculating the physical area of ​​the high-frequency anomaly region and normalizing it; second, determining the position weight coefficient based on the specific location of the region on the drive shaft (such as assigning higher weights to critical parts like bearing positions and gear meshing areas, and lower weights to non-critical parts); and finally, multiplying the normalized area by the position weight coefficient to obtain the regional impact index.

[0114] Finally, the temperature anomaly index, vibration anomaly index, and regional influence index are weighted and fused according to their importance in performance degradation (e.g., assigned weights of 0.4, 0.4, and 0.2 respectively) to generate a quantitative value of the overall performance degradation degree of the shaft segment in the range of 0-1.

[0115] In a specific case, the average temperature rise of a certain high-frequency anomaly region is 60℃, which is normalized to 0.6; the average gradient is 15℃ / cm, which is normalized to 0.5; the calculated temperature anomaly index is 0.6 * 0.6 + 0.5 * 0.4 = 0.56; the normalized resonance peak energy of this region is 0.8, and its center frequency is in the high-risk frequency band, with a frequency weighting factor of 1.2; the calculated vibration anomaly index is 0.8 * 1.2 = 0.96; the area of ​​this region is 15mm², which is normalized to 0.3; it is located at a critical bearing position, with a position weighting coefficient of 1.5; the calculated regional influence index is 0.3 * 1.5 = 0.45; finally, the quantified value of the degradation degree is: 0.56 * 0.4 + 0.96 * 0.4 + 0.45 * 0.2 = 0.722.

[0116] Dynamic change characteristics are extracted from the time series of inter-axis dynamic coupling degree values. The standard deviation of the coupling degree values ​​within the window is calculated using the sliding window method as the fluctuation amplitude to characterize the drastic change in coupling strength. The main frequency component of the coupling degree change is extracted by spectrum analysis as the change frequency to characterize the periodic law of coupling state change.

[0117] A three-input, single-output fuzzy logic system is established. The first input variable is the quantified value of the degradation degree of the performance degradation axis segment, with three fuzzy levels: "slight," "moderate," and "severe." The second input variable is the fluctuation amplitude of the dynamic coupling degree between axes, with three fuzzy levels: "stable," "fluctuating," and "violent." The third input variable is the change frequency of the dynamic coupling degree between axes, with three fuzzy levels: "low frequency," "medium frequency," and "high frequency." The output variable is the system stability evaluation coefficient, with three fuzzy levels: "stable," "warning," and "dangerous." A preset fuzzy rule base contains 27 inference rules, such as: "If the degradation degree is severe and the fluctuation is violent and the change is high frequency, then the system is dangerous"; "If the degradation degree is slight and the fluctuation is stable and the change is low frequency, then the system is stable." The centroid method is used for defuzzification calculation, and finally, the system stability evaluation coefficient in the range of 0-1 is generated.

[0118] In a specific implementation case, a temperature sensor detected a local temperature of 85℃ on a certain drive shaft (ambient temperature 25℃), with a temperature gradient of 15℃ / cm; a vibration sensor detected a vibration acceleration of 2.5g at a frequency of 800Hz (normal value <0.5g). This area was identified as a high-frequency abnormal area, and the calculated degradation degree quantification value was 0.72. At the same time, the standard deviation of the dynamic coupling degree between shafts fluctuated by 0.35 (fluctuation amplitude) within a 10s time window, and the main change frequency was 0.5Hz (change frequency). After inputting into the fuzzy logic system, through rule-based reasoning calculation, the final system stability evaluation coefficient was 0.68, and the system was in an "early warning" state.

[0119] Establish a spatial mapping relationship between the performance degradation shaft segment and the drive shaft control parameters, and construct a dynamic compensation factor matrix. The specific implementation is as follows:

[0120] The spatial location of the degraded shaft segment and its degradation degree quantification value are obtained. The spatial location is obtained by joint calibration of a high-precision encoder and a machine vision system. The geometric center position of the degraded shaft segment in the equipment space is accurately recorded using a three-dimensional coordinate system. The degradation degree quantification value comes from the normalized value in the range of 0-1 generated in the previous processing.

[0121] The compensation priority of the drive shaft control parameters is obtained by multiplying the reciprocal of the Euclidean distance from the spatial location of the degraded shaft segment to the key nodes of the mechanical structure of each drive shaft by the quantified value of the degradation degree of the degraded shaft segment. Then, the drive shafts are sorted from largest to smallest based on their spatial correlation factor values, and the sorting number is used as the compensation priority order for the corresponding drive shaft. The compensation intensity parameter of the drive shaft control parameters is obtained by multiplying the reciprocal of the Euclidean distance from the spatial location of the degraded shaft segment to the key nodes of the mechanical structure of each drive shaft by the quantified value of the degradation degree of the degraded shaft segment. The spatial correlation factor value is then input into a preset intensity mapping function. In practice, this intensity mapping function is implemented using a piecewise linear function. Taking a three-drive shaft system of a certain type of new energy vehicle motor assembly line as an example, when the spatial correlation factor value of a certain degraded shaft segment is in the range of 0-0.2, the compensation intensity is linearly mapped to the range of 0.1-0.3. For example, when the measured spatial correlation factor of drive shaft A is 0.15, the calculated compensation intensity is 0.1+. (0.15-0)×(0.3-0.1) / (0.2-0) = 0.25. When the spatial correlation factor value is in the range of 0.2-0.5, the compensation intensity is linearly mapped to the range of 0.3-0.6. When the measured spatial correlation factor of drive shaft B is 0.35, the calculated compensation intensity is 0.3 + (0.35-0.2)×(0.6-0.3) / (0.5-0.2) = 0.45. When the spatial correlation factor value is in the range of 0.5-1.0, the compensation intensity is linearly mapped to the range of 0.6-1.0. When the measured spatial correlation factor of drive shaft C is 0.75, the calculated compensation intensity is 0.6 + (0.75-0.5)×(1.0-0.6) / (1.0-0.5) = 0.80. This piecewise linear mapping relationship ensures that the larger the spatial correlation factor value, the higher the corresponding compensation intensity parameter value, accurately reflecting the degree of influence of the performance degradation shaft segment on the drive shaft. The compensation intensity of the drive shaft control parameters is determined based on the output value of the intensity mapping function. In specific implementation, a product model of distance weight coefficient and degradation degree is established to generate the compensation priority coefficient and compensation intensity parameter of the drive shaft control parameters corresponding to each drive shaft.

[0122] Based on the quantified value of the degradation degree of the performance degradation shaft segment and the compensation intensity parameter of the drive shaft control parameter, the compensation coefficient of the drive shaft control parameter is generated. The quantified value of the degradation degree of the performance degradation shaft segment is multiplied by the compensation intensity parameter of the drive shaft control parameter, and then multiplied by the system calibration coefficient to obtain the compensation coefficient of the control parameter of each drive shaft.

[0123] First, a global coordinate system is established using the geometric center coordinates of the performance degradation shaft segment as the reference point. The spatial coordinates of all drive shafts are then obtained, forming a spatially distributed topology network. An inverse distance weighted spatial interpolation algorithm is used to calculate the compensation coefficient weight of the drive shaft control parameters at each drive shaft position. The specific implementation process of this algorithm is as follows: The Euclidean distance from each drive shaft's spatial position to the reference point of the performance degradation shaft segment's spatial position is calculated. Weights are calculated based on the Euclidean distance, with drive shafts closer to the reference point receiving greater weights. The weights of all drive shafts are normalized to ensure the sum of the weights is 1. The normalized weights are multiplied by the drive shaft control parameter compensation coefficients to obtain the weighted drive shaft control parameter compensation coefficients for each drive shaft based on the performance degradation shaft segment's spatial position. A mapping table containing the spatial coordinates of each drive shaft and the compensation coefficients of the drive shaft control parameters is established. In this way, the performance degradation shaft segment's spatial position information is used to calculate the spatial weights, and the drive shaft control parameter compensation coefficients are used to generate the final weighted drive shaft control parameter compensation coefficients. Together, they constitute a spatial mapping relationship.

[0124] Based on the topology of multi-axis cooperative motion, the control parameter compensation coefficients of each drive axis are arranged in a matrix according to the spatial mapping relationship. The drive axis number is used as the row index and the control parameter type is used as the column index to construct a two-dimensional compensation coefficient matrix. Each element in the two-dimensional compensation coefficient matrix stores the control parameter compensation coefficient value of the corresponding drive axis, forming a preliminary compensation coefficient matrix.

[0125] The values ​​of each element in the preliminary compensation coefficient matrix are dynamically updated based on the real-time operating status of the system. By monitoring the changing trends of the system stability evaluation coefficient and the dynamic coupling degree between axes in real time, an adaptive filtering algorithm is used to adjust the preliminary compensation coefficient matrix online. When the system stability decreases or the coupling degree is abnormal, the compensation coefficient of the corresponding drive axis is increased according to the preset rules. When the system is running stably, the compensation coefficient is gradually reduced to avoid overcompensation, and finally a dynamic compensation factor matrix with adaptive capability is generated.

[0126] In a specific implementation case, a performance degradation axis segment was detected with spatial coordinates of (150, 75, 0) mm and a degradation degree quantification value of 0.72. The compensation priority coefficients for the three nearest drive axes were calculated to be 0.8, 0.6, and 0.4, respectively, with a compensation intensity parameter of 0.86. The generated control parameter compensation coefficients were 0.69, 0.52, and 0.34, respectively. The established spatial mapping table contained the correspondence between position coordinates and compensation coefficients. The resulting preliminary compensation coefficient matrix was a 3×2 matrix, containing the proportional gain and integral time compensation coefficients of the three drive axes. Based on the real-time operating status, when the system stability evaluation coefficient decreased by 0.15, the compensation coefficient of the corresponding drive axis increased by 20%, ultimately generating a dynamic compensation factor matrix.

[0127] Based on the changing trends of the system stability evaluation coefficients and the dynamic compensation factor matrix, the control parameters of each drive shaft are adjusted to ensure that the adjustment amount meets the quantified value of the system model parameter mismatch and the coordinated matching condition of the dynamic coupling degree between shafts. Specifically, the implementation is as follows:

[0128] The system stability evaluation coefficients are acquired in real time using a data acquisition system with a fixed sampling period (e.g., 100ms). The numerical trend of these coefficients is monitored, and the rate of change of the system stability evaluation coefficients per unit time is calculated using the sliding window method. This rate is calculated by dividing the difference between the current value and the previous value by the sampling period. Simultaneously, the acceleration characteristic of the rate of change is calculated, which is the difference between the rates of change in two consecutive time periods. This quantifies the numerical trend of the system stability evaluation coefficients. The update patterns of each element in the dynamic compensation factor matrix are monitored synchronously. The number of updates for each matrix element per unit time is counted, and the amplitude of these updates is recorded, i.e., the numerical change in each update. The gradient distribution characteristics of the matrix elements are analyzed, and the numerical gradients of the matrix in the row and column directions are calculated. The regions with the most significant gradient changes are identified, thereby comprehensively understanding the update patterns of the elements in the dynamic compensation factor matrix.

[0129] The percentage decrease of the system stability evaluation coefficient per unit time is calculated as the basis for the urgency of adjustment. The gradient distribution of the dynamic compensation factor matrix is ​​analyzed to determine the direction of change. The decrease rate of the system stability evaluation coefficient and the gradient change direction of the dynamic compensation factor matrix are jointly analyzed to generate the adjustment direction command and adjustment amplitude coefficient of each drive shaft control parameter.

[0130] Multiply the quantified value of the system model parameter mismatch by the system reference compensation coefficient to obtain the initial value of the basic compensation amount. Determine the positive or negative sign of the compensation amount according to the adjustment direction of the drive shaft control parameters. Scale the initial value of the basic compensation amount according to the adjustment amplitude coefficient of the adjustment direction of the drive shaft control parameters to obtain the basic compensation amount of the drive shaft control parameters.

[0131] First, a mapping relationship is established between the inter-axis dynamic coupling degree and the coupling compensation weight coefficient of multi-axis cooperative motion. This mapping relationship is implemented using a piecewise linear function: when the inter-axis dynamic coupling degree is in the range of 0-0.3, the coupling compensation weight coefficient of multi-axis cooperative motion is linearly mapped to the range of 0-0.2; when the value is in the range of 0.3-0.7, the coupling compensation weight coefficient of multi-axis cooperative motion is linearly mapped to the range of 0.2-0.6; and when the value is in the range of 0.7-1.0, the coupling compensation weight coefficient of multi-axis cooperative motion is linearly mapped to 0.6-1. The range is 0. Through this mapping relationship, it is ensured that the larger the coupling degree value, the higher the corresponding coupling compensation weight coefficient of the multi-axis coordinated motion, which accurately reflects the strength of the interaction between axes. Secondly, the coupling compensation weight coefficients of all multi-axis coordinated motions are normalized. The sum of the coupling compensation weight coefficients of all drive axis multi-axis coordinated motions is calculated. The coupling compensation weight coefficient of each multi-axis coordinated motion is divided by this sum to ensure that the sum of the normalized axis coordinated motion coupling compensation weight coefficients is 1. Finally, the coupling compensation weight of the multi-axis coordinated motion corresponding to each drive axis is obtained.

[0132] Using a fuzzy rule-based reasoning system, the basic compensation amount of the drive shaft control parameters is taken as the antecedent input, and the coupling compensation weight of multi-axis cooperative motion is taken as the consequent adjustment factor. Reasoning calculation is performed based on a preset cooperative matching rule base. The fuzzy set of the basic compensation amount (negative large, negative medium, negative small, zero, positive small, positive medium, positive large) and the fuzzy set of the coupling weight (low, medium, high) are set. A rule base containing multiple combination conditions is established (such as "if the basic compensation amount is positive large and the coupling weight is high, then the adjustment amount is positive very large"). The centroid method is used to defuzzify and generate the adjustment amount of the drive shaft control parameters that meet the cooperative matching conditions.

[0133] To verify whether the adjustment amount simultaneously meets the compensation requirements for the quantification value of the system model parameter mismatch degree and the optimization requirements for the dynamic coupling degree between axes, the generated adjustment amount is substituted into the system simulation model to calculate the parameter mismatch compensation effect index and the coupling optimization effect index. When both of the above indices reach the preset threshold, it is determined that the cooperative matching condition is met.

[0134] When the matching conditions are met, the final adjustment is applied to the control system of the corresponding drive shaft. By adjusting parameters such as the proportional gain of the current loop and the integral time, the adaptive adjustment of multi-axis coordinated motion is achieved. When the matching conditions are not met, the allocation ratio of the coupling compensation weights is readjusted, and the fuzzy rule reasoning and verification steps are iteratively executed until the matching conditions are met.

[0135] In a specific implementation case, the system stability evaluation coefficient was monitored to decrease at a rate of 0.15 per minute, and the gradient of the dynamic compensation factor matrix mainly pointed towards the drive axes A and B. Initial values ​​for the basic compensation were calculated: +0.35 for drive axis A, +0.28 for drive axis B, and -0.12 for drive axis C. The coupling weights were determined as: 0.4 for axis A, 0.35 for axis B, and 0.25 for axis C. Through fuzzy inference calculations, adjustment values ​​were generated: +0.42 for drive axis A, +0.31 for drive axis B, and -0.08 for drive axis C. Verification results showed that the parameter mismatch compensation effect index reached 0.85 (threshold 0.8), and the coupling optimization effect index reached 0.78 (threshold 0.75), satisfying the cooperative matching condition. Applying the adjustment values ​​to the control system reduced the system response time by 25% and improved synchronization accuracy by 30%.

[0136] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0137] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0138] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0139] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0140] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-drive cooperative control method for high-precision electric drive assembly equipment, characterized by ； S1, real-time acquisition of the motion state data of the multi-drive shaft in the new energy vehicle motor assembly line and the real-time data of the drive shaft control parameter, time sequence synchronization calibration of each shaft data, and generation of a multi-axis cooperative dynamic response sequence; S2, extracting the tracking error features and synchronization error features of each drive shaft from the multi-axis cooperative dynamic response sequence, calculating the inter-axis dynamic coupling degree to identify the system uncertainty mode; S3, based on the frequency domain features of the system uncertainty mode, determining the quantitative value of the system model parameter mismatch degree and the disturbance amplitude probability distribution; S4, collecting and analyzing the spatial distribution characteristics of the temperature field and vibration signal of each drive shaft, positioning the performance degradation shaft section according to the high-frequency abnormal area, and generating the system stability evaluation coefficient combined with the inter-axis dynamic coupling degree; S5, establishing the spatial mapping relationship between the performance degradation shaft section and the drive shaft control parameter, and constructing the dynamic compensation factor matrix; S6, according to the change trend of the system stability evaluation coefficient and the dynamic compensation factor matrix, adjusting the drive shaft control parameter of each drive shaft, so that the adjustment amount meets the cooperative matching condition of the quantitative value of the system model parameter mismatch degree and the inter-axis dynamic coupling degree; Based on the frequency domain features of the system uncertainty mode, the quantitative value of the system model parameter mismatch degree and the disturbance amplitude probability distribution are determined, and the specific steps include: Performing fast Fourier transform on the dynamic response data corresponding to the system uncertainty mode to obtain its frequency domain features; Extracting the power spectral density distribution of a specific frequency band from the frequency domain features, and obtaining the total energy value in the full frequency band by integrating the power spectral density distribution in the full frequency band, calculating the ratio of the characteristic energy value to the total energy value in the full frequency band, multiplying the ratio by the system calibration coefficient, and finally determining the quantitative value of the system model parameter mismatch degree; Analyze the statistical distribution characteristics of the power spectral density distribution amplitude, and use the Gaussian mixture model to fit to obtain the probability distribution model of the disturbance amplitude; According to the fitted probability distribution model, determine the occurrence probability of the disturbance amplitude in different intensity intervals, and generate the disturbance amplitude probability distribution; Combine the quantitative value of the system model parameter mismatch degree and the disturbance amplitude probability distribution to form the frequency domain features of the system uncertainty mode; Collect and analyze the spatial distribution characteristics of the temperature field and vibration signal of each drive shaft, position the performance degradation shaft section according to the high-frequency abnormal area, and generate the system stability evaluation coefficient combined with the inter-axis dynamic coupling degree, the specific steps include: Collecting the temperature field distribution data of each drive shaft through a distributed temperature sensor array, and collecting the vibration signal data of each drive shaft through a vibration acceleration sensor; Perform spatial gradient calculation on the temperature field distribution data to identify temperature mutation areas, and perform frequency spectrum analysis on the vibration signal data to extract high-frequency resonance features; According to the spatial coincidence degree of the temperature mutation area and the high-frequency resonance feature, locate the high-frequency abnormal area and determine the spatial position and degradation degree of the performance degradation shaft section; Quantify the degradation degree of the performance degradation shaft section and generate the degradation degree quantitative value of the performance degradation shaft section; Extract the dynamic change characteristics of the inter-axis dynamic coupling degree value, and calculate the fluctuation amplitude and change frequency; Based on the fuzzy logic system, the degradation degree quantitative value of the performance degradation shaft section is taken as the first input variable, the fluctuation amplitude of the inter-shaft dynamic coupling degree is taken as the second input variable, and the change frequency of the inter-shaft dynamic coupling degree is taken as the third input variable, inference calculation is carried out through the preset fuzzy rule base, and a system stability evaluation coefficient is generated.

2. The multi-drive cooperative control method for high-precision electric drive assembly equipment according to claim 1, characterized in that: Real-time acquisition of the motion state data of the multi-drive shaft in the new energy vehicle motor assembly line and the real-time data of the drive shaft control parameters, time sequence synchronization calibration of the shaft data, generation of a multi-shaft cooperative dynamic response sequence, the specific steps including: Through the high-precision encoder arranged on each drive shaft of the new energy vehicle motor assembly line, real-time acquisition of position data and speed data as motion state data, and through the servo driver, acquisition of current instruction data and torque instruction data as real-time data of the drive shaft control parameters; The torque instruction data is a control signal generated by the multi-shaft motion controller and sent to the servo driver; According to the system control cycle of the multi-shaft motion controller, a fixed time window is set, and under time sequence synchronization calibration, the motion state data of each drive shaft of the new energy vehicle motor assembly line and the real-time data of the drive shaft control parameters are combined into a multi-shaft cooperative dynamic response sequence with unified time coordinates and containing dynamic response data according to the fixed time window.

3. The multi-drive cooperative control method for high-precision electric drive assembly equipment according to claim 2, characterized in that: From the multi-shaft cooperative dynamic response sequence, the tracking error features and synchronization error features of each drive shaft are extracted, and the inter-shaft dynamic coupling degree is calculated to identify the system uncertainty mode, the specific steps including: From the multi-shaft cooperative dynamic response sequence, the actual position data and the instruction position data of each drive shaft are obtained, the tracking error data of each drive shaft is calculated, and the tracking statistical feature quantity of the tracking error data is extracted as the tracking error feature; From the multi-shaft cooperative dynamic response sequence, the relative position relationship data between each drive shaft is obtained, the synchronization error data between each drive shaft is calculated, and the synchronization statistical feature quantity of the synchronization error data is extracted as the synchronization error feature; The instruction position data is the target position information of each drive shaft generated by the multi-shaft motion controller according to the preset motion trajectory planning, and the instruction position data represents the theoretical position coordinates of each drive shaft expected to be reached in each control cycle in the form of a digital signal; According to the tracking error features and synchronization error features of each drive shaft, the inter-shaft coupling strength inter-shaft dynamic coupling degree value is calculated, and the dynamic change feature of the inter-shaft dynamic coupling degree value is extracted; Based on the comparison result of the inter-shaft dynamic coupling degree value and the preset coupling degree value threshold, the current uncertainty mode of the system is identified.

4. The multi-drive cooperative control method for high-precision electric drive assembly equipment according to claim 3, characterized in that: According to the tracking error features and synchronization error features of each drive shaft, the inter-shaft coupling strength inter-shaft dynamic coupling degree value is calculated, the specific steps including: The tracking error features and synchronization error features of each drive shaft are combined into a multi-dimensional feature vector; The covariance matrix between each component in the multi-dimensional feature vector is calculated; Through mathematical transformation, the covariance matrix is decomposed into a combination form of eigenvalues and eigenvectors, then the maximum eigenvalue is identified from all eigenvalues, and the eigenvector corresponding to the maximum eigenvalue is extracted as the principal eigenvector; The weight coefficients of each component in the principal feature vector are obtained by normalizing each component, and then the sum of squares of each weight coefficient is calculated, followed by square root operation on the sum of squares result, and finally the square root result is multiplied by the system calibration coefficient to obtain the final inter-axis dynamic coupling degree value of the inter-axis coupling strength; The inter-axis dynamic coupling degree value is used as the quantitative basis for system uncertainty pattern recognition.

5. The multi-drive cooperative control method for high-precision electric drive assembly equipment according to claim 4, characterized in that: The spatial mapping relationship between the performance degradation shaft section and the drive shaft control parameter is established, and the dynamic compensation factor matrix is constructed, including the following steps: Obtain the spatial position and degradation degree quantization value of the performance degradation shaft section; According to the spatial position and degradation degree quantization value of the performance degradation shaft section, determine the compensation priority and compensation intensity parameters of the corresponding drive shaft control parameter; Based on the degradation degree quantization value of the performance degradation shaft section and the compensation intensity parameters of the drive shaft control parameter, generate the compensation coefficient of the drive shaft control parameter; Establish the spatial mapping relationship between the spatial position of the performance degradation shaft section and the compensation coefficient of the drive shaft control parameter; According to the topology structure of multi-axis cooperative motion, arrange the compensation coefficients of the control parameters of each drive shaft in matrix form according to the spatial mapping relationship to form a preliminary compensation coefficient matrix; According to the real-time running state of the system, dynamically update the element values in the preliminary compensation coefficient matrix to generate a dynamic compensation factor matrix with adaptive ability.

6. The multi-drive cooperative control method for high-precision electric drive assembly equipment according to claim 5, characterized in that: The compensation priority of the drive shaft control parameter is obtained by calculating the reciprocal of the Euclidean distance from the spatial position of the performance degradation shaft section to the key nodes of the mechanical structure of each drive shaft, multiplying it by the degradation degree quantization value of the degradation shaft section to obtain the spatial correlation factor, and then sorting the spatial correlation factor values from large to small to obtain the compensation priority order of the corresponding drive shaft; The compensation intensity parameter of the drive shaft control parameter is obtained by calculating the reciprocal of the Euclidean distance from the spatial position of the performance degradation shaft section to the key nodes of the mechanical structure of each drive shaft, multiplying it by the degradation degree quantization value of the degradation shaft section to obtain the spatial correlation factor, and then inputting the spatial correlation factor value into a pre-set intensity mapping function to determine the compensation intensity of the drive shaft control parameter according to the output value of the intensity mapping function; The compensation coefficient of the drive shaft control parameter is generated by multiplying the degradation degree quantization value of the performance degradation shaft section by the compensation intensity parameter of the drive shaft control parameter, and then multiplying it by the system calibration coefficient.

7. The multi-drive cooperative control method for high-precision electric drive assembly equipment according to claim 6, characterized in that: According to the change trend of the system stability evaluation coefficient and the dynamic compensation factor matrix, adjust the drive shaft control parameters of each drive shaft so that the adjustment amount meets the collaborative matching condition of the quantization value of the system model parameter mismatch degree and the inter-axis dynamic coupling degree, including the following steps: Real-time monitor the value change trend of the system stability evaluation coefficient and the element update rule of the dynamic compensation factor matrix; According to the descending rate of the system stability evaluation coefficient and the gradient change direction of the dynamic compensation factor matrix, determine the adjustment direction and adjustment amplitude of the drive shaft control parameter; Based on the quantization value of the system model parameter mismatch degree, and combined with the adjustment direction and adjustment amplitude of the drive shaft control parameter, calculate the basic compensation amount of the drive shaft control parameter; According to the inter-axis dynamic coupling degree value, determine the coupling compensation weight of multi-axis cooperative motion; The basic compensation amount of the drive shaft control parameter is input as the antecedent, the coupling compensation weight of the multi-axis cooperative motion is adjusted as the consequent adjustment factor, inference calculation is performed based on the preset cooperative matching rule library, and the adjustment amount of the drive shaft control parameter meeting the cooperative matching condition is generated. Whether the adjustment amount meets the compensation demand of the quantized value of the system model parameter mismatch degree and the optimization demand of the dynamic coupling degree between the axes is verified to realize the judgment of the cooperative matching condition. When the cooperative matching condition is met, the final adjustment amount is applied to the control system of the corresponding drive shaft to realize the adaptive adjustment of the multi-axis cooperative motion. When the cooperative matching condition is not met, the coupling compensation weight is adjusted again, and the fuzzy rule inference and verification steps are iteratively executed until the cooperative matching condition is met.

Images