Servo motor speed control method and device based on prediction model optimization

Abstract

This invention discloses a servo motor speed control method and apparatus based on predictive model optimization, relating to the field of motor control technology. The method includes: acquiring gradient data by collecting servo motor operating state data at a high sampling rate using an electrical power device; analyzing the gradient data using a pre-trained predictive model to obtain peak gradient data; outputting a predicted transient adhesion resistance value of the servo motor based on the peak data and a dynamic detection threshold; inputting the predicted resistance value into a feedforward controller to generate a feedforward control command containing a compensation torque; monitoring the servo motor's operating state and adjusting the input parameters of the PID controller according to the operating state using a fuzzy algorithm to output an initial control command; and combining the feedforward control command and the initial control command to generate a speed control command. Through the organic combination of feedforward and feedback, the servo motor can achieve faster, smoother, and more precise speed control under high speed, high load, and complex operating conditions.

Description

Technical Field

[0001] This invention relates to the field of motor control technology, and specifically to a servo motor speed control method and device based on predictive model optimization. Background Technology

[0002] Servo motors are one of the main components of electrical equipment. The speed of servo motors needs to be controlled in real time according to the workload of the servo motor. The existing servo motors mainly use PID controllers to control the speed. When faced with sudden load disturbances, the PID controller adjusts according to its inherent feedback mechanism. It can only adjust after the error occurs in the drive operation of the servo motor, which results in a slow response of the servo motor working system and a large lag. As a result, the control of the PID controller cannot meet the speed regulation requirements of the servo motor in the high-load working environment. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides a solution that combines feedforward and feedback to enable servo motors to achieve faster, smoother, and more precise speed control under high speed, high load, and complex operating conditions. This effectively solves the problems of lag and slow adjustment in traditional PID control strategies, and significantly improves the overall performance and operational stability of the servo motor system.

[0004] This invention provides a servo motor speed control method based on predictive model optimization, the control method comprising:

[0005] The working status data of the servo motor is collected at a high sampling rate by the power equipment, and the second-order time derivative of the working status data is performed to obtain gradient data.

[0006] The gradient data is analyzed by a pre-trained prediction model to obtain the peak data of the gradient data. Based on the peak data and the dynamic detection threshold, the transient adhesion resistance prediction value of the servo motor is output.

[0007] The predicted resistance value is input into the feedforward controller to generate a feedforward control command containing a compensation torque proportional to the amplitude of the predicted transient adhesion resistance value.

[0008] The working status of the servo motor is monitored, and the input parameters of the PID controller are adjusted according to the working status using a fuzzy algorithm. The initial control command is then output through the PID controller.

[0009] By combining the feedforward control command and the initial control command, a speed control command is generated.

[0010] Furthermore, the step of acquiring the servo motor's operating status data at a high sampling rate using electrical equipment, and then performing second-order time derivative calculation on the operating status data to obtain gradient data includes:

[0011] A high-bandwidth Hall effect current sensor is used to detect the operating current data of the servo motor in real time.

[0012] The operating current data is encoded by an incremental encoder, and the encoded operating current data is input into the analysis and processing unit.

[0013] The analysis and processing unit obtains the change data of the servo motor operating current value during the detection time period based on the encoded operating current data.

[0014] Based on the current value change data and the detection time period, a second-order time derivative is performed to output the gradient data of the servo motor's operating current data.

[0015] Furthermore, the step of analyzing the gradient data using a pre-trained prediction model to obtain peak data of the gradient data, and outputting the predicted value of the transient adhesive resistance of the servo motor based on the peak data and a dynamic detection threshold, includes:

[0016] The pre-trained prediction model extracts the current gradient values ​​of the servo motor at several unit time nodes within the detection period based on the gradient data, and extracts several current gradient peaks from the several current gradient values.

[0017] An adaptive adjustment threshold is set according to the workload of the servo motor, and the peak values ​​of several current gradients are detected by the adaptive adjustment threshold. If the peak value of the current gradient is greater than the adaptive adjustment threshold, the transient adhesion resistance prediction value is calculated based on the peak value of the current gradient.

[0018] Furthermore, the step of inputting the predicted resistance value into the feedforward controller to generate a feedforward control command containing a compensation torque proportional to the amplitude of the predicted transient adhesion resistance value includes:

[0019] The torque amplitude of the transient adhesion resistance prediction value is obtained, and a compensation torque is calculated based on the torque amplitude using a preset proportional coefficient. The compensation torque is proportional to the torque amplitude of the transient adhesion resistance prediction value.

[0020] The pulse current with current adjustment amplitude and phase adjustment value is calculated based on the compensation torque, and a feedforward control command for generating the pulsating current is output.

[0021] Furthermore, the monitoring of the servo motor's operating status, the adjustment of the PID controller's input parameters based on the operating status using a fuzzy algorithm, and the output of initial control commands by the PID controller include:

[0022] The operating parameters of the servo motor are monitored and obtained by the status monitoring module, and the operating status of the servo motor is evaluated based on the operating parameters to determine whether transient fluctuations occur.

[0023] If transient fluctuations occur, the parameters of the PID controller are fine-tuned according to the amplitude and frequency of the transient fluctuations.

[0024] If no transient fluctuations occur, the input parameters of the PID controller are maintained.

[0025] Furthermore, the step of fine-tuning the parameters of the PID controller based on the amplitude and frequency of the transient fluctuations includes:

[0026] Acquire the change data of the amplitude and frequency of the transient fluctuation, and set the adjustment mode according to the correspondence between the input parameters of the PID controller and the change data of the transient fluctuation;

[0027] The input parameters of the PID controller are fine-tuned according to a preset adjustment step.

[0028] Furthermore, the monitoring of the servo motor's operating status, adjusting the PID controller's input parameters based on the operating status using a fuzzy algorithm, and outputting initial control commands through the PID controller also includes:

[0029] An adaptive adjustment rule is introduced into the fuzzy algorithm to obtain the change range of the working parameters of the servo motor under real-time working conditions.

[0030] The system detects whether the change amplitude is greater than a preset detection threshold. If the change amplitude is greater than the preset detection threshold, the PID parameters are adjusted based on the comparison data between the change amplitude and the preset detection threshold.

[0031] If the detected change is less than the preset detection threshold, the PID parameters of the PID controller in the current state are maintained.

[0032] Furthermore, the initial control command output via the PID controller includes:

[0033] The drive current and actual speed data of the servo motor are collected at a microsecond-level sampling rate;

[0034] High-frequency harmonic analysis is performed on the collected drive current and actual rotation speed data to extract the amplitude and phase of the high-frequency harmonic components;

[0035] The harmonic amplitude threshold and phase threshold are dynamically adjusted according to the operating status of the servo motor.

[0036] Determine whether the amplitude of the extracted high-frequency harmonic component exceeds the dynamically adjusted harmonic amplitude threshold, and determine whether the phase of the extracted high-frequency harmonic component exceeds the dynamically adjusted phase threshold;

[0037] If the amplitude of the high-frequency harmonic component exceeds the dynamically adjusted harmonic amplitude threshold, or if the phase of the high-frequency harmonic component exceeds the dynamically adjusted phase threshold, it is identified as a high-frequency micro-vibration anomaly.

[0038] If it is identified as a high-frequency micro-vibration anomaly, a sequence of instantaneous reverse-phase current pulses is generated;

[0039] The instantaneous reverse current pulse sequence is injected into the initial control command output by the PID controller.

[0040] Furthermore, the dynamic adjustment method for the harmonic amplitude threshold and phase threshold is as follows:

[0041] The operating status data of the load device of the servo motor is detected, and the temperature and vibration acceleration data of the mechanical joint of the load device are obtained from the operating status data;

[0042] Trend analysis was performed on the collected temperature and vibration acceleration data to obtain the cumulative degree of mechanical wear;

[0043] Based on the cumulative degree of mechanical wear, calculate the drift of the harmonic amplitude threshold and phase threshold;

[0044] The drift amount is superimposed on the preset harmonic amplitude threshold and phase threshold to obtain the updated harmonic amplitude threshold and the updated phase threshold.

[0045] The present invention also provides a servo motor speed control device based on predictive model optimization, the control device comprising:

[0046] Gradient data calculation module: used to collect the working status data of the servo motor at a high sampling rate through the power equipment, and to perform second-order time derivative on the working status data to obtain gradient data;

[0047] Prediction module: used to analyze the gradient data through a pre-trained prediction model, obtain the peak data of the gradient data, and output the transient adhesion resistance prediction value of the servo motor based on the peak data and the dynamic detection threshold.

[0048] Feedforward compensation module: Inputs the predicted resistance value into the feedforward controller and generates a feedforward control command containing a compensation torque proportional to the amplitude of the predicted transient adhesion resistance value;

[0049] Control calculation module: monitors the working status of the servo motor, adjusts the input parameters of the PID controller according to the working status using a fuzzy algorithm, and outputs the initial control command through the PID controller;

[0050] Control output module: Combines the feedforward control command and the initial control command to generate a speed control command.

[0051] This invention provides a servo motor speed control method and device based on predictive model optimization. By introducing a predictive model to predict the transient adhesion resistance during servo motor operation, and generating a feedforward compensation torque based on the prediction results, the lag of traditional PID control is effectively overcome. By acquiring operating state data at a high sampling rate and performing second-order time derivative, the invention can more sensitively capture minute changes in the operating state of the servo motor, providing high-quality input data for predicting transient adhesion resistance.

[0052] By inputting the predicted transient adhesion resistance value into the feedforward controller, a feedforward control command with a compensating torque proportional to the magnitude of the predicted value is generated, enabling early compensation for potential disturbances. This means that the system begins to apply a counter-torque to cancel out the transient adhesion resistance before it actually occurs, thereby significantly reducing speed fluctuations and overshoot.

[0053] Furthermore, by dynamically adjusting the input parameters of the PID controller based on the servo motor's operating state using a fuzzy algorithm, the PID controller can better adapt to different operating conditions and load variations, further improving the robustness and accuracy of the control system. This organic combination of feedforward and feedback enables the servo motor to achieve faster, smoother, and more precise speed control under high speed, high load, and complex operating conditions, effectively solving the problems of lag and slow adjustment in traditional PID control strategies, and significantly improving the overall performance and operational stability of the servo motor system. Attached Figure Description

[0054] Figure 1 This is a flowchart of the servo motor speed control method based on prediction model optimization in an embodiment of the present invention;

[0055] Figure 2 This is a schematic diagram of a servo motor speed control device based on prediction model optimization in an embodiment of the present invention. Detailed Implementation

[0056] 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.

[0057] Example 1:

[0058] Figure 1 A flowchart of a servo motor speed control method based on predictive model optimization in an embodiment of the present invention is shown. The control method includes:

[0059] S11: Collect the working status data of the servo motor at a high sampling rate through the power equipment, and perform second-order time derivative on the working status data to obtain gradient data;

[0060] A high-bandwidth Hall effect current sensor is used to detect the operating current data of the servo motor in real time. The high-bandwidth Hall effect current sensor is a device that can measure current in real time and accurately. Its high bandwidth characteristics ensure the ability to capture rapidly changing signal components in the operating current data of the servo motor, thereby providing high-quality raw data for subsequent transient analysis.

[0061] The operating current data is encoded by an incremental encoder, and the encoded operating current data is input into the analysis and processing unit. An incremental encoder can be understood as a device that converts analog signals into digital signals. It encodes the operating current data, converts it into a format that can be processed by a digital system, and transmits it to the analysis and processing unit.

[0062] The analysis and processing unit acquires the servo motor operating current value change data within the detection time period based on the encoded operating current data. Specifically, the analysis and processing unit receives the encoded operating current data and extracts the servo motor operating current value change data within a specific detection time period from this data according to a preset algorithm logic. This unit can be a microcontroller, digital signal processor (DSP), or application-specific integrated circuit (ASIC), and its purpose is to perform preliminary processing and calculation on the raw data.

[0063] Based on the current value change data and the detection time period, a second-order time derivative is performed to output the gradient data of the servo motor's operating current. The rate of change of the servo motor's operating current with time is differentiated twice. The first derivative represents the rate of change of current, i.e., the speed of current change, while the second derivative represents the rate of change of the current's amplitude. The current data changes obtained through second-order time derivatives can more sensitively capture minute and transient changes in the servo motor's operating state, thus enabling the output of the gradient value of the servo motor's current value change data.

[0064] By employing a high-bandwidth Hall effect current sensor to monitor the servo motor's operating current data in real time, the accuracy and real-time performance of the raw data are ensured. Subsequently, this high-precision current data is digitized using an incremental encoder, enabling effective processing by the analysis and processing unit. The analysis and processing unit is responsible for extracting key current value changes from the coded data. Finally, by performing second-order time derivatives on these changes, the transient trends of the servo motor's operating current are revealed, particularly the minute fluctuations related to transient adhesive resistance. This refined data acquisition and processing method allows the system to identify potential adhesive resistance earlier and more accurately, providing high-quality input for subsequent predictive models and thus optimizing the servo motor's speed control. This provides a solid data foundation for subsequent predictive model optimization, significantly improving the response speed and control accuracy of the servo motor's speed control and effectively suppressing speed fluctuations caused by transient adhesive resistance.

[0065] S12: Analyze the gradient data through a pre-trained prediction model to obtain the peak data of the gradient data, and output the transient adhesion resistance prediction value of the servo motor based on the peak data and the dynamic detection threshold.

[0066] The pre-trained prediction model extracts the current gradient values ​​of the servo motor at several unit time nodes within the detection period based on the gradient data, and extracts several current gradient peaks from these current gradient values. The pre-trained prediction model is configured to perform time-series analysis on the input gradient data. The pre-trained prediction model can employ deep learning architectures such as recurrent neural networks, long short-term memory networks, or gated recurrent units, or signal processing methods based on wavelet transform and Fourier transform to arrange several current gradient values ​​in the gradient data stream in chronological order, forming a gradient data curve. Local maxima points with significant variation characteristics are identified from the continuous gradient data stream, i.e., the peak points of the peaks in the curve. Through the pre-trained prediction model, key feature points related to transient adhesive resistance events can be accurately identified in the original gradient data.

[0067] An adaptive adjustment threshold is set according to the workload of the servo motor, and the peak values ​​of several current gradients are detected by the adaptive adjustment threshold. If the peak value of the current gradient is greater than the adaptive adjustment threshold, the transient adhesion resistance prediction value is calculated based on the peak value of the current gradient.

[0068] In this embodiment, the detection threshold required for the peak current gradient analysis of the servo motor is not a fixed value. It needs to be adaptively adjusted according to the working state of the servo motor. By dynamically adjusting the detection threshold, the authenticity and accuracy of the peak current gradient analysis can be improved. For example, when the servo motor is under heavy load, the allowable peak current gradient can be increased to meet the speed drive requirements of the load condition. At this time, the adaptive adjustment threshold will be increased accordingly to avoid misjudging normal load fluctuations as adhesion resistance.

[0069] Furthermore, the formula for calculating the adaptive adjustment threshold is as follows:

[0070] ;

[0071] in, To adaptively adjust the threshold, Let K be the standard deviation of the current gradient fluctuation under normal operating conditions of the servo motor, and K be the adjustment coefficient. The adjustment coefficient K is the ratio of the working load current of the servo motor to the working current value at the rated working frequency of the servo motor.

[0072] Furthermore, by using the ratio of the servo motor's operating load current to its rated current as the adjustment coefficient for the adaptive adjustment threshold, a mapping relationship can be established between the adaptive adjustment threshold and the servo motor's operating load. The adaptive adjustment threshold can be adaptively adjusted according to the servo motor's operating load, thereby improving the accuracy of current gradient peak detection processing.

[0073] Under light-load operating conditions of the servo motor, the adaptive adjustment threshold can be appropriately lowered to improve the system's sensitivity to minor adhesive resistance events of the servo motor under light-load conditions. This allows the peak detection process to better adapt to the characteristics of the servo motor under different operating conditions, reducing false alarms and missed alarms, and improving the accuracy of peak detection. Consequently, it enhances the reliability of the servo motor speed control system and improves the precision of servo motor speed control.

[0074] Furthermore, the pre-trained prediction model can accurately identify current gradient peaks related to adhesive resistance from the raw gradient data, laying the foundation for subsequent accurate judgment. Based on this, by dynamically adjusting the detection threshold according to the servo motor's workload, the system can distinguish between current fluctuations caused by normal load changes and abnormal peaks caused by transient adhesive resistance. By setting an adaptive detection mechanism, the prediction process becomes more intelligent and robust, significantly improving the accuracy and anti-interference capability for identifying transient adhesive resistance events.

[0075] Predictive analysis of servo motor operating current gradient peak data using a pre-trained prediction model with adaptively adjusted detection thresholds significantly improves the accuracy and reliability of servo motor transient adhesive resistance prediction. Compared to methods relying solely on a single dynamic detection threshold, this application refines the extraction of current gradient peaks and incorporates adaptive threshold adjustment based on workload, effectively avoiding false alarms or missed alarms under different operating conditions. Consequently, the prediction model can more accurately identify real transient adhesive resistance events, providing more precise compensation torque information to the feedforward controller. This enables the servo motor to achieve smoother and more precise speed control when facing transient adhesive resistance, improving the system's dynamic response performance and operational stability.

[0076] S13: Input the predicted resistance value into the feedforward controller to generate a feedforward control command containing a compensation torque proportional to the amplitude of the predicted transient adhesion resistance value;

[0077] The torque amplitude of the predicted transient adhesion resistance is obtained, and a compensation torque is calculated based on the torque amplitude using a preset proportional coefficient. The compensation torque is directly proportional to the torque amplitude of the predicted transient adhesion resistance, ensuring that the magnitude of the compensation torque directly reflects the magnitude of the predicted transient adhesion resistance. The preset proportional coefficient can be calibrated and optimized based on the specific characteristics of the servo motor, load conditions, and control accuracy requirements to ensure the effectiveness and accuracy of the compensation.

[0078] Specifically, a pulse current with current adjustment amplitude and phase adjustment values ​​is calculated based on the compensation torque. This pulse current is the actual execution form of the feedforward control command, and its amplitude and phase are precisely adjusted to generate the required compensation torque. The current adjustment amplitude is used to control the intensity of the pulse current, thereby affecting the magnitude of the compensation torque; the phase adjustment value is used to ensure the synchronization of the pulse current with the rotor position or speed of the servo motor, so as to apply the compensation torque at the correct time, thereby achieving precise speed control.

[0079] Therefore, a feedforward control command is output to generate the pulsating current. This command contains detailed parameters of the pulse current, such as amplitude, phase, frequency, and duration, which are used to drive the servo motor to counteract transient adhesion resistance, thereby maintaining the stability and accuracy of the servo motor speed.

[0080] The pulse current with current adjustment amplitude and phase adjustment value is calculated based on the compensation torque, and a feedforward control command for generating the pulsating current is output.

[0081] By converting the predicted transient adhesion resistance into a specific compensation torque, and further refining it into a feedforward control command in the form of a pulsed current with current and phase adjustment values, precise and real-time feedforward compensation for the transient adhesion resistance of the servo motor is achieved. This mechanism allows the system to apply a counteracting control effect before the adhesion resistance actually occurs, thereby effectively avoiding or significantly reducing the adverse effects of adhesion resistance on the speed stability of the servo motor. Through meticulous calculation and adjustment of the torque amplitude, proportional coefficient, current amplitude, and phase, the accuracy and timeliness of the compensation torque are ensured, enabling the feedforward control command to be precisely applied to the servo motor.

[0082] S14: Monitor the working status of the servo motor, adjust the input parameters of the PID controller according to the working status using a fuzzy algorithm, and output the initial control command through the PID controller;

[0083] Specifically, the operating parameters of the servo motor are monitored and acquired through the status monitoring module, and the operating status of the servo motor is evaluated based on the operating parameters to determine whether transient fluctuations occur. The status monitoring module can be a hardware or software unit integrated into the servo motor control system. The status monitoring module is used to collect and process the operating parameters of the servo motor in real time during operation, such as the main operating parameters under the working conditions, such as speed, current, voltage, temperature, and vibration. The operating parameters are quantitative indicators reflecting the current operating status of the servo motor.

[0084] Based on the changes in the operating parameters, i.e. the slope of the changes in the operating parameters within the corresponding time period, it is possible to assess whether the operating state of the servo motor exhibits transient fluctuations. By performing real-time analysis on the collected operating parameters and detecting abnormal amplitude or frequency components through signal processing techniques, it is possible to determine whether there are short-term, rapid, and non-periodic state deviations, i.e., transient fluctuations. The signal processing techniques can be Fourier transform, wavelet analysis, or other similar techniques.

[0085] If transient fluctuations occur, the parameters of the PID controller are fine-tuned according to the amplitude and frequency of the transient fluctuations. For transient fluctuations occurring within the detection period, the parameters of the PID controller are fine-tuned so that the input parameters of the PID controller can meet the dynamic detection requirements of the servo motor.

[0086] If no transient fluctuations occur, the input parameters of the PID controller are maintained.

[0087] The amplitude of transient fluctuations refers to the maximum extent to which the fluctuation deviates from the normal steady-state value, while the frequency reflects how quickly the fluctuation changes. Quantitative analysis of these characteristics can provide a basis for adjusting PID parameters. Fine-tuning refers to making small, precise adjustments to the original PID parameters to avoid system instability caused by large adjustments. For example, when high-amplitude, high-frequency transient fluctuations are detected, it may be necessary to appropriately reduce the proportional gain (Kp) to suppress overshoot, while increasing the derivative gain (Kd) to improve response speed and suppress oscillations.

[0088] Conversely, if the servo motor's operating status is assessed by the status monitoring module and no transient fluctuations are observed, indicating that the system is in a relatively stable operating state, the input parameters of the PID controller will remain unchanged. This is to avoid unnecessary parameter adjustments when the system is stable, thereby maintaining system stability and control accuracy.

[0089] This application's solution addresses the problem of insufficient response in traditional fuzzy PID control to rapidly changing transient conditions by introducing a real-time monitoring and evaluation mechanism for transient fluctuations in the servo motor. Specifically, the state monitoring module continuously acquires the servo motor's operating parameters and determines the presence of transient fluctuations. Once a transient fluctuation is identified, the system no longer relies solely on preset fuzzy rules for general adjustments but further analyzes the amplitude and frequency of the fluctuation. This refined fluctuation characteristic analysis makes PID parameter adjustments more targeted. For example, different fine-tuning strategies can be employed for transient fluctuations of different amplitudes or frequencies, ensuring that the PID controller can quickly and accurately adapt to the current operating conditions, effectively suppressing fluctuations and avoiding a decline in control performance. When the system is in a stable state, maintaining the PID parameters unchanged avoids unnecessary adjustments that introduce new disturbances, ensuring stable system operation.

[0090] Specifically, fine-tuning the parameters of the PID controller based on the amplitude and frequency of the transient fluctuations includes:

[0091] Acquire the change data of the amplitude and frequency of the transient fluctuation, and set the adjustment mode according to the correspondence between the input parameters of the PID controller and the change data of the transient fluctuation;

[0092] The input parameters of the PID controller are fine-tuned according to a preset adjustment step.

[0093] Specifically, acquiring data on the amplitude and frequency changes of transient fluctuations refers to continuously monitoring the operating status of the servo motor. When a transient fluctuation is detected, not only is its current amplitude and frequency recorded, but the trend or rate of change of these amplitudes and frequencies over time is further analyzed. For example, the increment or decrement of the fluctuation amplitude per unit time, as well as the frequency drift, can be calculated. This data provides deeper information about the dynamic characteristics of transient fluctuations, helping to more accurately determine the severity and development trend of the fluctuations.

[0094] In this context, setting the adjustment method based on the correspondence between the input parameters of the PID controller and the changes in instantaneous fluctuations can be understood as establishing a mapping rule or lookup table. This rule or lookup table defines the influence of different types of instantaneous fluctuation data, such as the rapid increase in amplitude and periodic changes in frequency, on the proportional (P), integral (I), and derivative (D) parameters of the PID controller.

[0095] Furthermore, the correspondence between the transient fluctuations and the input parameters can be trained and optimized based on empirical knowledge, simulation results, or machine learning models. This ensures that the adjustment of the PID parameters can accurately respond to the specific characteristics of the transient fluctuations, thereby achieving the optimal control effect.

[0096] Specifically, in practical applications, fine-tuning the input parameters of the PID controller is performed according to a preset adjustment step size. This means that after determining the adjustment direction, the input parameters of the PID controller are gradually adjusted with a pre-set, small increment or decrement to achieve fine-tuning. This adjustment is based on transient fluctuations, preventing excessively large changes in input parameters that could reduce the reliability of the PID controller. The adjustment step size can be set according to the system's sensitivity, control accuracy requirements, and the characteristics of transient fluctuations.

[0097] The fine-tuning methods for the input parameters of the PID controller are shown in the table below:

[0098] Table 1: PID Controller Input Parameter Fine-Tuning Direction Table

[0099]

[0100] The condition monitoring module collects operating parameters of the servo motor in real time, such as speed, current, and vibration acceleration. When an external load impact occurs, the speed data will show a rapid drop, the current data will show a momentary increase, and the vibration acceleration may also show a peak. The condition monitoring module detects these rapid changes in parameters exceeding the normal range through a preset algorithm, thereby assessing that the servo motor's current operating state has experienced transient fluctuations.

[0101] Furthermore, the system analyzes the amplitude and frequency of this transient fluctuation. For example, the speed drop is 500 rpm, and the oscillation frequency is 10 Hz. Based on these characteristics, the fuzzy algorithm, combined with preset fine-tuning rules, adjusts the parameters of the PID controller.

[0102] When a high-amplitude, medium-frequency transient fluctuation is detected, the system can fine-tune the proportional gain Kp by 5% to suppress overshoot; at the same time, it can fine-tune the derivative gain Kd by 10% to accelerate the suppression of fluctuations. The fine-tuning step size of each input parameter of the PID controller is set according to the experience of the staff, so that the PID controller can dynamically adjust the corresponding input parameters according to the occurrence of transient fluctuations by a preset step size.

[0103] In this way, the PID controller can quickly adjust its control strategy, generate more suitable initial control commands, thereby rapidly pulling the servo motor speed back to the set value and effectively suppressing subsequent oscillations, ensuring that the system can quickly recover stable operation after experiencing transient disturbances. If the servo motor's operating parameters remain within a stable range during other time periods and no transient fluctuations are detected, the PID controller parameters will remain unchanged to avoid unnecessary adjustments that could affect system stability.

[0104] In some embodiments described above, a method is proposed to fine-tune the parameters of the PID controller based on the amplitude and frequency of transient fluctuations in the servo motor. However, in its implementation, a lack of specific adjustment strategies and mechanisms may lead to poor adjustment results or inefficient adjustment processes. For example, if adjustments are made solely based on the current amplitude and frequency of the transient fluctuations, while ignoring their dynamic trends, it may be unable to respond promptly and accurately to rapidly changing operating conditions, thereby affecting the control accuracy and stability of the servo motor.

[0105] Specifically, the monitoring of the servo motor's operating status, adjusting the PID controller's input parameters based on the operating status using a fuzzy algorithm, and outputting initial control commands through the PID controller further includes:

[0106] Introducing adaptive adjustment rules into fuzzy algorithms allows for the acquisition of the variation range of operating parameters under real-time working conditions of the servo motor. This involves adding a mechanism to the fuzzy adjustment logic of the PID controller parameters that enables self-optimization based on the real-time operating status of the servo motor. This rule can be constructed based on historical data, expert experience, or machine learning methods. Its purpose is to enable the fuzzy algorithm to dynamically adjust its membership function, fuzzy rules, or output scaling factor to better adapt to constantly changing operating conditions.

[0107] Furthermore, obtaining the variation range of the operating parameters of the servo motor in real-time operation refers to continuously collecting various operating parameters of the servo motor at the current moment through the status monitoring module, such as speed, current, temperature, vibration, etc., and calculating the amount or rate of change of these parameters within a certain time window. This variation range can reflect the activity level or stability of the servo motor's operating status.

[0108] Specifically, the system detects whether the change amplitude exceeds a preset detection threshold. If the change amplitude exceeds the preset detection threshold, the PID parameters are adjusted based on the comparison between the change amplitude and the preset detection threshold. The preset detection threshold can be set according to the design requirements of the servo motor, the application scenario, and the control accuracy requirements, and is used to distinguish between normal fluctuations and significant changes that require parameter adjustment. Adjusting the PID parameters based on the comparison between the change amplitude and the preset detection threshold means that when the change amplitude of the operating parameters exceeds the preset threshold, it indicates that the operating state of the servo motor has changed significantly, and at this time, the parameters of the PID controller (such as the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd) need to be adjusted.

[0109] The PID controller can calculate a more precise adjustment amount through the adaptive adjustment rules of the fuzzy algorithm. The larger the change, the larger the adjustment amount. The adjustment coefficient of the relevant input parameters is adjusted through the adaptive adjustment rules, that is, the fine-tuning steps of the corresponding input parameters are adjusted, so that the adjustment direction and adjustment range are more accurate and conform to the actual working adjustment requirements of the servo motor.

[0110] If the detected change is less than the preset detection threshold, the PID parameters of the PID controller are maintained in the current state. This means that when the change in the operating parameters is within the preset threshold, the operating state of the servo motor is considered relatively stable, and there is no need to frequently adjust the PID parameters to avoid unnecessary control disturbances, thereby improving the stability of PID control.

[0111] By acquiring data on the amplitude and frequency changes of transient fluctuations, a more comprehensive and in-depth understanding of the dynamic changes in the servo motor's operating state can be achieved. It is precisely because of this detailed data that a more intelligent and adaptive adjustment strategy can be established based on the correspondence between the PID controller input parameters and the transient fluctuation data. On this basis, by fine-tuning the PID controller's input parameters according to preset adjustment steps, the smoothness and safety of the parameter adjustment process are ensured, avoiding system oscillations or instability that may be caused by sudden parameter changes. This step-by-step, fine-tuning mechanism effectively solves the problems of blindness or lag in parameter adjustment that may exist in traditional methods, enabling the PID controller to adapt to the servo motor's operating needs under transient fluctuations more promptly and accurately.

[0112] During the operation of a servo motor, its operating parameters such as current, speed, or temperature change in real time due to load variations or ambient temperature fluctuations. Traditional fuzzy PID control may only respond to these changes based on preset fuzzy rules. However, the fuzzy algorithm proposed in this invention, which introduces adaptive adjustment rules, can make more precise judgments based on the magnitude of these operating parameter changes. Through the adaptive adjustment mechanism, the PID controller can determine the intensity and frequency of adjustment based on the "activity" of the actual operating conditions, thereby ensuring control accuracy while avoiding oversensitivity or response lag.

[0113] For example, when the load on the servo motor suddenly increases, causing a significant increase in the fluctuation range of the operating current within a short period of time, exceeding the preset detection threshold, the adaptive adjustment rule dynamically adjusts the proportional coefficient Kp and integral coefficient Ki of the PID controller based on the comparison data between the current fluctuation range and the threshold. This allows the controller to respond to load changes more quickly and suppress speed drops. Conversely, if the fluctuation range of the operating parameters remains below the preset detection threshold, it indicates that the system is operating smoothly. In this case, the PID parameters will remain unchanged, avoiding unnecessary adjustments and thus improving system stability.

[0114] Specifically, a scheme involves adjusting the input parameters of a PID controller based on the operating state using a fuzzy algorithm, and then outputting initial control commands through the PID controller. However, in its implementation, even with adjusted PID parameters, the servo motor may still generate high-frequency micro-vibrations—small in amplitude but high in frequency—under specific operating conditions, such as high-speed operation, frequent start-stop, or sudden load changes. These micro-vibrations are difficult to completely eliminate using conventional PID adjustment. These micro-vibrations affect the positioning accuracy, operational stability, and lifespan of the servo motor. Traditional PID controllers, due to limitations in response speed and bandwidth, often struggle to achieve precise and real-time suppression of these high-frequency, transient micro-vibrations, potentially leading to decreased control accuracy or compromised system stability.

[0115] The initial control command output through the PID controller includes:

[0116] The system acquires drive current and actual speed data of the servo motor at a microsecond-level sampling rate, using high-precision, high-sampling-frequency sensors, such as high-speed current sensors and high-resolution encoders, to monitor the electrical and mechanical state of the servo motor in real time over an extremely short timescale. The aim is to capture high-frequency dynamic information that is easily overlooked under traditional low-sampling-rate conditions, providing sufficiently detailed raw data for subsequent high-frequency harmonic analysis.

[0117] High-frequency harmonic analysis is performed on the acquired drive current and actual rotational speed data to extract the amplitude and phase of the high-frequency harmonic components. The time-domain signal is then converted into a frequency-domain signal using Fourier transform, wavelet analysis, or other spectral analysis techniques, thereby identifying the high-frequency harmonic components in the drive current and actual rotational speed that are related to micro-vibrations. The amplitude and phase of these harmonic components can characterize the intensity and temporal characteristics of the micro-vibrations.

[0118] Based on the servo motor's operating status, the harmonic amplitude and phase thresholds are dynamically adjusted. Since the normal vibration characteristics and permissible harmonic levels of a servo motor vary under different loads, speeds, or temperatures, the system dynamically adjusts the harmonic amplitude and phase thresholds used to identify micro-vibration anomalies by real-time monitoring of the servo motor's operating parameters, such as load, speed, and temperature, and combining this with a preset empirical model or adaptive algorithm. The aim is to improve the accuracy and robustness of micro-vibration anomaly identification and avoid misjudgments or missed detections.

[0119] To determine whether the amplitude of the extracted high-frequency harmonic components exceeds the dynamically adjusted harmonic amplitude threshold, and whether the phase of the extracted high-frequency harmonic components exceeds the dynamically adjusted phase threshold, the amplitude and phase of the high-frequency harmonic components can be extracted using Fourier transform, wavelet analysis, or other spectral analysis techniques. This converts the time-domain signal into a frequency-domain signal, thereby identifying the high-frequency harmonic components related to micro-vibrations present in the drive current and actual rotational speed. The amplitude and phase of the extracted harmonic components characterize the intensity and temporal characteristics of the micro-vibrations, enabling analysis of the amplitude and phase of the extracted high-frequency harmonic components and detection of the vibration of the load equipment of the servo motor.

[0120] If the amplitude of the high-frequency harmonic component exceeds the dynamically adjusted harmonic amplitude threshold, or if the phase of the high-frequency harmonic component exceeds the dynamically adjusted phase threshold, it is identified as an abnormal high-frequency micro-vibration. If any indicator exceeds its corresponding threshold, it indicates that the servo motor may have abnormal high-frequency micro-vibration. By using dynamically adjusted thresholds for detection and analysis, the system can more accurately judge the high-frequency vibration of the servo motor's load equipment, which can better meet the actual operating requirements of the servo motor's load equipment.

[0121] If a high-frequency micro-vibration anomaly is identified, a sequence of instantaneous reverse-phase current pulses is generated. The amplitude, width, and timing of the instantaneous reverse-phase current pulse sequence are precisely calculated based on the amplitude and phase of the detected micro-vibration to ensure that it can generate an instantaneous torque that is opposite in direction and similar in magnitude to the micro-vibration.

[0122] The instantaneous reverse-phase current pulse sequence is injected into the initial control command output by the PID controller. This injection superimposes an instantaneous reverse-phase current pulse to counteract micro-vibrations onto the conventional control command generated by the PID controller, thereby actively suppressing micro-vibrations in a very short time.

[0123] This invention, by acquiring drive current and actual speed data of a servo motor at a microsecond-level sampling rate, can capture high-frequency dynamic details that are difficult to detect using traditional control methods. It is precisely this high-precision real-time data that makes high-frequency harmonic analysis of the drive current and actual speed data possible, thereby accurately extracting the amplitude and phase of the high-frequency harmonic components characterizing micro-vibration. Based on this, by dynamically adjusting the harmonic amplitude and phase thresholds according to the servo motor's operating state, the accuracy of micro-vibration anomaly identification is ensured, avoiding misjudgments or missed judgments caused by fixed thresholds under different operating conditions. Once a high-frequency micro-vibration anomaly is identified, the system can immediately generate a sequence of instantaneous current pulses in the opposite phase to the micro-vibration and inject it into the initial control command output by the PID controller. This direct and instantaneous opposite-phase current injection generates a compensating torque in the opposite direction to the torque generated by the micro-vibration, effectively canceling it out at the moment of occurrence and significantly reducing the amplitude of the micro-vibration.

[0124] When the load device driven by the servo motor moves at high speed, due to residual vibration of the mechanical structure or changes in load inertia, a high-frequency micro-vibration with a frequency of about 500Hz and a small amplitude is generated. The control system will sample the working parameters at a microsecond-level sampling rate and continuously collect the drive current and actual speed data of the servo motor.

[0125] The acquired data can be fed into a Digital Signal Processor (DSP). The DSP uses Fast Fourier Transform to perform real-time high-frequency harmonic analysis on the data. The analysis results show that harmonic components exist in the drive current, with amplitudes and phases of A and φ, respectively. Simultaneously, the system dynamically adjusts the harmonic amplitude and phase thresholds based on the current load and speed of the servo motor.

[0126] The control system precisely calculates and generates a sequence of instantaneous, opposite-phase current pulses based on the detected harmonic amplitude A and phase φ. This pulse sequence is designed to produce an instantaneous torque of opposite direction and equal magnitude when the micro-vibration reaches its peak. This instantaneous, opposite-phase current pulse sequence is superimposed on the initial control command output by the PID controller, acting together on the servo motor. In this way, high-frequency micro-vibrations are actively suppressed in their infancy, thereby ensuring rapid stabilization and high-precision positioning of the robotic arm after high-speed movement.

[0127] Specifically, the dynamic adjustment method for the harmonic amplitude threshold and phase threshold is as follows:

[0128] The operating status data of the load device of the servo motor is detected, and the temperature and vibration acceleration data of the mechanical joint of the load device are obtained from the operating status data;

[0129] Trend analysis was performed on the collected temperature and vibration acceleration data to obtain the cumulative degree of mechanical wear;

[0130] Based on the cumulative degree of mechanical wear, calculate the drift of the harmonic amplitude threshold and phase threshold;

[0131] The drift amount is superimposed on the preset harmonic amplitude threshold and phase threshold to obtain the updated harmonic amplitude threshold and the updated phase threshold.

[0132] Specifically, the load device refers to a mechanical system connected to and driven by a servo motor, such as a robot joint, a CNC machine tool table, or an actuator on an automated production line. The temperature and vibration acceleration data of the mechanical joint are key indicators reflecting the degree of mechanical wear. The trend analysis can employ methods such as time series analysis, regression analysis, or machine learning to predict the cumulative trend of mechanical wear by modeling historical data. The drift amount refers to a correction value calculated based on the cumulative degree of mechanical wear to a preset threshold. Its purpose is to enable the harmonic amplitude threshold and phase threshold to more accurately adapt to the current actual wear condition of the load device. By superimposing this drift amount onto the preset threshold, a more adaptive updated harmonic amplitude threshold and updated phase threshold can be obtained.

[0133] By incorporating consideration of the cumulative mechanical wear of the load equipment, the limitations of adjusting thresholds solely based on instantaneous operating conditions are overcome. By detecting and analyzing the temperature and vibration acceleration data of the mechanical joints of the load equipment, the cumulative degree of mechanical wear can be accurately assessed. Accumulated mechanical wear can cause servo motors to generate minute vibrations with specific harmonic characteristics even under normal operating conditions. Without considering this cumulative wear, preset thresholds or thresholds adjusted only based on instantaneous conditions may not effectively distinguish between vibrations caused by normal wear and genuine abnormal high-frequency micro-vibrations. By calculating the drift of the harmonic amplitude and phase thresholds based on the cumulative mechanical wear and adding this drift to the preset thresholds, the updated thresholds can dynamically adapt to the actual wear condition of the load equipment. Therefore, even when mechanical wear accumulates to a certain level, the system can more accurately identify genuine high-frequency micro-vibration anomalies, avoiding misjudgments or missed detections.

[0134] The above technical solution enables more precise and adaptive dynamic adjustment of harmonic amplitude and phase thresholds. This solution fully considers the cumulative effect of mechanical wear during long-term operation of the load equipment, allowing the thresholds to be proactively corrected as the degree of mechanical wear changes. Compared to methods that rely solely on instantaneous operating conditions for adjustment, the solution presented in this application significantly improves the accuracy and robustness of high-frequency micro-vibration anomaly identification, effectively avoiding false alarms or missed alarms caused by accumulated mechanical wear, thereby enhancing the overall performance and reliability of the servo motor speed control system and extending the equipment's service life.

[0135] S15: Combine the feedforward control command and the initial control command to generate a speed control command.

[0136] Specifically, the combination of feedforward control commands and initial control commands can be achieved through simple superposition, so that the compensation effect of feedforward control and the regulation effect of PID feedback control are combined to form a more complete control strategy.

[0137] Furthermore, the feedforward control command compensates in advance for the transient adhesive resistance generated by the servo motor's operating state, reducing the amount of disturbance that the PID controller needs to handle, thereby reducing the burden on the PID controller and improving the system's response speed. The PID controller, in conjunction with the vibration compensation control of the load equipment on the servo motor, ensures accurate speed tracking and stability. The final generated speed control command is sent to the servo motor driver, driving the servo motor to operate at the desired speed.

[0138] The servo motor speed control method based on predictive model optimization provided in this invention introduces a predictive model to predict the transient adhesive resistance during servo motor operation and generates a feedforward compensation torque based on the prediction results, thereby effectively overcoming the lag of traditional PID control. Traditional PID control strategies, when faced with sudden load disturbances, often require adjustment only after an error occurs due to their inherent feedback mechanism, resulting in slow system response, large overshoot, and difficulty in meeting the stringent dynamic performance requirements of high-end applications. This application, by acquiring operating state data at a high sampling rate and performing second-order time derivative calculation, can more sensitively capture minute changes in the servo motor's operating state, providing high-quality input data for predicting transient adhesive resistance. The pre-trained predictive model's analysis of gradient data can accurately identify the peak value of transient adhesive resistance and, combined with a dynamic detection threshold, output accurate predicted values.

[0139] Example 2:

[0140] Figure 2 This diagram illustrates the structure of a servo motor speed control device based on a predictive model optimization according to an embodiment of the present invention. The control device includes:

[0141] Gradient data calculation module 10: used to collect the working status data of the servo motor at a high sampling rate through the power equipment, and to perform second-order time derivative on the working status data to obtain gradient data;

[0142] The gradient data calculation module 10 is used to acquire the operating status data of the servo motor at a high sampling rate through a power supply device, and to perform second-order time differentiation on the operating status data to obtain gradient data. Specifically, by integrating a high-precision data acquisition circuit, such as an analog-to-digital converter (ADC) and a cache, it is used to acquire the operating parameters of the servo motor, such as current, voltage, speed, and position, in real time at a sampling rate of microseconds or nanoseconds. The power supply device can be understood as a power management unit or driver connected to the servo motor, which integrates a sensor interface for data acquisition. Subsequently, the digital signal processing unit or microcontroller inside this module performs second-order time differentiation on the acquired raw operating status data, for example, by calculating the rate of change of the rate of change of the data through a difference algorithm, thereby obtaining gradient data reflecting drastic changes in the operating status. This gradient data provides the basis for subsequent predictive analysis.

[0143] Prediction module 20: used to analyze the gradient data through a pre-trained prediction model, obtain the peak data of the gradient data, and output the transient adhesion resistance prediction value of the servo motor based on the peak data and the dynamic detection threshold.

[0144] The prediction module 20 may include an embedded processor and a memory storing a prediction model trained on a large amount of historical data, such as a machine learning model based on neural networks or support vector machines. The prediction module receives gradient data from the gradient data calculation module and uses the pre-trained prediction model to analyze this data in real time, identifying peak data indicating transient changes. Further, the module dynamically adjusts the detection threshold based on the servo motor's current workload, ambient temperature, and other operating conditions, and outputs a predicted value of the servo motor's transient adhesive resistance based on the comparison between the peak data and the dynamic detection threshold. This predicted value reflects the external disturbances or resistance that the servo motor may encounter in a very short time.

[0145] Feedforward compensation module 30: Inputs the predicted resistance value into the feedforward controller and generates a feedforward control command containing a compensation torque proportional to the amplitude of the predicted transient adhesion resistance value;

[0146] The feedforward compensation module 30 can generate compensation torque in a timely manner, effectively offsetting disturbances and overcoming the inherent lag of traditional feedback control. Simultaneously, the control calculation module 40 adaptively adjusts the PID controller parameters using a fuzzy algorithm, further enhancing the system's robustness and control accuracy under complex operating conditions. Finally, the control output module organically combines feedforward compensation and feedback regulation to generate precise speed control commands, thereby significantly improving the dynamic response and stability of servo motor speed control. This effectively solves the problems of insufficient control accuracy and slow response of servo motors under high-speed, high-load operation and sudden disturbances in existing technologies.

[0147] Control calculation module 40: monitors the working status of the servo motor, adjusts the input parameters of the PID controller according to the working status through a fuzzy algorithm, and outputs the initial control command through the PID controller;

[0148] The control calculation module 40 integrates a state monitoring unit, a fuzzy controller, and a PID controller. The state monitoring unit acquires real-time operating parameters of the servo motor, such as speed, position, and current. Based on this real-time operating data, the fuzzy controller adaptively adjusts the input parameters of the PID controller, such as the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd, using preset fuzzy rules. For example, when a large error or fluctuation is detected in the system, the fuzzy controller can dynamically optimize the PID parameters to improve control performance. Subsequently, the PID controller receives these adjusted parameters and calculates and outputs an initial control command based on the servo motor's speed error (the difference between the target speed and the actual speed). This command serves as the core of the feedback control.

[0149] Control output module 50: Combines the feedforward control command and the initial control command to generate a speed control command.

[0150] The control output module 50 receives feedforward control commands from the feedforward compensation module 30 and initial control commands from the control calculation module 40, and fuses them, for example, through simple superposition. Thus, the predictive compensation effect of feedforward control and the precise adjustment effect of PID feedback control are organically combined to form a comprehensive speed control command. This final speed control command is sent to the servo motor driver to precisely control the servo motor's speed, ensuring stable and efficient operation under various working conditions.

[0151] The servo motor speed control device based on predictive model optimization provided in this invention, through a gradient data calculation module and a prediction module, can capture minute changes in the servo motor's operating state in real time with high precision and proactively predict transient adhesion resistance. Therefore, the feedforward compensation module can generate and apply compensation torque in advance before the actual disturbance occurs, significantly reducing speed fluctuations and overshoot. Simultaneously, the control calculation module dynamically adjusts the PID controller parameters through a fuzzy algorithm, enabling the feedback control section to better adapt to different operating conditions, further improving the system's robustness and accuracy. This synergistic effect of feedforward and feedback allows the control device of this application to achieve faster, smoother, and more precise speed control under high speed, high load, and complex operating conditions, effectively solving the problems of lag and slow adjustment in existing technologies, and significantly improving the overall performance and operational stability of the servo motor system.

[0152] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0153] Furthermore, the above description of the servo motor speed control method and device based on predictive model optimization provided by the embodiments of the present invention has been detailed. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A servo motor speed control method based on predictive model optimization, characterized in that, The control method includes: The working status data of the servo motor is collected at a high sampling rate by the power equipment, and the second-order time derivative of the working status data is performed to obtain gradient data. The gradient data is analyzed by a pre-trained prediction model to obtain the peak data of the gradient data. Based on the peak data and the dynamic detection threshold, the transient adhesion resistance prediction value of the servo motor is output. The predicted resistance value is input into the feedforward controller to generate a feedforward control command containing a compensation torque proportional to the amplitude of the predicted transient adhesion resistance value. The working status of the servo motor is monitored, and the parameters of the PID controller are adjusted according to the working status using a fuzzy algorithm. The initial control command is output through the PID controller based on the speed error. By combining the feedforward control command and the initial control command, a speed control command is generated; The process of acquiring servo motor operating status data at a high sampling rate using electrical equipment, and then performing second-order time derivative calculation on the operating status data to obtain gradient data includes: A high-bandwidth Hall effect current sensor is used to detect the operating current data of the servo motor in real time. The operating current data is encoded by an incremental encoder, and the encoded operating current data is input into the analysis and processing unit. The analysis and processing unit obtains the change data of the servo motor operating current value during the detection time period based on the encoded operating current data. Based on the current value change data and the detection time period, the second-order time derivative is calculated to output the gradient data of the servo motor's operating current data; The step of analyzing the gradient data using a pre-trained prediction model to obtain peak gradient data, and then outputting the predicted transient adhesion resistance value of the servo motor based on the peak data and a dynamic detection threshold, includes: The pre-trained prediction model extracts the current gradient values ​​of the servo motor at several unit time nodes within the detection period based on the gradient data, and extracts several current gradient peaks from the several current gradient values. An adaptive adjustment threshold is set according to the workload of the servo motor, and the peak values ​​of several current gradients are detected by the adaptive adjustment threshold. If the peak value of the current gradient is greater than the adaptive adjustment threshold, the transient adhesion resistance prediction value is calculated based on the peak value of the current gradient.

2. The servo motor speed control method based on predictive model optimization as described in claim 1, characterized in that, The step of inputting the predicted resistance value into the feedforward controller to generate a feedforward control command containing a compensation torque proportional to the amplitude of the predicted transient adhesion resistance value includes: The torque amplitude of the transient adhesion resistance prediction value is obtained, and a compensation torque is calculated based on the torque amplitude using a preset proportional coefficient. The compensation torque is proportional to the torque amplitude of the transient adhesion resistance prediction value. Calculate the pulse current with current adjustment amplitude and phase adjustment value based on the compensation torque, and output the feedforward control command to generate the pulse current.

3. The servo motor speed control method based on predictive model optimization as described in claim 1, characterized in that, The monitoring of the servo motor's operating status, the adjustment of the PID controller parameters based on the operating status using a fuzzy algorithm, and the output of initial control commands by the PID controller include: The operating parameters of the servo motor are monitored and obtained by the status monitoring module, and the operating status of the servo motor is evaluated based on the operating parameters to determine whether transient fluctuations occur. If transient fluctuations occur, the parameters of the PID controller are fine-tuned according to the amplitude and frequency of the transient fluctuations. If no transient fluctuations occur, the parameters of the PID controller are maintained.

4. The servo motor speed control method based on predictive model optimization as described in claim 3, characterized in that, The step of fine-tuning the parameters of the PID controller based on the amplitude and frequency of the transient fluctuations includes: Acquire the change data of the amplitude and frequency of the transient fluctuation, and set the adjustment mode according to the correspondence between the parameters of the PID controller and the change data of the transient fluctuation; The parameters of the PID controller are fine-tuned according to the preset adjustment step.

5. The servo motor speed control method based on predictive model optimization as described in claim 4, characterized in that, The monitoring of the servo motor's operating status, adjusting the PID controller parameters based on the operating status using a fuzzy algorithm, and outputting initial control commands through the PID controller also includes: An adaptive adjustment rule is introduced into the fuzzy algorithm to obtain the change range of the working parameters of the servo motor under real-time working conditions. The system detects whether the change amplitude is greater than a preset detection threshold. If the change amplitude is greater than the preset detection threshold, the PID parameters are adjusted based on the comparison data between the change amplitude and the preset detection threshold. If the detected change is less than the preset detection threshold, the PID parameters of the PID controller in the current state are maintained.

6. The servo motor speed control method based on predictive model optimization as described in claim 3, characterized in that, The initial control command output through the PID controller includes: The drive current and actual speed data of the servo motor are collected at a microsecond-level sampling rate; High-frequency harmonic analysis is performed on the collected drive current and actual rotation speed data to extract the amplitude and phase of the high-frequency harmonic components; The harmonic amplitude threshold and phase threshold are dynamically adjusted according to the operating status of the servo motor. Determine whether the amplitude of the extracted high-frequency harmonic component exceeds the dynamically adjusted harmonic amplitude threshold, and determine whether the phase of the extracted high-frequency harmonic component exceeds the dynamically adjusted phase threshold; If the amplitude of the high-frequency harmonic component exceeds the dynamically adjusted harmonic amplitude threshold, or if the phase of the high-frequency harmonic component exceeds the dynamically adjusted phase threshold, it is identified as a high-frequency micro-vibration anomaly. If it is identified as a high-frequency micro-vibration anomaly, a sequence of instantaneous reverse-phase current pulses is generated; The instantaneous reverse current pulse sequence is injected into the initial control command output by the PID controller.

7. The servo motor speed control method based on predictive model optimization as described in claim 6, characterized in that, The dynamic adjustment method for the harmonic amplitude threshold and phase threshold is as follows: The operating status data of the load device of the servo motor is detected, and the temperature and vibration acceleration data of the mechanical joint of the load device are obtained from the operating status data; Trend analysis was performed on the collected temperature and vibration acceleration data to obtain the cumulative degree of mechanical wear; Based on the cumulative degree of mechanical wear, calculate the drift of the harmonic amplitude threshold and phase threshold; The drift amount is superimposed on the preset harmonic amplitude threshold and phase threshold to obtain the updated harmonic amplitude threshold and the updated phase threshold.

8. A servo motor speed control device based on predictive model optimization, characterized in that, The control device includes: Gradient data calculation module: used to collect the working status data of the servo motor at a high sampling rate through the power equipment, and to perform second-order time derivative on the working status data to obtain gradient data; A high-bandwidth Hall effect current sensor is used to detect the operating current data of the servo motor in real time. The operating current data is encoded by an incremental encoder, and the encoded operating current data is input into the analysis and processing unit. The analysis and processing unit obtains the change data of the servo motor operating current value during the detection time period based on the encoded operating current data. Based on the current value change data and the detection time period, the second-order time derivative is calculated to output the gradient data of the servo motor's operating current data; Prediction module: used to analyze the gradient data through a pre-trained prediction model, obtain the peak data of the gradient data, and output the transient adhesion resistance prediction value of the servo motor based on the peak data and the dynamic detection threshold. The pre-trained prediction model extracts the current gradient values ​​of the servo motor at several unit time nodes within the detection period based on the gradient data, and extracts several current gradient peaks from the several current gradient values. An adaptive adjustment threshold is set according to the workload of the servo motor, and the peak values ​​of the several current gradients are detected by the adaptive adjustment threshold. If the peak value of the current gradient is greater than the adaptive adjustment threshold, the transient adhesion resistance prediction value is calculated based on the peak value of the current gradient. Feedforward compensation module: Inputs the predicted resistance value into the feedforward controller and generates a feedforward control command containing a compensation torque proportional to the amplitude of the predicted transient adhesion resistance value; Control calculation module: monitors the working status of the servo motor, adjusts the parameters of the PID controller according to the working status through a fuzzy algorithm, and outputs the initial control command through the PID controller based on the speed error; Control output module: Combines the feedforward control command and the initial control command to generate a speed control command.

Images