Multi-stage ADRC control method of high-frequency servo driver and related equipment

Abstract

The invention relates to the technical field of motor control, and provides a multi-stage ADRC control method for a high-frequency servo driver and related equipment. The method comprises the following steps: performing multi-dimensional space vector decomposition and nonlinear differential processing on an obtained target three-phase current to obtain a state vector, performing multi-time scale disturbance observation on the state vector to obtain a disturbance estimation value, and performing fractional order tracking differential processing on the disturbance estimation value to obtain a compensation control quantity. And performing multi-target optimization processing on the compensation control quantity to obtain an optimal control quantity, performing PWM waveform prediction compensation processing on the optimal control quantity to obtain a switching sequence, and performing FPGA hardware mapping processing on the switching sequence to obtain a PWM driving signal. According to the invention, through multi-dimensional space vector decomposition, multi-time scale disturbance observation, fractional order tracking differential processing, multi-target optimization processing and FPGA hardware mapping processing, the control performance of the servo driver under high-frequency and complex working conditions is improved.

Description

Technical Field

[0001] The present application relates to the field of motor control technology, and in particular to a high-frequency servo drive multi-level ADRC control method and related equipment. Background Art

[0002] High-frequency servo drives play a vital role in modern automation, precision control and high-dynamic performance applications, especially in industrial robots, aerospace, precision machining and other fields. Their high-frequency drives can achieve higher control accuracy, shorter response time and lower energy consumption.

[0003] Existing high-frequency servo drives use methods based on traditional PID control or improved PID control to adjust the control quantity based on the system output error. Under high-frequency conditions, a single-loop control structure is usually used, which makes it impossible to handle multiple disturbance sources such as current, speed and temperature at the same time. Especially under high-frequency PWM modulation, the relationship between switching loss and current ripple increases exponentially. Traditional methods cannot provide sufficient response speed without sacrificing system efficiency. Summary of the invention

[0004] In view of this, the present application provides a multi-level ADRC control method and related equipment for a high-frequency servo drive to solve the problems of poor disturbance suppression and control under multi-physical field composite disturbances.

[0005] The first aspect of the present application provides a high-frequency servo drive multi-level ADRC control method, the method comprising: Perform multi-dimensional space vector decomposition and nonlinear differential processing on the obtained target three-phase current to obtain a state vector; Performing multi-time scale disturbance observation on the state vector to obtain a disturbance estimate; Performing fractional-order tracking differentiation processing on the disturbance estimation value to obtain a compensation control amount; Performing multi-objective optimization processing on the compensation control amount to obtain an optimal control amount; Performing PWM waveform prediction and compensation processing on the optimal control amount to obtain a switching sequence; The switching sequence is subjected to FPGA hardware mapping processing to obtain a PWM driving signal.

[0006] In an optional implementation, performing multi-dimensional space vector decomposition and nonlinear differential processing on the obtained target three-phase current to obtain the state vector includes: Performing Clarke-Park transformation on the target three-phase current to obtain a direct-axis current and a quadrature-axis current; Performing fast Fourier transform and high-frequency harmonic extraction processing on the direct-axis current and the quadrature-axis current to obtain high-frequency harmonic current components; Performing nonlinear differential processing on the direct-axis current, the quadrature-axis current and the high-frequency harmonic current component according to a preset dynamic sixth-order nonlinear differential equation to obtain an initial state vector; The initial state vector is normalized to obtain the state vector.

[0007] In an optional implementation, performing multi-time-scale disturbance observation on the state vector to obtain a disturbance estimate includes: Performing current expansion state observation processing on the direct-axis current, the quadrature-axis current, and the rotational speed in the state vector according to a preset fast-changing subsystem model to obtain a current loop disturbance estimation value; Performing speed expansion state observation processing on the quadrature axis current and the speed according to a preset medium speed variable subsystem model to obtain a speed loop disturbance estimation value; Performing temperature expansion state observation processing on the magnetic saturation disturbance and the temperature drift in the state vector according to a preset slow-varying subsystem model to obtain an estimated value of the temperature loop disturbance; The current loop disturbance estimation value, the speed loop disturbance estimation value and the temperature loop disturbance estimation value are weightedly fused through a preset sliding surface coupling model to obtain the disturbance estimation value.

[0008] In an optional implementation, performing fractional-order tracking differentiation processing on the disturbance estimation value to obtain the compensation control amount includes: Performing fractional differential processing on the disturbance estimate to obtain first-order differential data; Performing fractional-order differential processing on the first-order differential data to obtain second-order differential data; Performing control error gain compensation calculation on the first-order differential data and the second-order differential data to obtain an initial compensation control amount; The initial compensation control amount is subjected to data compensation processing according to the disturbance estimation value to obtain the compensation control amount.

[0009] In an optional implementation, performing multi-objective optimization processing on the compensation control amount to obtain the optimal control amount includes: Performing dynamic response error calculation on the compensation control amount to obtain a control error; Performing time-domain differential calculation on the state vector to obtain a rate of change; Calculate the dynamic change of error according to the control error and the change rate to obtain the error change rate; Performing weighted combination processing on the control error and the error change rate to obtain a multi-objective optimization cost function; Optimizing and calculating the compensation control amount by using the multi-objective optimization cost function to obtain an optimal control amount candidate solution set; The optimal control amount candidate solution set is subjected to entropy weight smoothing processing according to a preset smoothing processing method to obtain the optimal control amount.

[0010] In an optional implementation, performing PWM waveform prediction and compensation processing on the optimal control amount to obtain a switching sequence includes: Performing time domain analysis on the optimal control quantity to obtain an ideal output sequence of voltage waveform; Performing nonlinear transfer function calculation on the ideal output sequence through a preset inverter transfer function model to obtain waveform distortion data; By using a preset three-level optimization model, predictive control processing is performed according to the waveform distortion data and the ideal output sequence to obtain expected PWM waveform data in a preset prediction period; Perform error calculation based on the expected PWM waveform data and the detected actual PWM waveform data to obtain an initial compensation amount; Performing data combination processing on the initial compensation amount and the optimal control amount to obtain an optimized compensation amount; A switching sequence of a PWM wave control signal is generated according to the optimized compensation amount.

[0011] In an optional implementation, performing FPGA hardware mapping processing on the switch sequence to obtain a PWM drive signal includes: Performing timing analysis on the switching sequence to obtain timing control data for each switching cycle; Performing FPGA simulation modeling on a preset FPGA hardware architecture according to the timing control data to obtain an initial parallel control signal and an FPGA hardware model; Performing time optimization allocation processing on the initial parallel control signal to obtain a time domain optimized parallel control signal; Synchronously processing the time-domain optimized parallel control signal to obtain a synchronous control signal; The synchronous control signal is subjected to level conversion and signal amplification processing through the FPGA hardware model to obtain the PWM drive signal.

[0012] A second aspect of the present application provides a high-frequency servo drive multi-level ADRC control device, the device comprising: A state vector module is used to perform multi-dimensional space vector decomposition and nonlinear differential processing on the obtained target three-phase current to obtain a state vector; A disturbance observation module, used for performing multi-time scale disturbance observation on the state vector to obtain a disturbance estimation value; A tracking and differentiation module, used for performing a fractional-order tracking and differentiation process on the disturbance estimation value to obtain a compensation control amount; A target optimization module, used for performing multi-objective optimization processing on the compensation control amount to obtain an optimal control amount; A prediction and compensation module, used for performing PWM waveform prediction and compensation processing on the optimal control amount to obtain a switching sequence; The hardware mapping module is used to perform FPGA hardware mapping processing on the switch sequence to obtain a PWM drive signal.

[0013] A third aspect of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the high-frequency servo drive multi-stage ADRC control method as described above when executing the computer program.

[0014] A fourth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the high-frequency servo drive multi-stage ADRC control method as described above are implemented.

[0015] In summary, this application at least includes the following beneficial technical effects: 1. Based on disturbance observation models of different time scales, disturbance estimation is performed for the current loop, speed loop, and temperature loop respectively, so as to cope with complex working conditions and various types of disturbances for adjustment, to achieve precise control of the entire frequency band and improve the robustness of the servo control system.

[0016] 2. Through multi-objective optimization processing of the compensation control quantity, a balance is achieved among multiple objectives, effectively improving the overall performance of the servo control system. The use of multi-objective optimization cost function enables the control quantity to be reasonably balanced among multiple objectives, thereby achieving the optimal control effect.

[0017] 3. Through the combination of sliding surface coupling model, time domain analysis, error compensation and other observation and control strategies, the servo control system can be ensured to operate stably under complex and changing load conditions and reduce control failures caused by disturbances. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 is a flow chart of a multi-level ADRC control method for a high-frequency servo drive provided in an embodiment of the present application; Figure 2 It is a functional module diagram of a high-frequency servo drive multi-stage ADRC control device provided in an embodiment of the present application; Figure 3 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0021] The high-frequency servo drive multi-level ADRC control method provided in the embodiment of the present application is executed by an electronic device, and accordingly, the high-frequency servo drive multi-level ADRC control device runs in the electronic device. The high-frequency servo drive multi-level ADRC control method provided in the embodiment of the present application is described below from the perspective of the electronic device and in combination with the process of adjusting the PWM drive signal.

[0022] like Figure 1 FIG. 1 is a flowchart of a multi-level ADRC control method for a high-frequency servo drive according to an embodiment of the present application. The multi-level ADRC control method for a high-frequency servo drive according to an embodiment of the present application includes the following steps.

[0023] Step S1, performing multi-dimensional space vector decomposition and nonlinear differential processing on the obtained target three-phase current to obtain a state vector.

[0024] Among them, the target three-phase current is the input current signal of the three-phase motor. By decomposing the space vector of the three-phase current signal, the complexity caused by the motor rotation can be reduced and the difficulty of signal processing can be reduced. This application uses Clarke-Park transformation to convert the three-phase AC signal into direct-axis and quadrature-axis components. Through the converted direct-axis current and quadrature-axis current, the electromagnetic behavior of the motor can be analyzed more clearly and controlled and adjusted. The formula of Clarke-Park transformation is as follows: in, , , is the target three-phase current. is the Clarke-Park transformation matrix. It is the electrical angle of the motor, which can be calculated in real time through the motor encoder or sensor and is used to indicate the angle of the rotor relative to the stator. They are the direct-axis current and quadrature-axis current, respectively, which reflect the magnetic field state and electromagnetic torque in the motor. The three-phase current signal is mapped to a stationary reference frame through Clarke-Park transformation, making motor control easier, especially in the motor control algorithm, which can simplify the complex three-phase signal into two components (i.e., direct-axis current and quadrature-axis current).

[0025] There are high-frequency harmonic current components in the target three-phase current. The high-frequency components are usually generated by the switching devices driven by the motor, which will affect the performance of the motor. It is necessary to separate the high-frequency harmonic current components from the fundamental frequency signal so that the high-frequency components can be processed separately. After obtaining the direct-axis current and the quadrature-axis current, the following formula is used to perform fast Fourier transform processing on the direct-axis current and the quadrature-axis current to extract the high-frequency harmonic components in the target three-phase current.

[0026] in, Indicates that the signal is at frequency The spectral components on . is the input time domain signal, usually a current signal (i.e., or ). It is an imaginary unit used to represent complex numbers. It is a complex exponential function in Fourier transform, which is used to reflect the frequency domain components of the signal. Fast Fourier transform can convert time domain signals into frequency domain to help extract the frequency components. For high-frequency harmonics, especially those caused by switching devices, the use of fast Fourier transform can effectively separate these harmonic components and provide important information for subsequent analysis and control.

[0027] After obtaining the direct-axis current, quadrature-axis current, and high-frequency harmonic current components, nonlinear differential processing is performed through the dynamic sixth-order nonlinear differential equation shown below: in, are the direct-axis and quadrature-axis current components respectively. is the rotation speed. is the motor electrical angle. Magnetic saturation disturbance. is the temperature drift. They are motor parameters, resistance, inductance, etc. Mechanical parameters, pole pairs, moment of inertia, friction resistance, etc. Sat() is a saturation function used to indicate current limitation. Erf() is an error function, usually used to indicate thermal disturbance. The sixth-order nonlinear differential equation is used to describe the changes in state variables such as the motor's current and speed as a dynamic system, which can better reflect the real dynamics of the system. By solving the above equations, the initial state vector can be obtained .

[0028] Normalization is to adjust all state variables to the same dimension and value range for subsequent control and calculation. Normalization can eliminate the computational instability caused by dimensional differences and improve computational efficiency. Each component in the initial state vector is normalized so that its value is within the standard range (for example, [0, 1] or [-1, 1] interval). The normalization formula is as follows: in, is the initial state quantity. They are respectively the minimum and maximum values ​​of the preset state quantities. is the state vector obtained after normalization.

[0029] Step S2: Perform multi-time-scale disturbance observation on the state vector to obtain a disturbance estimate.

[0030] It should be understood that the current response speed in the servo control system is the fastest (microseconds), followed by mechanical speed changes (milliseconds), and temperature drift and other disturbances are the slowest (seconds). Based on the dynamic response differences, the servo control system is divided into the following three independent observation channel models: Fast-changing subsystem: Focus on the high-frequency harmonics of the current loop (such as the 5th and 7th harmonics), and set the time constant to s (corresponding to 10kHz bandwidth).

[0031] Medium speed subsystem: Tracking speed fluctuations (e.g., load changes), time constant s.

[0032] Slow-change subsystem: monitors temperature drift and magnetic saturation, time constant s.

[0033] Specifically, the fast-changing subsystem model uses an extended state observer to observe the direct-axis current, quadrature-axis current, and the speed related to the current loop to estimate the disturbance of the current loop and obtain the estimated value of the current loop disturbance. The extended state observer estimates the disturbance based on the mathematical model of the system and the relationship between the input signal (for example, current and speed) and the output signal (estimated disturbance value). The state observer of the fast-changing subsystem estimates the disturbance according to the following equation: in, is the tracking error of the observer, calculated as , represents the difference between the estimated current and the actual current. are the three estimated values ​​of the extended state observer, representing the states of current, speed, and disturbance respectively. is the gain parameter of ESO, which is used to adjust the response speed of the observer. is the control input of the current loop, is the inductance value. The disturbance of the current loop can be effectively estimated by the fast-changing subsystem model. , providing a basis for subsequent compensation. The current loop is the fastest changing part in the motor servo control system, and is usually subject to rapid disturbances from factors such as load changes and controller output. By performing disturbance observation through the fast-changing subsystem model, these rapid disturbances can be detected and compensated in real time, thereby improving the accuracy and response speed of the servo control system.

[0034] At the same time, the medium-speed subsystem uses an extended state observer to observe the quadrature axis current and speed to estimate the disturbance of the speed loop and obtain the estimated value of the speed loop disturbance. The state observer of the medium-speed subsystem estimates the disturbance according to the following equation: in, is an estimate of the rotational speed. is the actual speed of the system. is the ESO gain, which is used to adjust the observer's response to speed disturbances. is the number of pole pairs of the motor, is the moment of inertia, is the magnetic linkage constant, is the quadrature axis current. The medium speed subsystem equation corrects the control input of the speed loop by estimating the disturbance. The speed loop is a medium time scale disturbance, usually caused by factors such as load change and mechanical friction. By observing the speed and estimating the disturbance, the impact of speed change on the control performance can be effectively reduced, ensuring the stability of the servo control system.

[0035] At the same time, the slow-changing subsystem mainly deals with the disturbances of temperature drift and magnetic saturation. These disturbances change slowly, but have an important impact on the long-term performance and stability of the motor. The slow-changing subsystem model uses magnetic saturation disturbances and temperature drift for estimation to obtain the estimated value of the temperature loop disturbance. The state observer of the slow subsystem estimates the disturbance according to the following equation: in, is an estimate of the magnetic saturation disturbance. is the actual magnetic saturation disturbance. is the ESO gain, which is used to adjust the response to magnetic saturation perturbations. is the magnetic saturation constant, is the temperature change. The slow subsystem equation is used to estimate the disturbances of temperature and magnetic saturation, providing data support for subsequent temperature control and system optimization. Temperature and magnetic saturation are slow-changing disturbances in the system, and their impact on system performance is relatively long-lasting. By estimating the disturbance of the slow-changing subsystem, long-term stability and robustness can be guaranteed for the system.

[0036] In order to more comprehensively and accurately reflect the overall disturbance of the servo control system, this application uses a sliding surface coupling model to perform weighted fusion of the disturbance estimates from each subsystem (i.e., the current loop disturbance estimate, the speed loop disturbance estimate, and the temperature loop disturbance estimate) to obtain a total disturbance estimate. The fusion formula of the sliding surface coupling model is as follows: in, is the total disturbance estimate after fusion. are the disturbance estimates from the fast-varying subsystem, the medium-speed subsystem, and the slow-varying subsystem respectively. is the adaptive weight, which determines the weight of each subsystem disturbance estimate in the total disturbance estimate. It is the disturbance estimate ultimately used for control and can reflect the disturbance of the system at multiple time scales.

[0037] Step S3: performing fractional-order tracking differentiation processing on the disturbance estimation value to obtain a compensation control amount.

[0038] It should be understood that fractional-order differentials are a mathematical tool that can more accurately capture the dynamic changes of a system. In traditional integer-order differentials, the changes in the system are described as first-order derivatives, but in some complex dynamic systems, fractional-order differentials can provide a more accurate description. Fractional-order differentials can not only better track the high-frequency components of the signal, but also handle the nonlinear characteristics of the system. The disturbance estimate is processed by a fractional-order differential operator to obtain the first-order differential data. Among them, the order of the fractional-order differential is set to =1.25, we can get the first-order differential result. The fractional differential formula is as follows: in, represents the fractional differential operator, is the order of differentiation, usually chosen as a non-integer value between 0 and 1, representing the degree of response of the system to changes. is the disturbance estimate The result after fractional differentiation. It is the output of fractional order differential, that is, the first order differential data. Using fractional order differential can effectively reduce the phase lag in the system and capture the high-frequency dynamic change part of the disturbance signal, thus making the compensation control more accurate and responsive.

[0039] After obtaining the first-order differential data, it is necessary to perform fractional differential processing on the first-order differential data. The fractional differential operation is similar to the differential operation, which is a form of "accumulation" or "smoothing" of the signal. Fractional differential helps the system capture the long-term accumulation effect of disturbances, especially when facing low-frequency disturbances or slow-changing disturbances. This operation is also performed using fractional operators to ensure that the processing of the signal is not limited to immediate changes, but also takes into account past influences. The fractional differential formula is as follows: in, represents the fractional differential operator, is the order of the derivative (usually set to 0.75) and represents the cumulative effect of the derivative operation on the signal. It is the first-order data after fractional differentiation processing. It is the result of fractional differentiation, that is, the second-order differential data. Fractional differentiation is used to smooth disturbance signals, reduce the impact of high-frequency noise, provide responses to long-term changes in the system, and enhance the stability of the compensation effect.

[0040] After obtaining the first-order differential data and the second-order differential data, the system output is corrected by calculating the control error. Specifically, the initial compensation control amount is obtained by calculating the control error gain compensation. The control error usually refers to the difference between the target state and the actual state. The compensation control amount calculation formula is as follows: in, It is the preliminary compensation control quantity, which represents the correction value of the system output. and They are proportional gain and differential gain, which are used to adjust the error correction. The proportional gain controls the size of the error, and the differential gain controls the response of the error change rate. and They are fractional-order differential and differentiated data, respectively, reflecting the dynamic characteristics of the disturbance. and is the target state of the system, representing the target direct-axis current and target quadrature axis current By compensating and correcting the error, the system can quickly respond to external disturbances and maintain a stable control effect. The control error gain compensation calculation allows the system to make adaptive adjustments based on real-time changes.

[0041] The preliminary compensation control amount is further adjusted according to the disturbance estimation value to better adapt to the changes in the actual system. By removing the error in the preliminary compensation, the final control amount is more accurate, ensuring that the system can operate stably in the real environment. The data compensation formula is as follows: in, It is the final compensation control quantity, the result after data compensation. It is the compensation control quantity obtained by preliminary calculation. It is the disturbance estimation value from different observation channel models, which is used for further compensation. Data compensation processing further ensures the accuracy of the compensation control amount. By introducing the disturbance estimation value into the compensation calculation, the system can accurately respond to changes from the environment and load and maintain high-precision control.

[0042] Step S4: performing multi-objective optimization processing on the compensation control amount to obtain the optimal control amount.

[0043] The control error is the difference between the system output and the expected output. For any control system, error calculation is one of the core tasks, because the goal of the control system is to minimize this error so that the system response is as close to the expected value as possible. After obtaining the compensation control amount, the control error is obtained by calculating the difference between the compensation control amount and the target control amount of the detection feedback. The control error reflects the deviation between the actual control amount and the expected control amount. The performance of the control system is evaluated by minimizing the control error during the optimization process.

[0044] The rate of change indicates the speed at which the system state changes, that is, the degree of change in the system output. By calculating the time-domain differential of the system state vector, the rate of change of each state quantity can be obtained, which is crucial for capturing the dynamic response of the system. By performing differential calculations on the state vector, the rate of change of the system is obtained. In multi-objective optimization, the rate of change and the control error are used together as key optimization parameters to help evaluate the real-time dynamic characteristics of the system.

[0045] The error change rate refers to the rate at which the error changes over time, representing the changing trend of the system error. Error dynamic change calculation is a further analysis of the control error and its rate of change. Error dynamic change calculation is used to determine the changing trend of the control error, especially the rate of error change. The response of the control system is not only related to the current error, but also to the rate of error change, which determines the system's response speed to disturbances.

[0046] The multi-objective optimization cost function is the core of the multi-objective optimization algorithm. It comprehensively evaluates the performance of the control system by weighted combination of multiple optimization objectives. The control error and the error change rate are usually the two most important objectives, so these two objectives need to be weighted combined to obtain a single cost function. When calculating the cost function, each objective will be assigned a certain weight coefficient, and the size of the weight coefficient determines the importance of the objective in the optimization process. By adjusting these weights, a balance of control objectives can be achieved. The multi-objective optimization cost function is shown below: in, is the cost function, which represents the comprehensive performance of the system. Control error. Control quantity. is the rate of change of the control error. are weighted coefficients, which respectively represent the importance of control error, control amount and error change rate. The multi-objective optimization cost function is used to measure the overall performance of the system. By optimizing the cost function, the control error can be minimized, the change of control amount can be reduced, and the error fluctuation can be suppressed, thereby obtaining the optimal control amount.

[0047] The compensation control quantity is optimized and calculated using a multi-objective optimization cost function. Thus, multiple optimal control quantity candidate solutions are generated, which represent the control quantity obtained under different optimization conditions. The optimal control quantity candidate solution set is a set of multiple possible control quantity solutions. After obtaining the optimal control quantity candidate solution set, it is necessary to further select the optimal control quantity solution (i.e., the optimal control quantity) through an optimization algorithm. The candidate solution calculation formula is as follows: in, is the optimal control quantity, which means the control quantity obtained through optimization calculation. The cost function is optimized for multiple objectives and represents the overall performance of the system. By minimizing the cost function, the system can find the optimal control amount to ensure the most balanced response of the system, which can accurately track the target while reducing energy consumption and control errors.

[0048] Finally, the embodiment of the present application uses the entropy weight method to smooth the optimal control quantity candidate solution set. The entropy weight method is a method of smoothing the optimal control quantity by calculating the weight of the candidate solution, the purpose of which is to eliminate excessive fluctuations and make the control quantity more stable.

[0049] in, The weight of the candidate solution. is the cost function value for each candidate solution. is the control amount for each candidate solution. It is the smoothing coefficient in the entropy weight method, which is used to adjust the sensitivity of the weight. is the final optimal control quantity. Through entropy weight smoothing, a weight can be assigned to each solution according to the cost function value of the candidate solution, and finally a smooth and accurate optimal control quantity is obtained.

[0050] Step S5: performing PWM waveform prediction and compensation processing on the optimal control amount to obtain a switching sequence.

[0051] After obtaining the optimal control quantity, the optimal control quantity is analyzed in the time domain. First, the ideal output sequence of the voltage waveform needs to be obtained, so as to convert the optimal control quantity into a specific voltage waveform sequence in order to control the inverter. During the time domain analysis, the optimal control quantity is mapped to the voltage output of the inverter to generate an ideal PWM waveform (i.e., the ideal output sequence of the voltage waveform) through the circuit model. Although these waveforms represent the expected output of the inverter, they do not take into account the nonlinear effects in the actual hardware (for example, dead time, switching delay, etc.).

[0052] It should be understood that there are some nonlinear effects in the actual operation of the inverter, such as switching delay, dead time, etc. These non-ideal factors will cause distortion between the actual output voltage waveform and the ideal waveform. In order to accurately control the output of the inverter, it is necessary to calculate the influence of these nonlinear factors through the transfer function model of the inverter. The transfer function of the inverter usually takes into account factors such as the characteristics of the switching elements, dead time and switching delay. By inputting the ideal output sequence into the nonlinear transfer function model of the inverter, the waveform distortion data can be obtained, that is, the difference between the actual output waveform and the ideal output waveform. The inverter transfer function model can be expressed as the following formula: in, represents the transfer function of the inverter, where is the complex frequency variable in the Laplace transform. Dead time refers to the minimum time required for the inverter switching elements to switch states. Switching delay time refers to the time delay from the issuance of the control signal to the actual switching state.

[0053] The three-level optimization model optimizes the inverter switching process by predicting the waveform changes of multiple PWM cycles in the future. The model performs control processing based on waveform distortion data and ideal output sequence to generate expected PWM waveform data. The three-level optimization model takes into account the characteristics of actual hardware, including switching delay, dead time, etc., and minimizes the impact of these nonlinear effects on system output through optimized control. The three-level optimization model can be expressed as follows: in, is the optimal control quantity. is the PWM switching sequence, , where -1 means the lower tube is on, 0 means the switch is off, and 1 means the upper tube is on. is a DC voltage source. is the second norm, which represents the sum of squares of errors. is the switching loss. Weight coefficients are used to balance errors and losses. Through three-level optimization, the model can predict and correct waveform distortion for the next few cycles. The optimization objectives include minimizing the error and loss of the PWM waveform to achieve the best control effect.

[0054] After generating the expected PWM waveform data, the next step is to compare it with the actual detected PWM waveform data and calculate the error between them. The actual PWM waveform is obtained through the feedback signal of the actual output voltage of the inverter, which is usually measured by a current sensor and a voltage sensor. The error calculation is used to evaluate the difference between the expected waveform and the actual waveform to determine the compensation amount. The error will provide input to the compensation algorithm to further optimize the output waveform.

[0055] The final optimized compensation is obtained by combining the initial compensation with the optimal control. This compensation takes into account factors such as control error, waveform distortion, and switching delay to ensure that the generated PWM signal can accurately control the output of the inverter. Finally, the switching sequence of the final PWM wave control signal is generated based on the obtained optimized compensation. This switching sequence will be used to control the switching elements of the inverter to generate an output signal that meets the target voltage waveform. The switching sequence generation formula is as follows: in, : Optimized switch sequence, . The final optimized compensation amount. DC voltage source.

[0056] Step S6: Perform FPGA hardware mapping processing on the switch sequence to obtain a PWM drive signal.

[0057] Timing analysis is to obtain the timing data of the control signal by analyzing the switch sequence. The timing control data of each switching cycle refers to the state that the system needs to switch to at each moment (for example, the upper and lower tube states of the inverter switch). Thus, the discrete switching sequence is converted into control instructions in the time domain to ensure the accurate generation of the switching signal in each switching cycle. Through timing analysis, it can be ensured that the system switches the switch state in a predetermined order at each moment. This is an essential step in the control of high-frequency servo drive systems, which ensures that the control signal matches the response of the hardware.

[0058] It should be understood that FPGA simulation modeling is to model the hardware architecture through FPGA development tools to simulate the behavior of actual hardware. According to the timing control data, the FPGA simulation model will describe how the hardware responds to these timing control data to generate parallel control signals. Among them, parallel control signals mean that multiple control signals can be calculated and output at the same time, and the FPGA hardware architecture can achieve efficient parallel processing to meet high-frequency control requirements. Through FPGA simulation modeling, it can be verified whether the response of the hardware architecture to the timing control data meets the system requirements, thereby ensuring the timing accuracy and parallel processing capabilities of the system.

[0059] Time optimization allocation processing is to optimize the initially generated parallel control signals to ensure that the execution time of each control signal can meet the timing requirements of the system. Through time optimization, conflicts between multiple control signals can be avoided and the system can run smoothly at high frequencies. Specifically, the execution order and delay of the signal are adjusted based on the maximum clock cycle of the system and the response time of the control signal.

[0060] Synchronous processing is to ensure that multiple parallel control signals inside the FPGA run synchronously at the same time. The synchronous control signal is a key signal to ensure that all control signals change synchronously according to the predetermined timing. Synchronous triggers and clock management modules are often used in FPGA to achieve this function. Synchronous processing eliminates the timing inconsistency between different parallel modules and ensures signal synchronization. At the same time, through synchronous processing, it is ensured that there is no conflict between the control signals of various modules in the FPGA, and all signals are executed synchronously according to a unified timing, improving the stability and reliability of the system.

[0061] Level conversion and signal amplification are used to convert the synchronous control signal into a signal form suitable for driving the inverter. Since the output signal of the FPGA is usually a low-voltage signal, the signal needs to be adjusted to the drive voltage of the hardware (for example, ±15V) through a level conversion circuit. At the same time, signal amplification is to ensure that the driving capability of the PWM waveform is strong enough to drive the switch tube of the inverter. Through level conversion and signal amplification, it is ensured that the final PWM drive signal can meet the input requirements of the inverter and can drive the hardware to perform the corresponding switching operation.

[0062] The present application is applied to the field of motor control technology, by performing multi-dimensional space vector decomposition and nonlinear differential processing on the obtained target three-phase current to obtain a state vector, performing multi-time scale disturbance observation on the state vector to obtain a disturbance estimate, performing fractional-order tracking differential processing on the disturbance estimate to obtain a compensation control quantity, performing multi-objective optimization processing on the compensation control quantity to obtain an optimal control quantity, performing PWM waveform prediction compensation processing on the optimal control quantity to obtain a switching sequence, and performing FPGA hardware mapping processing on the switching sequence to obtain a PWM drive signal. The present application improves the control performance of the servo drive under high-frequency and complex working conditions through multi-dimensional space vector decomposition, multi-time scale disturbance observation, fractional-order tracking differential processing, multi-objective optimization processing, and FPGA hardware mapping processing.

[0063] like Figure 2 , which is a functional module diagram of a multi-stage ADRC control device for a high-frequency servo drive provided in an embodiment of the present application.

[0064] In some embodiments, the high-frequency servo drive multi-level ADRC control device 2 may include a plurality of functional modules composed of computer program segments. The computer programs of the various program segments in the high-frequency servo drive multi-level ADRC control device 2 may be stored in a memory of a server and executed by at least one processor to execute (see Figure 1 Description) Function of multi-level ADRC control method for high frequency servo drives.

[0065] In this embodiment, the high-frequency servo drive multi-level ADRC control device 2 can be divided into multiple functional modules according to the functions it performs. The functional modules may include: a state vector module 21, a disturbance observation module 22, a tracking differential module 23, a target optimization module 24, a prediction compensation module 25 and a hardware mapping module 26. The module referred to in the present invention refers to a series of computer program segments that can be executed by at least one processor and can complete fixed functions, which are stored in a memory. In this embodiment, the functions of each module will be described in detail in subsequent embodiments.

[0066] The state vector module 21 is used to perform multi-dimensional space vector decomposition and nonlinear differential processing on the obtained target three-phase current to obtain a state vector.

[0067] In an optional implementation, the state vector module 21 is specifically used for: Performing Clarke-Park transformation on the target three-phase current to obtain a direct-axis current and a quadrature-axis current; Performing fast Fourier transform and high-frequency harmonic extraction processing on the direct-axis current and the quadrature-axis current to obtain high-frequency harmonic current components; Performing nonlinear differential processing on the direct-axis current, the quadrature-axis current and the high-frequency harmonic current component according to a preset dynamic sixth-order nonlinear differential equation to obtain an initial state vector; The initial state vector is normalized to obtain the state vector.

[0068] The disturbance observation module 22 is used to perform multi-time scale disturbance observation on the state vector to obtain a disturbance estimation value.

[0069] In an optional implementation, the disturbance observation module 22 is specifically used for: Performing current expansion state observation processing on the direct-axis current, the quadrature-axis current, and the rotational speed in the state vector according to a preset fast-changing subsystem model to obtain a current loop disturbance estimation value; Performing speed expansion state observation processing on the quadrature axis current and the speed according to a preset medium speed variable subsystem model to obtain a speed loop disturbance estimation value; Performing temperature expansion state observation processing on the magnetic saturation disturbance and the temperature drift in the state vector according to a preset slow-varying subsystem model to obtain an estimated value of the temperature loop disturbance; The current loop disturbance estimation value, the speed loop disturbance estimation value and the temperature loop disturbance estimation value are weightedly fused through a preset sliding surface coupling model to obtain the disturbance estimation value.

[0070] The tracking and differentiation module 23 is used to perform fractional-order tracking and differentiation processing on the disturbance estimation value to obtain a compensation control amount.

[0071] In an optional implementation, the tracking differentiation module 23 is specifically used for: Performing fractional differential processing on the disturbance estimate to obtain first-order differential data; Performing fractional-order differential processing on the first-order differential data to obtain second-order differential data; Performing control error gain compensation calculation on the first-order differential data and the second-order differential data to obtain an initial compensation control amount; The initial compensation control amount is subjected to data compensation processing according to the disturbance estimation value to obtain the compensation control amount.

[0072] The target optimization module 24 is used to perform multi-objective optimization processing on the compensation control amount to obtain the optimal control amount.

[0073] In an optional implementation, the target optimization module 24 is specifically used to: Performing dynamic response error calculation on the compensation control amount to obtain a control error; Performing time-domain differential calculation on the state vector to obtain a rate of change; Calculate the dynamic change of error according to the control error and the change rate to obtain the error change rate; Performing weighted combination processing on the control error and the error change rate to obtain a multi-objective optimization cost function; Optimizing and calculating the compensation control amount by using the multi-objective optimization cost function to obtain an optimal control amount candidate solution set; The optimal control amount candidate solution set is subjected to entropy weight smoothing processing according to a preset smoothing processing method to obtain the optimal control amount.

[0074] The prediction and compensation module 25 is used to perform PWM waveform prediction and compensation processing on the optimal control amount to obtain a switching sequence.

[0075] In an optional implementation, the prediction compensation module 2 is specifically used for: Performing time domain analysis on the optimal control quantity to obtain an ideal output sequence of voltage waveform; Performing nonlinear transfer function calculation on the ideal output sequence through a preset inverter transfer function model to obtain waveform distortion data; By using a preset three-level optimization model, predictive control processing is performed according to the waveform distortion data and the ideal output sequence to obtain expected PWM waveform data in a preset prediction period; Perform error calculation based on the expected PWM waveform data and the detected actual PWM waveform data to obtain an initial compensation amount; Performing data combination processing on the initial compensation amount and the optimal control amount to obtain an optimized compensation amount; A switching sequence of a PWM wave control signal is generated according to the optimized compensation amount.

[0076] The hardware mapping module 26 is used to perform FPGA hardware mapping processing on the switch sequence to obtain a PWM drive signal.

[0077] In an optional implementation, the hardware mapping module 26 is specifically used to: Performing timing analysis on the switching sequence to obtain timing control data for each switching cycle; Performing FPGA simulation modeling on a preset FPGA hardware architecture according to the timing control data to obtain an initial parallel control signal and an FPGA hardware model; Performing time optimization allocation processing on the initial parallel control signal to obtain a time domain optimized parallel control signal; Synchronously processing the time-domain optimized parallel control signal to obtain a synchronous control signal; The synchronous control signal is subjected to level conversion and signal amplification processing through the FPGA hardware model to obtain the PWM drive signal.

[0078] It should be understood that the various variations and specific embodiments of the methods provided in the above embodiments are also applicable to the high-frequency servo drive multi-stage ADRC control device of the present embodiment. Through the above detailed description of the high-frequency servo drive multi-stage ADRC control method, those skilled in the art can clearly understand the implementation method of the high-frequency servo drive multi-stage ADRC control device in the present embodiment. For the sake of brevity of the specification, it will not be described in detail here.

[0079] like Figure 3 , which is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0080] In a preferred embodiment of the present invention, the electronic device 3 may include, but is not limited to: a memory 31 , at least one processor 32 and at least one communication bus 33 .

[0081] Those skilled in the art should understand that Figure 3 The structure of the electronic device 3 shown does not constitute a limitation of the embodiment of the present invention, and the electronic device 3 may also include more or less other hardware or software than shown in the figure, or a different arrangement of components.

[0082] In some embodiments, the electronic device 3 is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application-specific integrated circuits, programmable gate arrays, digital processors and embedded devices.

[0083] It should be noted that the electronic device 3 is only an example, and other existing or future electronic products that are suitable for the present application should also be included in the protection scope of the present application and included here by reference.

[0084] In some embodiments, the memory 31 stores a computer program, and when the computer program is executed by the at least one processor 32, all or part of the steps in the high-frequency servo drive multi-level ADRC control method as described above are implemented. The memory 31 includes a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data. Further, the computer-readable storage medium may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, and the like.

[0085] In some embodiments, the at least one processor 32 is the control core (ControlUnit) of the electronic device 3, and uses various interfaces and lines to connect various components of the entire electronic device 3, and executes various functions and processes data of the electronic device 3 by running or executing programs or modules stored in the memory 31, and calling data stored in the memory 31. For example, when the at least one processor 32 executes the computer program stored in the memory 31, it implements all or part of the steps of the high-frequency servo drive multi-level ADRC control method described in the embodiment of the present application; or implements all or part of the functions of the high-frequency servo drive multi-level ADRC control device. The at least one processor 32 can be composed of an integrated circuit, for example, it can be composed of a single packaged integrated circuit, or it can be composed of multiple integrated circuits with the same function or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips.

[0086] In some embodiments, the at least one communication bus 33 is configured to realize the connection and communication between the memory 31 and the at least one processor 32. Although not shown, the electronic device 3 may also include a power supply (such as a battery) for powering each component. Preferably, the power supply may be logically connected to the at least one processor 32 through a power management device, so as to realize the functions of managing charging, discharging, and power consumption management through the power management device. The power supply may also include any components such as one or more DC or AC power supplies, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, etc. The electronic device 3 may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0087] The above-mentioned integrated unit implemented in the form of a software function module can be stored in a computer-readable storage medium. The above-mentioned software function module is stored in a storage medium, including a number of instructions for enabling an electronic device (which can be a personal computer, electronic device, or network device, etc.) or a processor to execute part of the method described in each embodiment of the present application.

[0088] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0089] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0090] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A multi-level ADRC control method for a high-frequency servo drive, characterized in that: The method comprises: Perform multi-dimensional space vector decomposition and nonlinear differential processing on the obtained target three-phase current to obtain a state vector; Performing multi-time scale disturbance observation on the state vector to obtain a disturbance estimate; Performing fractional-order tracking differentiation processing on the disturbance estimation value to obtain a compensation control amount; Performing multi-objective optimization processing on the compensation control amount to obtain an optimal control amount; Performing PWM waveform prediction and compensation processing on the optimal control amount to obtain a switching sequence; The switching sequence is subjected to FPGA hardware mapping processing to obtain a PWM driving signal.

2. The high-frequency servo drive multi-level ADRC control method according to claim 1 is characterized in that: The performing multi-dimensional space vector decomposition and nonlinear differential processing on the obtained target three-phase current to obtain the state vector includes: Performing Clarke-Park transformation on the target three-phase current to obtain a direct-axis current and a quadrature-axis current; Performing fast Fourier transform and high-frequency harmonic extraction processing on the direct-axis current and the quadrature-axis current to obtain high-frequency harmonic current components; Performing nonlinear differential processing on the direct-axis current, the quadrature-axis current and the high-frequency harmonic current component according to a preset dynamic sixth-order nonlinear differential equation to obtain an initial state vector; The initial state vector is normalized to obtain the state vector.

3. The high-frequency servo drive multi-level ADRC control method according to claim 2, characterized in that: The performing multi-time-scale disturbance observation on the state vector to obtain a disturbance estimation value comprises: Performing current expansion state observation processing on the direct-axis current, the quadrature-axis current, and the rotational speed in the state vector according to a preset fast-changing subsystem model to obtain a current loop disturbance estimation value; Performing speed expansion state observation processing on the quadrature axis current and the speed according to a preset medium speed variable subsystem model to obtain a speed loop disturbance estimation value; Performing temperature expansion state observation processing on the magnetic saturation disturbance and the temperature drift in the state vector according to a preset slow-varying subsystem model to obtain an estimated value of the temperature loop disturbance; The current loop disturbance estimation value, the speed loop disturbance estimation value and the temperature loop disturbance estimation value are weightedly fused through a preset sliding surface coupling model to obtain the disturbance estimation value.

4. The high-frequency servo drive multi-level ADRC control method according to claim 1, characterized in that: The performing fractional-order tracking differentiation processing on the disturbance estimation value to obtain the compensation control amount comprises: Performing fractional differential processing on the disturbance estimate to obtain first-order differential data; Performing fractional-order differential processing on the first-order differential data to obtain second-order differential data; Performing control error gain compensation calculation on the first-order differential data and the second-order differential data to obtain an initial compensation control amount; The initial compensation control amount is subjected to data compensation processing according to the disturbance estimation value to obtain the compensation control amount.

5. The high-frequency servo drive multi-level ADRC control method according to claim 1, characterized in that: The performing multi-objective optimization processing on the compensation control amount to obtain the optimal control amount includes: Performing dynamic response error calculation on the compensation control amount to obtain a control error; Performing time-domain differential calculation on the state vector to obtain a rate of change; Calculate the dynamic change of error according to the control error and the change rate to obtain the error change rate; Performing weighted combination processing on the control error and the error change rate to obtain a multi-objective optimization cost function; Optimizing and calculating the compensation control amount by using the multi-objective optimization cost function to obtain an optimal control amount candidate solution set; The optimal control amount candidate solution set is subjected to entropy weight smoothing processing according to a preset smoothing processing method to obtain the optimal control amount.

6. The high-frequency servo drive multi-level ADRC control method according to claim 1, characterized in that: The performing PWM waveform prediction and compensation processing on the optimal control amount to obtain a switching sequence comprises: Performing time domain analysis on the optimal control quantity to obtain an ideal output sequence of voltage waveform; Performing nonlinear transfer function calculation on the ideal output sequence through a preset inverter transfer function model to obtain waveform distortion data; By using a preset three-level optimization model, predictive control processing is performed according to the waveform distortion data and the ideal output sequence to obtain expected PWM waveform data in a preset prediction period; Perform error calculation based on the expected PWM waveform data and the detected actual PWM waveform data to obtain an initial compensation amount; Performing data combination processing on the initial compensation amount and the optimal control amount to obtain an optimized compensation amount; A switching sequence of a PWM wave control signal is generated according to the optimized compensation amount.

7. The high-frequency servo drive multi-level ADRC control method according to claim 1, characterized in that: The performing FPGA hardware mapping processing on the switch sequence to obtain a PWM drive signal comprises: Performing timing analysis on the switching sequence to obtain timing control data for each switching cycle; Performing FPGA simulation modeling on a preset FPGA hardware architecture according to the timing control data to obtain an initial parallel control signal and an FPGA hardware model; Performing time optimization allocation processing on the initial parallel control signal to obtain a time domain optimized parallel control signal; Performing synchronous processing on the time-domain optimized parallel control signal to obtain a synchronous control signal; The synchronous control signal is subjected to level conversion and signal amplification processing through the FPGA hardware model to obtain the PWM drive signal.

8. A high-frequency servo drive multi-level ADRC control device, characterized in that: The device comprises: A state vector module is used to perform multi-dimensional space vector decomposition and nonlinear differential processing on the obtained target three-phase current to obtain a state vector; A disturbance observation module, used for performing multi-time scale disturbance observation on the state vector to obtain a disturbance estimation value; A tracking and differentiation module, used for performing a fractional-order tracking and differentiation process on the disturbance estimation value to obtain a compensation control amount; A target optimization module, used for performing multi-objective optimization processing on the compensation control amount to obtain an optimal control amount; A prediction and compensation module, used for performing PWM waveform prediction and compensation processing on the optimal control amount to obtain a switching sequence; The hardware mapping module is used to perform FPGA hardware mapping processing on the switch sequence to obtain a PWM drive signal.

9. An electronic device, characterized in that: The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the high-frequency servo drive multi-level ADRC control method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the high-frequency servo drive multi-level ADRC control method according to any one of claims 1 to 7 are implemented.

Images