Control design optimization method and system for permanent magnet synchronous motor

Abstract

The invention relates to the technical field of design optimization, in particular to a control design optimization method and system for a permanent magnet synchronous motor. The method comprises the following steps: obtaining and measuring a stator three-phase voltage value and a current value of the PMSM and a motor rotating speed in real time; constructing a d-q axis motor equivalent circuit model, and calculating an available boundary curve of the voltage of the weakening area by combining the three-phase voltage value of the stator and the rotating speed of the motor; determining a dynamic safety margin coefficient of a weakening area voltage-current working area based on the weakening area voltage available boundary curve; judging the motor rotating speed of the permanent magnet synchronous motor; by introducing voltage-current working area modeling and a dynamic safety margin coefficient, accurate current control and adaptive inverter compensation of the PMSM in a weak magnetic area are realized, and the problems of poor stability, insufficient voltage margin and incomplete compensation in traditional high-speed control are solved.

Description

Technical Field

[0001] This invention relates to the field of design optimization technology, and in particular to a control design optimization method and system for permanent magnet synchronous motors. Background Technology

[0002] Early PMSM control primarily relied on PI regulators to achieve decoupled control of the dq-axis current, but this resulted in problems such as field weakening control, insufficient voltage margin, and torque ripple in the high-speed region. Subsequently, advanced control algorithms such as Model Predictive Control (MPC), Sliding Mode Control (SMC), and Adaptive Control were gradually introduced to enhance system robustness and dynamic performance. Meanwhile, the impact of motor parameter variations on control performance has become increasingly prominent, prompting researchers to develop methods such as online parameter identification and nonlinear observers to improve control accuracy. Furthermore, the widespread adoption of digital controllers and high-performance microcontrollers has provided hardware support for the real-time implementation of complex control strategies. However, traditional field weakening control relies on empirically set negative q-axis current distribution above rated speed, which can easily lead to system instability in the field weakening region. Simultaneously, it is difficult to assess the dynamic voltage-current margin in the field weakening region in real time, potentially causing the inverter output to exceed the linear modulation range at high speeds, leading to overmodulation or overvoltage problems. Summary of the Invention

[0003] Therefore, it is necessary to provide a control design optimization method and system for permanent magnet synchronous motors to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a control design optimization method for permanent magnet synchronous motors is provided, the method comprising the following steps:

[0005] Step S1: Obtain and measure the stator three-phase voltage and current values ​​and motor speed of the PMSM in real time; construct the equivalent circuit model of the d-q axis motor, and calculate the usable boundary curve of the weakened region voltage by combining the stator three-phase voltage values ​​and motor speed;

[0006] Step S2: Determine the dynamic safety margin coefficient of the voltage-current working region of the weakened region based on the available boundary curve of the weakened region voltage; determine the motor speed of the permanent magnet synchronous motor. When the motor speed is greater than the preset rated speed, perform negative q-axis current distribution on the current value according to the voltage-current working region of the weakened region to obtain the optimal q-axis current.

[0007] Step S3: Input the optimal q-axis current and the d-axis current from the current value into the field-oriented controller, and combine the dynamic safety margin coefficient to limit and compensate the inverter output of the permanent magnet synchronous motor, generating inverter control compensation data;

[0008] Step S4: Perform control simulation on the inverter control compensation data to obtain simulation results; use the simulation results to dynamically adjust the dynamic safety margin coefficient to perform control design optimization of the permanent magnet synchronous motor.

[0009] This invention acquires real-time stator three-phase voltage, current, and speed data of a permanent magnet synchronous motor (PMSM) to construct an equivalent circuit model of the d-q axis motor. By combining this model with the available boundary curve of the weakened field region voltage, the voltage-current operating area and its safety margin coefficient are dynamically calculated, effectively identifying the operating boundary of the field weakening region. When the motor speed exceeds the rated speed, the optimal allocation of the negative q-axis current can be automatically performed based on the operating area, achieving fine-tuning of current control during the field weakening control process, thereby improving the motor's operating efficiency and control accuracy under high-speed conditions. Simultaneously, this scheme introduces a dynamic safety margin coefficient, which, along with the optimal d-q axis current, is input to the field-oriented controller to achieve dynamic limiting and voltage / current compensation control of the inverter output. This suppresses overvoltage and overcurrent problems at the source, enhancing the stability of the coordinated control between the inverter and the motor. Furthermore, by simulating and feeding back the control compensation data, the safety margin coefficient is dynamically updated, thereby constructing a closed-loop optimized control mechanism for operation in the field weakening region. This effectively solves the problems of inaccurate voltage boundaries, coarse current distribution, and system response lag in traditional magnetic field weakening strategies, significantly improving the system stability, safety, and energy efficiency ratio of the PMSM under high-speed field weakening operation, while extending the service life of the inverter and motor. It has significant engineering application value and promotional significance. Therefore, this invention, by introducing voltage-current operating region modeling and a dynamic safety margin coefficient, achieves precise current control and adaptive inverter compensation for the PMSM in the field weakening region, solving the problems of poor stability, insufficient voltage margin, and imperfect compensation in traditional high-speed control.

[0010] Preferably, step S1 includes the following steps:

[0011] Step S11: Acquire and measure the stator three-phase voltage and current values ​​of the PMSM in real time, and generate real-time three-phase electrical parameter data;

[0012] Step S12: Perform coordinate transformation on the real-time three-phase electrical parameter data, convert the three-phase stationary coordinate system data into d-q rotating coordinate system data, and generate d-q axis voltage and current data;

[0013] Step S13: Collect the current motor speed and filter and smooth the motor speed to generate stable motor speed data; based on the stable motor speed data and d-q axis voltage and current data, construct the d-q axis motor equivalent circuit model and obtain the equivalent circuit parameter data;

[0014] Step S14: Using equivalent circuit parameter data and stable motor speed data, perform voltage limit state analysis to generate the analytical expression of the usable boundary of the weakened region voltage; based on the analytical expression of the usable boundary of the weakened region voltage, perform space vector amplitude discretization calculation on the stator three-phase voltage values ​​to generate the usable boundary curve data of the weakened region voltage.

[0015] This invention generates high-frequency sampled three-phase electrical parameter data by real-time measurement of stator three-phase voltage and current, providing accurate raw input for subsequent modeling and significantly improving the response speed and data reliability of the motor control system under dynamic conditions. By introducing a transformation from a stationary coordinate system to a rotating coordinate system (such as the Clarke-Park transformation), the three-phase quantities are unified into a d-q axis form, enabling precise implementation of the field-oriented control logic in a rotating reference frame. This achieves a unified control coordinate system for voltage, current, and flux linkage, enhancing the controller model's expressive power. By acquiring motor speed and performing filtering and smoothing, speed fluctuations and measurement errors during high-speed operation are effectively eliminated, ensuring more stable and reliable speed data for modeling and improving the accuracy and stability of d-q axis equivalent circuit model parameter estimation. An analytical method is used to construct the usable boundary expression for the weakened region voltage, and based on this, discrete calculations of the space vector amplitude are performed to generate weakened region voltage boundary curve data. This is the first time that the safe operating boundary of the PMSM under high-speed field weakening conditions has been transformed from "empirical adjustment" to "mathematical model-driven," laying a theoretical foundation for the refinement of subsequent control strategies. This step chain realizes a closed-loop processing flow from raw data acquisition → coordinate system transformation → model construction → voltage boundary extraction, ensuring that the pre-judgment foundation of the entire field weakening control strategy is solid, the data is real-time, and the model is reliable. It is particularly suitable for the field weakening operation control requirements of PMSM under rapidly changing operating conditions.

[0016] Preferably, step S12 includes the following steps:

[0017] Step S121: Perform missing value correction and noise filtering on the real-time three-phase voltage and current data to generate stable three-phase electrical parameter data;

[0018] Step S122: Perform Clarke transformation based on the stabilized three-phase electrical parameter data to generate voltage and current data in the α–β stationary coordinate system;

[0019] Step S123: Obtain motor position data through the position sensor on the permanent magnet synchronous motor, calculate the motor angular velocity using the motor speed, integrate to obtain the current synchronous rotation angle, and generate synchronous rotation angle data;

[0020] Step S124: Construct a two-dimensional rotation matrix using synchronous rotation angle data; apply the rotation matrix to the voltage and current data in the α–β stationary coordinate system to perform Park transformation, generating d–q axis voltage and current data.

[0021] This invention addresses common issues in field sampling, such as sensor jitter, signal loss, and noise interference, by introducing missing value correction and noise filtering mechanisms. This ensures smoother and more stable input three-phase voltage and current data, improving the reliability of subsequent control calculations. Through the Clarke transformation in step S122, the three-phase system is reduced to α–β stationary coordinate system data, simplifying the computational complexity of the analysis model and making the electrical parameters more intuitive, facilitating rapid coordinate transformation and forward model establishment. Angular position information is collected by position sensors, and the current synchronization angle is calculated by combining dynamic speed integration, avoiding the time error and angle drift problems caused by traditional fixed sampling frequency integration. This provides a more accurate physical basis for constructing a synchronous rotating reference system. A two-dimensional rotation matrix is ​​constructed using the synchronization angle to accurately map the voltage and current data in the α–β stationary coordinate system to the d–q axis reference system, achieving the Park transformation. This process ensures that the current control remains synchronized with the rotor flux linkage, improving the stability and response speed of the field-oriented control (FOC). The entire S12 process not only establishes a full-chain coordinate transformation system from three phases to the d–q axis, but also introduces actual angle synchronization and real-time rotation reference, which significantly enhances the system's ability to dynamically track the direction of the magnetic field under high-speed operation and load fluctuations, and effectively improves the current decoupling accuracy and torque response rate.

[0022] Preferably, step S2 includes the following steps:

[0023] Step S21: Resample the discrete points of the available boundary curves of the voltage in the weakened region, extract the boundary points of the limit operation envelope, and generate a voltage-current envelope point set;

[0024] Step S22: Calculate the dynamic operating limit ellipse in the dq current coordinate system based on the voltage-current envelope point set data to construct a dynamic working region model of the weakened region;

[0025] Step S23: Filter the d-q axis current data in the d-q axis voltage and current data, and calculate the voltage-current safety margin distance under the current operating condition based on the d-q axis current data of the dynamic working area model of the weakened region, and generate the dynamic safety margin coefficient.

[0026] Step S24: Determine whether to enter the field weakening zone based on the current motor speed. When the motor speed is greater than the preset rated speed, enter the field weakening mode. After entering the field weakening mode, use the dynamic safety margin coefficient as a constraint condition to reconstruct the current value required by the current load torque demand using the q-d component, perform negative q-axis current distribution calculation, and obtain the optimal q-axis current value.

[0027] This invention addresses the issues of blurred boundaries and inaccurate approximations in traditional models by discretizing and resampling the voltage boundary curve in the weakened region to extract key operating envelope points. This reconstructs a more realistic "feasible operating boundary" that closely reflects the actual voltage limit, providing a geometric foundation for current regulation. By introducing an elliptical operating limit model in the d-q axis current coordinates based on envelope point set fitting, it achieves, for the first time, mathematical modeling of the dynamic operating limit region within the field weakening zone. This ensures that current regulation not only adheres to the maximum vector amplitude constraint but also couples with dynamic operating conditions in real time, enhancing the real-time interpretability of the control system for electrical limits. By calculating the distance between the current d-q axis current and its relative position within the elliptical model, a dynamic safety margin coefficient is obtained. This allows for a direct assessment of the safe distance from the voltage-current limit boundary under the current operating condition, preventing overvoltage and overcurrent. This provides a "soft limiting" mechanism, improving the safety redundancy of the system under field weakening control. After entering the field weakening mode, the dynamic safety margin coefficient is used as a physical constraint to reconstruct the q–d current component and optimally allocate the negative q-axis current. This upgrades the field weakening control from "coarse adjustment based on speed threshold" to "fine optimization based on real-time boundary awareness," improving the accuracy and efficiency of current control under high-speed load variations. The overall process enables the controller to maintain the accuracy of flux linkage control even in the high-speed over-rated region, avoiding problems such as excessive flux linkage weakening, voltage saturation, and system oscillation that are prone to occur in traditional field weakening strategies. This enhances the speed extension capability and safe operating range of the PMSM, making it particularly suitable for applications such as electric vehicles and wind power generation that require frequent operation in high-speed regions.

[0028] Preferably, step S22 includes the following steps:

[0029] Step S221: Based on the voltage-current envelope point set data, perform minimum circumscribed ellipse fitting, calculate the principal axis length, direction angle and center point coordinates of the fitted ellipse, and generate the limit ellipse geometric parameter data;

[0030] Step S222: Based on the geometric parameter data of the limit ellipse, construct the function expression of the limit ellipse in the d–q coordinate system and generate the expression data of the limit running ellipse;

[0031] Step S223: Generate the d-q current coordinate grid region using the limit operation ellipse expression data, and mark the interior of the ellipse envelope as the feasible operating domain of the weak magnetic region, and finally construct the dynamic working region model of the weakened region.

[0032] This invention, through minimum circumscribed ellipse fitting of sub-envelope points, not only effectively encompasses all voltage-current limit points, but also yields parameters such as principal axis length, direction angle, and center point with clear geometric and physical meanings. This provides a precise boundary closed region for modeling the weak magnetic field region, significantly outperforming traditional polygon or empirical formula approximations. The ellipse's functional expression (such as a standard quadratic equation) is constructed based on the fitted geometric parameters and can be directly used by the controller for operating point determination and margin analysis, avoiding computational discontinuities and error accumulation caused by numerical table lookups or discrete point approximations. The limit elliptic function is mapped onto a d-q current coordinate system grid, and the region within the ellipse's envelope is marked as the legal operating area, intuitively defining the current control range. This "graphical domain construction + formulaic evaluation" method can directly serve functions such as online judgment, graphical simulation, and limit collision warning, improving the intuitiveness and debuggability of the motor control system. By dynamically constructing an adjustable field-weakening operating region (rather than statically setting boundaries), the control system can adaptively adjust the elliptic parameters according to voltage, current, load, motor temperature rise, and other conditions. This provides a more flexible and safer operating margin under various high-speed operating conditions, improving system robustness and overall performance. The elliptic expression and operating domain formed in this step are not only the "physical constraints" for current control in the field-weakening region, but also the core boundary reference model for current scheduling calculations in subsequent steps (such as dynamic margin distance calculation and the optimal solution for the negative q-axis current).

[0033] Preferably, step S3 includes the following steps:

[0034] Step S31: Input the optimal q-axis current and the d-axis current from the current value into the field orientation controller, and perform field vector superposition calculation on the optimal q-axis current data and d-axis current data to generate comprehensive field current vector data;

[0035] Step S32: Based on the comprehensive magnetic field current vector data, the inverter voltage reference is nonlinearly limited by the dynamic safety margin coefficient to generate the limited inverter voltage control signal data.

[0036] Step S33: Perform inverter dead-time compensation on the limited inverter voltage control signal data to generate dead-time compensation correction control data;

[0037] Step S34: Perform dynamic carrier slope adjustment based on dead zone compensation correction control data to generate final inverter control compensation data.

[0038] This invention vector-superimposes the d-axis current with the optimal q-axis current to form a comprehensive magnetic field current vector. This allows current control to not only possess directionality (the q-axis determines torque, and the d-axis controls flux linkage) but also incorporates the optimally allocated q-axis value, further improving the accuracy of field-oriented control (FOC) and the motor's response to load torque changes. Based on the comprehensive current vector and dynamic safety margin coefficient, nonlinear limiting is applied to the inverter output voltage reference signal. This automatically limits the output voltage amplitude according to current operating conditions (such as high-speed field weakening zones), preventing the system from entering an overvoltage state and effectively improving the stability and safety margin of operation in the field weakening zone. Real-time compensation and correction are performed to address the dead-time effect generated by the inverter switching devices, avoiding output waveform distortion, phase voltage distortion, and torque fluctuations caused by dead-time voltage, ensuring accurate reproduction and transmission of the PWM signal on the inverter side. By dynamically adjusting the slope of the PWM carrier, the final generated control compensation signal maintains appropriate response bandwidth and modulation depth under different operating conditions (such as high speed, field weakening, and high-frequency load changes), thereby improving the tracking performance, smoothness, and spectral quality of the inverter control signal. The overall S3 process forms an integrated inverter control compensation chain that starts from the physical magnetic field vector → dynamic limiting → dead zone modeling → pulse width modulation optimization. This significantly improves the robustness and output linearity of the inverter system under complex operating conditions such as dynamic, high speed, and weak magnetic fields, effectively extending device life and reducing losses.

[0039] Preferably, step S33 includes the following steps:

[0040] Step S331: Perform edge delay compensation on the limited inverter voltage control signal data to generate edge-aligned control data;

[0041] Step S332: Perform positive and negative half-cycle symmetry reconstruction on the edge alignment control data to generate dead zone symmetry reconstruction data;

[0042] Step S333: Perform pulse width modulation processing on the dead-zone symmetric reconstruction data to generate pulse width expansion control signal data;

[0043] Step S334: Perform zero-crossing offset correction on the pulse width extension control signal data to generate compensated inverter voltage control signal data;

[0044] Step S335: Perform dynamic modeling of switching device hysteresis on the compensated inverter voltage control signal data to generate dynamic hysteresis correction model data; perform device conduction state mapping on the dynamic hysteresis correction model data to generate dead-zone compensation correction control data.

[0045] This invention effectively corrects the edge misalignment problem of the control signal caused by device switching delay or sampling timing error by compensating for the sampling edge delay of the inverter voltage control signal, thereby improving the edge synchronization of the PWM signal in the actual device and enhancing the timing control accuracy of the drive system. Performing positive and negative half-cycle symmetrical reconstruction on the signal can repair waveform distortion, DC component drift, and low-frequency oscillation caused by asymmetrical switching behavior under the influence of dead time, optimizing voltage output quality and reducing interference with torque control accuracy. Performing pulse width expansion operation in S333 not only maintains the required average output voltage in the presence of dead time but also provides greater modulation margin in the weak magnetic field region or high speed region, thus ensuring that the inverter's output capability in the critical voltage region is not suppressed. Performing zero-crossing offset correction on the PWM signal solves the problems of waveform jumps and phase voltage abrupt changes caused by zero-crossing offset in traditional control systems, enabling the control system to maintain precise synchronization of current and voltage during low speed, low current, or commutation processes. By performing hysteresis dynamic modeling and conduction state mapping on the control signal, and fully considering the actual conduction delay, charge accumulation, and turn-off response of the devices, an adaptive correction mechanism is established. This ensures that dead-time compensation is not merely a static offset, but a precise response with time-dynamic characteristics, thereby improving the overall linearity and hysteresis tolerance of the system. In summary, this mechanism significantly reduces output voltage errors, three-phase imbalances, current fluctuations, and torque ripple caused by dead-time effects, providing a more realistic, accurate, and dynamically matched PWM control foundation for subsequent inverter drive control.

[0046] Preferably, step S4 includes the following steps:

[0047] Step S41: Construct a model input for the inverter control compensation data, import it into the permanent magnet synchronous motor control simulation platform, and generate control simulation model input data;

[0048] Step S42: Perform dynamic electromagnetic response simulation on the input data of the control simulation model to obtain simulation results;

[0049] Step S43: Extract the current over-limit risk area from the simulation results and generate current margin fluctuation data;

[0050] Step S44: Perform error analysis on the current margin fluctuation data and dynamic safety margin coefficient, generate dynamic safety margin adjustment data, perform surface fitting and boundary reconstruction on the dynamic safety margin adjustment data, and output updated dynamic safety margin coefficient data to perform control design optimization work for permanent magnet synchronous motor.

[0051] This invention directly imports inverter control compensation data into the simulation platform and fuses it with motor model parameters to generate control simulation input data. This ensures that the simulation conditions perfectly match the actual control scenario, improving the accuracy of the control simulation model's response to real system behavior and the modeling precision. Performing dynamic electromagnetic response simulation allows for early evaluation of the phase current, voltage response, flux linkage changes, and torque output of the control strategy during actual operation. This is particularly suitable for rapid verification of control stability, linearity, and margin boundaries during high-speed, weak-field operation, enhancing the system's predictive control capabilities. By extracting the current over-limit region from the simulation results, the actual margin fluctuation behavior of the current control strategy under different d-q current combinations is quantified, helping to identify "hidden control danger zones" and "fluctuation blind zones," establishing a more robust dynamic safety judgment mechanism for the system. Through error analysis between current margin fluctuations and preset dynamic safety margin coefficients, a feedback closed loop is formed. Combined with surface fitting and boundary reconstruction methods, the dynamic safety margin coefficients on which the controller relies are automatically updated, realizing the intelligent evolution of the system's boundary perception and optimization capabilities during operation. The overall S4 process realizes a complete loop path from control output → simulation analysis → risk identification → margin callback → parameter redesign, endowing the controller with "learning ability" and "self-evolution ability", providing a highly adaptive control model update mechanism for PMSM to operate under complex conditions in the future.

[0052] Preferably, step S42 includes the following steps:

[0053] Step S421: Perform stator winding current simulation modeling on the input data of the control simulation model to generate stator current simulation data;

[0054] Step S422: Perform numerical calculation of the flux linkage closed path on the input data of the control simulation model to generate internal flux linkage response data of the motor;

[0055] Step S423: Perform time-series dynamic superposition of stator current simulation data and flux linkage response data to generate electromagnetic transient response composite data; identify energy dissipation paths from the electromagnetic transient response composite data and extract response energy distribution data;

[0056] Step S424: Identify critical operating conditions from the response energy distribution data, extract simulation boundary state data, and generate simulation results.

[0057] This invention, through simulation modeling of the stator winding current response of a motor, accurately simulates the dynamic effects of control signals during electromagnetic conversion, including high-frequency components, current abrupt changes, and harmonic characteristics. This provides a realistic scenario reference for subsequent control strategy limiting, compensation, and current margin judgment. By precisely solving the closed-loop path of the internal magnetic flux linkage of the motor, considering complex factors such as rotor position, winding distribution, and reluctance changes, more realistic magnetic flux linkage response data is obtained, overcoming the problem of rough estimation of magnetic flux linkage changes in traditional control systems and improving the physical consistency of field-oriented control. Introducing the temporal superposition of current and magnetic flux linkage reconstructs the complete electromagnetic transient behavior induced by the control signal in the permanent magnet synchronous motor. Simultaneously, energy dissipation path analysis is performed to identify energy transfer paths, dissipation links, and local power densities, facilitating accurate analysis of heating areas, voltage backlash, and efficiency fluctuation sources. By identifying critical states of energy distribution, boundary conditions such as instability, saturation, and power overload generated under the influence of control strategies are extracted, achieving simulation-driven "results-oriented" condition warnings. This helps control engineers identify and mitigate strategy risks during the design phase. The overall S42 process unifies and links the modeling of "current control behavior, magnetic flux physical evolution and energy distribution effect", forming a closed-loop structure of control simulation from the logic layer to the physical layer and then to the energy layer. This makes the simulation more comprehensive, more refined and more predictive, and is an important foundation for realizing digital twin-driven control optimization.

[0058] This specification provides a control design optimization system for permanent magnet synchronous motors, used to execute the aforementioned control design optimization method for permanent magnet synchronous motors. The control design optimization system for permanent magnet synchronous motors includes:

[0059] The region identification module is used to acquire and measure the stator three-phase voltage and current values ​​and motor speed of the PMSM in real time; construct the equivalent circuit model of the d-q axis motor; and calculate the usable boundary curve of the weakened region voltage by combining the stator three-phase voltage values ​​and motor speed.

[0060] The current distribution module is used to determine the dynamic safety margin coefficient of the voltage-current working area of ​​the weakened region based on the available boundary curve of the weakened region voltage; to determine the motor speed of the permanent magnet synchronous motor; when the motor speed is greater than the preset rated speed, the current value is distributed to the negative q-axis current according to the voltage-current working area of ​​the weakened region to obtain the optimal q-axis current.

[0061] The output compensation module is used to input the optimal q-axis current and the d-axis current from the current value to the field-oriented controller, and combine the dynamic safety margin coefficient to limit and compensate the inverter output of the permanent magnet synchronous motor, generating inverter control compensation data.

[0062] The dynamic adjustment module is used to perform control simulation on the inverter control compensation data and obtain simulation results. The dynamic safety margin coefficient is dynamically adjusted using the simulation results to perform control design optimization of the permanent magnet synchronous motor.

[0063] The beneficial effects of this invention are as follows: The region identification module measures the stator three-phase voltage, current, and speed in real time, and, combined with the d-q axis motor equivalent circuit model, accurately calculates the usable boundary curve of the weakened region voltage, ensuring that the system makes control judgments and formulates strategies based on real and dynamic operating data. The current distribution module dynamically calculates the safety margin coefficient based on the weakened region voltage-current boundary curve, reasonably judges whether the system has entered the field weakening region, and achieves optimal current distribution within the field weakening region through a negative q-axis current distribution strategy, enhancing the stability and efficiency of the motor at high speeds. The output compensation module inputs the optimal q-axis current and d-axis current into the field-oriented controller, and, combined with the safety margin coefficient, limits and compensates the inverter output signal, significantly reducing the risk of voltage overload and waveform distortion, and improving the system's dynamic response capability and control accuracy. The dynamic adjustment module, through simulation feedback of inverter control compensation data, promptly identifies current margin fluctuations and safety margin errors, dynamically adjusts the safety margin coefficient, forms a closed-loop optimization of the control design, and effectively improves the adaptive capability and long-term operational stability of the PMSM field weakening control. The overall system, through the organic collaboration between modules, achieves precise identification of the operating boundary in the field-weakening zone, optimized current allocation, inverter compensation output, and dynamic adjustment of the control strategy. This significantly improves the speed-up performance, control safety margin, and system stability of the permanent magnet synchronous motor (PMSM), adapting to complex loads and high-speed operation requirements. Therefore, this invention, by introducing voltage-current operating region modeling and a dynamic safety margin coefficient, achieves precise current control and adaptive inverter compensation for the PMSM in the field-weakening zone, solving the problems of poor stability, insufficient voltage margin, and incomplete compensation in traditional high-speed control. Attached Figure Description

[0064] Figure 1 This is a flowchart illustrating the steps of a control design optimization method for permanent magnet synchronous motors.

[0065] Figure 2 for Figure 1 A detailed flowchart illustrating the implementation steps of step S2.

[0066] Figure 3 for Figure 1 A detailed flowchart illustrating the implementation steps of step S3.

[0067] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0068] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0069] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0070] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0071] To achieve the above objectives, please refer to Figures 1 to 3 A control design optimization method for permanent magnet synchronous motors, the method comprising the following steps:

[0072] Step S1: Obtain and measure the stator three-phase voltage and current values ​​and motor speed of the PMSM in real time; construct the equivalent circuit model of the d-q axis motor, and calculate the usable boundary curve of the weakened region voltage by combining the stator three-phase voltage values ​​and motor speed;

[0073] Step S2: Determine the dynamic safety margin coefficient of the voltage-current working region of the weakened region based on the available boundary curve of the weakened region voltage; determine the motor speed of the permanent magnet synchronous motor. When the motor speed is greater than the preset rated speed, perform negative q-axis current distribution on the current value according to the voltage-current working region of the weakened region to obtain the optimal q-axis current.

[0074] Step S3: Input the optimal q-axis current and the d-axis current from the current value into the field-oriented controller, and combine the dynamic safety margin coefficient to limit and compensate the inverter output of the permanent magnet synchronous motor, generating inverter control compensation data;

[0075] Step S4: Perform control simulation on the inverter control compensation data to obtain simulation results; use the simulation results to dynamically adjust the dynamic safety margin coefficient to perform control design optimization of the permanent magnet synchronous motor.

[0076] This invention acquires real-time stator three-phase voltage, current, and speed data of a permanent magnet synchronous motor (PMSM) to construct an equivalent circuit model of the d-q axis motor. By combining this model with the available boundary curve of the weakened field region voltage, the voltage-current operating area and its safety margin coefficient are dynamically calculated, effectively identifying the operating boundary of the field weakening region. When the motor speed exceeds the rated speed, the optimal allocation of the negative q-axis current can be automatically performed based on the operating area, achieving fine-tuning of current control during the field weakening control process, thereby improving the motor's operating efficiency and control accuracy under high-speed conditions. Simultaneously, this scheme introduces a dynamic safety margin coefficient, which, along with the optimal d-q axis current, is input to the field-oriented controller to achieve dynamic limiting and voltage / current compensation control of the inverter output. This suppresses overvoltage and overcurrent problems at the source, enhancing the stability of the coordinated control between the inverter and the motor. Furthermore, by simulating and feeding back the control compensation data, the safety margin coefficient is dynamically updated, thereby constructing a closed-loop optimized control mechanism for operation in the field weakening region. This effectively solves the problems of inaccurate voltage boundaries, coarse current distribution, and system response lag in traditional magnetic field weakening strategies, significantly improving the system stability, safety, and energy efficiency ratio of the PMSM under high-speed field weakening operation, while extending the service life of the inverter and motor. It has significant engineering application value and promotional significance. Therefore, this invention, by introducing voltage-current operating region modeling and a dynamic safety margin coefficient, achieves precise current control and adaptive inverter compensation for the PMSM in the field weakening region, solving the problems of poor stability, insufficient voltage margin, and imperfect compensation in traditional high-speed control.

[0077] In this embodiment of the invention, reference Figure 1 The diagram shown is a flowchart illustrating the steps of a control design optimization method for a permanent magnet synchronous motor according to the present invention. In this example, the control design optimization method for a permanent magnet synchronous motor includes the following steps:

[0078] Step S1: Obtain and measure the stator three-phase voltage and current values ​​and motor speed of the PMSM in real time; construct the equivalent circuit model of the d-q axis motor, and calculate the usable boundary curve of the weakened region voltage by combining the stator three-phase voltage values ​​and motor speed;

[0079] Step S2: Determine the dynamic safety margin coefficient of the voltage-current working region of the weakened region based on the available boundary curve of the weakened region voltage; determine the motor speed of the permanent magnet synchronous motor. When the motor speed is greater than the preset rated speed, perform negative q-axis current distribution on the current value according to the voltage-current working region of the weakened region to obtain the optimal q-axis current.

[0080] Step S3: Input the optimal q-axis current and the d-axis current from the current value into the field-oriented controller, and combine the dynamic safety margin coefficient to limit and compensate the inverter output of the permanent magnet synchronous motor, generating inverter control compensation data;

[0081] Step S4: Perform control simulation on the inverter control compensation data to obtain simulation results; use the simulation results to dynamically adjust the dynamic safety margin coefficient to perform control design optimization of the permanent magnet synchronous motor.

[0082] In this embodiment of the invention, high-precision three-phase voltage and current sensors are used to collect the three-phase voltage and current values ​​on the stator side of the PMSM (Permanent Magnet Synchronous Motor), and the current motor speed is collected simultaneously. The voltage and current data sampling frequency is set to no less than 10kHz to ensure dynamic response accuracy. The collected voltage and current data are converted into equivalent voltage and current values ​​in the d-q axis coordinate system through Clark and Park transforms. Subsequently, the d-q axis motor equivalent circuit is constructed based on the standard PMSM mathematical model, including stator inductance, back electromotive force term, and stator resistance elements. Combining the converted voltage, current, and motor speed data, a voltage-current mapping relationship in the field-weakening region is constructed. Furthermore, based on the maximum modulation ratio constraint, the available voltage vector boundary is calculated, and the available voltage boundary curve of the field-weakening region is plotted. This curve is used for subsequent operating point scheduling. The available voltage boundary curve of the field-weakening region is discretized into several voltage-current operating blocks. For each block, the feasible operating boundary for each operating point at different motor speeds is calculated by combining the modulation ratio upper limit (e.g., 0.95) and the maximum voltage constraint. Then, a dynamic safety margin coefficient is calculated based on the tolerance width of each block. This coefficient is a proportional value between 0 and 1, reflecting the safe distance between the current operating point and the boundary. When the motor speed is detected to exceed the preset rated speed (e.g., 3000 rpm), the system dynamically adjusts the q-axis component of the current value in the negative direction according to the voltage-current boundary position of the current operating block, combined with the maximum torque per ampere (MTPA) strategy and the maximum speed limit, to allocate the optimal negative q-axis current value that meets the output requirements, ensuring stable motor operation without exceeding the inverter voltage limit. The optimal q-axis current obtained in the above steps, together with the converted d-axis current, is input as the current setpoint to the field-oriented controller (FOC). The field-oriented controller calculates the inverter output vector based on the input current value. Simultaneously, combined with the current dynamic safety margin coefficient, the amplitude of the inverter output vector is limited to avoid voltage overshoot. Simultaneously, fine-tuning compensation is performed using the inverter's voltage vector trajectory to generate inverter control compensation data. This control compensation data includes parameters such as three-phase reference voltage commands, modulation depth, and safety limits for triggering protection mechanisms, and is fed back to the inverter drive module as a real-time control signal. A complete PMSM control system simulation model is built in the MATLAB / Simulink environment using the generated inverter control compensation data. Time-domain response simulation of the control behavior is performed, with a simulation period of at least 5 seconds and a simulation time step of 10 μs. During the simulation, the changes in voltage margin in the field weakening region, inverter modulation ratio fluctuations, output torque change rate, and d-q axis current coupling degree are observed.Based on the simulation output of d-q axis current trajectories, voltage vector limit utilization, system stability margin, and other indicators, the calculation weights and boundary parameters of the dynamic safety margin coefficient are readjusted. For example, the q-axis current reduction ratio in the high-speed region is increased, or the safety window of the margin coefficient is adjusted. Ultimately, the iterative optimization of the permanent magnet synchronous motor control system design is completed. The entire process can be automated with the help of scripts to adjust parameters and automatically generate control optimization reports and parameter recommendation tables based on the simulation output results.

[0083] Preferably, step S1 includes the following steps:

[0084] Step S11: Acquire and measure the stator three-phase voltage and current values ​​of the PMSM in real time, and generate real-time three-phase electrical parameter data;

[0085] Step S12: Perform coordinate transformation on the real-time three-phase electrical parameter data, convert the three-phase stationary coordinate system data into d-q rotating coordinate system data, and generate d-q axis voltage and current data;

[0086] Step S13: Collect the current motor speed and filter and smooth the motor speed to generate stable motor speed data; based on the stable motor speed data and d-q axis voltage and current data, construct the d-q axis motor equivalent circuit model and obtain the equivalent circuit parameter data;

[0087] Step S14: Using equivalent circuit parameter data and stable motor speed data, perform voltage limit state analysis to generate the analytical expression of the usable boundary of the weakened region voltage; based on the analytical expression of the usable boundary of the weakened region voltage, perform space vector amplitude discretization calculation on the stator three-phase voltage values ​​to generate the usable boundary curve data of the weakened region voltage.

[0088] In this embodiment of the invention, a high-speed sampling module is used to measure the stator three-phase voltage and three-phase current values ​​generated during the operation of a permanent magnet synchronous motor (PMSM) in real time. Voltage and current sensors with a high sampling rate of at least 20kHz are connected to the U, V, and W three-phase input lines respectively, and the collected data constitutes real-time three-phase electrical parameter data. To ensure the stability and anti-interference capability of the measurement data, the sensor signals are filtered once by an analog low-pass filter, and then converted into digital signals by a 12-bit or higher precision analog-to-digital converter (ADC) module, and output synchronously in a timestamp manner to ensure the timing consistency of voltage and current sampling. A two-step transformation operation is performed on the acquired real-time three-phase electrical parameter data. First, the Clarke transform is used to convert the three-phase stationary coordinate system (a, b, c axes) data into two-phase stationary coordinate system (α, β axes) data. Then, the Park transform is further used to map the α and β axis data to the d–q axis coordinate system, which rotates synchronously with the rotor magnetic field, thereby obtaining the equivalent voltage and current values ​​in the d and q axis directions, and outputting the d–q axis voltage and current data. This step relies on the motor's current speed and electrical angle information. The rotor angle θ needs to be read using an encoder or Hall effect sensor, and the transformation formula parameters are updated in real time using sin(θ) and cos(θ) to ensure the dynamic accuracy of the transformation results. The motor's current speed data is collected at a sampling frequency of 1kHz or higher using a high-speed incremental photoelectric encoder with position encoding. To eliminate speed fluctuations caused by mechanical vibration or signal interference, the sampled raw speed data is smoothed using a weighted moving average filter with a filter window length of 50ms, outputting stable motor speed data. Next, combining the obtained stable motor speed data with d-q axis voltage and current data, an equivalent circuit model of the d-q axis is constructed according to the standard PMSM mathematical model formula. The model includes circuit parameters such as stator resistance, d-axis and q-axis inductance, and back electromotive force. These electrical parameters are calculated using a least-squares estimation method based on the steady-state motor equations, ultimately outputting the equivalent circuit parameter data. Using the equivalent circuit parameter data and the stable motor speed data, a voltage limit state analysis under PMSM operating conditions is performed. Based on the inverter's maximum output voltage vector constraint, a limit circular trajectory model is established, and the range of q-axis and d-axis current combinations at different speeds is derived. The usable d-q current space region under the voltage limit state is calculated. A mathematical analytical expression is then established, outputting the analytical formula for the usable voltage boundary in the weakened region. Subsequently, the stator voltage is uniformly discretized in the spatial vector domain using a space vector magnitude discretization method, with the sampling point interval controlled within 2°. Each discrete point is judged according to the analytical formula to determine whether it falls within the controllable range of the weakened region. Finally, complete usable voltage boundary curve data for the weakened region is plotted for subsequent negative q-axis current control and limiting strategy settings.All computation processes are executed using FPGA or DSP controllers, supporting edge real-time computing and data caching mechanisms.

[0089] Preferably, step S12 includes the following steps:

[0090] Step S121: Perform missing value correction and noise filtering on the real-time three-phase voltage and current data to generate stable three-phase electrical parameter data;

[0091] Step S122: Perform Clarke transformation based on the stabilized three-phase electrical parameter data to generate voltage and current data in the α–β stationary coordinate system;

[0092] Step S123: Obtain motor position data through the position sensor on the permanent magnet synchronous motor, calculate the motor angular velocity using the motor speed, integrate to obtain the current synchronous rotation angle, and generate synchronous rotation angle data;

[0093] Step S124: Construct a two-dimensional rotation matrix using synchronous rotation angle data; apply the rotation matrix to the voltage and current data in the α–β stationary coordinate system to perform Park transformation, generating d–q axis voltage and current data.

[0094] In this embodiment of the invention, the collected three-phase voltage and current data may experience short-term missing values ​​or noise fluctuations due to factors such as sensor jitter, electromagnetic interference, or communication packet loss. Therefore, missing value correction processing is first performed on each set of sampled data. A combination of forward interpolation and spline fitting is used to fill in the missing sampling points within the sampling period. Forward interpolation is used to fill in short-term (less than 3 sampling points) continuous missing intervals, while spline fitting is used to correct edge or discontinuous missing regions. Subsequently, a bandpass filter is applied to the completed data, with a filter bandwidth set to 10Hz to 5kHz. An FIR filter structure is used, with a filter order set to 64, to filter out high-frequency noise and DC offset terms, outputting stable three-phase voltage and current data. Clarke transform is performed on the stable three-phase voltage and current data obtained in step S121. The three-phase electrical quantities U, V, and W in the three-phase stationary coordinate system are mapped to two-phase orthogonal components on the α–β plane. The specific calculation method is based on a fixed transformation matrix. During this process, the Clarke transformation module is deployed within the vector transformation unit of the DSP controller. The transformation results are updated in real time, and the generated α-axis voltage and current and β-axis voltage and current are used as the coordinate rotation inputs for the next step. To obtain the precise position of the current motor, the rotary position encoder signal at the installation position of the permanent magnet synchronous motor is read. This encoder is an incremental photoelectric encoder with an output resolution of 2048 pulses / revolution. It reads the A, B, and Z channel signals in a differential manner to obtain the current rotor angular position. To improve the estimation accuracy of the angular velocity, the backward differential method is used to perform derivative calculations on the position data to calculate the current angular velocity of the motor; then, the synchronous rotation angle of the motor is obtained by integrating the angular velocity data. The integration uses the first-order trapezoidal integration method, with the integration period consistent with the sampling period (less than 100μs), and the output is the electrical angle data corresponding to each control cycle. Based on the synchronous rotation angle obtained in step S123, a two-dimensional rotation matrix is ​​constructed, which contains the sine and cosine values ​​of the current angle. This matrix is ​​used to convert the voltage and current vectors in the α–β coordinate system into the d–q axis coordinate system. The specific operation involves performing matrix multiplication on the voltage and current data along the α and β axes with the rotation matrix to obtain the equivalent voltage and current values ​​along the d-axis (in the same direction as the rotor flux linkage) and the q-axis (perpendicular to the rotor flux linkage). This transformation process is implemented through the CORDIC (Coordinate Rotation Numerical Calculator) module in the DSP core to ensure efficient sine and cosine calculations and rotation operations without a multiplier, ultimately outputting the d–q axis voltage and current data for subsequent field weakening control logic.

[0095] As an example of the present invention, reference is made to... Figure 2 As shown, step S2 in this example includes:

[0096] Step S21: Resample the discrete points of the available boundary curves of the voltage in the weakened region, extract the boundary points of the limit operation envelope, and generate a voltage-current envelope point set;

[0097] Step S22: Calculate the dynamic operating limit ellipse in the dq current coordinate system based on the voltage-current envelope point set data to construct a dynamic working region model of the weakened region;

[0098] Step S23: Filter the d-q axis current data in the d-q axis voltage and current data, and calculate the voltage-current safety margin distance under the current operating condition based on the d-q axis current data of the dynamic working area model of the weakened region, and generate the dynamic safety margin coefficient.

[0099] Step S24: Determine whether to enter the field weakening zone based on the current motor speed. When the motor speed is greater than the preset rated speed, enter the field weakening mode. After entering the field weakening mode, use the dynamic safety margin coefficient as a constraint condition to reconstruct the current value required by the current load torque demand using the q-d component, perform negative q-axis current distribution calculation, and obtain the optimal q-axis current value.

[0100] In this embodiment of the invention, the original boundary curve data is extracted from the available boundary curve of the weakened region voltage obtained in step S14. This data consists of continuous voltage-current boundary trajectory data points, representing the maximum allowable current operating range at different motor speeds. To facilitate subsequent modeling and calculation, the boundary curve is discretized using an equidistant resampling method. The sampling point spacing is set to a fixed d value (e.g., 0.1A), ensuring that the resampled point set covers all directions of the boundary contour. At each sampling point, the corresponding d-axis and q-axis current coordinate values ​​are extracted, and these are used to construct a voltage-current envelope point set. This envelope point set represents the maximum operating boundary of the current space in the form of a two-dimensional point cloud, used to characterize the limiting current constraint conditions of the motor in the weakened magnetic region. Geometric modeling is performed on the voltage-current envelope point set obtained in step S21. The least squares ellipse fitting method is used to model the boundary of the point set. During the fitting process, an ellipse model optimization method based on Algebraic Distance is used to ensure that the fitted curve remains closed, symmetrical, and stable. The ellipse parameters output by the fitting include the center coordinates, the major axis direction, the minor axis direction, and the lengths of the major and minor axes. The fitted ellipse defines the dynamic operating limit region of the current space in the d-q current plane, thus constructing a dynamic working region model for the weakened zone. The model results are expressed in elliptic analytical form and are updated synchronously with the current operating speed of the motor to reflect the dynamically adjusted safety boundary as the speed changes. The real-time d-axis and q-axis current values ​​obtained in step S124 are used as input data and substituted into the elliptic model expression generated in step S22 to calculate the minimum Euclidean distance from the operating point to the ellipse boundary, in amperes (A). This distance is the safety margin distance in the voltage-current space. Further normalization is applied, dividing this distance by the length of the minor axis of the ellipse to obtain a standardized dynamic safety margin coefficient. This safety margin coefficient is typically between 0 and 1. When the coefficient value is less than 0.2, it indicates that the current operating point is close to the limit boundary, prompting the control system to adjust the current command or reduce the load. The current operating speed of the motor is collected and compared with the preset rated speed, which is generally set at 4000 rpm. If the current speed is greater than this threshold, the system determines that it has entered the field weakening zone. Upon entering the weakening mode, based on the current load torque demand, the required total current vector magnitude is first calculated. Then, using the dynamic safety margin coefficient as a constraint, an iterative algorithm is invoked to perform q-d axis current decomposition calculation. The calculation prioritizes the q-axis current component that satisfies the load torque demand, and negative q-axis allocation is performed under the premise of satisfying the weakening boundary conditions, outputting the optimal q-axis current value. All calculations are executed in the real-time running thread of the motor controller, with a cycle not exceeding 100μs to ensure dynamic response capability. The final output optimal q-axis current value is used to drive the inverter for field modulation control, achieving the voltage limiting and torque control objectives of field weakening operation.

[0101] Preferably, step S22 includes the following steps:

[0102] Step S221: Based on the voltage-current envelope point set data, perform minimum circumscribed ellipse fitting, calculate the principal axis length, direction angle and center point coordinates of the fitted ellipse, and generate the limit ellipse geometric parameter data;

[0103] Step S222: Based on the geometric parameter data of the limit ellipse, construct the function expression of the limit ellipse in the d–q coordinate system and generate the expression data of the limit running ellipse;

[0104] Step S223: Generate the d-q current coordinate grid region using the limit operation ellipse expression data, and mark the interior of the ellipse envelope as the feasible operating domain of the weak magnetic region, and finally construct the dynamic working region model of the weakened region.

[0105] In this embodiment of the invention, the voltage-current envelope point set data obtained in step S21 is imported into a two-dimensional coordinate system, and the point set is visualized in the d-q axis coordinate plane. Then, the Khachiyan iterative algorithm is used to perform minimum bounding ellipse fitting on the point set. This algorithm is a convex optimization method that can solve for the minimum area ellipse containing all input points in two-dimensional space. During the calculation, the input is the d-axis and q-axis coordinate values ​​(in amperes) of all current envelope points, and the output is the geometric parameters of the fitted ellipse: including the principal axis lengths (the lengths of the major and minor axes, in amperes), the direction angle (the angle between the principal axis and the d-axis, in radians or degrees), and the coordinates of the ellipse center point (the center coordinates of the d-axis and q-axis directions). This process is achieved through matrix eigenvalue decomposition and weight update iteration, ultimately generating the limiting ellipse geometric parameter data, which is stored in the form of a structure, including fields such as center coordinates, principal axis lengths, and direction angles. Using the geometric parameters obtained in step S221, the limiting ellipse function expression in the d-q coordinate system is constructed. Specifically, the rotation angle, principal axis length, and center coordinates of the fitted ellipse in the d–q coordinates are substituted into the affine transformation expression of the standard ellipse equation to form a unified analytical function form. This expression is usually represented as a quadratic form of the rotated ellipse, used to describe whether any point (d, q) lies inside the ellipse. This function form facilitates subsequent region judgment and meshing processing. The final generated expression is saved as an analytical data structure, containing ellipse matrix coefficients and boundary judgment logic. Based on the limit ellipse expression from step S222, a uniformly distributed current coordinate mesh region is generated in the d–q plane, typically set to bidirectional equal-interval sampling between –I_max and +I_max, where I_max is usually set to the maximum allowable current value of the motor (e.g., 50A). At each mesh point, its d-axis and q-axis coordinate values ​​are substituted into the limit ellipse function expression to determine whether it falls inside the ellipse boundary. If the ellipse inequality condition is satisfied, the point is marked as a feasible operating point within the weakened magnetic region; otherwise, it is marked as an illegal or inoperable region. The set of all mesh points that meet the conditions is defined as the effective domain of the dynamic working region model of the weakened region. The final output of the region model data is in the form of a grid label matrix, which can be used for safety boundary constraint determination in real-time control algorithms. The entire model building process is preprocessed and generated during the initialization phase of the motor control system, or automatically reconstructed when the operating conditions change beyond a preset threshold, to ensure the real-time performance and accuracy of the boundary model.

[0106] As an example of the present invention, reference is made to... Figure 3 As shown, step S3 in this example includes:

[0107] Step S31: Input the optimal q-axis current and the d-axis current from the current value into the field orientation controller, and perform field vector superposition calculation on the optimal q-axis current data and d-axis current data to generate comprehensive field current vector data;

[0108] Step S32: Based on the comprehensive magnetic field current vector data, the inverter voltage reference is nonlinearly limited by the dynamic safety margin coefficient to generate the limited inverter voltage control signal data.

[0109] Step S33: Perform inverter dead-time compensation on the limited inverter voltage control signal data to generate dead-time compensation correction control data;

[0110] Step S34: Perform dynamic carrier slope adjustment based on dead zone compensation correction control data to generate final inverter control compensation data.

[0111] In this embodiment of the invention, the optimal q-axis current data calculated in step S24 is input together with the real-time d-axis current value from the current sampling module to the field-oriented controller (FOC control unit). Internally, the controller uses a vector calculation mechanism in the d-q axis coordinate system to combine the d-axis current (primarily controlling magnetic flux) and the q-axis current (primarily controlling torque) to form a two-dimensional current vector. A vector superposition algorithm is used: the d-axis current is taken as the real part, and the q-axis current as the imaginary part; their magnitude and phase in the complex plane are calculated to obtain the comprehensive field-current vector. This vector represents the direction and intensity of the total magnetic field generated by the stator and serves as the core input for the current loop control target, constructing the controller's real-time current feedback channel. Based on the comprehensive field-current vector data obtained in step S31, the inverter voltage output control quantity is calculated. Furthermore, the dynamic safety margin coefficient generated in step S23 is called to perform nonlinear limiting processing on the calculated inverter voltage reference value. The limiting processing employs a component-by-component limiting strategy, comparing the reference voltage value of each phase with the current vector projection, and dynamically adjusting the maximum allowable output amplitude according to the safety margin. The limiting method employs a bilateral sigmoid function for soft saturation control to avoid abrupt changes and outputs limited inverter voltage control signal data, keeping the inverter operating within a safe range and preventing overmodulation and saturation. Dead-time compensation is performed on the limited inverter voltage control signal data generated in step S32 using known dead-time parameters of the inverter drive circuit (typically 1.5μs to 2.5μs). The compensation method uses a logic control strategy based on current direction judgment: when the current phase current is positive, upper bridge arm drive delay correction is performed; when it is negative, lower bridge arm drive advance control is performed to compensate for the impact of dead time on the output voltage. The compensation amplitude for each phase is calculated based on the ratio of dead time to PWM period and is superimposed on the original voltage command using voltage feedforward compensation, ultimately generating dead-time compensation correction control data to improve the inverter's accuracy control under high-frequency switching. Based on the dead-time compensation correction control data obtained in step S33, the final PWM modulation waveform generation process is executed. The system calls the real-time carrier frequency scheduler to adjust the slope of the current PWM carrier. This process determines the rise and fall rates of the PWM triangular wave based on the phase change rate of the integrated magnetic field current vector and the trend of the dead-zone compensation amplitude, achieving dynamic slope adjustment. Specifically, when the current vector changes drastically, the carrier slope is increased to enhance the modulation response speed; when the change slows down, the carrier slope is decreased to reduce switching losses. The carrier frequency adjustment range is set to 8kHz to 20kHz. The resulting PWM waveform is used to control the conduction of the power devices in the three-phase inverter, outputting the final inverter control compensation data. This data is sent to the power drive module via a high-speed control interface, directly acting on the power conversion devices to achieve precise control of the motor's magnetic field and torque.

[0112] Preferably, step S33 includes the following steps:

[0113] Step S331: Perform edge delay compensation on the limited inverter voltage control signal data to generate edge-aligned control data;

[0114] Step S332: Perform positive and negative half-cycle symmetry reconstruction on the edge alignment control data to generate dead zone symmetry reconstruction data;

[0115] Step S333: Perform pulse width modulation processing on the dead-zone symmetric reconstruction data to generate pulse width expansion control signal data;

[0116] Step S334: Perform zero-crossing offset correction on the pulse width extension control signal data to generate compensated inverter voltage control signal data;

[0117] Step S335: Perform dynamic modeling of switching device hysteresis on the compensated inverter voltage control signal data to generate dynamic hysteresis correction model data; perform device conduction state mapping on the dynamic hysteresis correction model data to generate dead-zone compensation correction control data.

[0118] In this embodiment of the invention, the upper and lower bridge drive switching edges at the intersection of the PWM carrier and the voltage control signal are identified by sampling edge extraction processing of the limited inverter voltage control signal data. The extracted edge data includes the time point information of the rising and falling edges, in microseconds. Based on the typical edge response delay of the driver chip (e.g., 2μs), each edge position is shifted forward or backward by this fixed delay amount to achieve sampling edge alignment compensation. The compensated edge time points are reused to generate edge alignment control data to improve the synchronization of the PWM modulation process, ensure that the actual switching of the upper and lower bridge drive signals is consistent with the control command, and avoid misleading turn-on caused by edge offset. Based on the edge alignment control data in step S331, the positive and negative half-cycle symmetry is reconstructed. Specifically, for the positive half-cycle modulation signal in each PWM cycle, the corresponding negative half-cycle mirror is constructed using its voltage polarity information and edge change trend to ensure that the upper and lower bridge PWM drive signals have a symmetrical structure during the conduction period, thereby suppressing zero-sequence component and harmonic distortion caused by device switching asymmetry. This process constructs a mirror completion function based on the modulation signal polarity and dead time control logic, and the final output signal is the dead-time symmetrical reconstruction data. Pulse width modulation (PWM) processing is performed on the dead-time symmetrical reconstruction data generated in step S332. According to the preset dead-time compensation width coefficient in the controller (generally 1.05–1.15), a linear widening operation is performed on each conduction segment of the modulation signal, that is, the original conduction time is proportionally extended to keep the duty cycle of the modulation waveform within a safe threshold. The calculation logic of the PWM operation is based on the joint scheduling of the current drive frequency, dead time, and maximum duty cycle boundary controller to ensure that the PWM waveform does not exceed the limits after compensation, while providing redundant time for subsequent zero-crossing correction. The output of this operation is the pulse width expansion control signal data. Based on the pulse width expansion control signal data in step S333, zero-crossing offset correction processing is performed. This process mainly solves the misalignment problem between the current zero-crossing position and the voltage modulation signal, preventing cross-conduction or delayed conduction phenomena in the PWM waveform at the moment of zero crossing. The specific approach is as follows: The time information of the current phase current's zero-crossing point is acquired, and the control edges adjacent to it in the modulation signal are time-aligned and offset. If the zero-crossing point is found to be earlier than the modulation rising edge by more than a set threshold (e.g., 1.2μs), the control signal edge is advanced accordingly; otherwise, it is delayed. The output generated in this step is the compensated inverter voltage control signal data. Based on the compensated inverter voltage control signal data, a hysteresis dynamic model of the power switching devices is constructed. This model considers the turn-on delay, turn-off delay, and junction temperature influence factor of MOSFET or IGBT devices under different current directions. The typical modeling method is a first-order time-delay differential system, where the hysteresis time parameter is provided by the online device state detection module, in nanoseconds. This model maps the PWM signal input to the actual device conduction state sequence and generates dynamic hysteresis correction model data.Based on this, and combining the current conduction direction and hysteresis characteristics of each switching device, the final output voltage control signal is dynamically mapped to its conduction state, generating the final dead-time compensation correction control data. This data is directly transmitted to the power control unit to drive the three-phase inverter to achieve precise, symmetrical, and dead-time error-free drive control.

[0119] Preferably, step S4 includes the following steps:

[0120] Step S41: Construct a model input for the inverter control compensation data, import it into the permanent magnet synchronous motor control simulation platform, and generate control simulation model input data;

[0121] Step S42: Perform dynamic electromagnetic response simulation on the input data of the control simulation model to obtain simulation results;

[0122] Step S43: Extract the current over-limit risk area from the simulation results and generate current margin fluctuation data;

[0123] Step S44: Perform error analysis on the current margin fluctuation data and dynamic safety margin coefficient, generate dynamic safety margin adjustment data, perform surface fitting and boundary reconstruction on the dynamic safety margin adjustment data, and output updated dynamic safety margin coefficient data to perform control design optimization work for permanent magnet synchronous motor.

[0124] In this embodiment of the invention, inverter control compensation data is formatted to suit the input of the control simulation platform. This data includes three-phase voltage reference signals, dead-zone compensation correction data, and inverter switching state sequences, all of which are uniformly sampled and normalized by a digital signal processing module. The aforementioned signal data is imported into a permanent magnet synchronous motor (PMSM) control simulation platform, which is built on Matlab / Simulink or a similar high-precision dynamic simulation environment and includes a motor electromagnetic model, a mechanical load model, and a control algorithm module. The import process is completed through a simulation interface call, ensuring that the time series of the input data is synchronized with the simulation time axis, generating control simulation model input data that can be used for dynamic simulation. The control simulation model input data drives the simulation platform to perform electromagnetic response calculations. The simulation process employs a joint calculation method based on finite element analysis (FEA) and vector control algorithms to simulate the torque output, current changes, and flux linkage fluctuations of the PMSM under dynamic operating conditions. The simulation time step is set to the microsecond level to ensure accurate capture of high-speed changing characteristics. The simulation results include three-phase current waveforms, voltage response, speed, and torque output; all data are recorded in real time and output as a simulation result dataset. The current over-limit risk area is extracted from the simulation results. By setting a current threshold (e.g., the maximum allowable current is 1.2 times the rated current), the periods exceeding this threshold in the simulation are filtered and marked as current over-limit risk areas. The current fluctuation range within this area is statistically analyzed to calculate current margin fluctuation data, specifically including parameters such as fluctuation amplitude, duration, and frequency distribution. This data is used to assess the potential impact of current over-limit on motor performance and safety. Based on the current margin fluctuation data obtained in step S43 and the current dynamic safety margin coefficient, error analysis is performed. The error analysis uses multivariate regression and residual analysis methods to evaluate the deviation of the impact of current margin fluctuation on the safety margin coefficient, generating dynamic adjustment data for the safety margin. Subsequently, surface fitting is performed on the adjustment data, and polynomial regression or spline function fitting techniques are used to reconstruct the boundary of the safety margin coefficient change, ensuring that the adjustment coefficient changes smoothly within a reasonable operating range. Finally, the updated dynamic safety margin coefficient data is output, which is fed back into the control system to realize dynamic optimization and safety protection of the permanent magnet synchronous motor control design.

[0125] Preferably, step S42 includes the following steps:

[0126] Step S421: Perform stator winding current simulation modeling on the input data of the control simulation model to generate stator current simulation data;

[0127] Step S422: Perform numerical calculation of the flux linkage closed path on the input data of the control simulation model to generate internal flux linkage response data of the motor;

[0128] Step S423: Perform time-series dynamic superposition of stator current simulation data and flux linkage response data to generate electromagnetic transient response composite data; identify energy dissipation paths from the electromagnetic transient response composite data and extract response energy distribution data;

[0129] Step S424: Identify critical operating conditions from the response energy distribution data, extract simulation boundary state data, and generate simulation results.

[0130] In this embodiment of the invention, current simulation modeling is performed using a mathematical model of the stator winding current based on input data from a control simulation model. First, three-phase voltage signals and inverter control compensation data are collected as input. A set of differential equations for the stator winding current is constructed using the stator winding resistance, inductance, and back electromotive force parameters. The equations are then solved in the time domain using numerical integration methods (such as the fourth-order Runge-Kutta method) to calculate the three-phase stator current waveforms at various times, generating stator current simulation data. The simulation time resolution is set to the microsecond level to ensure high accuracy of the dynamic response. For the closed-loop magnetic flux linkage path within the motor, a flux linkage calculation model is constructed based on the motor's geometric parameters and the material's permeability. The finite element method is used to numerically solve the magnetic field distribution, calculating the flux linkage intensity and direction changes at different simulation time points. The flux linkage is then summed in a closed loop using the closed-loop integration method to obtain the motor's internal flux linkage response data. This process considers electromagnetic saturation effects and hysteresis characteristics to ensure the physical accuracy and dynamic realism of the flux linkage data. The stator current simulation data from step S421 and the flux linkage response data from step S422 are dynamically overlaid according to timestamps to form electromagnetic transient response composite data. Based on the composite data, an energy dissipation path identification algorithm is used to identify the distribution and transmission paths of electromagnetic energy between various magnetic circuits and windings using the principle of energy conservation, and to extract key response energy distribution data. This algorithm uses matrix decomposition and energy flow analysis techniques to reveal the energy transmission characteristics of the motor electromagnetic system. Based on the response energy distribution data, critical operating condition identification is performed. Threshold detection and pattern recognition techniques are used to screen abnormal states in the energy distribution that exceed the preset safety range, and the corresponding simulation boundary state data are extracted, including indicators such as current peak value, flux linkage limit, and energy fluctuation amplitude. Finally, these boundary state data are integrated to generate complete simulation results for subsequent performance evaluation and control parameter optimization.

[0131] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0132] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A control design optimization method for permanent magnet synchronous motors, characterized in that, Includes the following steps: Step S1: Obtain and measure the stator three-phase voltage and current values ​​and motor speed of the PMSM in real time; construct the equivalent circuit model of the d-q axis motor, and calculate the usable boundary curve of the weakened region voltage by combining the stator three-phase voltage values ​​and motor speed; Step S2: Determine the dynamic safety margin coefficient of the voltage-current working region of the weakened region based on the available boundary curve of the weakened region voltage; determine the motor speed of the permanent magnet synchronous motor. When the motor speed is greater than the preset rated speed, perform negative q-axis current distribution on the current value according to the voltage-current working region of the weakened region to obtain the optimal q-axis current. Step S3: Input the optimal q-axis current and the d-axis current from the current value into the field-oriented controller, and combine the dynamic safety margin coefficient to limit and compensate the inverter output of the permanent magnet synchronous motor, generating inverter control compensation data; Step S4: Perform control simulation on the inverter control compensation data to obtain simulation results; use the simulation results to dynamically adjust the dynamic safety margin coefficient to perform control design optimization of the permanent magnet synchronous motor.

2. The control design optimization method for permanent magnet synchronous motors according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Acquire and measure the stator three-phase voltage and current values ​​of the PMSM in real time, and generate real-time three-phase electrical parameter data; Step S12: Perform coordinate transformation on the real-time three-phase electrical parameter data, convert the three-phase stationary coordinate system data into d-q rotating coordinate system data, and generate d-q axis voltage and current data; Step S13: Collect the current motor speed and filter and smooth the motor speed to generate stable motor speed data; based on the stable motor speed data and d-q axis voltage and current data, construct the d-q axis motor equivalent circuit model and obtain the equivalent circuit parameter data; Step S14: Using equivalent circuit parameter data and stable motor speed data, perform voltage limit state analysis to generate the analytical expression of the usable boundary of the weakened region voltage; based on the analytical expression of the usable boundary of the weakened region voltage, perform space vector amplitude discretization calculation on the stator three-phase voltage values ​​to generate the usable boundary curve data of the weakened region voltage.

3. The control design optimization method for permanent magnet synchronous motors according to claim 2, characterized in that, Step S12 includes the following steps: Step S121: Perform missing value correction and noise filtering on the real-time three-phase voltage and current data to generate stable three-phase electrical parameter data; Step S122: Perform Clarke transformation based on the stabilized three-phase electrical parameter data to generate voltage and current data in the α–β stationary coordinate system; Step S123: Obtain motor position data through the position sensor on the permanent magnet synchronous motor, calculate the motor angular velocity using the motor speed, integrate to obtain the current synchronous rotation angle, and generate synchronous rotation angle data; Step S124: Construct a two-dimensional rotation matrix using synchronous rotation angle data; apply the rotation matrix to the voltage and current data in the α–β stationary coordinate system to perform Park transformation, generating d–q axis voltage and current data.

4. The control design optimization method for permanent magnet synchronous motors according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: Resample the discrete points of the available boundary curves of the voltage in the weakened region, extract the boundary points of the limit operation envelope, and generate a voltage-current envelope point set; Step S22: Calculate the dynamic operating limit ellipse in the dq current coordinate system based on the voltage-current envelope point set data to construct a dynamic working region model of the weakened region; Step S23: Filter the d-q axis current data in the d-q axis voltage and current data, and calculate the voltage-current safety margin distance under the current operating condition based on the d-q axis current data of the dynamic working area model of the weakened region, and generate the dynamic safety margin coefficient. Step S24: Determine whether to enter the field weakening zone based on the current motor speed. When the motor speed is greater than the preset rated speed, enter the field weakening mode. After entering the field weakening mode, use the dynamic safety margin coefficient as a constraint condition to reconstruct the current value required by the current load torque demand using the q-d component, perform negative q-axis current distribution calculation, and obtain the optimal q-axis current value.

5. The control design optimization method for permanent magnet synchronous motors according to claim 4, characterized in that, Step S22 includes the following steps: Step S221: Based on the voltage-current envelope point set data, perform minimum circumscribed ellipse fitting, calculate the principal axis length, direction angle and center point coordinates of the fitted ellipse, and generate the limit ellipse geometric parameter data; Step S222: Based on the geometric parameter data of the limit ellipse, construct the function expression of the limit ellipse in the d–q coordinate system and generate the expression data of the limit running ellipse; Step S223: Generate the d-q current coordinate grid region using the limit operation ellipse expression data, and mark the interior of the ellipse envelope as the feasible operating domain of the weak magnetic region, and finally construct the dynamic working region model of the weakened region.

6. The control design optimization method for permanent magnet synchronous motors according to claim 1, characterized in that, Step S3 includes the following steps: Step S31: Input the optimal q-axis current and the d-axis current from the current value into the field orientation controller, and perform field vector superposition calculation on the optimal q-axis current data and d-axis current data to generate comprehensive field current vector data; Step S32: Based on the comprehensive magnetic field current vector data, the inverter voltage reference is nonlinearly limited by the dynamic safety margin coefficient to generate the limited inverter voltage control signal data. Step S33: Perform inverter dead-time compensation on the limited inverter voltage control signal data to generate dead-time compensation correction control data; Step S34: Perform dynamic carrier slope adjustment based on dead zone compensation correction control data to generate final inverter control compensation data.

7. The control design optimization method for permanent magnet synchronous motors according to claim 6, characterized in that, Step S33 includes the following steps: Step S331: Perform edge delay compensation on the limited inverter voltage control signal data to generate edge-aligned control data; Step S332: Perform positive and negative half-cycle symmetry reconstruction on the edge alignment control data to generate dead zone symmetry reconstruction data; Step S333: Perform pulse width modulation processing on the dead-zone symmetric reconstruction data to generate pulse width expansion control signal data; Step S334: Perform zero-crossing offset correction on the pulse width extension control signal data to generate compensated inverter voltage control signal data; Step S335: Perform dynamic modeling of switching device hysteresis on the compensated inverter voltage control signal data to generate dynamic hysteresis correction model data; perform device conduction state mapping on the dynamic hysteresis correction model data to generate dead-zone compensation correction control data.

8. The control design optimization method for permanent magnet synchronous motors according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Construct a model input for the inverter control compensation data, import it into the permanent magnet synchronous motor control simulation platform, and generate control simulation model input data; Step S42: Perform dynamic electromagnetic response simulation on the input data of the control simulation model to obtain simulation results; Step S43: Extract the current over-limit risk area from the simulation results and generate current margin fluctuation data; Step S44: Perform error analysis on the current margin fluctuation data and dynamic safety margin coefficient, generate dynamic safety margin adjustment data, perform surface fitting and boundary reconstruction on the dynamic safety margin adjustment data, and output updated dynamic safety margin coefficient data to perform control design optimization work for permanent magnet synchronous motor.

9. The control design optimization method for permanent magnet synchronous motors according to claim 8, characterized in that, Step S42 includes the following steps: Step S421: Perform stator winding current simulation modeling on the input data of the control simulation model to generate stator current simulation data; Step S422: Perform numerical calculation of the flux linkage closed path on the input data of the control simulation model to generate internal flux linkage response data of the motor; Step S423: Perform time-series dynamic superposition of stator current simulation data and flux linkage response data to generate electromagnetic transient response composite data; identify energy dissipation paths from the electromagnetic transient response composite data and extract response energy distribution data; Step S424: Identify critical operating conditions from the response energy distribution data, extract simulation boundary state data, and generate simulation results.

10. A control design optimization system for permanent magnet synchronous motors, characterized in that, For executing the control design optimization method for a permanent magnet synchronous motor as described in claim 1, the control design optimization system for a permanent magnet synchronous motor comprises: The region identification module is used to acquire and measure the stator three-phase voltage and current values ​​and motor speed of the PMSM in real time; construct the equivalent circuit model of the d-q axis motor; and calculate the usable boundary curve of the weakened region voltage by combining the stator three-phase voltage values ​​and motor speed. The current distribution module is used to determine the dynamic safety margin coefficient of the voltage-current working area of ​​the weakened region based on the available boundary curve of the weakened region voltage; to determine the motor speed of the permanent magnet synchronous motor; when the motor speed is greater than the preset rated speed, the current value is distributed to the negative q-axis current according to the voltage-current working area of ​​the weakened region to obtain the optimal q-axis current. The output compensation module is used to input the optimal q-axis current and the d-axis current from the current value to the field-oriented controller, and combine the dynamic safety margin coefficient to limit and compensate the inverter output of the permanent magnet synchronous motor, generating inverter control compensation data. The dynamic adjustment module is used to perform control simulation on the inverter control compensation data and obtain simulation results. The dynamic safety margin coefficient is dynamically adjusted using the simulation results to perform control design optimization of the permanent magnet synchronous motor.

Images