A method for field weakening control based on adaptive parameter identification of permanent magnet synchronous motor

Abstract

The application provides a kind of weak magnetic control method based on permanent magnet synchronous motor adaptive parameter identification, it is related to motor control technical field, the method is by fusing the improved particle swarm optimization algorithm (CIPSO) of chaos inertia weight and Gaussian disturbance online identification stator resistance, cross-axis inductance and permanent magnet flux linkage, solve the problem of control performance decline caused by parameter time-varying;Below rated speed, maximum torque current ratio (MTPA) control based on real-time parameters is used, the convex pole effect is used to reduce the amplitude of stator current, improve the efficiency and dynamic response of low speed operation;Above rated speed, the voltage saturation criterion triggers the lead angle weak magnetic control, dynamically introduces the phase of lead angle correction current vector, makes the operating point move along the voltage limit ellipse boundary, breaks through the voltage constraint of inverter, while combining the parameter identification result carries out feedforward decoupling compensation, realizes the smooth switching from MTPA to weak magnetic and high speed area stable operation.

Description

Technical Field

[0001] This invention relates to the field of motor control technology, and specifically to a field weakening control method based on adaptive parameter identification of permanent magnet synchronous motors. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs), with their significant advantages of simple structure, small size, light weight, low loss, and high efficiency, are gradually replacing traditional asynchronous motors as the preferred solution for industrial drives and new energy vehicle power systems. In the current context of rapid development in industries such as new energy vehicles and high-end equipment manufacturing, unprecedentedly stringent requirements are being placed on the power density, speed range, and operating efficiency of drive motors. Among them, the integrated permanent magnet synchronous motor (IPMSM), through the salient pole effect formed by rotor magnetic circuit asymmetry, can generate a superposition effect of reluctance torque and permanent magnet torque, resulting in a torque density approximately 30% higher than that of traditional asynchronous motors within the same volume. Simultaneously, it possesses high mechanical strength and potential for speed expansion through field weakening, making it the preferred topology for passenger vehicle main drives, high-speed main shafts, and aerospace electric propulsion systems.

[0003] To unlock the full potential of IPMSMs, the industry commonly employs the Maximum Torque-to-Current Ratio (MTPA) vector control strategy. This strategy optimizes the current distribution along the quadrature and direct axes to achieve the minimum stator current for a given torque, thereby reducing copper losses, increasing driving range, and improving operating efficiency. However, in actual IPMSM operation, stator resistance experiences thermal drift due to temperature, quadrature and direct axis inductance exhibit nonlinear changes due to magnetic saturation, and permanent magnet flux linkages demagnetize with temperature rise. These time-varying parameters cause current distribution to deviate from the optimal trajectory, leading to problems such as torque ripple, reduced energy efficiency, and slow dynamic response.

[0004] Furthermore, IPMSMs face even more severe technical challenges in high-speed operation scenarios. As the speed increases, the back electromotive force (EMF) increases linearly with the speed. When approaching the limit of the inverter's DC bus voltage, the stator current regulation margin decreases sharply. Without an effective control mechanism, this will lead to a decrease in speed regulation capability or even loss of control. Field weakening control technology weakens the air gap magnetic field by adjusting the direct-axis demagnetizing current, thereby reducing the back EMF to overcome voltage constraints and becoming a key means to extend the motor's speed range. However, existing field weakening control strategies suffer from problems such as sensitivity to parameter fluctuations, uneven switching between different speed ranges, and insufficient torque stability in the high-speed domain, which restrict the realization of wide-speed-range high-precision control of IPMSMs. Traditional control strategies struggle to balance dynamic response, parameter robustness, and control accuracy; online parameter identification methods are either affected by noise interference or are computationally complex and cannot meet real-time control requirements; and the stability problem of field weakening control under voltage-current constraint dynamic coupling has not yet been effectively solved.

[0005] In view of the above, this application is hereby submitted. Summary of the Invention

[0006] This invention provides a field weakening control method based on adaptive parameter identification of permanent magnet synchronous motor, which can at least partially improve the above-mentioned problems.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] A field weakening control method based on adaptive parameter identification of a permanent magnet synchronous motor includes:

[0009] The acquisition motor is running in i d =0 and i d The actual voltage, current, and rotational speed under the two states <0, where i d The current is a direct axis current, and a preset parameter search range is obtained. Based on the parameter search range, a first-generation particle parameter set is randomly generated.

[0010] The first-generation particle parameter set is substituted into the identification model, and the optimal parameter set is obtained by combining the collected real voltage, current and rotation speed with CIPSO iterative processing.

[0011] The voltage vector amplitude is calculated in real time, and the voltage vector amplitude is judged to generate a judgment result. When the judgment result is low speed zone mode, the MTPA control mode is run. The optimal current command corresponding to the low speed zone is calculated based on the electromagnetic torque equation and Lagrange function, and the electrode is controlled according to the optimal current command.

[0012] When the judgment result is high-speed zone mode, the current vector angle is compensated based on the optimal parameter set. Based on the compensated current vector angle, the optimal current command corresponding to the high-speed zone is calculated, and the electrode is controlled according to the optimal current command.

[0013] In summary, this invention addresses the challenges of low-speed efficiency decline and high-speed voltage saturation instability in built-in permanent magnet synchronous motors caused by time-varying parameters. It proposes an integrated online identification-optimal current-field weakening speed-enhancing solution: A chaotic-Gaussian improved particle swarm optimization algorithm is used to capture accurate values ​​of resistance, inductance, and flux linkage in real time, providing a reliable model for full-speed domain control. Below the base speed, the analytical trajectory of the MTPA is solved based on this model, minimizing the stator current under the same torque, significantly improving system efficiency and dynamic response. When the voltage vector reaches the upper limit of the bus, a lead angle field weakening is instantaneously triggered via the "voltage saturation criterion" without the need for additional sensors, utilizing the dynamic field weakening generated by integration. The angle pushes the current vector toward the voltage limit ellipse boundary, continuously weakening the magnetic field and reducing the back electromotive force. Simultaneously, feedforward decoupling cancels coupling disturbances, achieving seamless integration between MTPA and field weakening modes. Under the same DC bus and inverter capacity constraints, this method increases the motor's maximum stable speed, maintaining low ripple and no step loss under high-speed, high-load conditions. All calculations can be completed within millisecond cycles of a conventional MCU, requiring no hardware modifications. This provides a simple, robust, and low-cost engineering approach for wide-speed-range, high-efficiency drive applications such as new energy vehicles and high-speed spindles. Attached Figure Description

[0014] Figure 1 This is a schematic flowchart of the field weakening control method based on adaptive parameter identification of permanent magnet synchronous motor provided in an embodiment of the present invention.

[0015] Figure 2 This is a schematic diagram of the framework of the field weakening control method based on adaptive parameter identification of permanent magnet synchronous motor provided in the embodiments of the present invention.

[0016] Figure 3 This is a block diagram of MTPA control based on the CIPSO parameter identification algorithm provided in an embodiment of the present invention.

[0017] Figure 4 This is a block diagram of the lead angle magnetic weakening system based on MTPA provided in an embodiment of the present invention.

[0018] Figure 5 This is a stator resistance identification curve provided in an embodiment of the present invention.

[0019] Figure 6 This is a permanent magnet flux linkage identification curve provided in an embodiment of the present invention.

[0020] Figure 7 This is a speed curve diagram under the MTPA control strategy provided in the embodiments of the present invention.

[0021] Figure 8 This is a speed curve diagram under the lead angle field weakening control strategy provided in the embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0023] refer to Figure 1 , Figure 2As shown, the first embodiment of the present invention discloses a field weakening control method based on adaptive parameter identification of a permanent magnet synchronous motor, which can be executed by a field weakening control device based on adaptive parameter identification of a permanent magnet synchronous motor (hereinafter referred to as the control device), specifically, by one or more processors within the control device, to implement the following method:

[0024] S1, collecting data on motor operation at i d =0 and i d The actual voltage, current, and rotational speed under the two states <0, where i d The current is a direct axis current, and a preset parameter search range is obtained. Based on the parameter search range, a first-generation particle parameter set is randomly generated.

[0025] The first generation of particle parameter set includes stator resistance R s Direct-axis inductor L d quadrature axis inductance L q Permanent magnet magnetic flux .

[0026] Specifically, in this embodiment, data acquisition causes the motor to... d =0 and i d The system operates in two states: <0 and <2, collecting real voltage, current, and rotational speed. Then, the algorithm is initialized, the search range of parameters is set, and the first generation of particles is randomly generated.

[0027] S2, Substitute the first-generation particle parameter set into the identification model, and combine it with the collected real voltage, current and rotation speed to perform CIPSO iterative processing to obtain the optimal parameter set;

[0028] Specifically, step S2 further includes: based on the collected real voltage, current, and speed, the discrete voltage equation is used as the identification model. Since the number of model equations is less than the number of parameters to be identified, the rank of the equation set is insufficient, making it impossible to obtain a unique solution. Therefore, a negative-sequence current is injected into the d-axis to obtain the full-rank discrete equations for the motor's dq-axis, the formula of which is: , , , Where k represents the k-th sample, For I d The actual measured value of the direct-axis voltage obtained during the k-th sampling in the control mode with =0. Let be the electric angular velocity at the k-th sampling. For in I d The quadrature-axis current sampled for the kth time in the =0 mode. For in I d The measured quadrature-axis voltage value obtained during the k-th sampling in the control mode with =0. For in I d Permanent magnet flux linkage sampled in the =0 mode for the kth time For in I d The direct-axis voltage sampled for the kth time in <0 mode. For in I d The direct-axis current sampled in the <0 mode, For in I d The quadrature-axis current sampled for the kth time in <0 mode, For in I d The quadrature-axis voltage sampled for the kth time in <0 mode;

[0029] Substituting the first-generation particle parameter set into the identification model yields the corresponding model-estimated voltage. The sum of squared errors between the model-estimated voltage and the actual voltage is then calculated to obtain the corresponding fitness value. It is then determined whether the fitness value is within a preset error range or whether the required number of iterations has been reached. The formula is as follows: n represents the number of samples, and k represents the k-th sample. , , , All are weighting coefficients. , For in I d In the control mode with =0, the actual sampled direct-axis voltage measurement value and quadrature-axis voltage measurement value obtained during the k-th sampling are... , For in I d In the control mode with =0, during the k-th sampling, the estimated direct-axis voltage and quadrature-axis voltage are calculated based on the current identification parameters. , For in I d In control mode <0, the actual sampled direct-axis voltage measurement value and quadrature-axis voltage measurement value obtained during the k-th sampling are... , For in I d In the control mode with <0, during the kth sampling, the estimated values ​​of direct-axis voltage and quadrature-axis voltage are calculated based on the current identification parameters;

[0030] If so, the set of particle parameters corresponding to the fitness value is taken as the optimal set of parameters. ;

[0031] If not, update the particle velocity and position using inertial weights, Gaussian perturbations, and asynchronous learning factors, and substitute the updated particle parameter set into the identification model to calculate the corresponding fitness value, perform error judgment, and repeat the above steps until the optimal parameter set is obtained. .

[0032] The velocity update incorporates a Gaussian perturbation, specifically a Gaussian perturbation term added to the individual cognitive component of the velocity update equation, to enhance the particle's ability to escape local optima; the particle velocity update formula is as follows: , The formula for updating the particle position is: , Let be the velocity of particle i in the g-th iteration. Let be the position of particle i in the g-th iteration. Let be the velocity of particle i in the (g+1)th iteration. Let be the position of particle i in the (g+1)th iteration. For inertial weights, , All are learning factors. , , , All numbers are random numbers within the interval [0,1]. This represents the optimal position in the history of an individual particle. The optimal position for the global population. Let be the Gaussian perturbation generated by particle i in the g-th iteration. The mean, Let Variance be the variance.

[0033] Inertia weight in the g-th iteration Dynamically adjusted by the Sine chaotic mapping, its formula is: , , , Learning factors , The learning factor is updated nonlinearly with each iteration, i.e., a nonlinear asynchronous learning factor is introduced. An exponential function is used to dynamically adjust the learning factor, and its formula is as follows: , In the early stages of iteration, a larger setting was made. and smaller To encourage particles to explore themselves, the size of the particle size distribution will be reduced later. Increase To promote information sharing and convergence within the group. Where g is the current iteration number. The maximum number of iterations. For the chaotic Sine map of the g-th iteration, For the chaotic Sine map of the (g-1)th iteration, This is the upper limit of the inertia weight. This is the lower bound of the inertia weight. This represents the upper limit of an individual's cognitive learning factor. This represents the lower limit of the individual's cognitive learning factor. This represents the lower limit of the social cognitive learning factor. denoted as the upper limit of the social cognitive learning factor, and e as the base of the natural logarithm.

[0034] In this embodiment, the CIPSO loop iterates, substituting the current particle into the discrete voltage equation (identification model), calculating the deviation (fitness), and subtracting the collected actual voltage from the theoretical voltage calculated by the model. This iteration continues until the stopping condition (minimum error) is met, outputting the final set of parameters. This approach aims to address the problems of traditional algorithms easily getting trapped in local optima and having slow convergence speed. The improved particle swarm optimization algorithm (CIPSO) significantly enhances the global optimization capability and convergence accuracy of the standard particle swarm optimization algorithm through three core improvement strategies.

[0035] Furthermore, this method evaluates the output difference between the actual model and the ideal model under the same input conditions by constructing a fitness function, and corrects the parameters to be identified by the CIPSO algorithm. When the particle fitness evaluation index is better than any or both optimal solutions, the substitution mechanism is triggered to update the optimal solution set. When the preset termination conditions (such as the iteration number threshold or convergence accuracy index) are met, the global optimal solution vector is output.

[0036] S3 calculates the voltage vector amplitude in real time, judges the voltage vector amplitude, and generates a judgment result. When the judgment result is low speed zone mode, the MTPA control mode is run. The optimal current command corresponding to the low speed zone is calculated based on the electromagnetic torque equation and Lagrange function, and the electrode is controlled according to the optimal current command.

[0037] Specifically, step S3 further includes: calculating the voltage vector magnitude. When it is determined When entering the low-speed mode, the electromagnetic torque equation is calculated based on the optimal parameter set. and Lagrange function Where p is the number of pole pairs of the motor. For ideal torque, For Lagrange operators, This is the maximum output voltage of the inverter;

[0038] Based on the electromagnetic torque equation and the Lagrange function, the current trajectory equation for the MTPA is obtained: By simultaneously solving the MTPA current trajectory equation and the electromagnetic torque equation, the optimal current command corresponding to the low-speed region is obtained. , , To identify the functional relationship between the parameters and the direct-axis current command, To identify the functional relationship between the parameters and the quadrature-axis current command, For direct-axis current, It is the quadrature-axis current.

[0039] Preferably, it further includes: during motor operation, it is subject to voltage-limiting elliptical constraints and current-limiting circular constraints; wherein, under the simplified condition of neglecting stator resistance voltage drop, the voltage constraint in the dq coordinate system satisfies the inequality... This constraint is characterized in the current plane as a center located at The ellipse is formed, and the ellipse contracts inward as the rotational speed increases; the current limiting circle constraint is that the stator current amplitude is limited by the inverter current capacity and must meet the following requirements. , This is represented in the current plane as a corresponding radius centered at the origin. The actual operating point of the motor must simultaneously lie within the intersection region enclosed by the aforementioned voltage-limiting ellipse and current-limiting circle.

[0040] Please see Figure 3 In this embodiment, the low-speed region lead angle field weakening module is not operational. The system operates entirely in MTPA control mode, optimizing the d-axis and q-axis current distribution. Specifically, during the low-speed region control phase, based on the identified parameters of MTPA, the voltage utilization rate is monitored; when it is less than U... max The system identifies the region as being in a low-speed zone. It then calculates the optimal current, inputs precise identification parameters and the user-defined target torque, substitutes these parameters into the MTPA current trajectory equation to calculate the minimum current combination required for the same torque, and finally inputs this value to the current loop drive motor. This method, by simultaneously solving the electromagnetic torque equation and the current trajectory equation, calculates the optimal current command corresponding to the low-speed zone based on the given torque command and sends it to the current loop PI controller to control the motor.

[0041] S4. When the judgment result is high-speed zone mode, the current vector angle is compensated based on the optimal parameter group. The optimal current command corresponding to the high-speed zone is calculated based on the compensated current vector angle, and the electrode is controlled according to the optimal current command.

[0042] Specifically, step S4 further includes: when it is determined that When entering high-speed zone mode, obtain the lead angle dynamically generated by the integrator. Among them, the leading angle The rate of change is directly proportional to the degree of voltage saturation. This is the proportionality coefficient;

[0043] Input the optimal parameter set into the MTPA calculation module to obtain the current vector angle. And based on the leading angle For current vector angle Make corrections, the correction formula is as follows ;

[0044] According to the corrected current vector angle The optimal current command corresponding to the high-speed region is calculated. , , This represents the stator current amplitude.

[0045] Preferably, it further includes: performing feedforward decoupling compensation, and the voltage formula after compensation is: , , , All of these are output voltage commands from a current PI regulator. It represents the electric angular velocity.

[0046] Please see Figure 4 In this embodiment, the voltage utilization rate is monitored, and when it is greater than or equal to U... max If the system detects a high-speed zone, it activates the lead angle weakening module. Through an integrator, it dynamically generates a lead angle based on the voltage difference. Simultaneously, it retrieves the optimal current vector angle from the MTPA module, adds the lead angle calculated by the field weakening module to this current vector angle, and uses this new angle to calculate a new dq-axis current combination. Finally, this new combination is input to the current loop drive motor. Specifically, when the system detects... When the motor is determined to be in a voltage saturation state, the advance angle field weakening module is immediately triggered. After field weakening is activated, the system automatically switches to voltage closed-loop control. By adjusting the operating point, the speed is increased by moving it along the boundary of the voltage limit ellipse. The advance angle field weakening module dynamically generates an advance angle through an integrator. This advance angle is used to correct the initial current vector angle calculated by the MTPA module, resulting in corrected d-axis and q-axis current commands. By introducing the advance angle, the current vector shifts towards the negative d-axis direction, thereby increasing the demagnetizing current component and decreasing the torque current component, reducing the back electromotive force, and causing the system operating point to move along the boundary of the voltage limit ellipse, achieving field weakening and speed increase.

[0047] Further addition of feedforward decoupling compensation can effectively offset the cross-coupling terms in the d-axis and q-axis voltage equations, improving the system stability under high-speed weak magnetic conditions.

[0048] Specifically, in this embodiment, a simulation model of parameter identification for a permanent magnet synchronous motor controlled by MTPA based on the CIPSO algorithm is built on the Matlab / Simulink platform to verify the effectiveness of the method. The target speed is set to 1500 r / min, and the motor load torque is 5 N·m. Under this condition, vector control is used, with parameters defined at I... d =0 and MTPA control, i.e., I d ≠0 (MTPA control steady state I) d In both cases (=-1.25A), the stator voltage U is collected. d U qCurrent i d i q and electric angular velocity The data serves as input information for the parameter identification module. To make the results comparable, the CIPSO algorithm is compared with the traditional particle swarm optimization (PSO) and IPSO algorithms. The simulation results are as follows: Figure 5 , Figure 6 As shown. By Figure 5 , Figure 6 It can be seen that for the stator resistance, the parameter identification results of the CIPSO algorithm, IPSO algorithm, and PSO algorithm are 0.97974Ω, 0.99588Ω, and 0.87194Ω, respectively. For the permanent magnet flux linkage, the parameter identification results of the CIPSO algorithm, IPSO algorithm, and PSO algorithm are 2.270Wb, 3.954Wb, and 8.983Wb, respectively. The actual stator resistance is 0.958Ω, and the actual permanent magnet flux linkage is 0.1827Wb. Therefore, the CIPSO algorithm has the best identification effect and the fastest convergence speed compared to the other two algorithms.

[0049] Furthermore, a simulation model of the lead angle field weakening control strategy based on MTPA was built on the Matlab / Simulink platform. The motor parameters were obtained using the results identified by an improved particle swarm optimization algorithm. The target speed was given as 2000 rpm, the simulation time was 1.5 s, the initial load torque of the motor was 5 N·m, the speed was increased to 6000 rpm after 0.4 s, and the load torque changed to 15 N·m at 0.8 s. To make the results comparable, the lead angle field weakening control strategy based on MTPA was compared with a control strategy using only the MTPA algorithm. The simulation results are as follows: Figure 7 , Figure 8 As shown. By Figure 7 , Figure 8 It can be seen that at low speeds, both are controlled by the MTPA algorithm, with an acceleration phase time of 0.068s and no steady-state error or overshoot. When the given speed is 6000rpm, adding field weakening control can increase the speed to 5685rpm, while using only the MTPA algorithm can only reach 4846rpm. Therefore, under the same bus voltage and current constraints, after using the field weakening control proposed in this invention, the maximum stable operating speed of the system is increased from 4846rpm to 5685rpm, extending the high-speed operating range by approximately 17.3%. This demonstrates the effectiveness of the lead angle field weakening control. When the load torque increases, the speed maintained by adding field weakening control is still 5386rpm, while the speed dropped to 4527rpm using only the MTPA algorithm. That is, the speed of the lead angle field weakening control is higher than that of the MTPA algorithm alone under different load torques. This verifies that the lead angle field weakening control proposed in this invention can effectively extend the high-speed operating range of the motor.

[0050] In summary, this invention aims to overcome the shortcomings of existing IPMSM control, such as poor time-varying adaptability of parameters, narrow field weakening speed range, and slow dynamic response. It provides a field weakening control method based on adaptive parameter identification, achieving high-precision real-time parameter identification through an improved particle swarm optimization algorithm. A mode switching strategy is combined to optimize the performance of low-speed MTPA control and high-speed field weakening control, ultimately achieving efficient and stable control over a wide speed range. First, online identification of motor parameters is performed using an improved particle swarm optimization algorithm (CIPSO) that integrates chaotic inertial weights and Gaussian perturbations. An identification model is established based on the discrete voltage equations of the permanent magnet synchronous motor. The fitness function values ​​of the actual output and the model output are calculated to dynamically update the individual optimal solution and the global optimal solution. After multiple rounds of iterative optimization, high-precision identification results for stator resistance, direct-axis and quadrature-axis inductance, and permanent magnet flux linkage are finally output.

[0051] Secondly, for operating conditions below rated speed, a maximum torque-to-current ratio (MTPA) control method based on parameter identification is adopted. This method uses the motor parameters obtained through online identification as a foundation, and solves the optimal current trajectory equation derived from the electromagnetic torque equation and Lagrange function. It then calculates and tracks in real time the combination of direct-axis and quadrature-axis currents that achieves the maximum torque-to-current ratio. Its core lies in the real-time calculation of the optimal operating point based on accurate motor parameters, utilizing the salient-pole characteristics of the integrated permanent magnet synchronous motor (IPMSM). This approach minimizes stator current amplitude while ensuring output torque, significantly improving efficiency and dynamic response performance in the low-to-medium speed operating range. Finally, above rated speed, a field weakening control strategy based on voltage saturation criteria is employed. This strategy determines the field weakening entry condition by real-time monitoring of voltage utilization, dynamically introducing a lead angle to adjust the phase angle of the dq-axis current command. This redistributes the current vector along the voltage limit ellipse boundary, weakening the air gap magnetic field by enhancing the direct-axis demagnetizing current component, thereby overcoming the inverter voltage limit. The control system integrates a current feedforward decoupling compensation stage and optimizes decoupling accuracy based on parameter identification results, achieving a smooth switch from maximum torque-to-current ratio control to field weakening control mode. This effectively extends the motor's high-speed operating range and maintains dynamic stability during sudden load changes.

[0052] In summary, this invention addresses the control misalignment issues of built-in permanent magnet synchronous motors (PMSMs) during full-speed operation caused by stator resistance thermal drift, quadrature-direct axis inductance nonlinear saturation, and permanent magnet demagnetization. It proposes an adaptive parameter identification-maximum torque-current ratio-lead angle field weakening continuum control approach: First, an improved particle swarm optimization algorithm incorporating chaotic inertial weights and Gaussian perturbations is used. The discrete voltage equation of the motor is used as the model, and the error between the measured voltage and the model output is used as the fitness factor. Real-time resistance, inductance, and flux linkage are iteratively output, ensuring the controller is always updated synchronously with temperature and magnetic state. Then, below the rated speed, the identification results are injected into the MTPA analytical expression to solve for the required parameters in real time. The minimum stator current combination required for a given torque significantly reduces copper losses, improves driving range, and accelerates dynamic response thanks to the IPMSM salient pole effect. When the voltage vector amplitude reaches the inverter limit, the system immediately activates the lead angle field weakening module. Through an integrator, a lead angle β proportional to the saturation level is dynamically generated, pushing the current vector along the voltage limit ellipse boundary towards the negative d-axis. This continuously weakens the air gap magnetic field without increasing the bus voltage, breaking through the back EMF blockade and achieving smooth acceleration. Simultaneously, feedforward compensation of the cross-coupling voltage is performed using parameter identification results, eliminating dynamic interference on the dq axis in the high-speed region, ensuring shock-free mode switching and no oscillation during load changes. The entire process requires no additional sensors or hardware modifications; all calculations are completed within millisecond cycles of a conventional control chip. Under the same DC bus and current constraints, the maximum stable motor speed is increased, maintaining low pulsation and no step loss even under high-speed, high-load conditions. This provides a simple, robust, and low-cost engineering technology path for wide-speed-range, high-efficiency drive applications such as electric vehicles and high-speed spindles.

[0053] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A field weakening control method based on adaptive parameter identification of a permanent magnet synchronous motor, characterized in that, include: The acquisition motor is running in i d =0 and i d The actual voltage, current, and rotational speed under the two states <0, where i d The current is a direct-axis current, and a preset parameter search range is obtained. Based on the parameter search range, a first-generation particle parameter set is randomly generated. The first-generation particle parameter set includes the stator resistance R. s Direct-axis inductor L d quadrature axis inductance L q Permanent magnet magnetic flux ; Substituting the first-generation particle parameter set into the identification model, and combining it with the collected real voltage, current, and rotational speed, CIPSO iterative processing is performed to obtain the optimal parameter set, specifically: Based on the collected real voltage, current, and rotational speed, the discrete voltage equation is used as the identification model, and its formula is as follows: , , , Where k represents the k-th sample, For I d The actual measured value of the direct-axis voltage obtained during the k-th sampling in the control mode with =0. Let be the electric angular velocity at the k-th sampling. For in I d The quadrature-axis current sampled for the kth time in the =0 mode. For in I d The measured quadrature-axis voltage value obtained during the k-th sampling in the control mode with =0. For in I d Permanent magnet flux linkage sampled in the =0 mode for the kth time For in I d The direct-axis voltage sampled for the kth time in <0 mode, For in I d The direct-axis current sampled in the <0 mode, For in I d The quadrature-axis current sampled for the kth time in <0 mode, For in I d The quadrature-axis voltage sampled for the kth time in <0 mode; Substituting the first-generation particle parameter set into the identification model yields the corresponding model-estimated voltage. The sum of squared errors between the model-estimated voltage and the actual voltage is then calculated to obtain the corresponding fitness value. It is then determined whether the fitness value is within a preset error range or whether the required number of iterations has been reached. The formula is as follows: n represents the number of samples, and k represents the k-th sample. , , , All are weighting coefficients. , For in I d In the control mode with =0, during the k-th sampling, the estimated direct-axis voltage and quadrature-axis voltage are calculated based on the current identification parameters. , For in I d In control mode <0, the actual sampled direct-axis voltage measurement value and quadrature-axis voltage measurement value obtained during the k-th sampling are... , For in I d In the control mode with <0, during the kth sampling, the estimated values ​​of direct-axis voltage and quadrature-axis voltage are calculated based on the current identification parameters; If so, the set of particle parameters corresponding to the fitness value is taken as the optimal set of parameters. ; If not, update the particle velocity and position using inertial weights, Gaussian perturbations, and asynchronous learning factors, and substitute the updated particle parameter set into the identification model to calculate the corresponding fitness value, perform error judgment, and repeat the above steps until the optimal parameter set is obtained. ; The voltage vector amplitude is calculated in real time, and the voltage vector amplitude is judged to generate a judgment result. When the judgment result is low speed zone mode, the MTPA control mode is run. The optimal current command corresponding to the low speed zone is calculated based on the electromagnetic torque equation and Lagrange function, and the motor is controlled according to the optimal current command. When the judgment result is high-speed zone mode, the current vector angle is compensated based on the optimal parameter set. Based on the compensated current vector angle, the optimal current command corresponding to the high-speed zone is calculated, and the motor is controlled according to the optimal current command.

2. The field weakening control method based on adaptive parameter identification of a permanent magnet synchronous motor according to claim 1, characterized in that, The formula for updating particle velocity is: , The formula for updating the particle position is: , Let be the velocity of particle i in the g-th iteration. Let i be the position of particle i in the g-th iteration. Let be the velocity of particle i in the (g+1)th iteration. Let be the position of particle i in the (g+1)th iteration. For inertial weights, , All are learning factors. , , , All numbers are random numbers within the interval [0,1]. This represents the optimal position in the history of an individual particle. The optimal position for the entire population. Let be the Gaussian perturbation generated by particle i in the g-th iteration. The mean, Let Variance be the variance.

3. The field weakening control method based on adaptive parameter identification of a permanent magnet synchronous motor according to claim 2, characterized in that, Inertia weight in the g-th iteration Dynamically adjusted by the Sine chaotic mapping, its formula is: , , , ; Learning factor , It updates nonlinearly with the number of iterations, and its formula is: , Where g is the current iteration number, The maximum number of iterations. For the chaotic Sine map of the g-th iteration, For the chaotic Sine map of the (g-1)th iteration, This is the upper limit of the inertia weight. This is the lower bound of the inertia weight. This represents the upper limit of an individual's cognitive learning factor. This represents the lower limit of the individual's cognitive learning factor. This represents the lower limit of the social cognitive learning factor. denoted as the upper limit of the social cognitive learning factor, and e as the base of the natural logarithm.

4. The field weakening control method based on adaptive parameter identification of a permanent magnet synchronous motor according to claim 3, characterized in that, The voltage vector magnitude is calculated in real time, and a judgment result is generated. When the judgment result indicates a low-speed zone mode, the MTPA control mode is run. Based on the electromagnetic torque equation and Lagrange function, the optimal current command corresponding to the low-speed zone is calculated, specifically: Calculate voltage vector magnitude When it is determined When entering the low-speed mode, the electromagnetic torque equation is calculated based on the optimal parameter set. and Lagrange function Where p is the number of pole pairs of the motor. For ideal torque, For Lagrange operators, This is the maximum output voltage of the inverter; Based on the electromagnetic torque equation and the Lagrange function, the current trajectory equation for the MTPA is obtained: By simultaneously solving the MTPA current trajectory equation and the electromagnetic torque equation, the optimal current command corresponding to the low-speed region is obtained. , , To identify the functional relationship between the parameters and the direct-axis current command, To identify the functional relationship between the parameters and the quadrature-axis current command, For direct-axis current, It is the quadrature-axis current.

5. The field weakening control method based on adaptive parameter identification of a permanent magnet synchronous motor according to claim 4, characterized in that, When the judgment result is high-speed mode, current vector angle compensation is performed based on the optimal parameter set. Based on the compensated current vector angle, the optimal current command corresponding to the high-speed mode is calculated, specifically: When judged When entering high-speed zone mode, obtain the lead angle dynamically generated by the integrator. Among them, the leading angle The rate of change is directly proportional to the degree of voltage saturation. This is the proportionality coefficient; Input the optimal parameter set into the MTPA calculation module to obtain the current vector angle. And based on the leading angle For current vector angle Make corrections, the correction formula is as follows ; According to the corrected current vector angle The optimal current command corresponding to the high-speed region is calculated. , , This represents the stator current amplitude.

6. The field weakening control method based on adaptive parameter identification of a permanent magnet synchronous motor according to claim 5, characterized in that, Also includes: After performing feedforward decoupling compensation, the voltage formula after compensation is: , , , All of these are output voltage commands from a current PI regulator. ω is the electric angular velocity.

7. The field weakening control method based on adaptive parameter identification of a permanent magnet synchronous motor according to claim 6, characterized in that, Also includes: During motor operation, it is subject to voltage-limiting elliptical constraints and current-limiting circular constraints. Specifically, under the simplified condition of neglecting stator resistance voltage drop, the voltage constraint in the dq coordinate system satisfies the inequality... This constraint is characterized in the current plane as a center located at The ellipse is formed, and the ellipse contracts inward as the rotational speed increases; the current limiting circle constraint is that the stator current amplitude is limited by the inverter current capacity and must meet the following requirements. , In the current plane, this is represented by a corresponding radius centered at the origin.

Images