High-precision torque closed-loop energy-saving control method and system of permanent magnet synchronous motor

Abstract

The invention relates to the technical field of permanent magnet synchronous motors, in particular to a high-precision torque closed-loop energy-saving control method and system for a permanent magnet synchronous motor, and the method comprises the steps: calculating and generating a disturbance feature vector at a sensing side, and describing the changes of flux linkage, current, speed and temperature rise in a differential form; on a target reconstruction side, generating a constrained torque reference according to the disturbance intensity and the non-balance thereof, and describing the current choice of energy saving and precision through an adjustable weight; on the modeling side, disturbance-guided modeling path selection and temperature rise memory response are introduced, condition updating is carried out on two-axis equivalent inductance and stator resistance, and mismatching of a fixed strategy under strong disturbance is avoided; and on the track side, an energy consumption model containing a copper loss item and a magnetization regular item is constructed, inter-axle coupling compensation is added in a torque error item, periodic-level tradeoff between energy consumption and precision is achieved through the energy-saving weight, and high-precision torque closed-loop energy-saving control adaptive to the complex working condition is achieved.

Description

Technical Field

[0001] This invention relates to the technical field of permanent magnet synchronous motors, and more particularly to a high-precision torque closed-loop energy control method and system for permanent magnet synchronous motors. Background Technology

[0002] Permanent magnet synchronous motors have become the core actuators of electric drive systems due to their high power density, high efficiency, and fast dynamic response. In engineering, vector control is commonly used, which achieves independent adjustment of torque and flux linkage by adjusting the current of the two axes separately in a rotating coordinate system.

[0003] However, the operating conditions in mass production environments and complex working conditions are far beyond what ideal models can fully cover. Fluctuations in the power grid, sudden changes in mechanical load, temperature rise caused by changes in heat transfer conditions, and parameter discrepancies and time-varying characteristics due to material magnetic properties and processing tolerances can all cause stator resistance, two-axis equivalent inductance, and equivalent flux linkage to deviate from their nominal values. This disrupts the decoupling relationship based on fixed parameters and manifests as inconsistencies between current command and electromagnetic response in typical operating conditions such as weak field, high speed, and low speed with high torque, directly inducing torque errors, oscillations, and increased energy consumption. To suppress these errors, conventional approaches rely on current loops with fixed parameters and empirically tuned weak field or lookup table paths. These methods work in the statically stable range, but when the perturbation structure changes, the original path no longer represents optimal energy efficiency and may even amplify current peaks and copper losses. Some studies have attempted to improve performance through stronger prediction or adaptive algorithms, but under the constraints of embedded computing power and sampling timing, the structural complexity of the prediction model, the fixed objective function, and the neglect of nonlinearity and coupling effects make it difficult to stably implement in the field. Engineering practice shows that the crux of the problem is not simply inaccurate parameters, but rather that the disturbances are multi-source and their structure changes constantly. The control objective needs to be conditionally adjusted under different disturbance intensities and forms. The model parameters need to be estimated and updated using a disturbance-guided strategy. Trajectory solving requires joint optimization of energy saving and accuracy under the current state with adjustable weights.

[0004] Without this integrated mechanism that connects disturbance quantization to target reconstruction and state modeling to trajectory generation, the torque closed loop will inevitably experience response lag and energy efficiency degradation under complex operating conditions, limiting the performance boundaries of permanent magnet synchronous motors in high-dynamic scenarios such as vehicles, servos, and electric spindles. Summary of the Invention

[0005] This invention establishes a closed-loop link from perception to execution around high-precision and energy-saving coordinated control driven by disturbance. The core is to quantify the current operating condition with the disturbance feature vector, transform it into a torque reference and energy-saving weight with engineering constraints, and use it to guide parameter modeling and trajectory solving to solve the problems existing in the prior art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] On one hand, a high-precision torque closed-loop energy control method for a permanent magnet synchronous motor is provided, including the following steps:

[0008] By acquiring the key state signals of the permanent magnet synchronous motor in real time, calculating the change in key state signals between the previous acquisition cycle and the current acquisition cycle, a disturbance vector is constructed to describe the disturbance characteristics.

[0009] Based on the multidimensional disturbance information in the disturbance vector, a disturbance intensity function is constructed to represent the disturbance intensity of the current permanent magnet synchronous motor. During periods with high disturbance intensity, the target torque is actively reduced so that the controller prioritizes stability. When the disturbance is small, the output expected by the user is kept unchanged. An energy-saving tolerance factor is calculated based on the disturbance intensity function to adjust the weight balance between the loss term and the torque error term in trajectory optimization.

[0010] The disturbance vector, target torque, energy-saving tolerance factor, and voltage, current, and temperature data of the permanent magnet synchronous motor collected in the current cycle are input into the historical temperature rise weighted memory model, and the structural parameter vector of the permanent magnet synchronous motor in the current state is output, including the estimated value of the d-axis inductance, the estimated value of the q-axis inductance, and the estimated value of the stator resistance in the current cycle.

[0011] By constructing an optimization objective function, the optimal current command pair is calculated in the dual-axis current space and used to send the controller to execute PWM modulation. This ensures torque control accuracy while dynamically finding the optimal energy consumption path.

[0012] Preferably, the key status signals include the current, voltage, angular displacement, speed, and temperature of the permanent magnet synchronous motor; the disturbance vector includes the flux linkage change amplitude, the difference in average current between two consecutive acquisition cycles, the speed difference between the current acquisition cycle and the previous acquisition cycle, and the temperature change between the current acquisition cycle and the previous acquisition cycle.

[0013] More preferably, the flux linkage variation amplitude is obtained by performing Clarke-Park transform on the voltage and current signals to obtain the voltage and current on the q-axis, and then using the voltage equation to estimate the flux linkage to obtain the flux linkage of the current period, which is obtained by subtracting the flux linkage of the previous acquisition period; wherein the flux linkage integral adopts the discrete integral form within a sliding window.

[0014] Preferably, a non-equilibrium regularization term of the disturbance structure is introduced into the disturbance intensity function to reflect the possible estimated risk of conventional disturbance superposition.

[0015] Preferably, the energy-saving tolerance factor maintains a low energy consumption weight when the disturbance intensity is small, and gradually increases the energy-saving tolerance when the disturbance intensity increases, so as to release more current buffer space.

[0016] Preferably, the energy-saving tolerance factor is also used to reduce energy consumption in periodic systems with small disturbances, while relaxing energy-saving constraints and releasing adjustment space when disturbances intensify.

[0017] Preferably, the historical temperature rise weighted memory model includes a perturbation-driven modeling confidence coefficient, which is used to adjust the response speed and confidence strength of parameter updates, thereby enhancing the confidence response sensitivity when the perturbation structure is unbalanced.

[0018] Preferably, the optimization objective function includes a loss control term and a torque accuracy term. The loss control term is used to suppress the d-axis energy consumption expansion trend under high-frequency weak magnetic conditions. The torque accuracy term is used to compensate for the torque deviation of the q-axis current under disturbance conditions.

[0019] Preferably, the optimal current command pair is obtained based on the objective function and used by the controller to generate a space vector PWM signal, which is then directly applied to the inverter to achieve control voltage distribution.

[0020] In another aspect of the present invention, a high-precision torque closed-loop energy control system for a permanent magnet synchronous motor is provided, comprising: sequentially connected:

[0021] The data acquisition and processing module is used to acquire key state signals of the permanent magnet synchronous motor in real time, calculate the change in key state signals between the previous acquisition cycle and the current acquisition cycle, and construct a disturbance vector to describe the disturbance characteristics.

[0022] The torque target adjustment module is used to construct a disturbance intensity function based on the multi-dimensional disturbance information in the disturbance vector to represent the disturbance intensity of the current permanent magnet synchronous motor. During periods with high disturbance intensity, the target torque is actively reduced so that the controller prioritizes stability. When the disturbance is small, the output expected by the user is kept unchanged. The energy-saving tolerance factor is calculated based on the disturbance intensity function to adjust the weight balance between the loss term and the torque error term in trajectory optimization.

[0023] The structural parameter vector prediction module is used to input the disturbance vector, target torque, energy-saving tolerance factor and voltage, current and temperature data of the permanent magnet synchronous motor collected in the current cycle into the historical temperature rise weighted memory model, and output the structural parameter vector of the permanent magnet synchronous motor in the current state, including the estimated value of the d-axis inductance in the current cycle, the estimated value of the q-axis inductance in the current cycle and the estimated value of the stator resistance in the current cycle.

[0024] The current command optimization module is used to calculate the optimal current command pair in the dual-axis current space by constructing an optimization objective function. This is then used to send the controller to execute PWM modulation, dynamically finding the optimal energy consumption path while ensuring torque control accuracy.

[0025] Compared with the prior art, the beneficial effects of the present invention are:

[0026] This invention establishes a closed-loop control system from sensing to execution, focusing on high-precision and energy-saving coordinated control driven by disturbances. The core of this system lies in quantifying the current operating condition using disturbance feature vectors, transforming them into torque references and energy-saving weights with engineering constraints, and using these to guide parameter modeling and trajectory solving. On the sensing side, disturbance feature vectors are calculated and generated within a cycle based on voltage, current, position, speed, and temperature, characterizing the changes in flux linkage, current, speed, and temperature rise in a differential form. On the target reconstruction side, a constrained torque reference is generated based on the disturbance intensity and its non-equilibrium, and adjustable weights describe the current trade-off between energy saving and accuracy. On the modeling side, disturbance-guided modeling path selection and temperature rise memory response are introduced to update the equivalent inductance and stator resistance of the two axes, avoiding mismatch of fixed strategies under strong disturbances. On the trajectory side, an energy consumption model including copper loss and magnetization regularization terms is constructed, and inter-axis coupling compensation is added to the torque error term. Through the aforementioned energy-saving weights, a cycle-level trade-off between energy consumption and accuracy is achieved, directly obtaining the two-axis current commands to drive the actuators. The aforementioned mechanism is characterized by the sequential coupling of disturbance quantization, target reconstruction, state modeling and trajectory optimization. It can operate within the embedded cycle and computing power boundary, enabling the torque closed loop to maintain stable accuracy under strong disturbances and time-varying parameters. At the same time, it actively seeks lower energy consumption current paths in the stable range, thereby achieving high-precision torque closed loop energy control adapted to complex working conditions. Attached Figure Description

[0027] Figure 1 This is a flowchart of a high-precision torque closed-loop energy control method for a permanent magnet synchronous motor according to a specific embodiment of the present invention;

[0028] Figure 2 This is a block diagram of a high-precision torque closed-loop energy control system for a permanent magnet synchronous motor, as described in a specific embodiment of the present invention. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] Please refer to Figure 1As shown, this application proposes a high-precision torque closed-loop energy control method for permanent magnet synchronous motors, comprising the following steps:

[0031] Step 1: By acquiring key state signals of the permanent magnet synchronous motor in real time, calculate the change in key state signals between the previous acquisition cycle and the current acquisition cycle, and construct a disturbance vector to describe the disturbance characteristics; specifically including:

[0032] This step aims to establish a disturbance identification method based on real-time operating data to capture various disturbances encountered by permanent magnet synchronous motors during operation. These disturbances may originate from load fluctuations, grid voltage changes, motor parameter drift caused by temperature rise, or flux disturbances. Traditional PI controllers typically can only correct errors after the fact and are unable to respond to these disturbances in advance. Therefore, this step calculates and constructs a disturbance vector to describe the characteristics of disturbances by real-time acquisition of key state signals such as current, voltage, angular displacement, speed, and temperature. This vector will subsequently serve as the input basis for dynamic control objectives and energy consumption regulation mechanisms.

[0033] The controller operates in each cycle (cycle length is...). (e.g., 50μs) will collect the following raw signals: three-phase stator current The three-phase stator voltage is acquired by a Hall sensor at a frequency of 20kHz and converted by an ADC. Rotor angle is acquired via isolation circuit. Sourced from a 12-bit incremental encoder; rotational speed Depend on Calculated; temperature The data is generated from a thermistor installed near the stator winding and is updated every 10ms.

[0034] Perform Clarke-Park transform on the voltage and current signals to obtain the voltage on the q-axis. With current The flux linkage was then estimated using the voltage equation. The flux linkage integral was performed using a discrete integral form within a sliding window, during the period... Internal sampling The calculation method for each point is as follows:

[0035] ;

[0036] in and They represent the first time. q-axis voltage and current values ​​at the next sampling point The sampling interval time. The stator resistance is initially set by the nameplate parameters and can be periodically updated during operation based on the temperature drift curve.

[0037] After completing the flux linkage estimation, the system calculates the changes in key state signals between the previous acquisition cycle and the current acquisition cycle, which are used to construct the disturbance vector. :

[0038] ;

[0039] in , indicating the magnitude of the change in magnetic flux; It is the difference between the average currents of two consecutive cycles; It is the difference between the current rotational speed and the previous rotational speed; It is a temperature change value used to identify possible parameter drift trends.

[0040] Step 2: Based on the multi-dimensional disturbance information in the disturbance vector, a disturbance intensity function is constructed to represent the current disturbance intensity of the permanent magnet synchronous motor. During periods of high disturbance intensity, the target torque is actively reduced to prioritize stability for the controller; while during periods of low disturbance, the user-desired output is maintained unchanged. An energy-saving tolerance factor is calculated based on the disturbance intensity function to adjust the weight balance between the loss term and the torque error term in trajectory optimization. Specifically, this includes:

[0041] This step aims to utilize the perturbation feature vector constructed in the previous step. Within the current control cycle, two key control variables are dynamically generated: the target torque adjustment value. With energy-saving tolerance factor , which serve as the target boundary and energy efficiency constraint weights for the subsequent trajectory optimizer.

[0042] In practical applications, permanent magnet synchronous motors often operate under complex conditions with frequent load changes, significant power supply fluctuations, or rapid temperature changes. These disturbances directly affect the stability of torque execution and the energy efficiency of the current distribution strategy. If the controller maintains a static target torque setting under different operating conditions, the following problems can easily occur: torque overshoot and current oscillations may occur during severe disturbances, affecting stability; while energy consumption may be too high when the load is stable, failing to fully utilize energy-saving potential.

[0043] To address the aforementioned issues, this step incorporates the following two key innovative directions in the structural design: First, the disturbance vector... The multidimensional perturbation information is used to construct the perturbation intensity function. Secondly, it guides the real-time fine-tuning of the torque target; and thirdly, it constructs an energy-saving tolerance factor based on disturbance information. Furthermore, a regularization term related to the "non-equilibrium" of the perturbation component is introduced into its computational structure to express the complexity of the perturbation structure, thereby assisting in the flexible switching between energy-saving and precision strategies.

[0044] The perturbation vector, after normalization, has a uniform amplitude level and can be directly used for weighted combination.

[0045] First, to construct a quantitative expression of the current perturbation state of the system, a perturbation intensity function is introduced. Its calculation structure is as follows:

[0046] ;

[0047] in, to The perturbation channel weighting coefficients are determined by tuning experiments to determine their priority. This is the coefficient of the regularization term, used to adjust the impact of inconsistencies between disturbance components on the control objective. The first four terms constitute the total strength score of the disturbance, and the fifth term is the degree of non-equilibrium in the disturbance structure. Its innovation lies in considering the "tension" between different disturbance components, such as in... In cases of sudden changes while other variables remain stable, the superposition of conventional disturbances may predict risk, while this regularization term can significantly reflect the deviation.

[0048] Next, we construct the torque target adjustment value for the current cycle. The basic idea of ​​target adjustment is to actively reduce the target torque during periods of high disturbance intensity, prioritizing stability for the controller; while maintaining the user-desired output when the disturbance is small. This relationship is expressed by a linear suppression function as follows:

[0049] ;

[0050] in, The target torque after dynamic adjustment in the current cycle serves as the task boundary for trajectory optimization. The desired torque value given by the upper-level scheduling system for the current cycle. This is the target adjustment sensitivity coefficient, used to control the range of influence of disturbance factors on target adjustment. A recommended value is... to This structure ensures that when When the current is increased, the target command will be compressed to a safe range to avoid current surges.

[0051] Subsequently, based on the same perturbation function Constructing an energy-saving tolerance factor This is used for energy efficiency weight control in subsequent trajectory optimization. Its structure employs a perturbation-enhanced Sigmoid modulation function. Maintain a low energy consumption weight when it is small, while Gradually increase the energy-saving tolerance during the rise to free up more current buffer space:

[0052] ;

[0053] in, Energy-saving tolerance factor, used to adjust the weight balance between loss term and torque error term in trajectory optimization; This is the Sigmoid slope control factor. Here, is the reference point for disturbance, the turning point of the adjustment function, and e is a constant. This expression structure achieves an automatic balance between energy saving and stability: in a periodic system with small disturbances, it tends to reduce energy consumption; while when the disturbance intensifies, it relaxes energy-saving constraints, giving more adjustment space to the torque tracking task.

[0054] Step 3: Input the disturbance vector, target torque, energy-saving tolerance factor, and the voltage, current, and temperature data of the permanent magnet synchronous motor collected in the current cycle into the historical temperature rise weighted memory model, and output the structural parameter vector of the permanent magnet synchronous motor in the current state, including the estimated value of the d-axis inductance, the estimated value of the q-axis inductance, and the estimated value of the stator resistance in the current cycle; specifically including:

[0055] The purpose of this step is to combine the perturbation feature vectors output from the previous two steps. Target torque With energy-saving tolerance factor Based on this, the structural parameter vector of the motor in its current state is estimated using the voltage and current data collected in the current cycle. This provides accurate model support for subsequent control trajectory optimization.

[0056] In the energy-saving control process of permanent magnet synchronous motors, the d-axis / q-axis inductance , and stator resistance These are the fundamental parameters that determine current path optimization and power loss calculation. However, in complex industrial scenarios, these parameters dynamically drift with changes in temperature, current amplitude, magnetic saturation, frequency, and material nonlinearity. If the controller relies on a static model for a long time, trajectory planning will deviate from the optimal point, making it difficult to achieve the goal of balancing torque accuracy and energy efficiency.

[0057] Unlike traditional methods that simply perform differential modeling based on sampled voltage and current, this step introduces two key innovative strategies in the parameter estimation process: first, it uses the perturbation vector from the previous step. The disturbance amplitude and structure are used as dynamic modeling weight adjustment factors to guide the update frequency and confidence of different parameters; secondly, in view of the d-axis / q-axis inductance asymmetry and temperature rise effect in permanent magnet synchronous motors, parameter difference regularization terms and resistance hysteresis response kernel functions are introduced to improve the model stability and traceability under dynamic disturbance environment.

[0058] Using the voltage and current values ​​obtained from sampling The voltage and current are the moving averages within the control period after Clarke-Park transformation; the terms in the disturbance vector are normalized to ensure comparability.

[0059] To achieve dynamic adaptability in parameter estimation, we first construct the modeling confidence coefficients driven by perturbation. This coefficient is used to adjust the response speed and confidence level of parameter updates. It takes into account the total disturbance strength and component non-equilibrium.

[0060] ;

[0061] in For the disturbance intensity, This is a regularization coefficient used to enhance the confidence response sensitivity when the disturbed structure is out of balance, reflecting the controller's rapid response mechanism to "local mutations". As a floating-point number in the range [0,1], it will be used as a step size weight factor in the parameter update logic.

[0062] Consider q-axis inductance The update path uses the voltage equation back-calculation method and integrates the previous period's estimate. The update formula is as follows:

[0063] ;

[0064] in, This is the estimated q-axis inductance value for the current sampling period. This is the estimated value of the q-axis inductance from the previous sampling period; This is the signal for the current period; To prevent the inhibitory factor from being too small in the denominator, it is usually taken as... ; This is the previously estimated resistance value. The update logic is... Control the update magnitude to maintain model stability under small disturbances and respond quickly under large disturbances.

[0065] To address the asymmetry issue of d-axis / q-axis inductance in the field weakening region, this step constructs a parameter consistency regularization term. To suppress physically inexplicable biases:

[0066] ;

[0067] in, This regularization term, used to estimate the d-axis inductance for the current sampling period, will be used as a penalty in the actual trajectory optimization objective function, but it is also used in the modeling process to adjust... Update the results to match The changing trend maintains physical consistency; It is an adjustment factor that can be set empirically based on the system's error identification.

[0068] Stator resistance The update path incorporates historical cumulative temperature rise and lag effects to avoid short-term oscillations in the model caused by sudden temperature changes. The update formula is as follows:

[0069] ;

[0070] in This is the initial nameplate value. This is the temperature rise sensitivity coefficient. Here, e represents the historical temperature rise coefficient. This structure constructs a historical temperature rise weighted memory model, exponentially preserving the recent period temperature rise trend. In engineering applications, it reflects the nonlinear response of resistance to heat accumulation, where e is a constant.

[0071] All computational structures can be implemented in DSP or ARM processors with a cycle latency of less than 10μs.

[0072] The structural parameter vector of the permanent magnet motor in its current state is obtained by summarizing. This will serve as the input condition for constructing the objective function and generating constraints in the next step of trajectory optimization.

[0073] Step 4: By constructing an optimization objective function, the optimal current command pair is calculated in the dual-axis current space. This pair is then sent to the controller to execute PWM modulation, dynamically finding the optimal energy consumption path while ensuring torque control accuracy. Specifically, this includes:

[0074] This step is used to generate the final control command for the permanent magnet synchronous motor under the current disturbance state. This is combined with the dynamic torque target output in step two. With energy-saving tolerance factor and the motor modeling state vector estimated in step three. This step aims to minimize energy consumption and torque error by calculating the optimal current command pair in the biaxial current space. It is used to send the controller to execute PWM modulation, thereby dynamically optimizing the energy consumption path while ensuring torque control accuracy.

[0075] Unlike traditional trajectory generation methods based on constant field weakening strategies or MTPA paths, this step does not employ a static current lookup table structure, nor does it assume constant motor parameters. Instead, it is based on a cycle-updated... , and A dynamically adjustable optimization objective function is constructed. This objective function fully reflects the uncertainty of the control objective, the dynamic nature of the parameter states, and the linkage between energy consumption strategy and disturbance perception, forming a trajectory calculation mechanism with structural flexibility. This mechanism supports the following three important characteristics:

[0076] Strict constraints on the torque target under the current disturbance; dynamic estimation and weighted adjustment of current path energy consumption; response control strategy for d / q axis control asymmetry.

[0077] The optimization objective function constructed in this step is as follows:

[0078] ;

[0079] The first part is the loss control item. This indicates copper loss caused by current. It is an energy coupling term based on the d-axis flux linkage, and its weighting coefficients are... The penalty intensity in the high magnetic field range is controlled to suppress the d-axis energy consumption expansion trend under high-frequency weak magnetic conditions. This part is multiplied by... This reflects the current system's focus on energy conservation. The higher the value, the more the system tends to reduce energy consumption paths.

[0080] The second part is the torque accuracy term. In the traditional control model, electromagnetic torque... by As a unique variable, it is expressed as However, in practice, due to the nonlinearity of the inductance and the change in the q-axis back electromotive force, under disturbance conditions... This introduces cross-coupling terms, leading to a decrease in the "torque efficiency" of the q-axis current. Therefore, this invention introduces... The term simulates this coupled disturbance structure, forming a torque deviation compensation term, which serves as an "efficiency reduction" factor for the q-axis current during disturbance. To prevent small quantities with a denominator of zero, the entire torque term is multiplied by... This reflects that when the disturbance is small and the system state is stable, the system pays more attention to accuracy, while when the disturbance is large and the energy saving weight is high, the system automatically relaxes the strict limit on accuracy.

[0081] The objective function can be viewed as a dual objective function of "energy consumption and accuracy" with dynamic switching of perturbation weights, automatically adjusting the optimization strategy as the perturbation changes in the current cycle. In engineering implementation, this structure can be achieved through a lookup table-local optimization approach: first construct... Track library, and then based on The current cost function is calculated, and gradient descent iterations are performed in the vicinity to obtain the optimal solution. The typical solution time is within 5μs, which meets the requirements of industrial motor control cycle.

[0082] Ultimately, the optimal current command pair is obtained. This is used by the controller to generate a space vector PWM signal, which is directly applied to the inverter to deliver the control voltage. Thus, the entire control path, from disturbance sensing, target construction, state modeling to trajectory generation, forms a closed loop.

[0083] Please refer to Figure 2 As shown, in a second aspect of this application, a high-precision torque closed-loop energy control system for a permanent magnet synchronous motor is also proposed, comprising:

[0084] The data acquisition and processing module is used to acquire key state signals of the permanent magnet synchronous motor in real time, calculate the change in key state signals between the previous acquisition cycle and the current acquisition cycle, and construct a disturbance vector to describe the disturbance characteristics.

[0085] The torque target adjustment module is used to construct a disturbance intensity function based on the multi-dimensional disturbance information in the disturbance vector to represent the disturbance intensity of the current permanent magnet synchronous motor. During periods with high disturbance intensity, the target torque is actively reduced so that the controller prioritizes stability. When the disturbance is small, the output expected by the user is kept unchanged. The energy-saving tolerance factor is calculated based on the disturbance intensity function to adjust the weight balance between the loss term and the torque error term in trajectory optimization.

[0086] The structural parameter vector prediction module is used to input the disturbance vector, target torque, energy-saving tolerance factor and voltage, current and temperature data of the permanent magnet synchronous motor collected in the current cycle into the historical temperature rise weighted memory model, and output the structural parameter vector of the permanent magnet synchronous motor in the current state, including the estimated value of the d-axis inductance in the current cycle, the estimated value of the q-axis inductance in the current cycle and the estimated value of the stator resistance in the current cycle.

[0087] The current command optimization module is used to calculate the optimal current command pair in the dual-axis current space by constructing an optimization objective function. This is then used to send the controller to execute PWM modulation, dynamically finding the optimal energy consumption path while ensuring torque control accuracy.

[0088] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A high-precision torque closed-loop energy control method for a permanent magnet synchronous motor, characterized in that, Includes the following steps: By acquiring the key state signals of the permanent magnet synchronous motor in real time, calculating the change in key state signals between the previous acquisition cycle and the current acquisition cycle, a disturbance vector is constructed to describe the disturbance characteristics. Based on the multidimensional disturbance information in the disturbance vector, a disturbance intensity function is constructed to represent the disturbance intensity of the current permanent magnet synchronous motor. During periods with high disturbance intensity, the target torque is actively reduced so that the controller prioritizes stability. When the disturbance is small, the output expected by the user is kept unchanged. An energy-saving tolerance factor is calculated based on the disturbance intensity function to adjust the weight balance between the loss term and the torque error term in trajectory optimization. The disturbance vector, target torque, energy-saving tolerance factor, and voltage, current, and temperature data of the permanent magnet synchronous motor collected in the current cycle are input into the historical temperature rise weighted memory model, and the structural parameter vector of the permanent magnet synchronous motor in the current state is output, including the estimated value of the d-axis inductance, the estimated value of the q-axis inductance, and the estimated value of the stator resistance in the current cycle. By constructing an optimization objective function, the optimal current command pair is calculated in the dual-axis current space and used to send the controller to execute PWM modulation. This ensures torque control accuracy while dynamically finding the optimal energy consumption path.

2. The high-precision torque closed-loop energy control method for a permanent magnet synchronous motor according to claim 1, characterized in that, The key status signals include the current, voltage, angular displacement, speed, and temperature of the permanent magnet synchronous motor; the disturbance vector includes the flux linkage change amplitude, the difference in average current between two consecutive acquisition cycles, the speed difference between the current acquisition cycle and the previous acquisition cycle, and the temperature change between the current acquisition cycle and the previous acquisition cycle.

3. The high-precision torque closed-loop energy control method for a permanent magnet synchronous motor according to claim 2, characterized in that, The flux linkage variation amplitude is obtained by performing Clarke-Park transform on the voltage and current signals to obtain the voltage and current on the q-axis, and then using the voltage equation to estimate the flux linkage to obtain the flux linkage of the current cycle, which is obtained by subtracting the flux linkage of the previous acquisition cycle. The flux linkage integral is expressed as a discrete integral within a sliding window.

4. The high-precision torque closed-loop energy control method for a permanent magnet synchronous motor according to claim 1, characterized in that, The disturbance intensity function is introduced with a non-equilibrium regularization term for the disturbance structure to reflect the possible estimated risk of superposition of conventional disturbances.

5. The high-precision torque closed-loop energy control method for a permanent magnet synchronous motor according to claim 1, characterized in that, The energy-saving tolerance factor maintains a low energy consumption weight when the disturbance intensity is small, and gradually increases the energy-saving tolerance when the disturbance intensity increases, so as to release more current buffer space.

6. The high-precision torque closed-loop energy control method for a permanent magnet synchronous motor according to claim 1, characterized in that, The energy-saving tolerance factor is also used to reduce energy consumption in periodic systems with small disturbances, while relaxing energy-saving constraints and releasing adjustment space when disturbances intensify.

7. The high-precision torque closed-loop energy control method for a permanent magnet synchronous motor according to claim 1, characterized in that, The historical temperature rise weighted memory model includes a perturbation-driven modeling confidence coefficient, which is used to adjust the response speed and confidence strength of parameter updates, thereby enhancing the confidence response sensitivity when the perturbation structure is unbalanced.

8. The high-precision torque closed-loop energy control method for a permanent magnet synchronous motor according to claim 1, characterized in that, The optimization objective function includes a loss control term and a torque accuracy term. The loss control term is used to suppress the d-axis energy consumption expansion trend under high-frequency weak magnetic conditions. The torque accuracy term is used to compensate for the torque deviation of the q-axis current under disturbance conditions.

9. A high-precision torque closed-loop energy control method for a permanent magnet synchronous motor according to claim 1, characterized in that, The optimal current command pair is obtained based on the objective function and used by the controller to generate a space vector PWM signal, which is directly applied to the inverter to realize the control voltage.

10. A high-precision torque closed-loop energy control system for a permanent magnet synchronous motor, characterized in that, Including those connected sequentially: The data acquisition and processing module is used to acquire key state signals of the permanent magnet synchronous motor in real time, calculate the change in key state signals between the previous acquisition cycle and the current acquisition cycle, and construct a disturbance vector to describe the disturbance characteristics. The torque target adjustment module is used to construct a disturbance intensity function based on the multi-dimensional disturbance information in the disturbance vector to represent the disturbance intensity of the current permanent magnet synchronous motor. During periods with high disturbance intensity, the target torque is actively reduced so that the controller prioritizes stability. When the disturbance is small, the output expected by the user is kept unchanged. The energy-saving tolerance factor is calculated based on the disturbance intensity function to adjust the weight balance between the loss term and the torque error term in trajectory optimization. The structural parameter vector prediction module is used to input the disturbance vector, target torque, energy-saving tolerance factor and voltage, current and temperature data of the permanent magnet synchronous motor collected in the current cycle into the historical temperature rise weighted memory model, and output the structural parameter vector of the permanent magnet synchronous motor in the current state, including the estimated value of the d-axis inductance in the current cycle, the estimated value of the q-axis inductance in the current cycle and the estimated value of the stator resistance in the current cycle. The current command optimization module is used to calculate the optimal current command pair in the dual-axis current space by constructing an optimization objective function. This is then used to send the controller to execute PWM modulation, dynamically finding the optimal energy consumption path while ensuring torque control accuracy.

Images