Motor optimization control method and system based on q-axis voltage control

Abstract

The invention relates to the technical field of motor control, and discloses a motor optimization control method and system based on q-axis voltage control, and the method comprises the steps: carrying out the Clark-Park conversion and sliding mode observation of a three-phase current signal and a three-phase voltage signal, and obtaining a d-axis current signal, a q-axis current signal, a d-axis flux linkage component and a q-axis flux linkage component; performing recursive least square identification processing to obtain a real-time salient pole ratio value; reluctance torque angle arc tangent calculation and d-q axis coupling quantity product operation are carried out to obtain reluctance torque coupling disturbance quantity; according to the method, the limitation that in the prior art, work is only carried out near an MTPA curve is broken through, the adverse effect of d-q-axis cross coupling on q-axis voltage control precision is effectively eliminated, and the control precision of the q-axis voltage is improved. And the torque output capability and the dynamic response performance of the motor under various working conditions are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of motor control, and in particular to a motor optimization control method and system based on q-axis voltage control. Background Art

[0002] Permanent magnet synchronous motors (PMSMs) offer significant advantages, including high power density, excellent efficiency, and robust control performance, making them widely used in electric vehicle drives. With increasing demands for motor control accuracy and dynamic response, traditional vector control strategies have become the mainstream technology for high-performance PMSM control. Q-axis voltage control, as the core component of torque control, directly impacts the motor's torque response characteristics and operational stability.

[0003] However, due to the structural characteristics of their embedded permanent magnets, permanent magnet synchronous motors (PMSMs) exhibit a significant salient pole effect, i.e., a inequality between the d-axis inductance Ld and the q-axis inductance Lq. This inductance difference leads to severe cross-coupling interference between the d-axis and q-axis currents. Specifically, the magnetic field generated by the d-axis current Id affects the control accuracy of the q-axis voltage, causing a deviation between the q-axis voltage command and the actual torque demand. This coupling effect significantly degrades the linearity and dynamic response of torque control, particularly under high-speed and heavy-load conditions, leading to increased torque ripple and decreased control accuracy. Summary of the Invention

[0004] The present invention provides a motor optimization control method and system based on q-axis voltage control. The present invention overcomes the limitation of the existing technology that it only works near the MTPA curve, effectively eliminates the adverse effects of dq-axis cross-coupling on the q-axis voltage control accuracy, and improves the torque output capability and dynamic response performance of the motor under various working conditions.

[0005] In a first aspect, the present invention provides a motor optimization control method based on q-axis voltage control, the motor optimization control method based on q-axis voltage control comprising: Based on the stator voltage equation of the permanent magnet synchronous motor, Clark-Park transformation and sliding mode observation are performed on the three-phase current and three-phase voltage signals to obtain the d-axis current signal, q-axis current signal, d-axis flux component, and q-axis flux component. According to the d-axis flux component, q-axis flux component, d-axis current signal and q-axis current signal, the d-axis inductance value and the q-axis inductance value are subjected to recursive least squares identification processing to obtain the real-time saliency ratio value; Based on the real-time salient pole ratio value, the d-axis current signal and the q-axis current signal are subjected to a reluctance torque angle arc tangent calculation and a dq-axis coupling amount product operation to obtain a reluctance torque coupling interference amount; A compensation gain calculation is performed on the reluctance torque coupling interference to obtain a q-axis voltage feedforward compensation, and a feedforward superposition process is performed on the original q-axis voltage command to generate a q-axis voltage control command.

[0006] In combination with the first aspect, in a first implementation of the first aspect of the present invention, the three-phase current signal and the three-phase voltage signal are subjected to Clark-Park transformation and sliding mode observation based on the stator voltage equation of the permanent magnet synchronous motor to obtain a d-axis current signal, a q-axis current signal, a d-axis flux component, and a q-axis flux component, including: Perform Clark transform and Park transform on the three-phase current signal and three-phase voltage signal of the permanent magnet synchronous motor respectively to obtain the d-axis voltage signal, q-axis voltage signal, d-axis current signal and q-axis current signal; According to the stator voltage equation of the permanent magnet synchronous motor, the d-axis voltage signal and the q-axis voltage signal are differentially transformed to obtain a d-axis magnetic flux differential component and a q-axis magnetic flux differential component; Performing sliding mode observation and integration operations on the d-axis magnetic flux differential and the q-axis magnetic flux differential to obtain a d-axis magnetic flux observation value and a q-axis magnetic flux observation value; The d-axis flux observation value and the q-axis flux observation value are filtered to obtain a d-axis flux component and a q-axis flux component.

[0007] In combination with the first aspect, in a second implementation of the first aspect of the present invention, performing recursive least squares identification processing on the d-axis inductance value and the q-axis inductance value based on the d-axis flux component, the q-axis flux component, the d-axis current signal, and the q-axis current signal to obtain a real-time saliency ratio value includes: performing a flux linkage inductance ratio calculation on the d-axis current signal according to the d-axis flux linkage component and the permanent magnet flux linkage to obtain a d-axis inductance identification value; performing a flux linkage inductance ratio calculation on the q-axis current signal according to the q-axis flux linkage component to obtain a q-axis inductance identification value; performing recursive least squares calculation on the d-axis inductance identification value and the q-axis inductance identification value respectively to obtain a d-axis inductance value and a q-axis inductance value; A ratio operation is performed on the d-axis inductance value and the q-axis inductance value to obtain a real-time salient pole ratio value.

[0008] In combination with the first aspect, in a third implementation of the first aspect of the present invention, performing recursive least squares calculation on the d-axis inductance identification value and the q-axis inductance identification value to obtain the d-axis inductance value and the q-axis inductance value respectively includes: A sampling recursive least squares algorithm is used to construct a regression matrix and organize observation vectors for the d-axis inductance identification value and the q-axis inductance identification value, respectively, to obtain a d-axis regression data matrix and a q-axis regression data matrix; Performing gain vector calculation and covariance matrix update on the d-axis regression data matrix and the q-axis regression data matrix respectively to obtain a d-axis gain vector and a q-axis gain vector; Based on the d-axis gain vector and the q-axis gain vector, respectively updating the parameters of the inductance value at the previous moment to obtain the d-axis inductance intermediate value and the q-axis inductance intermediate value at the current moment; Convergence judgment and numerical stability processing are performed on the intermediate value of the d-axis inductance and the intermediate value of the q-axis inductance respectively to obtain a d-axis inductance value and a q-axis inductance value.

[0009] In combination with the first aspect, in a fourth implementation of the first aspect of the present invention, the performing of a reluctance torque angle inverse tangent calculation and a dq-axis coupling amount product operation on the d-axis current signal and the q-axis current signal based on the real-time salient pole ratio value to obtain the reluctance torque coupling interference amount includes: Performing a difference calculation on the d-axis inductance value according to the real-time salient pole ratio value and the q-axis inductance value to obtain a dq-axis inductance difference; Based on the dq-axis inductance difference, performing a product operation on the d-axis current signal and the q-axis current signal to obtain a numerator term of the inverse tangent function of the reluctance torque angle; According to the permanent magnet flux and the q-axis current signal, a weighted sum is performed on the dq-axis inductance difference and the square term of the d-axis current signal to obtain the denominator term of the inverse tangent function of the reluctance torque angle; Performing magnetic resistance torque angle inverse tangent function calculation and angle range restriction on the numerator and denominator to obtain a magnetic resistance torque angle value; A dq-axis coupling product operation is performed on the rotor electrical angular velocity and the d-axis current signal according to the reluctance torque angle value to obtain the reluctance torque coupling interference amount.

[0010] In combination with the first aspect, in a fifth implementation of the first aspect of the present invention, the performing of the inverse tangent function calculation and angle range restriction on the numerator and denominator to obtain the reluctance torque angle value includes: Performing zero value detection and numerical protection on the denominator term to obtain a denominator protection value; Performing an inverse tangent function operation of a reluctance torque angle according to the numerator term and the denominator protection value to obtain a first reluctance torque angle value; performing angle range normalization and quadrant determination on the first reluctance torque angle value to obtain a second reluctance torque angle value; An angle continuity smoothing process is performed on the second reluctance torque angle value based on the reluctance torque angle at a previous moment to obtain a reluctance torque angle value.

[0011] In combination with the first aspect, in a sixth implementation of the first aspect of the present invention, performing a dq-axis coupling product operation on the rotor electrical angular velocity and the d-axis current signal according to the reluctance torque angle value to obtain the reluctance torque coupling interference amount includes: Performing low-pass filtering on the rotor electrical angular velocity to obtain filtered rotor electrical angular velocity; performing a cubic spline interpolation operation on a cross-inductance lookup table according to the d-axis current signal and the q-axis current signal to obtain a cross-inductance value; performing a product operation based on the filtered rotor electrical angular velocity, the d-axis inductance value, and the d-axis current signal, and performing a weighted summation in combination with the cross inductance value and the q-axis current signal to obtain a basic coupling amount; A cosine modulation operation is performed on the basic coupling amount according to the reluctance torque angle value to obtain the reluctance torque coupling interference amount.

[0012] In combination with the first aspect, in a seventh implementation of the first aspect of the present invention, the performing of compensation gain calculation on the reluctance torque coupling interference to obtain a q-axis voltage feedforward compensation, and performing feedforward superposition processing on the original q-axis voltage command to generate a q-axis voltage control command includes: Adaptively adjusting the proportional gain and the integral gain according to the real-time saliency ratio value to obtain adaptive compensator parameters; Inputting the reluctance torque coupling interference into a proportional-integral-differential compensator for adaptive gain adjustment to obtain a first compensation signal; performing speed adaptive gain attenuation and anti-integral windup processing on the first compensation signal based on the rotor electrical angular velocity to obtain a second compensation signal, and performing a change rate constraint on the second compensation signal to obtain a q-axis voltage feedforward compensation amount; The original q-axis voltage command is subjected to feedforward superposition processing according to the q-axis voltage feedforward compensation amount to generate a q-axis voltage control command.

[0013] In combination with the first aspect, in an eighth implementation of the first aspect of the present invention, performing feedforward superposition processing on the original q-axis voltage command according to the q-axis voltage feedforward compensation amount to generate the q-axis voltage control command includes: Performing frequency analysis and calculating compensation coefficients on the q-axis voltage feedforward compensation to obtain a frequency-domain weighted compensation; Calculating a speed feedforward term based on the rotor electrical angular velocity to obtain a speed feedforward compensation component; Based on the frequency domain weighted compensation amount and the speed feedforward compensation component, a superposition operation is performed on the original q-axis voltage command to obtain an initial voltage control command; A q-axis current tracking error root mean square value is calculated for the initial voltage control instruction to generate a q-axis voltage control instruction.

[0014] In a second aspect, the present invention provides a motor optimization control system based on q-axis voltage control, the motor optimization control system based on q-axis voltage control comprising: A transformation module is used to perform Clark-Park transformation and sliding mode observation on the three-phase current signal and the three-phase voltage signal based on the stator voltage equation of the permanent magnet synchronous motor to obtain a d-axis current signal, a q-axis current signal, a d-axis flux component, and a q-axis flux component; a processing module, configured to perform recursive least square identification processing on the d-axis inductance value and the q-axis inductance value according to the d-axis flux component, the q-axis flux component, the d-axis current signal, and the q-axis current signal, to obtain a real-time salient pole ratio value; a calculation module, configured to perform a reluctance torque angle arc tangent calculation and a dq-axis coupling amount product calculation on the d-axis current signal and the q-axis current signal based on the real-time salient pole ratio value to obtain a reluctance torque coupling interference amount; The compensation module is used to calculate the compensation gain of the magnetic resistance torque coupling interference to obtain the q-axis voltage feedforward compensation, and perform feedforward superposition processing on the original q-axis voltage instruction to generate a q-axis voltage control instruction.

[0015] The technical solution provided by the present invention utilizes an adaptive saliency ratio identification strategy based on flux linkage observation to accurately determine the dq-axis inductance during motor operation. This eliminates the traditional method's reliance on fixed inductance parameters and enables the control system to automatically adapt to changes in motor parameters due to factors such as temperature and magnetic saturation, significantly improving the accuracy and real-time performance of parameter identification. A real-time reluctance torque angle observation algorithm enables precise quantitative monitoring of reluctance torque throughout the motor's operating range, overcoming the limitation of existing technologies that only operate near the MTPA curve. A PID compensator design with adaptive parameter adjustment automatically adjusts compensator parameters based on the real-time saliency ratio, enabling adaptive optimization of the compensation strategy. This effectively addresses the technical challenge of fixed compensation values ​​being unable to adapt to changing operating conditions and significantly improves the accuracy and robustness of decoupled control. By implementing feedforward compensation in the q-axis voltage command generation phase, the complexity of parameter resetting caused by internal compensation in the controller is avoided, achieving complete decoupling control of reluctance torque and excitation torque, and effectively eliminating the adverse effects of dq-axis cross-coupling on q-axis voltage control accuracy. Based on precise decoupling compensation, the current vector angle control range is greatly expanded from the traditional MTPA curve to a wider working area, fully utilizing the potential of reluctance torque and improving the torque output capability and dynamic response performance of the motor under various working conditions.

[0016] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the present invention. The purposes and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0017] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 Schematic diagram of an embodiment of a motor optimization control method based on q-axis voltage control in an embodiment of the present invention; Figure 2 Schematic diagram of an embodiment of a motor optimization control system based on q-axis voltage control in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0020] The terms "including," "having," and any variations thereof, as used in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or device comprising a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to the process, method, product, or device.

[0021] To facilitate understanding of this embodiment, a motor optimization control method based on q-axis voltage control disclosed in an embodiment of the present invention is first described in detail. Figure 1 As shown, this method includes the following steps: 101. Based on the stator voltage equation of the permanent magnet synchronous motor, Clark-Park transformation and sliding mode observation are performed on the three-phase current signal and the three-phase voltage signal to obtain the d-axis current signal, q-axis current signal, d-axis flux component and q-axis flux component; It is understandable that the execution subject of the present invention can be a motor optimization control system based on q-axis voltage control, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking the server as the execution subject as an example.

[0022] Specifically, a Clark transform is performed on the three-phase current and voltage signals of the permanent magnet synchronous motor, converting the three-phase coordinate system into two-phase stationary coordinate system signals. This yields the current and voltage components on the α- and β-axes. A Park transform is then performed, mapping the αβ stationary coordinate system to the dq rotating coordinate system based on the current motor rotor position angle, yielding the d-axis voltage, q-axis voltage, d-axis current, and q-axis current. Based on the d- and q-axis voltage signals, a stator voltage model of the permanent magnet synchronous motor is introduced to isolate the contributions of voltage variations in the d- and q-axis directions to the flux variation. Numerically differentiating the voltage and current signals yields the d- and q-axis flux differentials. Sliding mode observations are then applied to the d- and q-axis flux differentials. A sliding mode observer is a structure suitable for state estimation of nonlinear systems. It dynamically corrects state estimation errors using a switching function and continuously accumulates differential values ​​using an integral operation to obtain smooth and continuous flux observations. To improve the quality of flux estimation and avoid short-term oscillations or discontinuous transitions under dynamic operating conditions, a filtering mechanism is introduced after the sliding mode observer output. This filtering employs a robust filtering structure, such as a filtering algorithm with a fixed observation gain. This method can filter out high-frequency interference while maintaining sufficient dynamic response speed. By smoothing and filtering the d-axis and q-axis flux observations, observation noise is effectively removed, resulting in d-axis and q-axis flux components that are closer to actual physical conditions.

[0023] 102. Perform recursive least square identification processing on the d-axis inductance value and the q-axis inductance value according to the d-axis flux component, the q-axis flux component, the d-axis current signal, and the q-axis current signal to obtain a real-time salient pole ratio value; Specifically, in a permanent magnet synchronous motor, the d-axis flux is composed of the fixed flux generated by the permanent magnets and the flux generated by the winding inductance. The fixed permanent magnet flux is subtracted from the d-axis flux to obtain the flux caused by the inductance component. This is then ratioed with the d-axis current signal to calculate the instantaneous d-axis inductance identification value. The q-axis flux, however, is generated solely by the winding inductance. Its flux component is directly ratioed with the q-axis current signal to obtain the instantaneous q-axis inductance identification value. Recursive least squares calculations are performed on the d-axis and q-axis inductance identification values, respectively. Based on the current inductance estimate, the algorithm combines historical estimates with observation errors to construct an incremental correction model. The inductance value is continuously updated within each control cycle, and the algorithm's adaptability to changing operating conditions is enhanced by setting a forgetting factor and an initial covariance matrix. This process yields the d-axis and q-axis inductance values, which are then ratioed to determine the real-time saliency ratio for the current motor operating state. The saliency ratio is a parameter that reflects the reluctance characteristics of a permanent magnet synchronous motor, and its value is related to the magnitude and direction of the reluctance torque component.

[0024] 103. Performing a reluctance torque angle arc tangent calculation and a dq-axis coupling product operation on the d-axis current signal and the q-axis current signal based on the real-time salient pole ratio value to obtain a reluctance torque coupling interference amount; Specifically, the difference between the d-axis and q-axis inductances is calculated based on the real-time saliency ratio and q-axis inductance. This difference reflects the degree of magnetic anisotropy in the motor structure, which is the source of the strength of the reluctance effect. The d-axis and q-axis inductance difference is multiplied with the current d-axis and q-axis current signals, respectively, to form the numerator of the inverse tangent function for the reluctance torque angle. The value of this numerator represents the energy distribution structure of the reluctance coupling caused by the mismatch between the d-axis and q-axis inductances. To construct the denominator of the inverse tangent function, the q-axis current signal is multiplied with the fixed flux provided by the permanent magnets. The inductance difference is then weighted and summed with the square of the d-axis current signal to form the complete denominator. This denominator describes the combined effect of the permanent magnet flux and the inductive coupling term on the total flux. After obtaining the numerator and denominator, a nonlinear ratio mapping is performed on them using the inverse tangent function to obtain an instantaneous estimate of the reluctance torque angle. Because the inverse tangent function can produce amplitude fluctuations greater than 180 degrees or less than −180 degrees in actual calculations, this process incorporates an angle range limitation mechanism to ensure that the reluctance torque angle remains within a stable and controllable engineering angle range. The reluctance torque angle is introduced as a modulation factor into the reluctance coupling interference modeling process. This product is combined with the real-time rotor electrical angular velocity and the d-axis current signal to generate a coupling interference quantity that reflects the influence of the reluctance disturbance. This coupling interference quantity represents the actual degree of interference caused by the d-axis current on the q-axis torque through the reluctance effect in the current state.

[0025] 104. Perform compensation gain calculation on the reluctance torque coupling interference to obtain a q-axis voltage feedforward compensation, and perform feedforward superposition processing on the original q-axis voltage command to generate a q-axis voltage control command.

[0026] Specifically, with the real-time salient pole ratio as the core adjustment factor, the proportional gain and integral gain in the compensation controller are dynamically corrected to obtain the optimal control response parameters under the current operating state. This adaptive parameter adjustment mechanism can dynamically modify the response strength of the compensator in response to the nonlinear interference caused by the change of inductance ratio with the working conditions. The proportional gain is used to quickly track the compensation error, and the integral gain is used to correct the long-term steady-state error, so that the compensator can maintain a compatible balance between accuracy and stability in different torque loads and speed ranges. The reluctance torque coupling interference is input into the proportional-integral-differential compensator with adaptive gain for processing. In the compensator structure, the real-time salient pole ratio is used as a regulating variable to participate in the dynamic scaling of the gain, thereby generating a first compensation signal, which preliminarily reflects the correction component of the coupling interference in the q-axis voltage command. Considering that motors operating at low speeds or under variable acceleration and deceleration conditions are prone to control overshoot due to error accumulation, a speed adaptive mechanism based on the rotor's electrical angular velocity is introduced to perform gain reduction on the first compensation signal. This automatically reduces the compensator output when the speed falls below a certain threshold to prevent signal amplification distortion. Furthermore, to improve system stability, an anti-integral windup strategy is integrated within the compensator. This strategy automatically stops the integral accumulation process when the compensation output reaches a preset voltage upper limit, preventing continued amplification of the integral term and control signal overflow. After this processing, a rate constraint is applied to the adjusted second compensation signal to prevent sudden or rapid changes in the voltage compensation term in the time domain, which could cause incoherent dynamic responses in the power drive train. A maximum slope threshold is set to limit the increase in the q-axis compensation voltage per unit time, improving the smoothness of the compensation channel and ensuring dynamic consistency with the main control loop. The original q-axis voltage command is feedforward-superimposed based on the q-axis voltage feedforward compensation and combined with other voltage components, such as the speed feedforward, to generate the q-axis voltage control command.

[0027] In a specific embodiment, the process of executing step 101 may specifically include the following steps: Perform Clark transform and Park transform on the three-phase current signal and three-phase voltage signal of the permanent magnet synchronous motor respectively to obtain the d-axis voltage signal, q-axis voltage signal, d-axis current signal and q-axis current signal; According to the stator voltage equation of the permanent magnet synchronous motor, the d-axis voltage signal and the q-axis voltage signal are differentially transformed to obtain a d-axis magnetic flux differential component and a q-axis magnetic flux differential component; Performing sliding mode observation and integration operations on the d-axis magnetic flux differential and the q-axis magnetic flux differential to obtain a d-axis magnetic flux observation value and a q-axis magnetic flux observation value; The d-axis flux observation value and the q-axis flux observation value are filtered to obtain a d-axis flux component and a q-axis flux component.

[0028] Specifically, the three-phase stator current and voltage signals of the permanent magnet synchronous motor are acquired during operation. These signals are acquired in real time at a high-frequency sampling rate using current transformers and voltage sampling modules and fed into the control system for coordinate transformation. Because the motor's three-phase windings are symmetrically distributed 120 degrees in space, directly processing these three-phase signals is difficult to achieve a concise and effective mathematical description in the control algorithm. Therefore, the Clark transform, also known as the αβ transform, is applied. This transform converts the three-phase AC quantities into equivalent two-phase orthogonal components, compressing the signals originally distributed in a three-dimensional spatial coordinate system onto a two-dimensional coordinate plane, forming current and voltage components along the α and β axes. The Park transform, also known as the dq transform, is applied to the αβ components. Based on the current rotor position angle, the αβ components in the stationary reference frame are rotationally transformed into a rotating coordinate system referenced to the rotor. The external observation angle rotates synchronously with the rotor in real time, mapping the motor's dynamic behavior into components that vary in a fixed direction, achieving a decoupled representation of the voltage and current quantities along the d and q axes. This transformation yields the d-axis voltage, q-axis voltage, d-axis current, and q-axis current signals, respectively. The d-axis corresponds to the magnetic pole direction of the permanent magnet, while the q-axis is perpendicular to the d-axis and is the primary direction of torque generation. Based on the stator voltage equation for a permanent magnet synchronous motor, the d-axis and q-axis voltage signals are differentially transformed. This voltage equation mathematically describes the relationship between voltage, current, resistance, and flux within the motor from the perspective of energy conservation. By analyzing its d- and q-axis components, an expression is constructed with the voltage and current as known quantities and the flux rate as an unknown quantity. Based on this, differential transformations are applied to the d- and q-axis voltage signals, respectively. Combined with the current d- and q-axis currents, the d- and q-axis flux differentials are indirectly calculated. These differentials numerically represent the instantaneous increment of flux over time and are key intermediate quantities in the formation of the dynamic flux trajectory. A sliding mode observer is introduced to perform nonlinear state estimation of the flux differentials. As an observation structure that is highly robust to system uncertainties and external disturbances, the sliding mode observer can quickly drive the system back to a stable track in the presence of state errors by setting a reasonable switching surface and sliding mode gain, thereby achieving a high-confidence reconstruction of the internal state of the motor. In this application, the sliding mode observer uses the flux differential as input, combines the error correction mechanism to model and predict the flux change trend, and reconstructs the continuous and stable flux observation values ​​by accumulating the differential trajectory through integral operations, obtaining the d-axis flux observation value and the q-axis flux observation value respectively. In order to improve the stability and anti-interference ability of the flux estimation results, a filtering process is set after the sliding mode observation. The filter design takes into account both response speed and smoothness, ensuring that it can quickly follow dynamic changes and effectively suppress high-frequency interference during steady state.The observation values ​​are processed using a fixed-gain low-pass filter or a simplified Kalman filter, making the output flux data not only numerically continuous but also highly resistant to jitter. The filtered d-axis flux observation value and q-axis flux observation value form the d-axis flux component and the q-axis flux component.

[0029] In a specific embodiment, the process of executing step 102 may specifically include the following steps: performing a flux linkage inductance ratio calculation on the d-axis current signal according to the d-axis flux linkage component and the permanent magnet flux linkage to obtain a d-axis inductance identification value; performing a flux linkage inductance ratio calculation on the q-axis current signal according to the q-axis flux linkage component to obtain a q-axis inductance identification value; performing recursive least squares calculation on the d-axis inductance identification value and the q-axis inductance identification value respectively to obtain a d-axis inductance value and a q-axis inductance value; A ratio operation is performed on the d-axis inductance value and the q-axis inductance value to obtain a real-time salient pole ratio value.

[0030] Specifically, the d-axis and q-axis flux components are used as physical inputs, and the known permanent magnet flux values ​​within the motor are introduced as fixed parameters for modeling. In the d-axis direction, since the flux is derived not only from the inductor current but also from the permanent magnet's own magnetic flux, the permanent magnet flux contained in the d-axis flux component is removed during the calculation process, retaining the flux formed by the inductor's response to current. This allows for a pure flux-to-current ratio calculation, which is physically equivalent to the d-axis inductance identification value. In the q-axis direction, since there is no permanent magnet flux component, the q-axis flux component is directly proportional to the q-axis current signal to obtain the q-axis inductance identification value. Because both the instantaneous flux component and the current signal are affected by electromagnetic fluctuations, load disturbances, and sampling errors, a stable modeling method with dynamic response capabilities and the ability to utilize historical data is employed. Within this framework, a recursive least squares algorithm incrementally updates the input and observation errors at each moment and dynamically adjusts the estimated values ​​based on historical sample weights. This algorithm maintains computational efficiency while minimizing the impact of short-term errors on modeling accuracy. The identified d-axis and q-axis inductance values ​​are used as input signals, respectively. The estimation matrix and gain coefficients are adjusted using the minimum squared error criterion based on the previous cycle's estimation results. Recursively updating the values ​​based on the current input yields the current d-axis and q-axis inductance values. The recursive calculation sets the initial inductance estimation range and covariance matrix, and uses a forgetting factor to control sensitivity to historical data, thereby avoiding extreme misjudgments when the motor's operating state changes suddenly. As the runtime progresses, the identified values ​​gradually stabilize and become adaptable to operating point changes, enabling real-time tracking of the inductance response characteristics under varying speeds and loads. Once stable estimates of the d- and q-axis inductances are achieved, these two time-varying variables are input into the inductance ratio calculation module to obtain the real-time saliency ratio. This ratio reflects the degree of electromagnetic anisotropy between the motor rotor along the d-axis and the q-axis and is a direct measure of the motor's reluctance torque generation capability. When the d-axis inductance is significantly greater than the q-axis inductance, the saliency ratio increases significantly, indicating a stronger potential for reluctance torque modulation. Conversely, a lower saliency ratio indicates weaker magnetic anisotropy and less pronounced reluctance characteristics. To enhance the stability and real-time performance of this ratio, the entire computational chain implements high-frequency data acquisition and low-latency recursive updates at the hardware level using a DSP or FPGA. At the software level, a sliding window mechanism is introduced to dynamically smooth the ratio curve, thereby counteracting the effects of input noise and maintaining the physical plausibility of saliency ratio variations. To accommodate short-term nonlinear fluctuations caused by sudden loads or voltage disturbances, a fault detection mechanism is designed to temporarily suspend updates when an abnormal rate of change is triggered, resuming identification and updates after the system stabilizes.

[0031] In a specific embodiment, the step of performing recursive least squares calculation on the d-axis inductance identification value and the q-axis inductance identification value to obtain the d-axis inductance value and the q-axis inductance value may specifically include the following steps: A sampling recursive least squares algorithm is used to construct a regression matrix and organize observation vectors for the d-axis inductance identification value and the q-axis inductance identification value, respectively, to obtain a d-axis regression data matrix and a q-axis regression data matrix; Performing gain vector calculation and covariance matrix update on the d-axis regression data matrix and the q-axis regression data matrix respectively to obtain a d-axis gain vector and a q-axis gain vector; Based on the d-axis gain vector and the q-axis gain vector, respectively updating the parameters of the inductance value at the previous moment to obtain the d-axis inductance intermediate value and the q-axis inductance intermediate value at the current moment; Convergence judgment and numerical stability processing are performed on the intermediate value of the d-axis inductance and the intermediate value of the q-axis inductance respectively to obtain a d-axis inductance value and a q-axis inductance value.

[0032] Specifically, based on the d-axis and q-axis inductance identification values, a data structure containing the mapping relationship between input and output is established. Continuously sampled current, voltage, and flux data are organized, with the current signal at each sampling moment used as the regression input and the corresponding flux identification value as the observed output. This results in the formation of d-axis and q-axis regression data matrices. These matrices, respectively, store the linear relationship models between multiple input variables and target parameters at current and historical times and are used within the algorithm to recursively update core parameters. As new sampled data is continuously input, the system dynamically expands or updates these two regression matrices, ensuring that they retain historical information while promptly reflecting current system state changes. Based on the regression matrices, an associated gain vector is calculated. This gain vector reflects the contribution of the current input data to the target inductance parameter correction. In the recursive least squares algorithm, the gain vector not only directly influences the weight distribution for parameter updates but also correlates with the current level of uncertainty in the system. A covariance matrix is ​​introduced as an intermediate factor for gain adjustment and is dynamically updated during each recursive step. The covariance matrix records the distribution of the estimation error across various dimensions, and its changes reflect the algorithm's adjustment of its confidence in the current inductance estimate. By combining the current regression matrix with historical covariance data, the current d-axis gain vector and q-axis gain vector are gradually updated. These two gain vectors numerically reflect the parameter model's response sensitivity to the input information and the estimated direction. Based on these gain vectors, the original inductance value is used as the estimate for the previous moment and is corrected and updated by multiplying the current gain vector by the observed error to form the new d-axis and q-axis inductance intermediate values. Convergence judgment and numerical stability processing are performed on the d-axis and q-axis intermediate inductance values, respectively. To prevent the inductance value from drifting continuously or falling into local extremes under specific operating conditions, two constraint mechanisms are introduced: one is an incremental constraint mechanism, which determines whether the change in inductance value between the current moment and the previous moment is within a physically reasonable range. If the rate of change exceeds the limit, an update pause or rollback logic is triggered. The other is a long-term stability analysis mechanism, which assesses whether the valuation trend has converged to a stable range based on the fluctuation of the inductance estimate within a sliding time window. If oscillation or sharp fluctuations persist, the forgetting factor is appropriately increased, reducing the weight of historical data, thereby improving the model's response to the latest data. A fusion filter module is introduced to the convergence judgment, low-pass filtering the intermediate values. This eliminates high-frequency disturbances while preserving the true dynamic trend, further improving the physical reliability of the estimate. Once the convergence conditions are met and the intermediate values ​​pass the stability check, they are used as the final d-axis and q-axis inductance values ​​for control.

[0033] In a specific embodiment, the process of executing step 103 may specifically include the following steps: Performing a difference calculation on the d-axis inductance value according to the real-time salient pole ratio value and the q-axis inductance value to obtain a dq-axis inductance difference; Based on the dq-axis inductance difference, performing a product operation on the d-axis current signal and the q-axis current signal to obtain a numerator term of the inverse tangent function of the reluctance torque angle; According to the permanent magnet flux and the q-axis current signal, a weighted sum is performed on the dq-axis inductance difference and the square term of the d-axis current signal to obtain the denominator term of the inverse tangent function of the reluctance torque angle; Performing magnetic resistance torque angle inverse tangent function calculation and angle range restriction on the numerator and denominator to obtain a magnetic resistance torque angle value; A dq-axis coupling product operation is performed on the rotor electrical angular velocity and the d-axis current signal according to the reluctance torque angle value to obtain the reluctance torque coupling interference amount.

[0034] Specifically, based on the current real-time saliency ratio and q-axis inductance, these two parameters are algebraically manipulated to infer the corresponding d-axis inductance. This difference is then subtracted from the currently derived d-axis inductance estimate. This difference represents the inductance asymmetry between the d-axis and q-axis due to structural magnetic anisotropy, i.e., the dq-axis inductance difference. This reflects the strength of the motor's reluctance effect and is the core physical basis for its ability to generate reluctance torque. This inductance difference is then substituted into the reluctance torque angle modeling framework for function construction. The numerator of the inverse tangent function is constructed by multiplying the d-axis current signal by the q-axis current signal and multiplying it by the aforementioned inductance difference to form a numerically proportional term for the reluctance torque angle. Physically, this numerator describes the asymmetric magnetic field distribution induced by the coupling between the d-axis and q-axis currents, which is the primary source of reluctance torque. Because the current signal is constantly fluctuating, this product term exhibits significant time-varying characteristics, and its contribution to the reluctance torque angle adjusts in real time with changing operating conditions. The denominator of the inverse tangent function for the reluctance torque angle is constructed. This component balances the fixed permanent magnet flux and the inductance-modulated flux in the flux distribution, providing a normalized baseline for function calculation. In this process, the current q-axis current signal is taken and multiplied by the permanent magnet flux value preset in the motor parameters to form the first component of the denominator, which represents the flux contribution directly provided by the permanent magnet. The second component of the denominator is derived from the product of the squared d-axis current and the inductance difference. This component reflects the flux modulation effect caused by the inductance difference under the influence of the d-axis current. The weighted sum of these two sub-terms forms the complete denominator of the reluctance torque angle. The value of this denominator reflects the magnitude of the current non-uniformity in the total flux distribution. The inverse tangent function of the reluctance torque angle is calculated on the numerator and denominator to obtain the real-time value of the reluctance torque angle. Considering the jumps and discontinuities of the inverse tangent function at the boundaries of its definition domain, the obtained reluctance torque angle is subject to an angle range constraint to ensure system stability. A compression function or a sign-limiting mechanism, limiting the angle to between −90 and 90 degrees, is used to avoid sudden changes in the output and ensure that the obtained angle value remains reasonable in both physical and engineering terms. After the reluctance torque angle is calculated, a quantitative representation of the reluctance torque coupling interference is constructed. The reluctance torque angle value is multiplied with the current rotor electrical angular velocity and d-axis current signal to form the reluctance torque interference modeling term. In this operation, the electrical angular velocity, as a time rate parameter, introduces a modulation effect in the frequency dimension, while the d-axis current signal, as a driving component, reflects the actual contribution of the magnetic field to its spatial distribution. This product structure models the dynamic interference effect of the reluctance torque on the q-axis voltage control channel under spatial and temporal coupling. The resulting reluctance torque coupling interference has directionality and amplitude adjustability, and can reflect the actual motor operating conditions in real time, making it suitable for constructing a feedforward control path in the compensator module.

[0035] In a specific embodiment, the step of performing the reluctance torque angle inverse tangent function calculation and angle range restriction on the numerator and denominator to obtain the reluctance torque angle value may specifically include the following steps: Performing zero value detection and numerical protection on the denominator term to obtain a denominator protection value; Performing an inverse tangent function operation of a reluctance torque angle according to the numerator term and the denominator protection value to obtain a first reluctance torque angle value; performing angle range normalization and quadrant determination on the first reluctance torque angle value to obtain a second reluctance torque angle value; An angle continuity smoothing process is performed on the second reluctance torque angle value based on the reluctance torque angle at a previous moment to obtain a reluctance torque angle value.

[0036] Specifically, a zero-value detection and numerical protection mechanism is implemented for the denominator term to ensure that the entire reluctance torque angle calculation process is stable at the numerical level. Since the reluctance torque angle is essentially the result of modeling the asymmetric distribution of magnetic flux based on the inverse tangent function structure, its denominator term is composed of terms related to the difference between the permanent magnet magnetic flux and the inductance, and under special current conditions, transient conditions close to zero or extremely small values ​​may occur, which will cause instability such as sudden changes in the function output, angle jumps, or numerical overflows. In order to eliminate these abnormal risks, a threshold judgment logic is introduced for the denominator term. When it is detected that the denominator value is less than the preset zero-point proximity threshold, the set lower limit protection value is automatically used to replace the original denominator to form a denominator protection value, thereby avoiding the error of dividing by zero in a mathematical sense and ensuring the differentiability and controllability of the function curve. After completing the denominator protection processing, this protection value and the numerator term calculated in the previous sequence together constitute the input parameters of the inverse tangent function, and the standard inverse tangent function operation is performed to obtain the first reluctance torque angle value. This angle physically reflects the degree of magnetic field distortion induced by inductance asymmetry in the spatial distribution of the motor's internal flux linkage and is a key intermediate variable for determining the directionality of the reluctance torque. Due to the mathematical properties of the inverse tangent function, its output range exhibits multi-quadrant overlap, meaning the function's value can cross the ±180-degree boundary or experience sudden jumps when the input perturbation is small. To address this issue, the first reluctance torque angle value is range-normalized and subjected to quadrant judgment to ensure that the final angle value always falls within a closed and continuous target range, limited to ±90 or ±180 degrees. Angle jumps are corrected through modulo, sign mapping, or phase envelope reconstruction. The mapping between positive and negative signs and amplitude is adjusted in conjunction with quadrant judgment logic to form the second reluctance torque angle value. Under high-speed operation, rapid current perturbations, or extremely small inductance differences, the reluctance torque angle exhibits a tendency to fluctuate dramatically from point to point, causing disruptive impacts on the control system's response. To this end, an angle continuity smoothing mechanism is introduced based on the second reluctance torque angle value, performing a constrained smooth transition between it and the previously calculated and stored reluctance torque angle value. During this process, a sliding average, exponential weighting, or an interpolation function with a maximum rate of change limit is used to temporally fuse the current angle value with the previous angle value. This ensures that the angle variation per unit time does not exceed the system's acceptable dynamic threshold, while also ensuring that the angle trajectory exhibits a continuous derivative and a stable growth trend on short timescales. To avoid sign misalignment and multi-cycle jumps in the angle value data structure, a first-order difference monitoring mechanism is introduced. This mechanism calculates the angle difference between two adjacent moments and compares the absolute value of the difference. When the difference exceeds a set variation limit, an anomaly correction logic is automatically triggered, reverting to the most recent continuous angle state. This enhances the controllability and anti-interference capability of the angle value in the time domain.

[0037] In a specific embodiment, the step of performing a dq-axis coupling product operation on the rotor electrical angular velocity and the d-axis current signal according to the reluctance torque angle value to obtain the reluctance torque coupling interference value may specifically include the following steps: Performing low-pass filtering on the rotor electrical angular velocity to obtain filtered rotor electrical angular velocity; performing a cubic spline interpolation operation on a cross-inductance lookup table according to the d-axis current signal and the q-axis current signal to obtain a cross-inductance value; performing a product operation based on the filtered rotor electrical angular velocity, the d-axis inductance value, and the d-axis current signal, and performing a weighted summation in combination with the cross inductance value and the q-axis current signal to obtain a basic coupling amount; A cosine modulation operation is performed on the basic coupling amount according to the reluctance torque angle value to obtain the reluctance torque coupling interference amount.

[0038] Specifically, the rotor electrical angular velocity signal is low-pass filtered. A first- or second-order low-pass filter is used to limit its frequency range, remove high-frequency interference, and retain only the physically significant components of the rotational speed, resulting in the filtered rotor electrical angular velocity. The cross-inductance is modeled to account for the cross-coupling effects caused by the non-ideal flux structure in the motor. Cross-inductance is the mutual inductance between the d-axis and q-axis due to winding position, magnetic field distortion, and material nonlinearity. It significantly impacts the coupling term in the voltage control link during high-speed operation or under field-weakening conditions. Because the cross-inductance value cannot be directly calculated using a simple formula, it is calculated offline at multiple d-axis and q-axis current operating points using finite element analysis during the design phase. The values ​​are then structured and stored in a two-dimensional lookup table. To obtain the cross-inductance value at the current operating point in real time within the control system, the lookup table is interpolated. In the specific implementation, cubic spline interpolation is used to interpolate the two-dimensional cross-inductance lookup table. Using the current d-axis and q-axis current signals as the horizontal and vertical coordinates, a smooth transition value of the inductance near that operating point is obtained. Compared to linear interpolation or simple nearest neighbor interpolation, cubic spline interpolation achieves smooth connections across function surfaces while maintaining accuracy. This avoids sudden angle changes at interpolation boundaries or intermediate locations, effectively improving modeling continuity and response quality, and obtaining the cross-inductance value for the current operating condition. The first coupling term, reflecting the effect of d-axis inductance changes on the q-axis direction, is obtained by multiplying the filtered rotor electrical angular velocity, d-axis inductance, and d-axis current signal. The cross-inductance term is then weighted and multiplied with the current q-axis current signal to form the coupling effect component caused by the cross-inductance term. The physical meaning of these two sub-terms corresponds to the coupling interference sources caused by the self-inductance term and the cross-mutual inductance term in the motor, respectively. These two sub-terms are directional, amplitude-dependent, and operating-condition-dependent. The summation of these two coupling terms yields the basic coupling term, which directly reflects the interference contribution of the interaction between the d-axis and q-axis inductance and current to the voltage control loop during the current control cycle. The effective interference degree of the coupling term is also modulated by the space vector angle parameter, the reluctance torque angle. The reluctance torque angle reflects the spatial position of the reluctance torque relative to the motor's stator reference frame. Its value determines the strength of the reluctance term's projection along the q-axis. Therefore, a cosine modulation operation is applied to the basic coupling quantity to reflect the angular effect on the directionality of the coupling component. In this step, the basic coupling quantity is multiplied by the cosine value of the reluctance torque angle. This cosine term essentially maps the spatial rotation angle to the orthogonal projection ratio along the q-axis. Therefore, the modulated result truly describes the effective impact of the coupling interference along the q-axis. The value obtained through this modulation process is the reluctance torque coupling interference quantity.

[0039] In a specific embodiment, the process of executing step 104 may specifically include the following steps: Adaptively adjusting the proportional gain and the integral gain according to the real-time saliency ratio value to obtain adaptive compensator parameters; Inputting the reluctance torque coupling interference into a proportional-integral-differential compensator for adaptive gain adjustment to obtain a first compensation signal; performing speed adaptive gain attenuation and anti-integral windup processing on the first compensation signal based on the rotor electrical angular velocity to obtain a second compensation signal, and performing a change rate constraint on the second compensation signal to obtain a q-axis voltage feedforward compensation amount; The original q-axis voltage command is subjected to feedforward superposition processing according to the q-axis voltage feedforward compensation amount to generate a q-axis voltage control command.

[0040] Specifically, the real-time saliency ratio is used as the basis for adaptive parameter adjustment. Combined with preset proportional gain and integral gain base values, the saliency ratio is dynamically mapped to a function to adjust the parameters of the proportional-integral controller in real time. A larger saliency ratio indicates a stronger motor state with greater magnetic anisotropy and higher reluctance torque contribution. In this case, the system faces stronger coupling disturbances. Therefore, the proportional gain is increased to enhance response speed, while the integral gain is correspondingly increased to strengthen steady-state suppression capability. Conversely, when the saliency ratio decreases, the system's reluctance influence weakens, and the controller gain is appropriately reduced to avoid over-amplification of the system response and oscillation. Through a gain adaptive mechanism based on the real-time saliency ratio, the optimal compensator parameter configuration for the current operating conditions is determined, thereby constructing a controller internal structure with intelligent adjustment capabilities. The reluctance torque coupling disturbance is introduced as an input signal into the proportional-integral-differential compensator. This input signal is adjusted in real time using the updated gain, outputting a first compensation signal. This first compensation signal represents the control system's initial response to the current disturbance intensity. Its numerical structure reflects the proportional amplification relationship and integral regulation trend between the q-axis voltage compensation and the disturbance term. Considering the time-varying motor operating conditions, especially at low speeds or sudden torque loads, the system's ability to accept compensation has significant nonlinear limitations. Therefore, adaptive speed modulation is performed based on the first compensation signal. The current rotor electrical angular velocity is used as a modulation factor to construct a speed-adaptive gain attenuation function. When the electrical angular velocity falls below a set threshold, the compensation gain is automatically linearly attenuated to prevent excessive compensator output and q-axis voltage overshoot at low speeds. Furthermore, the system incorporates an anti-windup mechanism. This mechanism automatically stops integral accumulation when the accumulated integral term within the compensator exceeds the upper limit of the compensation amplitude, preventing output convergence failure due to continued error accumulation. These two modulation mechanisms provide speed protection and integral safety margin control for the first compensation signal, resulting in a second compensation signal. The second compensation signal is rate-constrained, constraining the rising slope of the compensation value in the time domain to ensure that its unit-time variation does not exceed the system's acceptable dynamic variation limit, thereby ensuring a smooth output transition across the entire compensation path. By setting a maximum rate-of-change threshold and employing a differential comparison between the current signal and the previous cycle's signal, the system automatically limits the signal to within an allowable range when excessive changes are detected, enabling flexible regulation of the signal growth slope. This yields a q-axis voltage feedforward compensation, reflecting the voltage component that requires correction under the system's current magnetoresistive coupling interference state. The system exhibits multiple characteristics, including responsiveness, boundary control, and dynamic compliance. This feedforward compensation is then superimposed in real time with the original q-axis voltage command generated by the PI controller or velocity feedforward module in the main control channel to form the final q-axis voltage control command output.

[0041] In a specific embodiment, the step of performing feedforward superposition processing on the original q-axis voltage command according to the q-axis voltage feedforward compensation amount to generate the q-axis voltage control command may specifically include the following steps: Performing frequency analysis and calculating compensation coefficients on the q-axis voltage feedforward compensation to obtain a frequency-domain weighted compensation; Calculating a speed feedforward term based on the rotor electrical angular velocity to obtain a speed feedforward compensation component; Based on the frequency domain weighted compensation amount and the speed feedforward compensation component, a superposition operation is performed on the original q-axis voltage command to obtain an initial voltage control command; A q-axis current tracking error root mean square value is calculated for the initial voltage control instruction to generate a q-axis voltage control instruction.

[0042] Specifically, the q-axis voltage feedforward compensation is subjected to frequency analysis. Because this compensation is affected by multidimensional signals such as torque disturbances, inductance variations, and speed fluctuations during the reluctance-torque coupling modeling process, its spectrum contains significant mid- and low-frequency components as well as some high-frequency spurious responses. Therefore, the compensation is subjected to frequency deconstruction using techniques such as Fourier transform, sliding window spectrum analysis, or bandpass filtering. Its energy distribution in specific frequency bands is extracted and normalized, resulting in a complete characterization of the signal's frequency characteristics. Based on the frequency analysis results, a compensation coefficient calculation mechanism is established. Specifically, different weighting factors are applied to the compensation in different frequency bands according to the power density distribution or frequency weight response function in the frequency domain. In low-frequency bands, the compensation maintains full-amplitude response to enhance the ability to correct slow-varying disturbances, while an attenuation factor is applied in high-frequency bands to avoid amplifying high-frequency noise. This results in a frequency-domain weighted compensation with frequency selectivity and response flexibility. The speed feedforward term is calculated based on the rotor electrical angular velocity. This calculation is based on the current rotor electrical angular velocity and establishes a feedforward control channel based on the speed-voltage response relationship in the motor model. Leveraging the direct proportional relationship between speed and induced voltage, combined with given system parameters such as stator inductance and permanent magnet flux, a forward-looking estimate of speed trends is generated to form a speed feedforward compensation component. This component, independent of closed-loop current error feedback, preemptively adjusts the voltage command based on the system's dynamic trend. This effectively improves q-axis current tracking accuracy in high-speed ranges or acceleration conditions, reduces current hysteresis, and enhances the overall system dynamic response efficiency. The frequency-domain weighted compensation and the speed feedforward compensation component are combined with the original q-axis voltage command. The original command, derived from the output of a PI regulator, primarily regulates the error between the actual and desired q-axis current values. Summing these three voltage components yields an initial voltage control command that incorporates error feedback correction, magnetic reluctance disturbance suppression, and speed response pre-adjustment, forming a multi-channel fused composite control signal. A quantitative evaluation mechanism for q-axis current tracking error allows for performance evaluation and correction of the initial command. The error between the actual q-axis current and its target reference value is monitored in real time, and the RMS value of the error is calculated using a sliding window structure. This calculation filters out the effects of transient pulse errors and reflects the steady-state and dynamic tracking capabilities of the current control system within the current control cycle. This RMS error is dynamically assessed. When it falls below a preset stability threshold, the controller is performing well under the current parameter configuration, and the initial voltage control command is directly output as the final q-axis voltage control command. If the error deviates too far, the system automatically triggers an optimization module, such as adjusting the compensator gain, resetting the velocity feedforward coefficient, or adjusting the frequency weighting strategy, to gradually optimize the voltage command output path in the next control cycle. This generates the q-axis voltage control command.

[0043] The above describes the motor optimization control method based on q-axis voltage control in the embodiment of the present invention. The following describes the motor optimization control system based on q-axis voltage control in the embodiment of the present invention. Figure 2 In one embodiment of the present invention, a motor optimization control system based on q-axis voltage control includes: A transformation module 201 is configured to perform Clark-Park transformation and sliding mode observation on the three-phase current signal and the three-phase voltage signal based on the stator voltage equation of the permanent magnet synchronous motor to obtain a d-axis current signal, a q-axis current signal, a d-axis flux component, and a q-axis flux component; The processing module 202 is configured to perform recursive least squares identification processing on the d-axis inductance value and the q-axis inductance value based on the d-axis flux component, the q-axis flux component, the d-axis current signal, and the q-axis current signal to obtain a real-time saliency ratio value. A calculation module 203 is configured to perform a reluctance torque angle arc tangent calculation and a dq-axis coupling product calculation on the d-axis current signal and the q-axis current signal based on the real-time salient pole ratio value to obtain a reluctance torque coupling interference value; The compensation module 204 is configured to calculate a compensation gain for the reluctance torque coupling interference to obtain a q-axis voltage feedforward compensation, and perform feedforward superposition processing on the original q-axis voltage command to generate a q-axis voltage control command.

[0044] Through the collaborative efforts of these components, an adaptive saliency ratio identification strategy based on flux linkage observation enables real-time acquisition of accurate values ​​for the d / q axis inductances during motor operation. This eliminates the traditional method's reliance on fixed inductance parameters and enables the control system to automatically adapt to changes in motor parameters due to factors such as temperature and magnetic saturation, significantly improving the accuracy and real-time performance of parameter identification. A real-time reluctance torque angle observation algorithm enables precise quantitative monitoring of reluctance torque across the entire motor operating range, overcoming the limitation of existing technologies that only operate near the MTPA curve. A PID compensator design with adaptive parameter adjustment automatically adjusts compensator parameters based on the real-time saliency ratio, enabling adaptive optimization of the compensation strategy. This effectively addresses the technical challenge of fixed compensation values ​​being unable to adapt to changing operating conditions and significantly improves the accuracy and robustness of decoupled control. Feedforward compensation is implemented in the q-axis voltage command generation phase, eliminating the complexity of parameter resetting required by traditional controller internal compensation methods. This achieves complete decoupling of reluctance torque from excitation torque, effectively eliminating the adverse effects of d / q axis cross-coupling on q-axis voltage control accuracy. Based on precise decoupling compensation, the current vector angle control range is greatly expanded from the traditional MTPA curve to a wider working area, fully utilizing the potential of reluctance torque and improving the torque output capability and dynamic response performance of the motor under various working conditions.

[0045] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0046] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0047] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A motor optimization control method based on q-axis voltage control, characterized in that: include: Based on the stator voltage equation of the permanent magnet synchronous motor, Clark-Park transformation and sliding mode observation are performed on the three-phase current and three-phase voltage signals to obtain the d-axis current signal, q-axis current signal, d-axis flux component, and q-axis flux component. According to the d-axis flux component, q-axis flux component, d-axis current signal and q-axis current signal, the d-axis inductance value and the q-axis inductance value are subjected to recursive least squares identification processing to obtain the real-time saliency ratio value; Based on the real-time salient pole ratio value, the d-axis current signal and the q-axis current signal are subjected to a reluctance torque angle arc tangent calculation and a dq-axis coupling amount product operation to obtain a reluctance torque coupling interference amount; A compensation gain calculation is performed on the reluctance torque coupling interference to obtain a q-axis voltage feedforward compensation, and a feedforward superposition process is performed on the original q-axis voltage command to generate a q-axis voltage control command.

2. The motor optimization control method based on q-axis voltage control according to claim 1, characterized in that: The three-phase current signal and the three-phase voltage signal are subjected to Clark-Park transformation and sliding mode observation based on the stator voltage equation of the permanent magnet synchronous motor to obtain the d-axis current signal, the q-axis current signal, the d-axis flux component and the q-axis flux component, including: Perform Clark transform and Park transform on the three-phase current signal and three-phase voltage signal of the permanent magnet synchronous motor respectively to obtain the d-axis voltage signal, q-axis voltage signal, d-axis current signal and q-axis current signal; According to the stator voltage equation of the permanent magnet synchronous motor, the d-axis voltage signal and the q-axis voltage signal are differentially transformed to obtain a d-axis magnetic flux differential component and a q-axis magnetic flux differential component; Performing sliding mode observation and integration operations on the d-axis magnetic flux differential and the q-axis magnetic flux differential to obtain a d-axis magnetic flux observation value and a q-axis magnetic flux observation value; The d-axis flux observation value and the q-axis flux observation value are filtered to obtain a d-axis flux component and a q-axis flux component.

3. The motor optimization control method based on q-axis voltage control according to claim 1, characterized in that: The recursive least squares identification processing is performed on the d-axis inductance value and the q-axis inductance value according to the d-axis flux component, the q-axis flux component, the d-axis current signal, and the q-axis current signal to obtain a real-time saliency ratio value, including: performing a flux linkage inductance ratio calculation on the d-axis current signal according to the d-axis flux linkage component and the permanent magnet flux linkage to obtain a d-axis inductance identification value; performing a flux linkage inductance ratio operation on the q-axis current signal according to the q-axis flux linkage component to obtain a q-axis inductance identification value; performing recursive least squares calculation on the d-axis inductance identification value and the q-axis inductance identification value respectively to obtain a d-axis inductance value and a q-axis inductance value; A ratio operation is performed on the d-axis inductance value and the q-axis inductance value to obtain a real-time salient pole ratio value.

4. The motor optimization control method based on q-axis voltage control according to claim 3, characterized in that: The performing recursive least squares calculation on the d-axis inductance identification value and the q-axis inductance identification value respectively to obtain the d-axis inductance value and the q-axis inductance value includes: A sampling recursive least squares algorithm is used to construct a regression matrix and organize observation vectors for the d-axis inductance identification value and the q-axis inductance identification value, respectively, to obtain a d-axis regression data matrix and a q-axis regression data matrix; Performing gain vector calculation and covariance matrix update on the d-axis regression data matrix and the q-axis regression data matrix respectively to obtain a d-axis gain vector and a q-axis gain vector; Based on the d-axis gain vector and the q-axis gain vector, respectively updating the parameters of the inductance value at the previous moment to obtain the d-axis inductance intermediate value and the q-axis inductance intermediate value at the current moment; Convergence judgment and numerical stability processing are performed on the intermediate value of the d-axis inductance and the intermediate value of the q-axis inductance respectively to obtain a d-axis inductance value and a q-axis inductance value.

5. The motor optimization control method based on q-axis voltage control according to claim 1, characterized in that: The performing of a magnetic resistance torque angle arc tangent calculation and a dq axis coupling amount product operation on the d-axis current signal and the q-axis current signal based on the real-time salient pole ratio value to obtain a magnetic resistance torque coupling interference amount includes: Performing a difference calculation on the d-axis inductance value according to the real-time salient pole ratio value and the q-axis inductance value to obtain a dq-axis inductance difference; Based on the dq-axis inductance difference, performing a product operation on the d-axis current signal and the q-axis current signal to obtain a numerator term of the inverse tangent function of the reluctance torque angle; According to the permanent magnet flux and the q-axis current signal, a weighted sum is performed on the dq-axis inductance difference and the square term of the d-axis current signal to obtain the denominator term of the inverse tangent function of the reluctance torque angle; Performing magnetic resistance torque angle inverse tangent function calculation and angle range restriction on the numerator and denominator to obtain a magnetic resistance torque angle value; A dq-axis coupling product operation is performed on the rotor electrical angular velocity and the d-axis current signal according to the reluctance torque angle value to obtain the reluctance torque coupling interference amount.

6. The motor optimization control method based on q-axis voltage control according to claim 5, characterized in that: The calculation of the inverse tangent function of the reluctance torque angle and the angle range limitation of the numerator and denominator to obtain the reluctance torque angle value includes: Performing zero value detection and numerical protection on the denominator term to obtain a denominator protection value; Performing an inverse tangent function operation of a reluctance torque angle according to the numerator term and the denominator protection value to obtain a first reluctance torque angle value; performing angle range normalization and quadrant determination on the first reluctance torque angle value to obtain a second reluctance torque angle value; An angle continuity smoothing process is performed on the second reluctance torque angle value based on the reluctance torque angle at a previous moment to obtain a reluctance torque angle value.

7. The motor optimization control method based on q-axis voltage control according to claim 6, characterized in that: The step of performing a dq-axis coupling product operation on the rotor electrical angular velocity and the d-axis current signal according to the reluctance torque angle value to obtain the reluctance torque coupling interference amount includes: Performing low-pass filtering on the rotor electrical angular velocity to obtain filtered rotor electrical angular velocity; performing a cubic spline interpolation operation on a cross-inductance lookup table according to the d-axis current signal and the q-axis current signal to obtain a cross-inductance value; performing a product operation based on the filtered rotor electrical angular velocity, the d-axis inductance value, and the d-axis current signal, and performing a weighted summation in combination with the cross inductance value and the q-axis current signal to obtain a basic coupling amount; A cosine modulation operation is performed on the basic coupling amount according to the reluctance torque angle value to obtain the reluctance torque coupling interference amount.

8. The motor optimization control method based on q-axis voltage control according to claim 1, characterized in that: The method of performing compensation gain calculation on the reluctance torque coupling interference to obtain a q-axis voltage feedforward compensation, and performing feedforward superposition processing on the original q-axis voltage instruction to generate a q-axis voltage control instruction includes: Adaptively adjusting the proportional gain and the integral gain according to the real-time saliency ratio value to obtain adaptive compensator parameters; Inputting the reluctance torque coupling interference into a proportional-integral-differential compensator for adaptive gain adjustment to obtain a first compensation signal; performing speed adaptive gain attenuation and anti-integral windup processing on the first compensation signal based on the rotor electrical angular velocity to obtain a second compensation signal, and performing a change rate constraint on the second compensation signal to obtain a q-axis voltage feedforward compensation amount; The original q-axis voltage command is subjected to feedforward superposition processing according to the q-axis voltage feedforward compensation amount to generate a q-axis voltage control command.

9. The motor optimization control method based on q-axis voltage control according to claim 8, characterized in that: The performing feedforward superposition processing on the original q-axis voltage instruction according to the q-axis voltage feedforward compensation amount to generate a q-axis voltage control instruction includes: Performing frequency analysis and calculating compensation coefficients on the q-axis voltage feedforward compensation to obtain a frequency-domain weighted compensation; Calculating a speed feedforward term based on the rotor electrical angular velocity to obtain a speed feedforward compensation component; Based on the frequency domain weighted compensation amount and the speed feedforward compensation component, a superposition operation is performed on the original q-axis voltage command to obtain an initial voltage control command; A q-axis current tracking error root mean square value is calculated for the initial voltage control instruction to generate a q-axis voltage control instruction.

10. A motor optimization control system based on q-axis voltage control, characterized in that: The method for executing the motor optimization control method based on q-axis voltage control according to any one of claims 1 to 9 comprises: A transformation module is used to perform Clark-Park transformation and sliding mode observation on the three-phase current signal and the three-phase voltage signal based on the stator voltage equation of the permanent magnet synchronous motor to obtain a d-axis current signal, a q-axis current signal, a d-axis flux component, and a q-axis flux component; a processing module, configured to perform recursive least square identification processing on the d-axis inductance value and the q-axis inductance value according to the d-axis flux component, the q-axis flux component, the d-axis current signal, and the q-axis current signal, to obtain a real-time salient pole ratio value; a calculation module, configured to perform a reluctance torque angle arc tangent calculation and a dq-axis coupling amount product calculation on the d-axis current signal and the q-axis current signal based on the real-time salient pole ratio value to obtain a reluctance torque coupling interference amount; The compensation module is used to calculate the compensation gain of the magnetic resistance torque coupling interference to obtain the q-axis voltage feedforward compensation, and perform feedforward superposition processing on the original q-axis voltage instruction to generate a q-axis voltage control instruction.

Images