A high-order sliding mode control system and method for permanent magnet synchronous motor

Abstract

The application relates to a high-order sliding mode control system and method for a permanent magnet synchronous motor. The system comprises a data acquisition and processing module, a speed outer loop control module, an adaptive generalized super-twisting extended state observer module, a current inner loop control module and a motor drive control module. The speed outer loop control module takes a speed tracking error as a state variable, constructs a second-order integral non-singular terminal sliding mode surface and an improved composite exponential reaching law, and obtains a q-axis reference current; the adaptive generalized super-twisting extended state observer is constructed based on a lumped disturbance model, and obtains a lumped disturbance estimation value and an actual mechanical angular velocity estimation value; the current inner loop control module obtains a d-axis reference voltage and a q-axis reference voltage; and the motor drive control module generates a drive signal to control the operation of the permanent magnet synchronous motor. The application improves the dynamic response speed, steady-state control precision and anti-disturbance robustness of the speed loop, and reduces high-frequency chattering in the sliding mode control.

Description

Technical Field

[0001] This application relates to the technical field of vector control of permanent magnet synchronous motors, and in particular to a high-order sliding mode control system and method for permanent magnet synchronous motors. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) are characterized by high efficiency, high power density, low noise, and reliable structure, and are widely used in electric vehicles, aerospace, high-precision machine tools, and industrial servo systems. Because PMSM drive systems are multivariable, strongly coupled, nonlinear, and susceptible to parameter variations and external load disturbances, their speed control performance directly determines the dynamic response, steady-state accuracy, and disturbance rejection capabilities of the drive system.

[0003] Currently, the PI control method commonly used in engineering applications is simple in structure and easy to implement. However, it is essentially a linear control method, and its adaptability to motor parameter perturbations, sudden changes in load torque, inverter non-ideal factors, and external disturbances is limited. When the system operating conditions change, fixed PI parameters cannot simultaneously ensure speed, stability, and disturbance rejection, and are prone to problems such as increased overshoot, prolonged speed recovery time, and aggravated steady-state fluctuations.

[0004] To overcome the technical limitations of PI control, sliding mode control, with its advantages of strong robustness and flexible design, has become the mainstream improved algorithm for PMSM speed control. However, traditional sliding mode control has significant shortcomings. Relying on large switching gain to suppress disturbances easily leads to high-frequency chattering, causing current ripple, torque pulsation, and increased mechanical vibration. While reducing the gain can suppress chattering, it weakens the system's disturbance rejection stability. At the same time, existing terminal and fixed-time sliding mode control technologies generally suffer from zero-point singularity risks, convergence speed constrained by initial state, complex parameter tuning, and a prominent contradiction between convergence and chattering suppression.

[0005] Existing technologies often introduce disturbance observers and extended state observers to compensate for disturbances and reduce sliding mode gain dependence, but still have many shortcomings. Traditional disturbance observers rely on accurate motor models, and the estimation accuracy drops significantly under time-varying parameters and unknown disturbances; linear extended state observers have slow convergence speed and weak nonlinear disturbance estimation capabilities, while high-order and cascaded observers have complex structures, severe parameter coupling, high engineering implementation costs, and poor practicality. Summary of the Invention

[0006] To address the issues of chattering and insufficient disturbance estimation accuracy in existing sliding mode control methods, this invention proposes a high-order sliding mode control system and method for permanent magnet synchronous motors. By using an adaptive generalized super-twisted extended state observer to estimate the lumped disturbance of the motor in real time, the estimated lumped disturbance value is fed forward to compensate the high-order sliding mode control law, reducing the sliding mode controller's dependence on high switching gain. This reduces high-frequency chattering while ensuring strong robustness and fast convergence performance.

[0007] To achieve the above-mentioned technical effects, the technical solution of the present invention is as follows: A high-order sliding mode control system for a permanent magnet synchronous motor (PMSM) is provided. The control system employs a cascaded double-loop structure composed of a speed outer loop and a current inner loop. In this structure, the speed outer loop and the current inner loop are connected in series. The data acquisition and processing module is used to convert the three-phase stator current of the permanent magnet synchronous motor into d-axis current and q-axis current, and calculate the speed tracking error in the outer speed loop based on the given mechanical angular velocity and the actual mechanical angular velocity of the permanent magnet synchronous motor. The speed outer loop control module is used to construct a second-order integral non-singular terminal sliding surface and an improved composite exponential reaching law using the speed tracking error as the state variable. Based on the speed tracking error, the parameters of the permanent magnet synchronous motor, the second-order integral non-singular terminal sliding surface, and the improved composite exponential reaching law, it obtains the q-axis reference current. The improved composite exponential reaching law includes a speed-changing term related to the second-order integral non-singular terminal sliding surface, a variable exponential term related to the speed tracking error, and a switching function term. An adaptive generalized super-twisted extended state observation module is provided, comprising an adaptive generalized super-twisted extended state observer. The observer is constructed based on a lumped disturbance model, which is established based on the q-axis current and the actual mechanical angular velocity. The module estimates the lumped disturbance and the actual mechanical angular velocity, obtaining estimated values ​​for both. These estimated values ​​are then fed forward to the outer velocity control module to compensate for the disturbance. The current inner loop control module is used to compare the d-axis reference current and q-axis reference current with the d-axis current and q-axis current respectively to obtain the current error signal input to the current inner loop, and obtain the d-axis reference voltage and q-axis reference voltage through the current inner loop; the d-axis reference current is 0; The motor drive control module is used to generate a drive signal based on the d-axis reference voltage and the q-axis reference voltage, and the drive signal is used to control the operation of the permanent magnet synchronous motor.

[0008] Preferably, the data acquisition and processing module acquires the actual mechanical angular velocity during the operation of the permanent magnet synchronous motor. and three-phase stator current , , By using Clarke and Park transformations, the three-phase stator currents of a permanent magnet synchronous motor in a three-phase stationary coordinate system are converted into d-axis currents in a two-phase rotating coordinate system. With q-axis current ; the given mechanical angular velocity Compared with actual mechanical angular velocity By comparison, the speed tracking error is obtained, and the expression for the speed tracking error is: .

[0009] Preferably, in the speed outer loop control module, the process of constructing the second-order integral nonsingular terminal sliding surface is as follows: Using the speed tracking error as a state variable, a first-order integral sliding surface is constructed; The expression for the first-order integral sliding surface is: ,in, Denotes the first-order integral sliding surface. Represents state variables, Indicates the weighting coefficient. Indicates time, Indicates a lower-order power exponent. Indicates a higher-order power exponent; Based on the first-order integral sliding surface, a second-order integral non-singular terminal sliding surface is constructed. The expression for the second-order integral non-singular terminal sliding surface is: ,in, This indicates that the sliding surface of the second-order integral is a non-singular terminal. The derivative of the first-order integral sliding surface is given by . This represents the sliding surface gain coefficient. 、 Indicates a variable power exponent. , , , , , , This represents the adjustment parameter for the variable power exponent.

[0010] Preferably, in the speed outer loop control module, the expression for the improved composite exponential reaching law is: in, Indicates the variable speed term. Indicates a variable exponent term; Indicates the switch adjustment coefficient. Represents the switching function term. Represents a symbolic function. Represents the hyperbolic tangent function. The boundary layer thickness is indicated; the boundary layer represents a continuous transition interval set on both sides of the sliding surface, and the boundary layer thickness parameter is a preset positive constant.

[0011] Preferably, the process of obtaining the q-axis reference current satisfies the expression: ,in, This represents the estimated value of the lumped disturbance. Represents the torque constant. Indicates the moment of inertia. This represents the gain of the improved composite exponential reaching law.

[0012] Preferably, the expression for the lumped disturbance model is: ,in, Indicates lumped disturbance; Represents mechanical torque. This represents the viscous damping coefficient.

[0013] Preferably, an extended state variable D is introduced. , The derivative of the extended state variable is represented by the expression for the adaptive generalized super-twisted extended state observer: ,in, This indicates the error in estimating the mechanical angular velocity. This represents the lumped disturbance estimation error, used to estimate the lumped disturbance and the actual mechanical angular velocity. and They represent and The estimated value, , This represents the generalized super-torsion nonlinear correction term. , Indicates observation gain; The expression for the generalized super-torsion nonlinear correction term is: in, This represents the first nonlinear correction term, used to perform nonlinear correction on the mechanical angular velocity estimation error; This represents the second nonlinear correction term, used to perform nonlinear correction on the lumped disturbance estimation error; This indicates an adjustable gain parameter.

[0014] Preferably, the expression for the adjustable gain parameter is: ,in, This indicates the given speed of the permanent magnet synchronous motor. This represents the observation rotation speed of the adaptive generalized super-twisted extended state observer. This indicates that a threshold value is set. Indicates the original gain. This indicates the maximum gain.

[0015] Preferably, in the motor drive control module, a drive signal is generated based on the d-axis reference voltage and the q-axis reference voltage using inverse Park transformation and SVPWM modulation. The drive signal serves as the switching signal for the inverter connected to the permanent magnet synchronous motor.

[0016] This invention also proposes a high-order sliding mode control method for permanent magnet synchronous motors. The method is based on the aforementioned system and includes the following steps: The three-phase stator current of the permanent magnet synchronous motor is converted into d-axis current and q-axis current. Based on the given mechanical angular velocity and the actual mechanical angular velocity of the permanent magnet synchronous motor, the speed tracking error is calculated in the outer speed loop. Using the speed tracking error as a state variable, a second-order integral non-singular terminal sliding surface and an improved composite exponential reaching law are constructed. Based on the speed tracking error, the parameters of the permanent magnet synchronous motor, the second-order integral non-singular terminal sliding surface, and the improved composite exponential reaching law, the q-axis reference current is obtained. The improved composite exponential reaching law includes a speed-changing term related to the second-order integral non-singular terminal sliding surface, a variable exponential term related to the speed tracking error, and a switching function term. A lumped disturbance model is established based on the q-axis current and the actual mechanical angular velocity. An adaptive generalized super-twisted extended state observer is constructed based on the lumped disturbance model. The lumped disturbance and the actual mechanical angular velocity are estimated using the adaptive generalized super-twisted extended state observer to obtain the lumped disturbance estimate and the actual mechanical angular velocity estimate. The lumped disturbance estimate is then fed forward to compensate for the process of obtaining the q-axis reference current. The d-axis reference current and q-axis reference current are compared with the d-axis current and q-axis current respectively to obtain the current error signal input to the inner current loop. The d-axis reference voltage and q-axis reference voltage are obtained through the inner current loop; the d-axis reference current is 0. Based on the d-axis reference voltage and the q-axis reference voltage, a drive signal is generated, which is used to control the operation of the permanent magnet synchronous motor.

[0017] Compared with the prior art, the beneficial effects of the technical solution of the present invention are: This invention proposes a high-order sliding mode control system and method for permanent magnet synchronous motors. By combining a second-order integral nonsingular terminal sliding mode surface with an improved composite exponential reaching law, a fast and smooth approach process is achieved in the outer speed loop, effectively suppressing sliding mode chattering and improving the system's robustness and responsiveness to load abrupt changes and parameter fluctuations. An adaptive generalized super-twisted extended state observer designed based on a lumped disturbance model can estimate the actual mechanical angular velocity and lumped disturbances, and feedforward the disturbance estimate to the q-axis reference current generation process, enhancing the system's anti-interference capability and adaptability. Combined with d-axis zero-current vector control and precise tracking of the inner current loop, the control accuracy and steady-state performance of speed and torque are further improved. This invention has a clear control hierarchy, achieving rapid convergence of motor speed errors, real-time estimation and compensation for load disturbances and parameter changes, and effective suppression of sliding mode chattering. Attached Figure Description

[0018] Figure 1 A block diagram illustrating the high-order sliding mode control system for permanent magnet synchronous motors proposed in this embodiment of the invention; Figure 2 This diagram illustrates the control block diagram of the high-order sliding mode control system for permanent magnet synchronous motors proposed in this embodiment of the invention. Figure 3 A comparison chart showing the output characteristics of the three types of switching functions proposed in the embodiments of the present invention; Figure 4 This diagram illustrates the structure of the speed outer loop module proposed in this embodiment of the invention. Figure 5 This diagram illustrates the structure of AGSTESO as proposed in this embodiment of the invention. Figure 6 This diagram illustrates the control block diagram of the experimental equipment proposed in this embodiment of the invention. Figure 7 This diagram illustrates the application environment of the experimental equipment proposed in the embodiments of the present invention. Figure 8 This is a comparison chart of speed tracking curves for a sudden speed increase from 500 rpm to 1200 rpm under half-load conditions, as proposed in the embodiments of the present invention. Figure 9 This is a comparison of the current response curves under half-load conditions when the speed is suddenly increased from 500 rpm to 1200 rpm in the embodiments of the present invention. Figure 10 This is a comparison chart of speed tracking curves for a sudden drop from 1200 rpm to 800 rpm under half-load conditions, as presented in the embodiments of the present invention. Figure 11 This is a comparison chart of the current response curves for a sudden speed drop from 1200 rpm to 800 rpm under half-load conditions as proposed in the embodiments of the present invention. Figure 12 This diagram shows a comparison of the response of the motor proposed in this invention under a sudden half-load condition during steady-state operation at 600 r / min. Figure 13 This diagram shows a comparison of the response of the motor in the embodiment of the present invention under a sudden half-load condition during steady-state operation at 600 r / min. Figure 14 This figure shows the speed tracking curve of the PI controller under half-load conditions with a DC biased sinusoidal signal as proposed in the embodiments of the present invention. Figure 15 This diagram shows the speed tracking curve of the TISMC controller under half-load conditions according to the embodiment of the present invention for a sinusoidal signal with DC bias. Figure 16 This is a graph showing the speed tracking curve of the ASOSMC-GSTEB0 controller with DC bias sinusoidal signal under half-load conditions proposed in this embodiment of the invention. Figure 17 The graph shows the speed tracking error curves of the three controllers under half-load conditions proposed in the embodiments of the present invention for the DC biased sinusoidal signal. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0020] Example 1 like Figure 1 As shown, this embodiment proposes a high-order sliding mode control system for permanent magnet synchronous motors. The high-order sliding mode control system for permanent magnet synchronous motors adopts a system based on an outer speed loop and an inner current loop. Figure 2 The cascaded double-loop structure shown is based on a speed outer loop and a current inner loop. The speed outer loop and the current inner loop are connected in series. It includes: a data acquisition and processing module, a speed outer loop control module, an adaptive generalized super-twist extended state observation module, a current inner loop control module, and a motor drive control module.

[0021] The data acquisition and processing module is used to convert the three-phase stator current of the permanent magnet synchronous motor into d-axis current and q-axis current, and calculate the speed tracking error in the outer speed loop based on the given mechanical angular velocity and the actual mechanical angular velocity of the permanent magnet synchronous motor.

[0022] The speed outer loop control module is used to construct a second-order integral non-singular terminal sliding surface and an improved composite exponential reaching law, with the speed tracking error as the state variable. Based on the speed tracking error, the parameters of the permanent magnet synchronous motor, the second-order integral non-singular terminal sliding surface, and the improved composite exponential reaching law, the q-axis reference current is obtained. The improved composite exponential reaching law includes a speed change term related to the second-order integral non-singular terminal sliding surface, a variable exponential term related to the speed tracking error, and a switching function term.

[0023] The adaptive generalized super-twisted extended state observation module includes an adaptive generalized super-twisted extended state observer, which is constructed based on a lumped disturbance model. The lumped disturbance model is established based on the q-axis current and the actual mechanical angular velocity. The adaptive generalized super-twisted extended state observation module is used to estimate the lumped disturbance and the actual mechanical angular velocity, obtain the estimated values ​​of the lumped disturbance and the actual mechanical angular velocity, and feed forward the estimated values ​​of the lumped disturbance to the speed outer loop control module to perform feedforward cancellation of the disturbance.

[0024] The inner current loop control module compares the d-axis and q-axis reference currents with the d-axis and q-axis currents respectively to obtain the current error signal input to the inner current loop. The inner current loop then generates the d-axis and q-axis reference voltages; the d-axis reference current is 0. The motor drive control module generates drive signals based on the d-axis and q-axis reference voltages. These drive signals are used to control the operation of the permanent magnet synchronous motor.

[0025] In this embodiment, the data acquisition and processing module includes Figure 2 Analog-to-digital sampling (A / D Sampling) and Clarke transform ( Park transformation (current-related) The system includes a position and speed detection unit. ASOSMC represents the speed outer loop module. AGSTESO represents the adaptive generalized super-twisted extended state observation module. The current inner loop module contains two proportional-integral (PI) controllers. The motor drive control module includes an inverse Park transform (voltage-dependent). ), Space Vector Pulse Width Modulation (SVPWM) and Voltage Source Inverter.

[0026] See Figure 2 At the entrance of the outer speed loop module, speed error is generated, based on the given mechanical angular velocity. As the command input, it indicates the target speed of the permanent magnet synchronous motor. Simultaneously, the position and speed detection unit acquires the motor rotor position signal via an encoder or resolver, and calculates the rotor electrical angle. Compared with actual mechanical angular velocity ;wherein the rotor electrical angle Provides a rotational reference for the Park transform and inverse Park transform, based on the actual mechanical angular velocity. The data is uploaded along the feedback path to the subtraction node of the speed outer loop module.

[0027] At the subtraction node and Difference to generate speed tracking error This error is the sole input to the entire speed outer loop controller and serves as the common starting point for subsequent construction of the second-order integral nonsingular terminal sliding surface, adjustment of the improved composite exponential reaching law (ICERL reaching law), and the final control law solution. If e > 0, it means the actual speed is lower than the given value, and the motor needs to accelerate; if e < 0, it needs to decelerate.

[0028] Simultaneously with the generation of speed errors, coordinate transformation is performed. An analog-to-digital (A / D) sampling unit acquires the three-phase stator current of the permanent magnet synchronous motor. , , These three are AC quantities in the natural coordinate system and cannot be directly used for vector control. First, through Clarke transformation, the three-phase stationary coordinate system abc is mapped to the two-phase stationary coordinate system α-β, resulting in... Then, after Park transformation, the rotor position angle provided by the sensor is used. As the orientation reference of the rotating coordinate system, the α-β stationary coordinate system is rotated to the dq rotating coordinate system synchronized with the rotor magnetic field, and the d-axis current is fed back. and q-axis current feedback At this point, the three-phase alternating current has been converted into two direct currents. and These two feedback currents will be compared with their reference values ​​in the subsequent current inner loop.

[0029] In the outer velocity loop module ASOSMC, the velocity error e enters the adaptive second-order sliding mode controller. Inside this module: firstly, a first-order integral sliding surface is constructed with e as the state variable. Based on this, a second-order sliding surface s is further constructed, and two sets of power terms, one low-order and one high-order, are introduced. The low-order terms dominate convergence when e is close to zero (near the equilibrium point), while the high-order terms accelerate the pull back when e is large (far from the equilibrium point). The variable power terms enable the convergence rate to be adaptively adjusted according to the system state.

[0030] Secondly, ASOSMC internally implements the ICERL improved composite exponential reaching law, which determines the variation law based on the current value and sign of the sliding surface s. When |s| is large (the system is far from the sliding surface), the variable exponential term dominates, accelerating the convergence at a rate higher than that of the traditional exponential reaching law; when |s| approaches zero, the variable exponential term naturally decays, and the variable speed term takes over, smoothly converging to the sliding surface with an adaptively reduced gain; when the system enters the boundary layer, ... Replace discontinuous Eliminate the switching transition.

[0031] In the Adaptive Generalized Super-Twist Extended State Observation Module (AGSTESO), the Adaptive Generalized Super-Twist Extended State Observer receives two inputs: the measured mechanical angular velocity. and current control quantity (Output from the previous control cycle), two estimates are output in real time: speed estimate. and lumped disturbance estimate .

[0032] AGSTESO is based on a lumped perturbation model, extending the lumped perturbation into a second state variable. When the estimation error is large, the adaptive gain automatically switches to a large gain, while the super-distorted nonlinear term provides a convergence rate higher than linear, making... It rapidly approximates the true perturbation value; when the estimation error shrinks to within the threshold, it switches back to a small gain, the nonlinear term degenerates into linearity, the accuracy of steady-state observation is maintained without amplifying high-frequency noise, and the output lumped perturbation estimate is... It will enter the speed controller in the next step as a feedforward compensation quantity. In the control quantity, it is equal to the lumped disturbance with the opposite sign, and physically realizes the purpose of using the observed disturbance quantity to cancel the actual disturbance.

[0033] In the inner current loop, the q-axis reference current output by the outer velocity loop. It crosses the inter-loop boundary and enters the inner current loop. Meanwhile, the d-axis reference current follows... The control strategy and The actual values ​​fed back from the coordinate transformation branches respectively and The comparison at the subtraction node generates d-axis current error and q-axis current error. These two error signals are fed into the inner current loop PI controller (two independent PI regulators), and after proportional-integral operation, the d-axis reference voltage is output. and q-axis reference voltage .

[0034] Current inner loop output and It is still a DC voltage quantity in a rotating coordinate system. To drive the actual PMSM stator windings, it needs to be converted back to an AC voltage signal in a three-phase stationary coordinate system. First, it undergoes an inverse Park transformation, utilizing the rotor position angle provided by the sensor. Mapping the dq rotating coordinate system back to the α-β two-phase stationary coordinate system, we obtain... and Subsequently, space vector pulse width modulation is based on The duty cycles of the six PWM switching signals are calculated to control the on / off timing of the six power switching devices (IGBT / MOSFET) in the three-phase inverter. The inverter chops and modulates the DC bus voltage into a three-phase AC voltage with variable amplitude and frequency, which is applied to the stator windings of the PMSM to generate a rotating magnetic field that drives the rotor to rotate.

[0035] After the PMSM rotates, the sensor mounted on the rotor shaft re-collects data. and Three-phase current sensor data acquisition , , . Returning to the outer velocity loop subtraction node along the feedback path, and... The comparison generates a new speed error e; the three-phase current returns to the inner current loop subtraction node via Clarke / Park transformation, and is compared with the new... Comparison. Once one control cycle is completed, the next cycle begins immediately.

[0036] Example 2 This embodiment improves upon Embodiment 1 by providing a high-order sliding mode control system for a permanent magnet synchronous motor. The data acquisition and processing module collects the actual mechanical angular velocity during the operation of the permanent magnet synchronous motor. Three-phase stator current , , Including DC bus voltage and other status information, the three-phase stator current of the permanent magnet synchronous motor in the three-phase stationary coordinate system is converted into the d-axis current in the two-phase rotating coordinate system through Clarke and Park transformations. With q-axis current ; the given mechanical angular velocity Compared with actual mechanical angular velocity By comparison, the speed tracking error is obtained, and the expression for the speed tracking error is: .

[0037] In this embodiment, the speed tracking error It serves as the basic input variable for subsequent second-order integral nonsingular terminal sliding surfaces, improved composite exponential reaching laws, and generalized super-twisted extended state observers.

[0038] In an optional embodiment, the process of constructing the second-order integral nonsingular terminal sliding surface in the speed outer loop control module is as follows: A first-order integral sliding surface is constructed using the velocity tracking error as the state variable; The expression for the first-order integral sliding surface is: ,in, Denotes the first-order integral sliding surface. Represents state variables, Indicates the weighting coefficient. Indicates time, Indicates a lower-order power exponent. Indicates a higher-order power exponent; Based on the first-order integral sliding surface, a second-order integral nonsingular terminal sliding surface is constructed. The expression for the non-singular terminal sliding surface of the second-order integral is: ,in, This indicates that the sliding surface of the second-order integral is a non-singular terminal. The derivative of the first-order integral sliding surface is given by . This represents the sliding surface gain coefficient. 、 Indicates a variable power exponent. , , , , , , This represents the adjustment parameter for the variable power exponent.

[0039] In this embodiment, to improve the dynamic convergence performance of the PMSM velocity error and avoid the singularity problem that may occur in traditional terminal sliding mode control, this embodiment constructs a second-order integral non-singular terminal sliding surface based on a first-order integral sliding surface. In the expression for the first-order integral sliding surface, , , , and It is a very small constant. In the expression for the second-order integral nonsingular terminal sliding surface, lower-order power terms are used to improve the convergence speed of the system near the equilibrium point, while higher-order power terms are used to improve the convergence speed when the system is far from the equilibrium point. The integral term is used to reduce steady-state error and enhance low-frequency disturbance suppression capability. Simultaneously, the use of the integral function solves the singularity problem of traditional terminal sliding surfaces. Furthermore, the second-order integral nonsingular terminal sliding surface possesses fast convergence and fixed-time stability, and the variable power term can adaptively adjust the approach speed according to the state between the system and the equilibrium point. Through the above design, the constructed sliding surface combines fast convergence, nonsingularity, and steady-state error suppression capability.

[0040] In an optional embodiment, in the speed outer loop control module, the expression for the improved composite exponential reaching law is: in, Indicates the variable speed term. Indicates a variable exponent term; Indicates the switch adjustment coefficient. Represents the switching function term. Represents a symbolic function. Represents the hyperbolic tangent function. This indicates the boundary layer thickness; the boundary layer represents a continuous transition zone set on both sides of the sliding surface, and the boundary layer thickness parameter is a preset positive constant.

[0041] In this embodiment, the improved composite exponential reaching law (novel ICERL) introduces a power function related to the system state variables, and sets the exponential term in the power function as an adjustable parameter, establishing the relationship between the reaching velocity and the system state evolution during the system transition phase. An adaptive gain function is introduced into the variable velocity term to dynamically adjust the reaching velocity, while employing... Replace discontinuous sign functions . Figure 3 Showing , and the methods used in this embodiment The comparison chart of the output characteristics of the three types of switching functions shows that... A better balance is achieved between smooth switching performance and approaching dynamic performance. When the system state trajectory deviates from the sliding surface, ICERL approaches the sliding surface at two rates: variable speed and variable exponential. The second term plays a dominant role, and the improved variable exponential gain... Greater than the gain of the traditional exponential term The convergence speed is faster. When the system state's trajectory approaches the sliding surface, the exponential term gradually approaches 0, and the velocity term plays a dominant role, ensuring a certain convergence speed while reducing sliding mode chattering. When the system state enters the boundary layer, the sign function... Hyperbolic tangent function The replacement effectively reduced the chattering.

[0042] The structural diagram of the speed outer loop module is as follows: Figure 3As shown, the overall design adopts a nested second-order sliding mode architecture, integrating a second-order integral non-singular terminal sliding mode surface, an improved composite exponential reaching law (ICERL), and a lumped disturbance feedforward compensation structure. Given the mechanical angular velocity and the actual mechanical angular velocity, the two are subtracted to obtain the speed tracking error x, which serves as the core state variable of the sliding mode control. The speed tracking error x is fed into three parallel nonlinear computation units, which sequentially calculate the signed low-order power terms, high-order power terms, and linear terms. The results of the three calculations are aggregated by an adder, and the amplitude is adjusted by the integral gain coefficient k before being fed into the integration stage to complete the integration calculation. The integral output and the original speed error x together constitute the first-order integral sliding mode surface variable. This structure takes into account the convergence speed of near and far equilibrium points through multi-order power terms, eliminates steady-state errors through integral terms, and avoids the singularity problem of traditional terminal sliding mode. The first-order sliding surface variable is output in two paths: one path yields its first derivative through a differential operation (s); the other path feeds into two nonlinear operation units, which calculate two variable power terms. The first derivative term and the two variable power terms are then summed in an adder to obtain the second-order sliding surface variable. The adaptive variable power term automatically switches according to the sliding state amplitude, giving the system a fixed-time convergence characteristic, with the upper limit of convergence time independent of the initial error magnitude. The second-order sliding surface variable and the velocity tracking error x are input together to the improved composite exponential reaching law unit, whose internal structure is as follows: Figure 3 As shown in the enlarged dashed box at the bottom center, the circuit contains two parallel approaching branches. The first is a variable-speed switching branch, which takes the second-order sliding mode variable as input. It first undergoes an adaptive gain term to dynamically adjust the gain according to the error amplitude, then completes the switching operation through a piecewise switching function, and finally multiplies by the base gain to output the variable-speed approaching component. The second is a variable-exponential state branch, which takes the speed tracking error x as input. After exponential operation, it is multiplied by the second-order sliding mode variable and then by the gain coefficient, outputting a variable-exponential approaching component related to the system state. The results of the two branches are summed by an adder to obtain the approaching rate of the second-order sliding surface, which is then output to the subsequent integration stage. This design achieves an adaptive effect of "high-speed approaching when far from the sliding surface and smooth convergence when close to the sliding surface," while suppressing sliding mode chattering through a piecewise smoothing switching function. The control quantity synthesis and the approach rate of the output approach law are calculated by the integrator and together with the switching term of the second-order sliding surface to form the basic sliding mode control quantity. At the control quantity synthesis node, a lumped disturbance estimate is introduced for feedforward compensation to offset the effects of load disturbances and parameter perturbations in the system, reducing the sliding mode controller's dependence on high switching gain. The synthesized total control quantity is converted from the mechanical domain control quantity to the current domain command through dimension conversion gain, and finally outputs the q-axis reference current as the given input for the inner current loop.

[0043] In an optional embodiment, the process of obtaining the q-axis reference current satisfies the expression: ,in, This represents the estimated value of the lumped disturbance. Represents the torque constant. Indicates the moment of inertia. This represents the gain of the improved composite exponential reaching law.

[0044] In this embodiment, a lumped disturbance estimate is introduced. By using feedforward cancellation to compensate for system disturbances, the sliding mode controller no longer relies entirely on a large switching gain to overcome external disturbances and parameter uncertainties. Therefore, this system can reduce the high-frequency switching component in the control input while maintaining robustness, thereby reducing current ripple and speed chattering.

[0045] In an optional embodiment, the expression for the lumped disturbance model is: ,in, Indicates lumped disturbance; Represents mechanical torque. This represents the viscous damping coefficient.

[0046] In this embodiment, the voltage equation, electromagnetic torque equation, and dynamic equation of the permanent magnet synchronous motor in the dq coordinate system are established. Uncertainties existing in actual operation are uniformly equivalent to lumped disturbances, and the expression for the lumped disturbance model is rewritten. The specific process is as follows: The voltage equation for a surface-mounted PMSM in the dq coordinate system is expressed as follows: ,in, , , and These are the stator voltage and current along the d and q axes, respectively; and These are the stator inductances along the d and q axes, respectively. Stator resistance; For permanent magnet flux linkage; Let be the electric angular velocity, satisfying = ,in and These are the number of pole pairs of the motor and the mechanical angular velocity, respectively.

[0047] The expression for the dynamic equation of the surface-mounted PMSM is: ,in, It is the moment of inertia; It is mechanical torque; It is electromagnetic torque; It is the viscous damping coefficient. Considering the surface-mounted PMSM... The electromagnetic torque equation can be simplified to: The motor dynamics equations are reformulated as: ,in, The torque constant is Furthermore, considering factors such as sudden changes in load torque, rotational inertia perturbation, changes in friction coefficient, changes in flux linkage parameters, and unmodeled dynamics in actual operation, these uncertainties are uniformly equivalent to a lumped disturbance. The motor dynamics equations can be further rewritten as follows: ,in, This represents the reference current along the q-axis. For lumped disturbances, the expression for a lumped disturbance is: .

[0048] By addressing the uncertainties in the actual operation of permanent magnet synchronous motors, complex uncertainty problems can be transformed into total disturbance estimation and compensation problems, providing a foundation for the subsequent AGSTESO design.

[0049] In an alternative embodiment, an extended state variable D is introduced. , The derivative of the extended state variable is represented by the expression for the adaptive generalized super-twisted extended state observer: ,in, This indicates the error in estimating the mechanical angular velocity. This represents the lumped disturbance estimation error, used to estimate the lumped disturbance and the actual mechanical angular velocity. and They represent and The estimated value, , This represents the generalized super-torsion nonlinear correction term. , Indicates observation gain; The expression for the generalized super-torsion nonlinear correction term is: in, This represents the first nonlinear correction term, used to perform nonlinear correction on the mechanical angular velocity estimation error; This represents the second nonlinear correction term, used to perform nonlinear correction on the lumped disturbance estimation error; This indicates an adjustable gain parameter.

[0050] The expression for the adjustable gain parameter is: ,in, This indicates the given speed of the permanent magnet synchronous motor. This represents the observation rotation speed of the adaptive generalized super-twisted extended state observer. This indicates that a threshold value is set. Indicates the original gain. This indicates the maximum gain.

[0051] In this embodiment, and It is the observation gain, which is usually determined by the bandwidth polynomial. , , Observation bandwidth. Adjustable gain parameter. Based on the expression for the adaptive generalized super-twisted extended state observer, the estimation error system of AGSTESO can be represented by the following definition: .

[0052] AGSTESO introduces a nonlinear term based on traditional ESO. When the error is large, the nonlinear term can provide a faster convergence speed. This can be achieved by adjusting the parameters of the nonlinear term. This ensures that the error converges to zero within a finite time. To further improve the observation performance of the observer under dynamic conditions, this embodiment modifies the original fixed parameters. Designed as a state-dependent adjustable gain parameter. The threshold value is an adjustable parameter. .parameter The change in bandwidth can be equivalent to the change in observer bandwidth gain, thus ensuring the dynamic and steady-state observation performance of the observer. When the observation error is large, at this time... With a large gain, the effect of the nonlinear term is enhanced, accelerating the convergence of observation errors; when the difference between the given rotational speed and the observed rotational speed is less than hour, Switching back to low gain maintains the observer's original steady-state performance, avoiding the chattering amplification and increased steady-state observation error caused by excessive gain. The block diagram of AGSTESO is shown below. Figure 4As shown, the actual mechanical angular velocity of the motor and the q-axis reference current are used as inputs to output real-time estimates of the mechanical angular velocity and lumped disturbance. The difference between the estimated mechanical angular velocity output from the observer and the actual mechanical angular velocity of the motor is calculated to obtain the speed observation error. This error serves as the common input to the nonlinear correction unit, providing an error benchmark for the state correction of the entire observer. Using the speed observation error as a unified input, two parallel nonlinear correction branches are included, corresponding to two sets of generalized super-twisted nonlinear functions, replacing the linear correction terms of the traditional linear extended state observer and achieving finite-time convergence of the observation error. The first correction branch: The observation error is fed into two signed fractional power arithmetic modules, which sequentially calculate the 1 / 2 power term and the 2 / 3 power term. The 2 / 3 power term is weighted by adjustable parameters and then summed with the 1 / 2 power term in an adder to obtain the first nonlinear correction function. After the correction function is output, its amplitude is adjusted by the first observation gain and then output to the angular velocity state update stage. The second correction branch: The observation errors are fed into three signed fractional power arithmetic modules, which sequentially calculate the 0th power, 1 / 6th power, and 1 / 3th power terms. These three power terms are weighted by their corresponding coefficients and then summed in an adder to obtain the second nonlinear correction function. The output of the correction function is then amplitude-adjusted by the second observation gain. The input signal is integrated to obtain a lumped disturbance estimate. This estimate serves as the core disturbance output of the observer, feedforward to the speed loop sliding mode controller to offset system uncertainties, and is also fed back to the angular velocity state update stage to participate in the prediction and correction of the rotational speed state. The adder node receives three input signals: one is the model prediction term obtained by normalizing the q-axis reference current through torque coefficients and moment of inertia; the second is the lumped disturbance estimate; and the third is the nonlinear correction quantity output from the first correction branch. The sum of these three signals yields the rate of change of the angular velocity estimate, which is then integrated to output the final mechanical angular velocity estimate. One path of this angular velocity estimate serves as the rotational speed output of the observer, while the other path is used for error generation, forming a closed-loop observation circuit.

[0053] In one optional embodiment, in the motor drive control module, a drive signal is generated based on the d-axis reference voltage and the q-axis reference voltage using inverse Park transformation and SVPWM modulation. The drive signal serves as the switching signal for the inverter connected to the permanent magnet synchronous motor.

[0054] In this embodiment, the speed loop controller outputs a q-axis reference current. In surface-mounted PMSM control, the following can be used: Control strategies to improve the simplicity of the control structure. and Relative to the actual dq axis currents , The comparison yields the current error signal, which is then input into the inner current loop controller to generate the dq-axis reference voltage. , Subsequently, the inverter switching signal is generated through the inverse Park transform and SVPWM modulation module to achieve drive control of the PMSM. Thus, this system forms a complete closed-loop control structure consisting of "speed error calculation - second-order sliding mode control - disturbance observation compensation - current loop adjustment - SVPWM drive".

[0055] Example 3 This embodiment proposes a high-order sliding mode control method for permanent magnet synchronous motors. The method is based on the system proposed in Embodiment 1 or Embodiment 2 and includes the following steps: The three-phase stator current of the permanent magnet synchronous motor is converted into d-axis current and q-axis current. Based on the given mechanical angular velocity and the actual mechanical angular velocity of the permanent magnet synchronous motor, the speed tracking error is calculated in the outer speed loop. Using speed tracking error as the state variable, a second-order integral non-singular terminal sliding surface and an improved composite exponential reaching law are constructed. Based on the speed tracking error, the parameters of the permanent magnet synchronous motor, the second-order integral non-singular terminal sliding surface, and the improved composite exponential reaching law, the q-axis reference current is obtained. The improved composite exponential reaching law includes a speed-changing term related to the second-order integral non-singular terminal sliding surface, a speed-changing exponential term related to the speed tracking error, and a switching function term. Based on the q-axis current and the actual mechanical angular velocity, a lumped disturbance model is established; based on the lumped disturbance model, an adaptive generalized super-twisted extended state observer is constructed; the lumped disturbance and the actual mechanical angular velocity are estimated using the adaptive generalized super-twisted extended state observer to obtain the lumped disturbance estimate and the actual mechanical angular velocity estimate, and the lumped disturbance estimate is fed forward to compensate for the process of obtaining the q-axis reference current; The d-axis reference current and q-axis reference current are compared with the d-axis current and q-axis current respectively to obtain the current error signal input to the inner current loop. The d-axis reference voltage and q-axis reference voltage are obtained through the inner current loop; the d-axis reference current is 0. Based on the d-axis reference voltage and the q-axis reference voltage, a drive signal is generated, which is used to control the operation of the permanent magnet synchronous motor.

[0056] To demonstrate the feasibility and effectiveness of the proposed control method, comparative experiments were conducted on a surface-mounted PMSM motor control experimental platform. The configuration of the experimental platform is as follows: Figure 6 and Figure 7 As shown in Table 1, the main control unit is a dSPACE DS1103 real-time control system, integrating Control Desk host computer software for data visualization, acquisition, and storage. The drive unit uses a three-phase voltage source inverter with 5μs hardware dead-time protection. The controller sampling frequency is 10kHz. Motor parameters are shown in Table 1. Magnetic field orientation control; both d-axis and q-axis current loops use PI controllers, with control parameters uniformly selected as follows: , .

[0057] Table 1 Parameters of Surface-Mounted PMSM Motors

[0058] The parameter selection criteria for the proposed control method are as follows. Parameters , , , Increasing the value can speed up convergence, but setting it too high may cause control input saturation. Additionally, the parameter... , , The convergence speed is directly related to the chattering suppression effect, therefore, it is necessary to coordinate the relationship between chattering suppression and convergence speed. Because It is the thickness of the boundary layer. A higher value means better chatter suppression; however, Excessively high values ​​will reduce dynamic performance and system robustness. At the same time, larger values... Values ​​that can accelerate convergence and reduce tracking error should be noted, however, overshoot issues should be considered. After evaluating these factors, the parameters listed in Table 2 are selected to achieve optimal experimental performance.

[0059] Table 2 Parameters of High-Order Sliding Mode Control Method for Permanent Magnet Synchronous Motors

[0060] Meanwhile, in order to better compare the performance of the proposed sliding mode speed controller, three speed controllers were used for comparison in the experiment: PI controller, traditional integral sliding mode controller (TISMC), and ASOSMC proposed in this embodiment.

[0061] The expression for a traditional integral sliding mode controller is: The control parameters selected for TISMC are: , , .

[0062] To simulate acceleration and deceleration processes in real industrial scenarios, this group of experiments conducted step acceleration tests from 500 rpm to 1200 rpm under half-load operating conditions. Figure 8 and Figure 9 As shown; and the step drop experiment from 1200 rpm to 800 rpm, as shown. Figure 10 and Figure 11As shown in the figure, during acceleration, the PI control exhibits significant speed overshoot, reaching as high as 255 rpm, and the adjustment process is relatively slow. While TISMC improves the response speed, it introduces large transient fluctuations. In contrast, ASOSMC+GSTESO achieves faster convergence and significantly reduces overshoot, with virtually no noticeable oscillations after the step. Similar patterns can be observed in the deceleration experiment, demonstrating that the proposed method effectively suppresses speed downshoot and shortens the recovery time. Furthermore, the corresponding current response shows that ASOSMC+GSTESO has a smaller transient current peak and a smoother current change process. These results indicate that the proposed controller effectively reduces current surges while improving transient tracking performance, verifying its superior dynamic adjustment capability and robustness under speed step conditions. The deceleration experiment further verifies the controller's rapid braking capability. When faced with a deceleration command, both PI and TISMC exhibit severe speed drop-offs, with overshoots reaching as high as 155 rpm and 210 rpm, respectively. In comparison, the ASOSMC proposed in this embodiment exhibits a good dynamic convergence advantage: its descent overshoot is 48 rpm, which is reduced by 69.0% and 77.1% compared to PI and TISMC, respectively; at the same time, ASOSMC also achieves the optimal descent recovery time, which is about 26.4% to 31.1% faster than the traditional control strategy.

[0063] Figure 12 and Figure 13 The response of the motor under sudden half-load and sudden half-load removal conditions during steady-state operation at 600 r / min is compared. Under the sudden half-load condition, the load disturbance caused a significant speed drop under PI and TISMC control. The PI control exhibited relatively large speed deviation and recovery time, with a speed drop of 18 rpm that took 170 ms to recover to the given speed. While TISMC improved the disturbance response speed, some transient fluctuations still existed. In contrast, ASOSMC+GSTESO more effectively suppressed the speed drop and recovered to near the given value in a shorter time. Under the sudden half-load removal condition, both PI and TISMC showed significant speed overshoot and recovery, while the proposed method exhibited lower speed peaks and faster steady-state recovery, with a speed fluctuation of only 38 rpm and a recovery time maintained at 121 ms. The current response shows that, at the moment of load change, the current regulation process of ASOSMC+GSTESO was smoother, suppressing current surges and subsequent fluctuations. The above results show that the proposed control strategy can effectively compensate for the impact of load disturbances on the speed loop, and exhibits stronger anti-disturbance capability and better dynamic adjustment performance under both sudden load increase and sudden load decrease conditions.

[0064] In this experiment, under half-load operation, the target speed command for the motor was set to a sinusoidal signal with DC bias, i.e. Waveform diagram Figure 14-17When tracking a 5Hz sinusoidal signal, both the PI and TISMC controllers exhibited significant phase lag and amplitude attenuation. The PI controller, essentially a linear lag element, lacks the ability to predict the rate of change of the command and can only rely on the integral of the lag error for passive correction. Its steady-state tracking error exhibits sinusoidal fluctuations at the same frequency as the fundamental wave, with an absolute error value as high as ±20 rpm. Although the TISMC theoretically possesses extremely strong robustness, its tracking performance is even worse than the PI controller when facing continuously changing reference commands. The error curves show that the TISMC's speed error range spreads to ±25 rpm. This is because the TISMC uses a fixed switching gain; when the system state continuously crosses the sliding mode surface at the peaks and troughs of the sinusoidal wave, the excessively large hard-switching gain triggers violent high-frequency switching of the control law, which in turn worsens the continuous tracking accuracy. Combining the error curves, it can be seen that the ASOSMC compresses the dynamic tracking error to within ±12 rpm, significantly reducing it by 40.0% and 52.0% compared to the PI and TISMC, respectively. This achieves a balance between dynamic response and smooth tracking.

[0065] The embodiments described are merely examples to clearly illustrate the present invention and are not intended to limit the implementation of the invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively describe all possible implementations. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A high-order sliding mode control system for a permanent magnet synchronous motor, wherein the high-order sliding mode control system for the permanent magnet synchronous motor adopts a cascaded double-loop structure based on a speed outer loop and a current inner loop, wherein the speed outer loop and the current inner loop are connected in series in the cascaded double-loop structure, characterized in that, include: The data acquisition and processing module is used to convert the three-phase stator current of the permanent magnet synchronous motor into d-axis current and q-axis current, and calculate the speed tracking error in the outer speed loop based on the given mechanical angular velocity and the actual mechanical angular velocity of the permanent magnet synchronous motor. The speed outer loop control module is used to construct a second-order integral non-singular terminal sliding surface and an improved composite exponential reaching law using the speed tracking error as the state variable. Based on the speed tracking error, the parameters of the permanent magnet synchronous motor, the second-order integral non-singular terminal sliding surface, and the improved composite exponential reaching law, it obtains the q-axis reference current. The improved composite exponential reaching law includes a speed-changing term related to the second-order integral non-singular terminal sliding surface, a variable exponential term related to the speed tracking error, and a switching function term. An adaptive generalized super-twisted extended state observation module is provided, comprising an adaptive generalized super-twisted extended state observer. The observer is constructed based on a lumped disturbance model, which is established based on the q-axis current and the actual mechanical angular velocity. The module estimates the lumped disturbance and the actual mechanical angular velocity, obtaining estimated values ​​for both. These estimated values ​​are then fed forward to the outer velocity control module to compensate for the disturbance. The current inner loop control module is used to compare the d-axis reference current and q-axis reference current with the d-axis current and q-axis current respectively to obtain the current error signal input to the current inner loop, and obtain the d-axis reference voltage and q-axis reference voltage through the current inner loop; the d-axis reference current is 0; The motor drive control module is used to generate a drive signal based on the d-axis reference voltage and the q-axis reference voltage, and the drive signal is used to control the operation of the permanent magnet synchronous motor.

2. The high-order sliding mode control system for a permanent magnet synchronous motor according to claim 1, characterized in that, The data acquisition and processing module collects the actual mechanical angular velocity during the operation of the permanent magnet synchronous motor. and three-phase stator current , , By using Clarke and Park transformations, the three-phase stator currents of a permanent magnet synchronous motor in a three-phase stationary coordinate system are converted into d-axis currents in a two-phase rotating coordinate system. With q-axis current ; the given mechanical angular velocity Compared with actual mechanical angular velocity By comparison, the speed tracking error is obtained, and the expression for the speed tracking error is: .

3. The high-order sliding mode control system for a permanent magnet synchronous motor according to claim 1, characterized in that, In the speed outer loop control module, the process of constructing the second-order integral nonsingular terminal sliding surface is as follows: Using the speed tracking error as a state variable, a first-order integral sliding surface is constructed; The expression for the first-order integral sliding surface is: ,in, Denotes the first-order integral sliding surface. Represents state variables, Indicates the weighting coefficient. Indicates time, Indicates a lower-order power exponent. Indicates a higher-order power exponent; Based on the first-order integral sliding surface, a second-order integral non-singular terminal sliding surface is constructed. The expression for the second-order integral non-singular terminal sliding surface is: ,in, This indicates that the sliding surface of the second-order integral is a non-singular terminal. The derivative of the first-order integral sliding surface is given by . This represents the sliding surface gain coefficient. 、 Indicates a variable power exponent. , , , , , , This represents the adjustment parameter for the variable power exponent.

4. A high-order sliding mode control system for a permanent magnet synchronous motor according to claim 3, characterized in that, In the speed outer loop control module, the expression for the improved composite exponential reaching law is: in, Indicates the variable speed term. Indicates a variable exponent term; Indicates the switch adjustment coefficient. Represents the switching function term. Represents a symbolic function. Represents the hyperbolic tangent function. The boundary layer thickness is indicated; the boundary layer represents a continuous transition interval set on both sides of the sliding surface, and the boundary layer thickness parameter is a preset positive constant.

5. A high-order sliding mode control system for a permanent magnet synchronous motor according to claim 4, characterized in that, The process of obtaining the q-axis reference current satisfies the expression: ,in, Indicates the q-axis reference current. This represents the estimated value of the lumped disturbance. Represents the torque constant. Indicates the moment of inertia. This represents the gain of the improved composite exponential reaching law.

6. A high-order sliding mode control system for a permanent magnet synchronous motor according to claim 1, characterized in that, The expression for the lumped disturbance model is: ,in, Indicates lumped disturbance; Represents mechanical torque. This represents the viscous damping coefficient.

7. A high-order sliding mode control system for a permanent magnet synchronous motor according to claim 1, characterized in that, Introduce extended state variable D, , The derivative of the extended state variable is represented by the expression for the adaptive generalized super-twisted extended state observer: ,in, This indicates the error in estimating the mechanical angular velocity. This represents the lumped disturbance estimation error, used to estimate the lumped disturbance and the actual mechanical angular velocity. and They represent and The estimated value, , This represents the generalized super-torsion nonlinear correction term. , Indicates observation gain; The expression for the generalized super-torsion nonlinear correction term is: in, This represents the first nonlinear correction term, used to perform nonlinear correction on the mechanical angular velocity estimation error; This represents the second nonlinear correction term, used to perform nonlinear correction on the lumped disturbance estimation error; This indicates an adjustable gain parameter.

8. A high-order sliding mode control system for a permanent magnet synchronous motor according to claim 7, characterized in that, The expression for the adjustable gain parameter is: ,in, This indicates the given speed of the permanent magnet synchronous motor. This represents the observation rotation speed of the adaptive generalized super-twisted extended state observer. This indicates that a threshold value is set. Indicates the original gain. This indicates the maximum gain.

9. A high-order sliding mode control system for a permanent magnet synchronous motor according to claim 1, characterized in that, In the motor drive control module, based on the d-axis reference voltage and the q-axis reference voltage, a drive signal is generated using inverse Park transformation and SVPWM modulation. The drive signal serves as the switching signal for the inverter connected to the permanent magnet synchronous motor.

10. A high-order sliding mode control method for a permanent magnet synchronous motor, characterized in that, The method is implemented based on the system described in any one of claims 1 to 9, and includes the following steps: The three-phase stator current of the permanent magnet synchronous motor is converted into d-axis current and q-axis current. Based on the given mechanical angular velocity and the actual mechanical angular velocity of the permanent magnet synchronous motor, the speed tracking error is calculated in the outer speed loop. Using the speed tracking error as a state variable, a second-order integral non-singular terminal sliding surface and an improved composite exponential reaching law are constructed. Based on the speed tracking error, the parameters of the permanent magnet synchronous motor, the second-order integral non-singular terminal sliding surface, and the improved composite exponential reaching law, the q-axis reference current is obtained. The improved composite exponential reaching law includes a speed-changing term related to the second-order integral non-singular terminal sliding surface, a variable exponential term related to the speed tracking error, and a switching function term. A lumped disturbance model is established based on the q-axis current and the actual mechanical angular velocity. An adaptive generalized super-twisted extended state observer is constructed based on the lumped disturbance model. The lumped disturbance and the actual mechanical angular velocity are estimated using the adaptive generalized super-twisted extended state observer to obtain the lumped disturbance estimate and the actual mechanical angular velocity estimate. The lumped disturbance estimate is then fed forward to compensate for the process of obtaining the q-axis reference current. The d-axis reference current and q-axis reference current are compared with the d-axis current and q-axis current respectively to obtain the current error signal input to the inner current loop. The d-axis reference voltage and q-axis reference voltage are obtained through the inner current loop; the d-axis reference current is 0. Based on the d-axis reference voltage and the q-axis reference voltage, a drive signal is generated, which is used to control the operation of the permanent magnet synchronous motor.

Images