Sliding mode observation method and system for permanent magnet synchronous motor non-inductive control

Abstract

The invention discloses a sliding mode observation method and a sliding mode observation system for permanent magnet synchronous motor non-inductive control, which are used for improving numerical stability and observation precision of a discrete system and realizing permanent magnet synchronous motor non-inductive closed-loop control with good robustness and stability. The method comprises the following steps: discretizing a current state equation according to a sampling period to obtain an initial discrete sliding mode observation equation; determining a proportional adjustment coefficient k based on an inductance parameter and a sampling period of the motor, performing weighted adjustment on a voltage related item and a sliding-mode control quantity item in the initial discrete sliding-mode observation equation by using the proportional adjustment coefficient k, and constructing an improved sliding-mode observation equation; extended counter electromotive force is extracted based on the output of the sliding mode control law, wherein the contribution proportion of the high-frequency buffeting component output by the sliding mode control law in the extended counter electromotive force is weakened through a proportion adjustment coefficient; and inputting the extended back electromotive force into a phase-locked loop module to calculate the rotor position and the rotating speed of the motor so as to realize closed-loop control of the motor.

Description

Technical Field

[0001] This application relates to the field of motor control technology, and in particular to a sliding mode observation method and system for sensorless control of permanent magnet synchronous motors. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) are widely used in industrial automation, electric vehicles, and home appliances due to their high efficiency, high power density, and excellent control performance. Existing PMSMs typically use position sensors mounted on the motor shaft to detect rotor position and achieve closed-loop control. However, the introduction of position sensors increases system cost, complicates the motor structure, and places higher demands on installation accuracy and operating environment.

[0003] To reduce system costs and simplify the structure, existing technologies have proposed sensorless control methods for permanent magnet synchronous motors, which estimate rotor position and speed through algorithms without using mechanical position sensors. A commonly used sensorless control scheme consists of a sliding mode observer and a phase-locked loop (PLL). The sliding mode observer is used to obtain the extended back electromotive force (EMF) of the motor, and the PLL is used to calculate the motor's electrical angle and speed from the extended back EMF.

[0004] However, due to their inherent control characteristics, existing sliding mode observers typically contain strong high-frequency chattering components in the control input, and the extended back EMF is usually directly extracted from these control inputs, resulting in a significant chattering component in the obtained extended back EMF as well. To suppress this chattering, strong filtering is often required, but this introduces phase delay and noise amplification problems, affecting the accuracy of the phase-locked loop (PLL) calculation. Furthermore, in practical applications, the sliding mode observation algorithm needs to be discretized. When the sampling period is large or the motor inductance parameter is small, some coefficients in the discrete state equation may be too large, making it difficult for the observation to converge, thus greatly reducing the stability and robustness of the sensorless control. Summary of the Invention

[0005] This application provides a sliding mode observation method and system for sensorless control of permanent magnet synchronous motors, which improves the numerical stability and observation accuracy of discrete systems and realizes sensorless closed-loop control of permanent magnet synchronous motors with good robustness and stability.

[0006] The first aspect of this application provides a sliding mode observation method for sensorless control of a permanent magnet synchronous motor, including: Obtain the actual current and voltage components of the motor in a two-phase stationary coordinate system; A current state equation based on the physical parameters of the motor is established, and the current state equation is discretized according to the sampling period to obtain the initial discrete sliding mode observation equation. Based on the inductance parameters of the motor and the sampling period, a proportional adjustment coefficient k is determined. The voltage-related terms and sliding mode control terms in the initial discrete sliding mode observation equation are weighted and adjusted using the proportional adjustment coefficient k to construct an improved sliding mode observation equation. The proportional adjustment coefficient k is a value greater than 0 and less than 1. Based on the improved sliding mode observation equation, the current observation is iteratively calculated based on the observed current at the previous moment, the actual voltage component, and the sliding mode control law, and the observed current is driven to track the actual current component. The extended back electromotive force is extracted based on the output of the sliding mode control law, wherein the contribution ratio of the high-frequency chattering component of the sliding mode control law output to the extended back electromotive force is weakened by the proportional adjustment coefficient. The extended back EMF input phase-locked loop module calculates the rotor position and speed of the motor, and generates control signals accordingly to achieve closed-loop control of the motor.

[0007] Optionally, determining the proportional adjustment coefficient k based on the inductance parameters of the motor and the sampling period includes: Calculate the ratio of the sampling period to the inductance parameter of the motor to obtain the initial weighting coefficients corresponding to the sliding mode control quantity in the discrete sliding mode observation equation; Based on the influence of the initial weighting coefficient on the observed current change amplitude or the extended back EMF waveform during the discrete iterative calculation, determine whether the initial weighting coefficient exceeds the numerical convergence boundary. If so, a value greater than 0 and less than 1 is selected as the scaling factor k, and the initial weight coefficient is scaled using the scaling factor k so that the scaled weight coefficient is constrained within the numerical convergence boundary.

[0008] Optionally, after selecting a value greater than 0 and less than 1 as the scaling factor k, and using the scaling factor k to scale the initial weight coefficient, the sliding mode observation method further includes: Based on the fact that the discrete sliding mode observation equation satisfies numerical convergence, the value of the proportional adjustment coefficient k is further reduced to reduce the high-frequency chattering component contained in the extended back electromotive force until the amplitude of the high-frequency chattering component is lower than the preset signal noise tolerance threshold.

[0009] Optionally, the step of extracting the extended back electromotive force based on the output of the sliding mode control law includes: Obtain the real-time output value of the sliding mode control law in the improved sliding mode observation equation, as well as the actual voltage components in the two-phase stationary coordinate system; The extended back electromotive force is calculated by weighting and summing the actual voltage component and the real-time output value of the sliding mode control law using the proportional adjustment coefficient k, so that the main amplitude of the extended back electromotive force is provided by the actual voltage component without chattering. The formula for the weighted summation calculation satisfies: , ; in, and They are respectively shaft and The extended back electromotive force of the shaft, , They are respectively shaft and The actual voltage component of the shaft, , This is the real-time output value of the sliding mode control law; Before the extended back EMF input phase-locked loop module calculates the rotor position and speed of the motor, the sliding mode observation method further includes: The extended back EMF is input to a first-order low-pass filter for filtering, and the cutoff frequency of the first-order low-pass filter is higher than the fundamental frequency of the motor.

[0010] Optionally, the initial discrete sliding mode observation equation is: ; in, , The sampling period is... For motor resistance, , These are the equivalent inductance parameters of the motor along the d-axis and q-axis in a synchronous rotating coordinate system, respectively. Electric angular velocity, , They are respectively shaft and The actual voltage component of the shaft, , They are respectively shaft and The observed current of the axis, , is the real-time output value of the sliding mode control law, and n is used to represent the discrete time step index.

[0011] Optionally, the improved sliding mode observation equation is: ; in, k is the proportional adjustment coefficient. The sampling period is... For motor resistance, , These are the equivalent inductance parameters of the motor along the d-axis and q-axis in a synchronous rotating coordinate system, respectively. Electric angular velocity, , They are respectively shaft and The actual voltage component of the shaft, , They are respectively shaft and The observed current of the axis, , is the real-time output value of the sliding mode control law, and n is used to represent the discrete time step index.

[0012] Optionally, the step of calculating the rotor position and speed of the motor by inputting the extended back EMF into the phase-locked loop module, and generating a control signal accordingly to achieve closed-loop control of the motor, includes: The angle error signal characterizing the rotor position estimation error is calculated based on the extended back electromotive force; The angle error signal is input into the proportional-integral controller, and the steady-state error is eliminated by the integral action of the proportional-integral controller to obtain the estimated speed of the motor. The rotor position electrical angle of the motor is obtained by performing discrete integration on the estimated rotational speed value. By using the rotor position electrical angle, a rotating coordinate transformation is performed on the actual current components in the two-phase stationary coordinate system to obtain the current feedback value in the synchronous rotating coordinate system. Based on the difference between the preset current command value and the current feedback value, the voltage command value in the synchronous rotating coordinate system is calculated by the current loop controller; The voltage command value is subjected to inverse coordinate transformation and space vector pulse width modulation operation using the rotor position electrical angle to generate a PWM control signal for driving the motor.

[0013] Optionally, the sliding mode control law may employ a saturation function, a sign function, a continuous function, or a hyperbolic tangent function.

[0014] The second aspect of this application provides a sliding mode observation system for sensorless control of a permanent magnet synchronous motor, comprising: The acquisition unit is used to acquire the actual current component and actual voltage component of the motor in a two-phase stationary coordinate system. Discrete unit, used to establish current state equation based on the physical parameters of the motor, and to discretize the current state equation according to the sampling period to obtain the initial discrete sliding mode observation equation; The construction unit is used to determine the proportional adjustment coefficient k based on the inductance parameters of the motor and the sampling period, and to use the proportional adjustment coefficient k to perform weighted adjustment on the voltage-related terms and sliding mode control terms in the initial discrete sliding mode observation equation to construct the improved sliding mode observation equation. The calculation unit is used to iteratively calculate the current observation based on the observed current at the previous moment, the actual voltage component, and the sliding mode control law according to the improved sliding mode observation equation, and drive the observed current to track the actual current component. An extraction unit is used to extract extended back electromotive force based on the output of the sliding mode control law, wherein the contribution ratio of the high-frequency chattering component of the sliding mode control law output to the extended back electromotive force is weakened by the proportional adjustment coefficient. The control unit is used to input the extended back EMF into the phase-locked loop module to calculate the rotor position and speed of the motor, and generate control signals accordingly to realize closed-loop control of the motor.

[0015] A third aspect of this application provides a sliding mode observation device for sensorless control of a permanent magnet synchronous motor, the device comprising: Processor, memory, input / output units, and bus; The processor is connected to the memory, the input / output unit, and the bus; The memory stores a program, which the processor calls to execute the first aspect and any one of the first aspects of the sliding mode observation method for sensorless control of a permanent magnet synchronous motor.

[0016] The fourth aspect of this application provides a computer-readable storage medium storing a program that, when executed on a computer, performs the sliding mode observation method for sensorless control of a permanent magnet synchronous motor, as described in the first aspect and any optional method of the first aspect.

[0017] As can be seen from the above technical solutions, this application has the following advantages: Since the stability analysis of sliding mode observers is usually based on continuous system models, while actual control systems need to be discretely implemented, the numerical behavior of discrete systems can easily deviate from the continuous system assumptions when the sampling period is large or the motor inductance parameter is small, manifesting as observation non-convergence or significant high-frequency chattering. Based on this, this invention introduces a proportional adjustment coefficient k, determined according to the sampling period and motor inductance parameter, on the basis of the discretization of the current state equation. This coefficient synergistically weights and adjusts the voltage-related terms and sliding mode control terms in the discrete sliding mode observer, thereby effectively constraining the variation amplitude within a unit sampling period during the discrete state update process without changing the sampling period. This makes the numerical behavior of the discrete sliding mode observer closer to the stability characteristics under continuous system conditions. Under this synergistic adjustment mechanism, the proportional adjustment coefficient suppresses the high-frequency chattering amplitude introduced by the sliding mode control law during discrete implementation, and simultaneously reduces the proportion of high-frequency chattering components in the extended back EMF. This makes the input signal of the phase-locked loop constructed based on the extended back EMF smoother and more stable, thereby improving the noise immunity and reliability of the rotor position and speed calculation process and avoiding the amplification and transmission of observation chattering to the closed-loop control link.

[0018] By introducing a proportional adjustment coefficient based on the discretization of the current state equation, and coordinating the weighted adjustment of the voltage-related terms and sliding mode control terms in the discrete sliding mode observer, this invention can significantly improve the numerical stability and observation accuracy of the discrete system without increasing the computational complexity or relying on shortening the sampling period, thereby realizing sensorless closed-loop control of permanent magnet synchronous motors with good robustness and stability. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A schematic flowchart of an embodiment of the sliding mode observation method for sensorless control of permanent magnet synchronous motors provided in this application; Figure 2 A schematic flowchart illustrating an embodiment of the sliding mode observation method for sensorless control of a permanent magnet synchronous motor provided in this application for determining the proportional adjustment coefficient k; Figure 3 A schematic flowchart of another embodiment of the sliding mode observation method for sensorless control of permanent magnet synchronous motors provided in this application; Figure 4-a for Actual shaft current & traditional sliding mode observer tracking shaft current; Figure 4-b For observations by a traditional sliding mode observer ; Figure 5-a for Actual shaft current & tracking of the improved sliding mode observer in this application shaft current; Figure 5-b For observations by the improved sliding mode observer ; Figure 6-a For actual electrical angle & observed electrical angle; Figure 6-b The electrical angle error of the observation; Figure 6-c For actual speed and observed speed; Figure 6-d The velocity tracking effect was observed at 50Hz. Figure 7 A schematic diagram of an embodiment of the sliding mode observation system for sensorless control of permanent magnet synchronous motors provided in this application; Figure 8 This is a schematic diagram of an embodiment of the sliding mode observation device for sensorless control of permanent magnet synchronous motors provided in this application. Detailed Implementation

[0021] This application provides a sliding mode observation method and system for sensorless control of permanent magnet synchronous motors, which improves the numerical stability and observation accuracy of discrete systems and realizes sensorless closed-loop control of permanent magnet synchronous motors with good robustness and stability.

[0022] It should be noted that the sliding mode observation method for sensorless control of permanent magnet synchronous motors provided in this application can be implemented by various types of control devices. For example, the method can be integrated directly into the servo driver or controller of the permanent magnet synchronous motor itself as firmware, executed by its internal microcontroller or digital signal processor to achieve in-machine self-calibration and sensorless closed-loop control of the motor. Alternatively, the method can be executed by an external host computer, programmable logic controller, or dedicated production line control equipment. In this case, the external device can send real-time or constant-speed operation commands to the motor driver through a communication interface and obtain operating data such as current and voltage from the driver. Iterative calculations and model processing are performed on the external device to obtain the observed value of the proportional adjustment coefficient k or the extended back electromotive force. The calculation results are then sent back to the driver or controller for real-time sensorless control. For ease of explanation, the embodiments of this application are described with a controller as the background, but this example does not constitute a limitation on the scope of protection of this application.

[0023] Please see Figure 1 , Figure 1An embodiment of the sliding mode observation method for sensorless control of permanent magnet synchronous motors provided in this application includes: 101. Obtain the actual current components and actual voltage components of the motor in a two-phase stationary coordinate system; To achieve sensorless control of a permanent magnet synchronous motor, the controller first needs to acquire the motor's three-phase current signals and bus voltage signals. The three-phase currents refer to the three phase currents of the motor's stator windings, which collectively reflect the motor's instantaneous operating state. The bus voltage is the DC voltage supplied to the motor, used to control its torque output. Subsequently, the controller performs coordinate transformation on these signals, mapping the three-phase static currents and voltages to a two-phase stationary coordinate system to form intuitive measurements independent of the motor's rotating magnetic field.

[0024] The two-phase stationary coordinate system, or α-β coordinate system, is a physical model in which the three-phase stator windings of a motor are spatially distributed with a 120-degree difference between each other. This model is mathematically transformed into a mathematical model with two mutually perpendicular and stationary virtual axes. Its purpose is to eliminate time-varying parameters and simplify the mathematical calculations of the control system. The actual current and voltage components, on the other hand, refer to the vector components that mathematically reflect the current electromagnetic state of the motor after the analog signals collected by physical sensors are transformed using the aforementioned coordinate system. In practical applications, this transformation can be achieved through the Clarke transform, which maps the three-phase current and voltage into two sets of mutually perpendicular components through a fixed linear combination. This allows subsequent observers to directly use these two sets of signals for state estimation without relying on rotor position sensors.

[0025] 102. Establish the current state equation based on the physical parameters of the motor, and discretize the current state equation according to the sampling period to obtain the initial discrete sliding mode observation equation. After obtaining the actual current and voltage components of the motor, a mathematical model describing the motor's electromagnetic behavior, namely the current state equation, needs to be constructed to estimate the motor's internal state. This current state equation is based on the motor's inherent physical parameters, mainly stator resistance and stator inductance, which objectively determine the motor's current response dynamics under specific voltage excitation. Since the actual control system of the motor uses a digital controller for calculation, its computation is based on discrete time steps, making it impossible to directly solve continuous-time differential equations. Therefore, a sampling period, i.e., the controller's computation period, needs to be introduced as a time reference to discretize the continuous current state equation.

[0026] In practical implementation, numerical analysis methods such as the forward Euler method can be used to transform the differential terms in the current state equation into a difference form suitable for digital processor computation. Alternatively, methods such as the backward Euler method or bilinear transformation can be used for discretization; the specific method is not limited here. This processing yields the initial discrete sliding mode observation equation, which describes the mathematical logic of how to recursively predict the current value at the next moment using the current value at the previous moment and the current voltage value.

[0027] 103. Based on the inductance parameters of the motor and the sampling period, determine the proportional adjustment coefficient k. Use the proportional adjustment coefficient k to weight and adjust the voltage-related terms and sliding mode control terms in the initial discrete sliding mode observation equation to construct the improved sliding mode observation equation. The proportional adjustment coefficient k is a value greater than 0 and less than 1. After obtaining the initial discrete sliding mode observation equations, to address the issue of divergence or high-frequency chattering that traditional sliding mode observers are prone to under actual operating conditions, this embodiment introduces a crucial proportional adjustment coefficient k in step 103. In digital control systems, the iteration step size of the current observer is mainly determined by two physical quantities: one is the sampling period of the controller, which limits the time interval of discrete calculations; the other is the inductance parameter of the motor, which reflects the inertial response of the current to voltage changes. In practical engineering, if a low-inductance motor is used or the sampling period is large due to hardware performance limitations, the coefficient before the control quantity in the discrete equations, i.e., the ratio of the sampling period to the inductance parameter, may increase significantly. In this case, the iteration step size exceeds the stable region of the numerical algorithm, which, from a physical perspective, manifests as significant overshoot and oscillation of the observed current in each iteration, causing the observed value to fail to converge to the true current, affecting the accuracy and stability of sensorless control.

[0028] To address this issue, this embodiment introduces a proportional adjustment coefficient k into both the voltage-related term reflecting the feedforward effect and the sliding mode control term reflecting the feedback correction effect in the initial discrete sliding mode observation equation. This coefficient k is then weighted and adjusted, with the value of k ranging from 0 to 1. This weighted adjustment essentially introduces a damping element at the input of the mathematical model, effectively limiting the current variation during iterative calculations and thus reshaping the dynamic characteristics of the observer. This improvement effectively controls the gain amplitude of a single iteration, maintains the phase response characteristics of the observer, and confines the closed-loop poles of the system to a stable region within the unit circle, thereby significantly improving numerical convergence and chattering resistance under digital discrete conditions.

[0029] 104. Based on the improved sliding mode observation equation, the observed current at the current moment is iteratively calculated based on the observed current, actual voltage components and sliding mode control law of the previous moment, and the observed current is driven to track the actual current components. After obtaining the improved sliding mode observation equation, the controller iteratively calculates the current observed current based on the observed current from the previous moment, the actual voltage components in the two-phase stationary coordinate system, and the output of the sliding mode control law. This sliding mode control law is a nonlinear control algorithm used to incorporate the observer's error feedback into the observation equation, enabling rapid tracking of the actual motor current. Specifically, once a deviation is detected, the sliding mode control law immediately generates a virtual control quantity with a reverse correction function to correct the deviation. In each iteration of the practical application, the controller combines the observed current from the previous moment with the real-time acquired voltage signal and calculates the current observed current value using the improved sliding mode observation equation. The controller then uses this observed current as an estimation result to drive the output signal of the sliding mode observation equation to track the actual current components, achieving closed-loop control of the motor current.

[0030] 105. The extended back electromotive force is extracted based on the output of the sliding mode control law, wherein the contribution ratio of the high-frequency chattering component of the sliding mode control law output to the extended back electromotive force is weakened by the proportional adjustment coefficient. After obtaining the observed current at the current moment, the controller can further extract the extended back electromotive force (EMF) using the output signal of the sliding mode control law. The extended back EMF is a key physical quantity used in sensorless motor control to estimate rotor position and speed; it reflects the instantaneous changes in the electromagnetic induction within the motor. In actual operation, the signal output by the sliding mode control law contains two parts: one is the effective correction to the actual current deviation of the motor, and the other is the high-frequency chattering component introduced by discrete iteration and control nonlinearity. If the original sliding mode control law output is directly used as the extended back EMF, the high-frequency chattering will be superimposed on the estimated signal, causing noise when the phase-locked loop calculates the rotor angle, affecting the stability and accuracy of the sensorless control.

[0031] Based on this, this embodiment utilizes the aforementioned proportional adjustment coefficient k to attenuate and weaken the high-frequency chattering component in the sliding mode control law output. Specifically, when generating the extended back EMF, the sliding mode control law output is scaled according to k. Since k is a value greater than 0 and less than 1, this operation objectively attenuates the amplitude of the high-frequency chattering component in the control quantity, significantly reducing the proportion of chattering noise in the final extracted extended back EMF signal while maintaining the integrity of the effective current correction signal. In this way, the extended back EMF can effectively suppress numerical noise generated during discrete iteration while preserving rotor position information and dynamic response characteristics.

[0032] 106. The extended back EMF input phase-locked loop module calculates the rotor position and speed of the motor, and generates control signals accordingly to achieve closed-loop control of the motor.

[0033] The controller inputs the extended back EMF to the phase-locked loop (PLL) module to calculate the rotor position and speed. This PLL module continuously tracks the phase of the extended back EMF vector. Specifically, the PLL calculates the phase error signal between the extended back EMF vector phase and the internally estimated phase based on arctangent calculation or a small-angle approximation algorithm. The phase error is then adjusted by a closed-loop regulator to update the speed estimate. Subsequently, the speed estimate is integrated to obtain the electrical angle corresponding to the rotor position. Because the high-frequency chattering component in the sliding mode control law output has been effectively weakened by the proportional adjustment coefficient k in the preceding steps, the extended back EMF signal input to the PLL module has high smoothness and a high signal-to-noise ratio, enabling the PLL to achieve stable phase locking without relying on a large time constant filtering stage. This signal characteristic improves the dynamic response capability of the PLL and reduces noise interference during phase estimation, thereby enhancing the accuracy of rotor position and speed calculations.

[0034] The controller feeds back the calculated rotor position electrical angle to the coordinate transformation stage of the vector control system. This transforms the acquired stator current from a two-phase stationary coordinate system to a synchronous rotating coordinate system, thereby achieving decoupled control of the torque and flux components. Simultaneously, the calculated rotational speed is used as feedback for the speed loop in closed-loop regulation. Based on this, the controller calculates the stator voltage command through the current and speed loops and generates a corresponding pulse width modulation signal using space vector pulse width modulation technology. This signal drives the inverter to apply control voltage to the permanent magnet synchronous motor, thus achieving sensorless closed-loop vector control operation.

[0035] In this embodiment, since the stability analysis of sliding mode observers is usually based on continuous system models, while actual control systems need to be discretely implemented, the numerical behavior of discrete systems is prone to deviating from the continuous system assumptions when the sampling period is large or the motor inductance parameter is small, manifesting as observation non-convergence or significant high-frequency chattering. Based on this, this invention introduces a proportional adjustment coefficient k, determined according to the sampling period and motor inductance parameter, on the basis of the discretization of the current state equation. This coefficient synergistically weights and adjusts the voltage-related terms and sliding mode control terms in the discrete sliding mode observer, thereby effectively constraining the variation amplitude within a unit sampling period during the discrete state update process without changing the sampling period. This makes the numerical behavior of the discrete sliding mode observer closer to the stability characteristics under continuous system conditions. Under this synergistic adjustment mechanism, the proportional adjustment coefficient, on the one hand, suppresses the high-frequency chattering amplitude introduced by the sliding mode control law during discrete implementation, and on the other hand, simultaneously reduces the proportion of high-frequency chattering components in the extended back EMF, making the input signal of the phase-locked loop constructed based on the extended back EMF smoother and more stable. This improves the noise immunity and reliability of the rotor position and speed calculation process, avoiding the amplification and transmission of observation chattering to the closed-loop control link.

[0036] By introducing a proportional adjustment coefficient based on the discretization of the current state equation, and coordinating the weighted adjustment of the voltage-related terms and sliding mode control terms in the discrete sliding mode observer, this invention can significantly improve the numerical stability and observation accuracy of the discrete system without increasing the computational complexity or relying on shortening the sampling period, thereby realizing sensorless closed-loop control of permanent magnet synchronous motors with good robustness and stability.

[0037] Please see Figure 2 According to some embodiments of the present invention, in step S103, the proportional adjustment coefficient k is determined based on the inductance parameters of the motor and the sampling period, which may specifically include, but is not limited to, the following: 201. Calculate the ratio of the sampling period to the inductance parameter of the motor to obtain the initial weighting coefficients corresponding to the sliding mode control quantity in the discrete sliding mode observation equation; The controller calculates the ratio between the current sampling period and the motor's inductance parameters to obtain the initial weighting coefficient corresponding to the sliding mode control quantity in the discrete sliding mode observation equation. This initial weighting coefficient reflects the strength of the voltage effect on the current change within a unit sampling period, and its magnitude directly determines the amplitude of the observed current change in each discrete iteration. In practical engineering, when the sampling period is relatively large or the motor inductance is small, this initial weighting coefficient often increases, thus significantly amplifying the current change in a single iteration.

[0038] 202. Based on the influence of the initial weighting coefficients on the amplitude of the observed current change or the waveform of the extended back electromotive force during the discrete iterative calculation, determine whether the initial weighting coefficients exceed the numerical convergence boundary. After obtaining the initial weighting coefficients, their stability needs to be evaluated. In the iterative mechanism of the discrete sliding mode observer, there exists a theoretical upper limit to the gain required to maintain numerical convergence, which is the numerical convergence boundary in this application. If the initial weighting coefficients are too large and exceed this boundary, it means that the correction magnitude of the control quantity within a single sampling period exceeds the physical limit or mathematical stability region of the system's dynamic response. This excessive gain manifests in practical physical phenomena as the observed current exhibiting drastic amplitude jumps and high-frequency oscillations during the iteration process, making it impossible to smoothly track the actual current; or as severe distortion and discontinuity in the extracted extended back EMF waveform. Based on this, the stability criterion for discrete systems in control theory can be used to determine whether the current initial weighting coefficients are within a stable and usable range by analyzing the system's dynamic response that the coefficients may cause, such as whether they induce divergence or excessive chattering.

[0039] 203. If so, select a value greater than 0 and less than 1 as the scaling factor k, and use the scaling factor k to scale the initial weight coefficient so that the scaled weight coefficient is constrained within the numerical convergence boundary. When the judgment result shows that the initial weighting coefficient exceeds the numerical convergence boundary, it indicates that the current hardware conditions cannot directly support the stable operation of the traditional sliding mode observer, such as limited sampling frequency or insufficient motor inductance. In this case, a positive number between 0 and 1 can be selected as the proportional adjustment coefficient k. The controller uses this coefficient k to perform multiplicative scaling operations on the excessively large initial weighting coefficient to calculate the corrected effective weighting coefficient. The physical essence of this operation is to reduce the feedback gain in the discrete iteration process, forcibly pulling the originally overflowing weighting coefficient back within the numerical convergence boundary. Through this active constraint at the software level, the improved observer mathematical model is ensured to meet the stability conditions of the discrete system, enabling the system to avoid numerical divergence and achieve stable convergence of the observed current even under unfavorable hardware conditions.

[0040] 204. On the basis that the discrete sliding mode observation equation satisfies numerical convergence, further reduce the value of the proportional adjustment coefficient k to reduce the high-frequency chattering component contained in the extended back electromotive force until the amplitude of the high-frequency chattering component is lower than the preset signal noise tolerance threshold.

[0041] After ensuring the sliding mode observer meets the numerical convergence condition by initially setting the proportional gain coefficient k, although the divergence risk can be eliminated and the current can be stably tracked, the inherent high-frequency switching characteristics of sliding mode control inevitably introduce high-frequency chattering into the control quantity. To further improve the observation accuracy, a secondary optimization of the proportional gain coefficient k can be performed while satisfying the stability constraints. Since the extended back EMF in the improved observation model is synthesized from the actual voltage component and the k-weighted sliding mode control quantity, the magnitude of the coefficient k directly determines the proportion of high-frequency noise in the control quantity transmitted to the final back EMF result. Therefore, the controller can adopt a strategy of further reducing the value of the proportional gain coefficient k, using a smaller weighting coefficient to physically suppress the high-frequency chattering component in the output of the sliding mode control law. This adjustment process aims to optimize signal quality by reducing the value of k so that the amplitude of the residual high-frequency chattering component in the extended back EMF continues to decay until it is lower than the preset signal noise tolerance threshold. This strategy, while ensuring system stability, sacrifices a very small amount of dynamic gain to achieve a significant improvement in the signal-to-noise ratio of the observed signal, thereby providing a quieter input source for subsequent phase-locked loop (PLL) calculations.

[0042] The sliding mode observation method for sensorless control of permanent magnet synchronous motors provided in this application is explained in detail below with specific formulas and principles. Please refer to [link to relevant documentation]. Figure 3 , Figure 3Another embodiment of the sliding mode observation method for sensorless control of permanent magnet synchronous motors provided in this application includes: 301. Obtain the actual current components and actual voltage components of the motor in a two-phase stationary coordinate system; In this embodiment, step 301 is similar to step 101 in the previous embodiment, and will not be described again here.

[0043] 302. Establish the current state equation based on the physical parameters of the motor, and discretize the current state equation according to the sampling period to obtain the initial discrete sliding mode observation equation. The following is the current state equation for a permanent magnet synchronous motor: ,in, , In the formula For motor resistance, , These are the d-axis and q-axis inductances, respectively. Electric angular velocity, , , , They are respectively axis Shaft current and voltage and This refers to the EMF to be observed.

[0044] Therefore, the traditional sliding mode observer is designed as follows: ,in , The observed current value, , This is the control variable. Its control law is: Where G is the gain. The symbolic function can be replaced with a saturated function, a continuous function, or a hyperbolic tangent function, etc., as needed. No specific restrictions are imposed here. Once the observed current catches up with the actual current, then = , = Thus obtain , value.

[0045] In this embodiment, the observer needs to be discretized into the following recursive form to obtain the initial discrete sliding mode observation equation: , in, , The sampling period, which is also the algorithm's computation period, For motor resistance, , These are the equivalent inductance parameters of the motor along the d-axis and q-axis in a synchronous rotating coordinate system, respectively. Electric angular velocity, , They are respectively shaft and The actual voltage component of the shaft, , They are respectively shaft and The observed current of the axis, , is the real-time output value of the sliding mode control law, and n is used to represent the discrete time step index.

[0046] As can be seen from the discrete form, if the operation period Too large or motor inductance Too small a value will lead to a decrease in the coefficient before the control quantity. If the value is too large, it amplifies the chattering in the control variable, leading to drastic changes in the observed current and difficulty in convergence. Even optimizations such as changing the sign function in the control law to a saturation function can only alleviate chattering to a certain extent. Furthermore, because... = , = This introduces the chattering of the control variables entirely into the EMF, resulting in significant chattering within the EMF itself, which complicates the subsequent phase-locked loop (PLL) algorithm. This is the problem with traditional sliding mode observers.

[0047] 303. Based on the inductance parameters of the motor and the sampling period, determine the proportional adjustment coefficient k, and use the proportional adjustment coefficient k to perform weighted adjustment on the voltage-related terms and sliding mode control terms in the initial discrete sliding mode observation equation to construct the improved sliding mode observation equation. As can be seen from the discrete form in step 302, if the operation period Too large or motor inductance Too small a value will lead to a decrease in the coefficient before the control quantity. If the value is too large, it amplifies the chattering in the control variable, leading to drastic changes in the observed current and difficulty in convergence. Even optimizations such as changing the sign function in the control law to a saturation function can only alleviate chattering to a certain extent. Furthermore, because... = , = This introduces the chattering of the control variables entirely into the EMF, resulting in significant chattering within the EMF itself, which complicates the subsequent phase-locked loop (PLL) algorithm. This is the problem with traditional sliding mode observers.

[0048] To address the aforementioned issues, this embodiment first determines the proportional control coefficient k based on the motor's inductance parameters and the sampling period. Then, an improved sliding mode observation equation is constructed using k, where k is a decimal between 0 and 1, and the control law remains unchanged. The improved sliding mode observation equation is constructed as follows: ; As shown in the equation, this embodiment adds a proportional adjustment coefficient before the voltage-related term and the sliding mode control term. Used for adjustment, can be based on , Determine the actual value situation The value ensures that the overall coefficients are within an appropriate range, thus affecting... , There are no longer any restrictions on the range of values, and it can adapt to motors with any calculation cycle and any inductance.

[0049] 304. Based on the improved sliding mode observation equation, the observed current at the current moment is iteratively calculated based on the observed current, actual voltage components and sliding mode control law of the previous moment, and the observed current is driven to track the actual current components. In this embodiment, step 304 is similar to step 104 in the previous embodiment, and will not be described again here.

[0050] 305. Obtain the real-time output value of the sliding mode control law in the improved sliding mode observation equation, as well as the actual voltage components in the two-phase stationary coordinate system; At the end of each control cycle of the observer's iterative operation, the controller needs to read the current real-time output value of the sliding mode control law, i.e., the control quantity, from the improved sliding mode observation equation. , This variable contains a high-frequency switching signal used to correct for observed current errors. Simultaneously, the controller synchronously acquires the actual voltage component reconstructed in the two-phase stationary coordinate system, i.e. , These two variables represent the feedback correction term and the feedforward input term, respectively, forming the basic data for the subsequent synthesis of the back electromotive force signal.

[0051] 306. The extended back electromotive force is obtained by weighted summation of the actual voltage component and the real-time output value of the sliding mode control law using the proportional adjustment coefficient k, so that the main amplitude of the extended back electromotive force is provided by the chatter-free actual voltage component. After acquiring the basic data, the controller performs signal synthesis calculations based on the proportional adjustment coefficient k. Unlike traditional sliding mode observers that directly equate noisy control quantities to back electromotive force (EMF), this embodiment uses the improved mathematical structure of the sliding mode observation equation to perform a weighted summation of the actual voltage component and the output value of the sliding mode control law, ensuring that the main amplitude of the extended back EMF is provided by the chatter-free actual voltage component. The formula for this weighted summation calculation satisfies: , ; in, and They are respectively shaft and The extended back electromotive force of the shaft, , They are respectively shaft and The actual voltage component of the shaft, , This represents the real-time output value of the sliding mode control law. It can be seen that only a portion of the EMF value comes from control variables with chattering characteristics, and these variables have been modified... Scaling by a factor of 1, when When the value is small, this scaling is very significant, which greatly reduces the amount of jitter in the EMF, thereby obtaining an EMF value with very high signal quality.

[0052] Overall, the high-frequency chattering component originating from the sliding mode control law in the extended back EMF is significantly scaled and attenuated by a ratio k, while the fundamental component containing the true rotor information is fully preserved through the voltage component. This mechanism fundamentally alters the composition of the observed signal, resulting in a significantly reduced chattering noise level in the final extended back EMF while preserving dynamic information, thus enabling the acquisition of observations with extremely high signal quality.

[0053] 307. The extended back EMF input is filtered by a first-order low-pass filter. The cutoff frequency of the first-order low-pass filter is higher than the fundamental frequency of the motor. After weighted calculation, the signal-to-noise ratio of the extended back EMF has been significantly improved. To further filter out residual quantization noise or subharmonics, the controller inputs the signal into a digital first-order low-pass filter for processing. Since the jitter amplitude has been effectively suppressed using the proportional adjustment coefficient k in the previous steps, the noise energy in the signal is now at an extremely low level. Therefore, it is no longer necessary to rely on a low-cutoff frequency filter to forcibly filter out noise. Thus, the cutoff frequency of this first-order low-pass filter can be configured to be higher than the fundamental frequency of the motor. This high cutoff frequency configuration strategy, while ensuring signal smoothness, greatly reduces the phase lag introduced by the filter, ensuring the phase accuracy of the extended back EMF signal, thereby avoiding rotor position estimation errors caused by filtering delay.

[0054] 308. Calculate the angle error signal characterizing the rotor position estimation error based on the extended back electromotive force; After obtaining the filtered, high-quality extended back EMF, the controller enters the position calculation stage. Since the back EMF vector lags behind the rotor flux linkage vector by 90 degrees (or is orthogonal) in space, its two components contain the rotor's current absolute position information. The controller uses trigonometric function operations or small-angle approximation algorithms to process the extended back EMF components along the α and β axes, calculating the phase angle of the current extended back EMF vector. This phase angle is then compared with the rotor position angle estimated at the previous moment, or the back EMF is directly projected onto the d-axis of the estimated rotating coordinate system using a coordinate transformation based on the estimated angle, thus obtaining an angular error signal characterizing the deviation between the actual rotor position and the estimated position.

[0055] 309. Input the angle error signal into the proportional-integral controller. The steady-state error is eliminated by the integral action of the proportional-integral controller, and the estimated value of the motor speed is obtained. To achieve zero steady-state error tracking of the rotor position, the controller employs a software phase-locked loop (PLL) structure. The calculated angle error signal is input to the proportional-integral (PI) regulator within the PLL. In this closed-loop regulation, the proportional element responds to rapidly changing phase errors, providing the system with dynamic tracking capability; while the integral element accumulates historical errors, using its zero steady-state error characteristic to force the steady-state angle error to approach zero. As the regulation process progresses, the PI regulator's output is continuously corrected, eventually converging to a stable value. This value physically corresponds to the motor's current electrical angular velocity, i.e., the estimated motor speed.

[0056] 310. Perform discrete integration on the estimated rotational speed to obtain the electrical angle of the motor rotor position; After obtaining the converged speed estimate, the controller performs integration in the discrete-time domain. Specifically, the controller multiplies the current speed estimate by the sampling period and adds it to the rotor position angle from the previous moment to recursively obtain the current rotor position electrical angle. This electrical angle is a key variable for achieving vector control orientation, and its accuracy directly determines the smoothness and efficiency of the motor torque output. The integration process typically includes angle normalization, constraining the angle value within a period range of 0 to 2π to meet the continuous rotation control requirements of the motor.

[0057] 311. Using the rotor position electrical angle, perform a rotating coordinate transformation on the actual current components in the two-phase stationary coordinate system to obtain the current feedback value in the synchronous rotating coordinate system. The controller performs a rotating coordinate transformation on the actual current components in the two-phase stationary coordinate system obtained in step 301. This mathematical transformation uses a trigonometric function matrix to project the AC current signal in the stationary coordinate system onto the dq coordinate system, which rotates synchronously with the rotor, converting it into a DC current feedback value. At this point, the d-axis current component corresponds to the excitation component, and the q-axis current component corresponds to the torque component, thus achieving linearization and decoupling of the AC motor control object.

[0058] 312. Based on the difference between the preset current command value and the current feedback value, the voltage command value in the synchronous rotating coordinate system is calculated by the current loop controller. In the synchronous rotating coordinate system, the controller subtracts the preset current command value (typically 0 for the d-axis to achieve maximum torque-to-current ratio control, and the q-axis command determined by the speed loop output) from the actual current feedback value obtained in step 311 to obtain the current tracking error. This error is sent to the current loop PI controller for adjustment. Based on the magnitude and direction of the error, the current loop controller calculates the control voltage vector required to eliminate the error, i.e., the d-axis voltage command value and the q-axis voltage command value in the synchronous rotating coordinate system.

[0059] 313. Utilize the rotor position electrical angle to perform inverse coordinate transformation and space vector pulse width modulation calculation on the voltage command value to generate a PWM control signal for driving the motor.

[0060] To drive the three-phase inverter, the controller needs to convert the DC voltage command in the rotating coordinate system back to the control signal in the stationary coordinate system. Using the rotor position electrical angle obtained in step 310, the controller performs an inverse Park transformation on the d-axis and q-axis voltage command values ​​to obtain the voltage vectors in the two-phase stationary coordinate system, i.e. , Subsequently, the SVPWM algorithm calculates the conduction time of the power switching devices in each arm of the inverter based on the sector and amplitude of the voltage vector, and generates corresponding PWM control signals. These PWM signals are ultimately applied to the power drive circuit to control the voltage of the motor stator windings, forming a circular rotating magnetic field to drive the motor, thus completing the entire sensorless control process from current sampling and state observation to closed-loop drive.

[0061] The following section compares the improved sliding mode observer and the traditional sliding mode observer in the sliding mode observation method for sensorless control of permanent magnet synchronous motors provided in this application in terms of stability conditions, to verify that the introduction of the proportional adjustment coefficient k does not adversely affect the stability of the system. It should be noted that both sliding mode observers are based on the same Lyapunov stability criterion. The design process includes, The main difference lies in the relative weights of the parameters in the stability inequality under discrete implementation conditions. Therefore, the stability requirement is: Traditional sliding mode observer: ; The improved sliding mode observer of this application: ; For the sliding mode gain parameter G, its value does not need to be precise; it only needs to be as small as possible while satisfying the stability condition. Therefore, when comparing the stability requirements of the two observers, the dominant factors can be analyzed from an engineering application perspective. In practice, compared to the voltage term later... , The maximum possible value (approximately 2 / 3 of the bus voltage). , The voltage is often much smaller (possibly by more than two orders of magnitude), therefore the dominant term is the voltage-related term that follows. This can be seen in the following motor voltage equation. ,in It is a differential operator; In practice, , These are all very small values, especially for surface-mounted permanent magnet synchronous motors, where the differences in their axial parameters are negligible. Therefore, the motor terminal voltage and the extended back electromotive force are of the same order of magnitude in both amplitude and direction. Thus, considering only the differences between the two stability requirements, it can be roughly assumed from the voltage equation that… = , = That is, the motor voltage and the extended back EMF are of the same order of magnitude and in the same direction. Then we have... , .

[0062] Therefore, the above stability condition can be simplified as follows: Traditional sliding mode observer: ; The improved sliding mode observer of this application: ; Although this invention increases the preceding current term The gain G is multiplied by a factor of k, but as mentioned earlier, its value is much smaller than the subsequent voltage term. Therefore, even after increasing it, it is still smaller than the voltage term, and the dominant term for G remains the voltage term. Thus, the gain G required for the stability condition of this invention is comparable to that of a traditional sliding mode observer, and the introduction of the adjustment factor k does not significantly increase G.

[0063] The following experiments will further verify this, using an actual resistance R=45. ,inductance = =44 Comparative verification was performed on the joint motor, with the algorithm sampling period being T=50. All have a gain of G=30, and the control law is in saturation function form. Then the coefficients... = 1.136. For the traditional sliding mode algorithm (i.e., k=1), the coefficients are too large at this point, resulting in severe chattering during current tracking, making convergence difficult. The EMF observations jump back and forth between G values, rendering the results unusable. For example... Figure 4-a and Figure 4-b As shown, where Figure 4-a for Actual shaft current & traditional sliding mode observer tracking shaft current; Figure 4-b For observations by a traditional sliding mode observer .

[0064] For the improved sliding mode observer of this invention, when k=0.02, its current can track the actual current very well, the observed EMF curve is smooth, the jitter is very small, and the signal quality is very high. Figure 5-a and Figure 5-b As shown, where Figure 5-a for Actual shaft current & tracking of the improved sliding mode observer in this application shaft current; Figure 5-b For observations by the improved sliding mode observer .

[0065] Then, by slightly filtering the EMF value of this invention, for example, setting the filter cutoff frequency to 2300 rad / s in this experiment, it can be input into the next PLL stage to obtain the motor angle and speed. Figure 6-a ,6-b As shown in 6-c and 6-d, Figure 6-a For actual electrical angle & observed electrical angle; Figure 6-b The electrical angle error of the observation; Figure 6-c For actual speed and observed speed; Figure 6-d The velocity tracking effect is observed at 50Hz.

[0066] As can be seen, the observed electrical angle almost coincides with the actual electrical angle (the actual electrical angle was measured by the encoder, which has been calibrated and time-compensated). Calculations show that the mechanical angle observed by this invention (obtained by dividing the electrical angle by the number of pole pairs; the motor used in this experiment has 10 pole pairs) has an average absolute error of only 0.025 degrees compared to the actual angle, with a maximum deviation of only 0.11 degrees. The observed speed is consistent with the actual speed, with a noise standard deviation of 0.78 rad / s, while the speed tracking bandwidth at this point can reach over 100 Hz. It is particularly important to note that this data is a simplified measurement only to illustrate the advantages of the sliding mode observer in this application; the sliding mode control law and phase-locked loop have not been optimized. In practical applications, the sliding mode control law and phase-locked loop can be deeply optimized simultaneously for better application results.

[0067] The above experiments have fully demonstrated that, under the same sampling period, the same sliding mode gain, and the same control law form, traditional sliding mode observers are prone to current chattering and unusable observation results under low inductance motor conditions. However, the improved sliding mode observer of this application effectively suppresses numerical oscillations in the discrete iteration process by introducing a proportional adjustment coefficient k, enabling the observed current to stably track the actual current and extending the back EMF waveform to be smooth and reliable, thereby providing high-quality input for subsequent phase-locked loop calculation of rotor position and speed.

[0068] Finally, it should be noted that this application only introduces a proportional adjustment coefficient k into the voltage-related terms and sliding mode control terms in the discrete sliding mode observation equation to synchronously weight and adjust the related terms, thereby achieving overall constraint on the discrete iterative characteristics of the observer. However, the core idea of ​​this application is not limited to using only a single adjustment factor or only acting on the aforementioned specific terms. Without deviating from the basic principles of this application, those skilled in the art can also introduce one or more adjustment factors into other terms of the observation equation, or use different adjustment factors to weight and adjust the terms corresponding to different physical quantities, based on specific motor parameters, sampling periods, control objectives, or numerical stability requirements, to further refine the constraint and optimization of the dynamic characteristics of the discrete system. For example, independent adjustment coefficients can be set for the current state term, resistance-related term, or coupling term, or the proportional adjustment coefficient can be extended to an adaptive parameter that changes with the operating state. All of the above variations are equivalent extensions of the technical ideas of this application.

[0069] The sliding mode observation system for sensorless control of permanent magnet synchronous motors provided in this application is described in detail below. Please refer to [link / reference]. Figure 7 , Figure 7 An embodiment of the sliding mode observation system for sensorless control of a permanent magnet synchronous motor provided in this application includes: The acquisition unit 701 is used to acquire the actual current component and the actual voltage component of the motor in a two-phase stationary coordinate system. Discrete unit 702 is used to establish the current state equation based on the physical parameters of the motor, and to discretize the current state equation according to the sampling period to obtain the initial discrete sliding mode observation equation. The construction unit 703 is used to determine the proportional adjustment coefficient k based on the inductance parameters of the motor and the sampling period. The proportional adjustment coefficient k is used to weight and adjust the voltage-related terms and sliding mode control terms in the initial discrete sliding mode observation equation to construct the improved sliding mode observation equation. The calculation unit 704 is used to iteratively calculate the current observation based on the observed current, actual voltage components and sliding mode control law of the previous moment according to the improved sliding mode observation equation, and drive the observed current to track the actual current components. Extraction unit 705 is used to extract extended back electromotive force based on the output of sliding mode control law, wherein the contribution ratio of the high-frequency chattering component of the sliding mode control law output in the extended back electromotive force is weakened by a proportional adjustment coefficient. The control unit 706 is used to input the extended back EMF into the phase-locked loop module to calculate the rotor position and speed of the motor, and generate control signals accordingly to realize closed-loop control of the motor.

[0070] In this embodiment, the functions of each unit are the same as described above. Figures 1 to 3 The steps in the method embodiments shown correspond to those in the examples, and will not be repeated here.

[0071] This application also provides a sliding mode observation device for sensorless control of permanent magnet synchronous motors. Please refer to [link to relevant documentation]. Figure 8 , Figure 8 One embodiment of the sliding mode observation device for sensorless control of permanent magnet synchronous motors provided in this application includes: Processor 801, memory 802, input / output unit 803, bus 804; The processor 801 is connected to the memory 802, the input / output unit 803, and the bus 804; The memory 802 stores a program, and the processor 801 calls the program to execute any of the sliding mode observation methods for sensorless control of permanent magnet synchronous motors as described above.

[0072] This application also relates to a computer-readable storage medium storing a program that, when run on a computer, causes the computer to execute any of the sliding mode observation methods for sensorless control of permanent magnet synchronous motors described above.

[0073] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0074] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between apparatuses or units, and may be electrical, mechanical, or other forms.

[0075] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0076] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0077] 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 this application, in essence, or the part 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 to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A sliding mode observation method for sensorless control of permanent magnet synchronous motors, characterized in that, The sliding mode observation method includes: Obtain the actual current and voltage components of the motor in a two-phase stationary coordinate system; A current state equation based on the physical parameters of the motor is established, and the current state equation is discretized according to the sampling period to obtain the initial discrete sliding mode observation equation. Based on the inductance parameters of the motor and the sampling period, a proportional adjustment coefficient k is determined. The voltage-related terms and sliding mode control terms in the initial discrete sliding mode observation equation are weighted and adjusted using the proportional adjustment coefficient k to construct an improved sliding mode observation equation. The proportional adjustment coefficient k is a value greater than 0 and less than 1. Based on the improved sliding mode observation equation, the current observation is iteratively calculated based on the observed current at the previous moment, the actual voltage component, and the sliding mode control law, and the observed current is driven to track the actual current component. The extended back electromotive force is extracted based on the output of the sliding mode control law, wherein the contribution ratio of the high-frequency chattering component of the sliding mode control law output to the extended back electromotive force is weakened by the proportional adjustment coefficient. The extended back EMF input phase-locked loop module calculates the rotor position and speed of the motor, and generates control signals accordingly to achieve closed-loop control of the motor.

2. The sliding mode observation method according to claim 1, characterized in that, The determination of the proportional adjustment coefficient k based on the inductance parameters of the motor and the sampling period includes: Calculate the ratio of the sampling period to the inductance parameter of the motor to obtain the initial weighting coefficients corresponding to the sliding mode control quantity in the discrete sliding mode observation equation; Based on the influence of the initial weighting coefficient on the observed current change amplitude or the extended back EMF waveform during the discrete iterative calculation, determine whether the initial weighting coefficient exceeds the numerical convergence boundary. If so, a value greater than 0 and less than 1 is selected as the scaling factor k, and the initial weight coefficient is scaled using the scaling factor k so that the scaled weight coefficient is constrained within the numerical convergence boundary.

3. The sliding mode observation method according to claim 2, characterized in that, After selecting a value greater than 0 and less than 1 as the scaling factor k, and using the scaling factor k to scale the initial weight coefficient, the sliding mode observation method further includes: Based on the fact that the discrete sliding mode observation equation satisfies numerical convergence, the value of the proportional adjustment coefficient k is further reduced to reduce the high-frequency chattering component contained in the extended back electromotive force until the amplitude of the high-frequency chattering component is lower than the preset signal noise tolerance threshold.

4. The sliding mode observation method according to claim 1, characterized in that, The extraction of extended back electromotive force based on the output of the sliding mode control law includes: Obtain the real-time output value of the sliding mode control law in the improved sliding mode observation equation, as well as the actual voltage components in the two-phase stationary coordinate system; The extended back electromotive force is calculated by weighting and summing the actual voltage component and the real-time output value of the sliding mode control law using the proportional adjustment coefficient k, so that the main amplitude of the extended back electromotive force is provided by the actual voltage component without chattering. The formula for the weighted summation calculation satisfies: ， ； in, and They are respectively shaft and The extended back electromotive force of the shaft, , They are respectively shaft and The actual voltage component of the shaft, , This is the real-time output value of the sliding mode control law; Before the extended back EMF input phase-locked loop module calculates the rotor position and speed of the motor, the sliding mode observation method further includes: The extended back EMF is input to a first-order low-pass filter for filtering, and the cutoff frequency of the first-order low-pass filter is higher than the fundamental frequency of the motor.

5. The sliding mode observation method according to claim 1, characterized in that, The initial discrete sliding mode observation equation is: ； in, , The sampling period is... For motor resistance, , These are the equivalent inductance parameters of the motor along the d-axis and q-axis in a synchronous rotating coordinate system, respectively. Electric angular velocity, , They are respectively shaft and The actual voltage component of the shaft, , They are respectively shaft and The observed current of the axis, , is the real-time output value of the sliding mode control law, and n is used to represent the discrete time step index.

6. The sliding mode observation method according to claim 5, characterized in that, The improved sliding mode observation equation is as follows: ； in, k is the proportional adjustment coefficient. The sampling period is... For motor resistance, , These are the equivalent inductance parameters of the motor along the d-axis and q-axis in a synchronous rotating coordinate system, respectively. Electric angular velocity, , They are respectively shaft and The actual voltage component of the shaft, , They are respectively shaft and The observed current of the axis, , is the real-time output value of the sliding mode control law, and n is used to represent the discrete time step index.

7. The sliding mode observation method according to any one of claims 1 to 6, characterized in that, The step of calculating the rotor position and speed of the motor by inputting the extended back EMF into the phase-locked loop module, and generating control signals accordingly to achieve closed-loop control of the motor, includes: The angle error signal characterizing the rotor position estimation error is calculated based on the extended back electromotive force; The angle error signal is input into the proportional-integral controller, and the steady-state error is eliminated by the integral action of the proportional-integral controller to obtain the estimated speed of the motor. The rotor position electrical angle of the motor is obtained by performing discrete integration on the estimated rotational speed value. By using the rotor position electrical angle, a rotating coordinate transformation is performed on the actual current components in the two-phase stationary coordinate system to obtain the current feedback value in the synchronous rotating coordinate system. Based on the difference between the preset current command value and the current feedback value, the voltage command value in the synchronous rotating coordinate system is calculated by the current loop controller; The voltage command value is subjected to inverse coordinate transformation and space vector pulse width modulation operation using the rotor position electrical angle to generate a PWM control signal for driving the motor.

8. The sliding mode observation method according to any one of claims 1 to 6, characterized in that, The sliding mode control law adopts a saturation function, a sign function, a continuous function, or a hyperbolic tangent function.

9. A sliding mode observation system for sensorless control of a permanent magnet synchronous motor, characterized in that, The sliding mode observation system includes: The acquisition unit is used to acquire the actual current component and actual voltage component of the motor in a two-phase stationary coordinate system. Discrete unit, used to establish current state equation based on the physical parameters of the motor, and to discretize the current state equation according to the sampling period to obtain the initial discrete sliding mode observation equation; The construction unit is used to determine the proportional adjustment coefficient k based on the inductance parameters of the motor and the sampling period, and to use the proportional adjustment coefficient k to perform weighted adjustment on the voltage-related terms and sliding mode control terms in the initial discrete sliding mode observation equation to construct the improved sliding mode observation equation. The calculation unit is used to iteratively calculate the current observation based on the observed current at the previous moment, the actual voltage component, and the sliding mode control law according to the improved sliding mode observation equation, and drive the observed current to track the actual current component. An extraction unit is used to extract extended back electromotive force based on the output of the sliding mode control law, wherein the contribution ratio of the high-frequency chattering component of the sliding mode control law output to the extended back electromotive force is weakened by the proportional adjustment coefficient. The control unit is used to input the extended back EMF into the phase-locked loop module to calculate the rotor position and speed of the motor, and generate control signals accordingly to realize closed-loop control of the motor.

10. A sliding mode observation device for sensorless control of a permanent magnet synchronous motor, characterized in that, The sliding mode observation device includes: Processor, memory, input / output units, and bus; The processor is connected to the memory, the input / output unit, and the bus; The memory stores a program, which the processor invokes to perform the method as described in any one of claims 1 to 8.

Images