Adaptive sliding-mode observer control method based on novel composite approach rate

Abstract

The invention discloses a self-adaptive sliding-mode observer control method based on a novel composite approach rate, and belongs to the field of direct-current synchronous motor control. According to the method, a composite reaching law is constructed based on an exponential function and a power function, and when a system state is far away from a sliding mode surface, the composite reaching law converges towards the sliding mode surface through an exponential reaching item and a power tangent saturation reaching item; when the system state is close to the sliding mode surface, the composite reaching law converges towards the sliding mode surface through an exponential reaching item and a power hyperbolic tangent reaching item, and the method can restrain buffeting of the motor under the high-speed working condition and improve the convergence speed and the control precision. Furthermore, the invention further discloses a full-speed-section position-sensorless compound control method, in the low-speed operation stage of the permanent magnet synchronous motor, pulsating high-frequency voltage injection control is utilized, and the self-adaptive sliding-mode observer is adopted for control in the medium-high-speed stage. Meanwhile, a weighted average switching strategy is used for designing an optimal switching interval between a medium-high speed algorithm and a low-speed algorithm, and efficient and stable control over the full-speed section is achieved.

Description

Technical Field

[0001] The invention belongs to the field of direct current synchronous motor control, and in particular relates to an adaptive sliding mode observer control method based on a novel composite approach rate. Background Art

[0002] Permanent magnet synchronous motor (PMSM) plays an important role in electric vehicles, robots, rail transit and other fields due to its advantages such as low inertia, high performance, high power density and economy. However, traditional sensory control strategies face many challenges in harsh working environments, such as poor anti-interference ability, high cost and fragility, which limit the widespread deployment of motors in high-performance applications. Therefore, position sensorless control technology has gradually been favored by researchers because it does not rely on external sensors.

[0003] In order to improve the control performance of PMSM and overcome various interference factors, researchers have proposed a variety of nonlinear control strategies, such as model predictive control, adaptive control, neural network control, fuzzy control, and anti-disturbance control. Among these strategies, the sliding mode observer (SMO) has been widely used in the field of motor control because of its easy implementation, low requirement for model accuracy, and excellent robustness. However, the jitter problem in the SMO algorithm, that is, the high-frequency fast switching caused by the discontinuity of the switching function, not only generates ripples in the back electromotive force of the motor system, but also causes errors in the rotor position and speed estimation. To solve this problem, the academic community has proposed a variety of improved SMO forms to improve control accuracy and system stability. Scholars have explored different variants such as high-order SMO, non-singular terminal SMO, and fractional-order SMO, hoping to overcome the jitter problem and improve motor control accuracy and convergence speed through algorithm innovation, but they often require complex system design and a lot of calculations, which makes it difficult to implement in practical applications. Therefore, future research needs to consider how to simplify the design and reduce the computational burden while ensuring algorithm performance, so as to promote the practicality and popularity of these advanced control strategies.

[0004] In recent years, in the research of sliding mode observer (SMO), improved reaching law has been widely used because of its simplicity, efficiency and easy implementation. These improvements are aimed at solving the chattering problem in the traditional SMO algorithm and accelerating the convergence speed of the system. Gao et al. proposed an exponential reaching law (ERL) to improve the convergence speed, but when the system state is far away from the sliding mode surface, the convergence time is long, which limits its application in high-precision control. Devika et al. further proposed the power rate exponential reaching law (PRERL), which has made certain progress in reducing chattering and shortening the convergence time. Wang et al. further improved the convergence speed by introducing the power reaching law. Due to the existence of the sign switching function, the chattering problem is still not effectively suppressed. Iqbal et al. proposed a reaching law containing a sigmoid function, which effectively reduced the chattering phenomenon, but its convergence speed needs to be improved. Zhu et al. designed a terminal sliding mode control algorithm based on the reaching law, which significantly improved the convergence speed when the system is far away from the equilibrium point. However, when the system approaches the equilibrium point, the convergence process slows down and the system design becomes more complicated. Hou et al. proposed a hypertorsion function reaching law, which improves the robustness of the system by reducing the switching amplitude, but its performance is greatly affected by the selection of its parameters and is not easy to implement. Wang et al. proposed a sliding mode speed controller based on a new exponential reaching law, and designed a control method based on an extended sliding mode disturbance observer to improve the anti-interference ability of the system. In summary, although some progress has been made in improving the reaching law of SMO, there are still certain limitations. Therefore, how to find a better balance between improving the convergence speed and reducing jitter to achieve a more efficient and reliable SMO control strategy is an urgent problem to be solved. Summary of the invention

[0005] In view of the defects of the prior art, the present invention proposes a new composite reaching law (New Composite Reaching Law, NCRL) based on piecewise function, and combines this reaching law with a phase-locked loop to design a fast adaptive sliding mode observer (Fast Adaptive Sliding Mode Observer, FASMO) to resolve the contradiction between system convergence speed and jitter and enhance system robustness.

[0006] In order to achieve the above object, the present invention provides an adaptive sliding mode observer control method based on a novel composite approach rate, comprising the following steps:

[0007] Construct a composite reaching law based on exponential function and power function. When the absolute value of the error between the estimated state variable and the actual state variable is greater than the switching parameter, that is, when the system state is far from the sliding surface, the composite reaching law converges to the sliding surface with an exponential approaching term and a power tangent saturation approaching term; conversely, when the system state is close to the sliding surface, the composite reaching law converges to the sliding surface with an exponential approaching term and a power hyperbolic tangent approaching term.

[0008] Furthermore, the composite reaching law based on exponential function and power function is as follows:

[0009]

[0010] where: c is the linear gain and c > 0; K is the switching gain and K > 0; h is the adjustment parameter, 0 < h < 1; s is the error between the estimated state variable and the actual state variable; cs is the exponential approaching term, K|s h M(s) is the power approaching term; M(s) is the switching function;

[0011] The switching function is:

[0012]

[0013] where: a is the switching parameter and a > 0; H(s) is the tangent saturation function; tanh(s) is the hyperbolic tangent function;

[0014] Therefore:

[0015]

[0016] When the system state is far from the sliding surface, the reaching law is where the exponential approaching term cs plays a key role, accelerating the convergence of the system state to the sliding surface through its fast response characteristic; and when the system state is close to the sliding surface, that is, s → 0, the reaching law is where the power approaching term plays a key role, slowing down the convergence speed to the sliding surface through the hyperbolic tangent function.

[0017] Furthermore, the piecewise switching function is specifically:

[0018]

[0019] where: e is the boundary layer thickness and e > 0; λ = arctan(1).

[0020] Furthermore, the mathematical model of the permanent magnet synchronous motor is:

[0021]

[0022] where: R is the phase resistance; L s is the phase inductance; iα 、i β are the stator currents in the α-β coordinate system; E α 、E β are the stator back electromotive force on the α-axis and β-axis respectively; u α 、u β are the stator voltages on the α-axis and β-axis, ω e is the speed information; θ e is the rotor position; ψ f is the permanent magnetic flux linkage;

[0023] The sliding surface expression is:

[0024]

[0025] in: ix. are the observed value, actual value and estimated error value of the stator current of the permanent magnet synchronous motor respectively; wherein the value of x is α, β;

[0026] For the sliding surface Based on the composite reaching law based on the exponential function and the power function, we get:

[0027]

[0028] in,

[0029] When the system state variable approaches the sliding surface, that is, s→0, According to the equivalent control principle of sliding mode control, The expanded back electromotive force is:

[0030]

[0031] According to the mathematical model of the permanent magnet synchronous motor and the extended back electromotive force, the rotor position and speed information of the permanent magnet synchronous motor can be obtained.

[0032] Furthermore, the adaptive sliding film mode observer adopts a phase-locked loop (PLL) to replace the original arctan(x) to obtain the permanent magnet synchronous motor rotor position and speed information.

[0033] Further, the phase-locked loop includes a phase detector, a loop filter and a voltage-controlled oscillator connected in sequence, and the loop filter is composed of a PI controller;

[0034] The output of the voltage-controlled oscillator is the speed information of the motor.

[0035] The speed information of the motor After the integrator, the rotor position of the motor is obtained

[0036] The phase-locked loop transfer function is:

[0037]

[0038] in: G is the gain coefficient of the phase detector; K i , K p is the proportional coefficient of the PI controller.

[0039] The present invention also provides a full-speed position sensorless composite control method based on a novel composite reaching law, based on the adaptive sliding mode observer control method based on the novel composite reaching rate as described above, comprising the following steps:

[0040] (1) When the permanent magnet synchronous motor is running at low speed, pulse high-frequency voltage injection control is used to estimate the rotor position and speed information of the permanent magnet synchronous motor;

[0041] (2) When the permanent magnet synchronous motor is running at medium and high speed, the adaptive sliding mode observer control based on the new composite reaching rate is used to complete the estimation of the permanent magnet synchronous motor rotor position and speed information;

[0042] (3) The weighted average switching method is used to design the optimal switching intervals of the two algorithms in step (1) and step (2) to realize position sensorless control of the permanent magnet synchronous motor in the full speed range.

[0043] Furthermore, the step (1) is specifically as follows:

[0044] (1.1) A high-frequency voltage signal is injected into the direct axis of the estimated two-phase rotating coordinate system. Under the excitation of the high-frequency voltage signal, a high-frequency current component is generated, and the high-frequency current component of the quadrature axis (q axis) is extracted.

[0045] (1.2) High-frequency current component With the same frequency sinusoidal high frequency signal sin(ω h t) multiply to perform signal modulation;

[0046] (1.3) The modulated high-frequency current component passes through a frequency of ω h A band-pass filter (BPF) is used to filter out the carrier frequency signal and the fundamental frequency signal;

[0047] (1.4) The signal after bandpass filtering is low-pass filtered (LPF) to extract the low-frequency signal related to the rotor position and speed, and output the error function f v (△θ e ), after the rotor position is tracked, the rotor position and speed information of the permanent magnet synchronous motor are estimated.

[0048] Further, the rotor position tracking includes a PI regulator and an integrator;

[0049] The error function f v (△θ e ) The motor speed information is obtained through the PI regulator

[0050] The speed information of the motor After the integrator, the rotor position of the motor is obtained

[0051] Furthermore, the step (3) is specifically as follows:

[0052] (3.1) Define the switching interval [ω1, ω2], where: ω1 is the lower speed limit of the transition zone switching, ω2 is the upper speed limit of the transition zone switching;

[0053] (3.2) The switching function is defined as:

[0054]

[0055] They are respectively the permanent magnet synchronous motor speed information and rotor position estimated by the pulse high frequency voltage injection method in step (1); They are respectively the permanent magnet synchronous motor speed information and rotor position estimated by the adaptive sliding mode observer in step (2).

[0056] Beneficial effects of the present invention:

[0057] 1. The present invention establishes a composite convergence law of a piecewise function. When the system error of the composite convergence law is greater than a switching parameter, the composite convergence law line converges to the sliding surface with an exponential convergence term and a power tangent saturation function. When the system error of the composite convergence law is less than or equal to the switching parameter, the composite convergence law converges to the sliding surface with an exponential convergence term and a power hyperbolic tangent convergence term. When the system state switches on the sliding surface, the tangent change of the switching point is small and gradual, so that even if the switching is fast, it is not easy to cause chattering. Therefore, the present invention can greatly alleviate the chattering phenomenon of the system when the state is adjusted, thereby improving the overall performance of the system.

[0058] 2. The adaptive sliding mode observer control method proposed in the present invention also adopts a phase-locked loop to replace the original arctan (x) method to obtain the motor rotor position and speed, further avoiding the error amplification caused by high-frequency jitter and improving the observation accuracy of the sliding mode observer.

[0059] 3. The present invention also provides a composite control strategy for the full-speed section. A pulse high-frequency voltage injection method is used for the low-speed section, and an adaptive sliding mode observer is used for the medium and high-speed sections to alleviate jitter and improve control accuracy. At the same time, a weighted average switching method is used to design the optimal switching interval between the medium and high-speed sections, so that the switching process is smoother, and the speed and rotor position estimation will not have large signal jumps, and the estimation accuracy is higher, thereby realizing position sensorless control of the permanent magnet synchronous motor in the full-speed section. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 The schematic diagram of the control structure of an adaptive sliding mode observer based on a novel composite reaching law according to an embodiment of the present invention.

[0061] Figure 2 Schematic diagram of a tangent saturation function curve according to an embodiment of the present invention.

[0062] Figure 3 Schematic diagram of various switching function curves according to embodiments of the present invention.

[0063] Figure 4 It is a schematic diagram of a segmented switching function curve according to an embodiment of the present invention.

[0064] Figure 5 The present invention is a schematic diagram of a phase-locked loop structure in an adaptive sliding mode observer according to an embodiment of the present invention.

[0065] Figure 6 FIG. 4 is a rotation speed estimation curve diagram of an embodiment of the present invention.

[0066] Figure 7 Graph showing rotor position and position error estimation according to an embodiment of the present invention.

[0067] Figure 8 It is a schematic diagram of a full-speed position sensorless composite control structure based on a novel composite reaching law according to an embodiment of the present invention.

[0068] Fig. 9 Schematic diagram of the pulse high-frequency voltage injection control structure according to an embodiment of the present invention.

[0069] Fig.10 It is a schematic diagram of the rotor position tracking structure in the pulse high-frequency voltage injection control according to an embodiment of the present invention.

[0070] Fig.11 The figure is a schematic diagram of the principle of the weighted average switching method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0071] In order to make the technical means, creative features, achieved objectives and effects of the present invention easy to understand, the following embodiments will specifically elaborate on the adaptive sliding mode observer control method, storage medium and system of the present invention in conjunction with the accompanying drawings. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation to the present invention.

[0072] An embodiment of the present invention provides an adaptive sliding mode observer control method based on a novel composite reaching law. As Figure 1 shown, the adaptive sliding mode observer consists of a sliding mode observer based on a composite reaching law of an exponential function and a power function and a phase-locked loop.

[0073] When the permanent magnet synchronous motor runs at medium and high speeds, in order to achieve fast convergence and reduce chattering, an exponential function and a power function are introduced into the sliding mode control framework in this embodiment of the present invention to construct a novel composite reaching law NCRL.

[0074] The novel composite reaching law is in the form of:

[0075]

[0076] where: c is a linear gain, and c>0; K is a switching gain, and K>0; h is an adjustment parameter, 0<h<1; s is the error between the estimated state variable and the actual state variable; cs is the exponential reaching term, K|s h M(s) is the power reaching term; M(s) is the switching function.

[0077] In order to solve the problem of chattering in the system caused by the high-frequency switching of the switching function sign(s) in the traditional sliding mode reaching law, some continuous switching functions are used to weaken the chattering phenomenon, such as the sigmoid(s) function and the tanh(s) hyperbolic tangent function, etc. However, the above functions achieve the suppression of chattering at the cost of the convergence time. In order to simultaneously satisfy fast convergence and reduce chattering, this embodiment of the present invention proposes a piecewise switching function M(s) based on the tangent saturation function H(s) and the hyperbolic tangent function tanh. This design aims to improve the stability and output smoothness of the system while maintaining a fast response.

[0078]

[0079] Substituting (2) into (1) gives:

[0080]

[0081] where: a is a switching parameter, and a>0; H(s) is the tangent saturation function; tanh(s) is the hyperbolic tangent function; K|shtanh(s) is the power hyperbolic tangent reaching term, and K|shH(s) is the power tangent saturation reaching term.

[0082] Furthermore, the switch function is specifically:

[0083]

[0084] Where: λ=arctan(1); ε is the boundary layer thickness, and e>0.

[0085] The switching function designed in this way can smooth the system output and reduce the system jitter. Figure 2 As shown in FIG. 1 , the characteristic curve of H(s) reveals the influence of boundary layer thickness on system performance. When the value of e is larger, the control curve is smoother and the system has stronger vibration suppression capability. On the contrary, when the value of e is smaller, the control curve becomes steeper and the convergence speed of the system state variables is faster. Therefore, the boundary layer thickness e needs to be balanced between control accuracy and convergence speed. Based on this analysis, the boundary layer thickness e=1.5 is selected in the embodiment of the present invention.

[0086] During the sliding film control process, the dynamic behavior of the system state variables can be divided into two main stages: away from and close to the sliding film surface. When the system state is in the stage away from the sliding surface, that is, when the absolute value of the error between the estimated state variables and the actual state variables is greater than the switching parameter (|s|>a), the reaching law is adopted. To accelerate the convergence speed of the system. This reaching law contains the exponential reaching term cs, whose remarkable fast response characteristics help to promote the system state to converge quickly to the sliding surface. At the same time, the power tangent saturation reaching term K|s h The introduction of H(s) not only ensures the rapidity of the convergence process, but also effectively reduces the chattering phenomenon of the system. As the system state gradually approaches the sliding surface, that is, when the absolute value of the error between the estimated state variable and the actual state variable is less than or equal to the switching parameter (|s|≤a), the convergence law is adjusted to At this stage, the exponential approach term cs continues to converge faster, and the power hyperbolic tangent approach term K|s h The role of tanh(s) becomes particularly critical. It helps to smooth the control process, reduce system jitter, and improve control accuracy by appropriately slowing down the convergence speed. In addition, the adjustment parameter h introduced in the power approach term provides a flexible adjustment mechanism for the control gain. By adjusting the value of h, the control logic can be simplified and more precise control of the system state can be achieved. Figure 3 Describes the function curves of various switching functions sign(s), H(s), tanh(s), sigmoid(s), Figure 4 The function curve of the segmented switching function M(s) is given. Figure 3As can be seen from the figure, as s→0, sign(s) changes in a step-like manner. This sudden change easily causes the system to vibrate. H(s) can quickly converge to 0 in the boundary layer, but its smoothness needs to be improved. Tanh(s) and sigmoid(s) can converge smoothly to 0 in the boundary layer, but their convergence speed will gradually slow down as the distance from the sliding membrane surface increases. Therefore, from Figure 4 It can be seen that the piecewise switching function M(s) is proposed based on the H(s) and tanh(s) functions, which can not only ensure the convergence within a finite time, but also effectively suppress the chattering phenomenon.

[0087] In the embodiment of the present invention, a positive real number switching parameter a is introduced to realize the dynamic adjustment of the composite convergence law (NCRL). Figure 3 As shown in the figure, when e = 1.5, the horizontal coordinate of the intersection of the tanh(s) function and H(s) is s = 1.377, which means that the switching parameter a = 1.377. Therefore, the new piecewise reaching law (NCRL) can be further divided into three stages according to the distance between the system state variable and the synovial surface: the early stage away from the synovial surface, the late stage away from the synovial surface, and the stage close to the synovial surface. The expression of the new piecewise reaching law (NCRL) is:

[0088]

[0089] Taking s≥0 as an example, the same is true for s<0. In the early stage away from the synovial front, that is, when |s|≥1.5, the convergence law is defined as At this time, the exponential approach term cs and the power approach term K|s h The superposition effect effectively accelerates the convergence speed of the system state variables. When entering the late stage away from the synovial surface, that is, 1.377≤|s|≤1.5, the convergence law changes to The exponential approach term cs continues to accelerate convergence, while the power tangent saturation approach term This ensures fast convergence while slowing down the convergence speed, effectively reducing the system chattering phenomenon. When the system state variable approaches the sliding membrane surface, that is, |s|<1.377, the reaching law is adjusted to The exponential approximation term cs maintains the fast convergence property, while the power hyperbolic tangent approximation term K|s h tanh(s) further slows down the convergence speed, improves the smooth control process, and significantly reduces system chattering.

[0090] This new piecewise convergence law (NCRL) fully utilizes the advantages of the piecewise switching function based on tanh(s) and H(s), which not only ensures the rapid convergence of the system, but also effectively suppresses the chattering phenomenon and improves the system control performance, providing a new solution for high-speed sliding film observer control in motors.

[0091] The chattering boundary width of the new composite reaching law (NCRL) is analyzed below and compared with the exponential reaching law (ERL) and the exponential sigmoid reaching rate (ESRL) to evaluate the performance of NCRL when the system state variables are close to the sliding surface.

[0092] When the system state variables are close to the sliding surface, that is, s→0, NCRL can be simplified as:

[0093]

[0094] In order to analyze the chattering level of NCRL, its discretization expression is:

[0095] s(t+1)-s(t)≈-K|s h tanh(s(t))T s -cs(t)T s (7)

[0096] Where: 0 <T s <1 is the sampling period.

[0097] When s(t)→0 + When , the next cycle can be derived as:

[0098] s(t+1)≤-K|s h tanh(0 + )T s -c·(0 + )·T s (8)

[0099] Similarly, when s(t)→0 - hour:

[0100] s(t+1)<-K|s h tanh(0 - )T s -c·(0 - )·T s (9)

[0101] Based on the above analysis, the chattering boundary width of NCRL is calculated and expressed as:

[0102] △ c1 =2K|s h tanh(0 + )T s +2c·(0 + )·T s (10)

[0103] Similar derivations are performed for ERL and ESRL. When s→0, ERL can be simplified to:

[0104]

[0105] The chattering boundary width of ERL is:

[0106] △ c2 =2KT s (12)

[0107] Similarly, the chattering boundary width of ESRL can be derived as:

[0108] △ c3 =2Ksigmoid(0 + )T s (13)

[0109] By comparing formula (10) and (13), it can be concluded that NCRL performs better than ESRL in terms of stable convergence of system state. At the same time, by observing formula (12), it is found that the chattering boundary width of ERL is fixed, in (-KT s ,KT s ) interval and cannot converge within a limited time.

[0110] The mathematical model of permanent magnet synchronous motor in static coordinate system is:

[0111]

[0112] Where, R is the phase resistance; L S is the phase inductance; θ e is the rotor angular position; i α 、i β are the stator currents in the α-β coordinate system; E α 、E β are the stator back electromotive force on the α-axis and β-axis respectively; u α 、u β are the stator voltages on the α-axis and β-axis respectively; ω e is the rotor angular velocity; ψ f It is permanent magnetic flux linkage.

[0113] The sliding surface expression is:

[0114]

[0115] in: i x , They are the observed value, actual value and estimated error value of the stator current of the permanent magnet synchronous motor respectively; among which the values ​​of x are α and β.

[0116] The mathematical model of the adaptive sliding mode observer is constructed based on the mathematical model of the permanent magnet synchronous motor and the composite reaching law:

[0117]

[0118] in:

[0119]

[0120] Sliding surface Based on the current estimation error, formulas (16)-(14) are obtained:

[0121]

[0122] When the system state variables are close to the sliding surface, that is, s→0, According to the equivalent control principle of sliding mode control, The expanded back electromotive force is:

[0123]

[0124] In traditional sliding mode observers, the inverse trigonometric function method arctan(x) is widely used to estimate the state of the motor rotor. However, this method has a division operation in the calculation, which increases the system's sensitivity to jitter, especially in the high-frequency range, and also amplifies the error, affecting the accuracy of the rotor position estimation. To address this problem, an embodiment of the present invention also proposes an improved solution based on a phase-locked loop. Phase-locked loop technology has excellent frequency and phase tracking capabilities, and can accurately estimate the rotor position and speed of the motor. Compared with the traditional sliding mode observer, the improved system can automatically adjust and lock the actual phase of the motor rotor, reducing the need for phase compensation.

[0125] In the embodiment of the present invention, the phase-locked loop (PLL) system is mainly composed of a phase detector (PD), a voltage-controlled oscillator (VOC) and a loop filter (LF). Figure 5 shown.

[0126] Back electromotive error △E:

[0127]

[0128] Among them, θ e is the motor rotor position, To estimate the motor rotor position, ω e is the motor angular velocity, To estimate the motor angular velocity.

[0129] When △E is small enough, that is, It can be considered that In this case, formula (20) can be simplified as:

[0130]

[0131] in:

[0132] The output of the voltage-controlled oscillator is the motor speed information ω e ; Motor speed information ω e After passing through the integrator, the rotor position θ of the motor is obtained e ;The overall transfer function of the phase-locked loop is:

[0133]

[0134] in: G is the gain coefficient of the phase detector; K i , K p is the proportional coefficient of the PI controller.

[0135] In order to verify the effectiveness of the synovial observer control method based on the new composite reaching rate, a simulation comparison is made between it and the synovial observer control method based on the exponential reaching law and the synovial observer control method based on the exponential sigmoid reaching rate. The SMO based on the new composite reaching law NCRL+PLL is named NCSMO, the SMO based on the exponential reaching law ERL+arctan is named ESMO, and the SMO based on the exponential sigmoid reaching rate ESRL+arctan is named ESSMO.

[0136] In the simulation, the parameters of PMSM are shown in Table 1.

[0137] Table 1

[0138]

[0139] A. Motor starting and load speed test

[0140] This simulation tests the no-load start response performance of the three control strategies at a reference speed of 1000r / min, and adds a load torque of 8Nm at t=0.05s. The estimated speed of the motor is as follows Figure 6 The detailed comparison results of the response parameters of the three control methods are listed in Table 2. By comparison, NCSMO exhibits fast adaptation, smaller vibration range, efficient overshoot constraint and strong anti-interference ability.

[0141] Table 2

[0142]

[0143] B. Motor rotor position test

[0144] This experiment tested the rotor position and rotor position error of three control strategies at a reference speed of 1000r / min. Figure 7 (a)-(c) show the rotor position measurement results of the three control strategies ESMO, ESSMO, and NCSMO, respectively. In the early stage of motor startup, the rotor position estimation of ESMO and ESSMO shows large fluctuations, while the waveform of the actual rotor position of NCSMO is highly consistent with the estimated position. Further analysis of the rotor position estimation error in the figure shows that in the motor startup stage, the rotor position estimation error of NCSMO is more accurate and the error fluctuation frequency is more stable. Therefore, NCSMO has a greater advantage in the accuracy of rotor position estimation.

[0145] In summary, by comparing the simulation experimental results of the ESMO and ESSMO control strategies on the speed and rotor position under motor startup and sudden load conditions, NCSMO shows a higher advantage. Specifically, NCSMO not only has a fast convergence speed and a small amplitude range, but also has high overshoot constraint and anti-interference performance. In addition, it also performs well in speed and position tracking accuracy. These characteristics together prove the effectiveness of NCSMO in motor control.

[0146] like Figure 8 As shown, the embodiment of the present invention also provides a full-speed position sensorless composite control method based on a new composite reaching law, comprising the following steps:

[0147] S101. When the permanent magnet synchronous motor runs at low speed, the pulse high-frequency voltage injection method is used to complete the permanent magnet synchronous motor rotor position and speed estimation.

[0148] like Fig. 9 As shown in the figure, the pulse high-frequency voltage injection method is specifically:

[0149] (1) A high-frequency voltage signal is injected into the direct axis of the estimated two-phase rotating coordinate system. Under the excitation of the high-frequency voltage signal, two high-frequency current components are generated on the direct axis (d axis) and the quadrature axis (q axis). The d-axis current is aligned with the magnetic flux, and the q-axis current is directly related to the motor torque. Therefore, the high-frequency current component of the quadrature axis (q axis) is extracted.

[0150] (2) High-frequency current component With the same frequency sinusoidal high frequency signal sin(ω h t) are multiplied to perform signal modulation.

[0151] (3) The modulated high-frequency current component passes through a frequency of ω h A band-pass filter (BPF) is used to filter out the carrier frequency signal and the fundamental frequency signal.

[0152] (4) The signal after bandpass filtering is low-pass filtered (LPF) to extract the low-frequency signal related to the rotor position and speed, and output the error function f v (△θ e ), after the rotor position is tracked, the rotor position and speed information of the permanent magnet synchronous motor are estimated.

[0153] The error function is:

[0154]

[0155] When the error between the estimated rotor position and the actual position is small, that is, △θ e When sin(2△θ approaches zero infinitely e )≈2△θ e , we can get:

[0156]

[0157] Where: L = (L d +L q ) / 2 is the average inductance, △L=(L d -L q ) / 2 is the semi-differential inductance, L d , L q is the dq axis inductance component of the permanent magnet synchronous motor; is the rotor error angle, θ e is the actual position angle of the motor rotor, is the estimated position angle; ω h , U mh are the frequency and amplitude of the injected voltage signal respectively.

[0158] like Fig.10 As shown, the rotor position tracking includes a PI regulator and an integrator; the error function f v (△θ e ) The motor speed information is obtained through the PI regulator Motor speed information After the integrator, the rotor position of the motor is obtained

[0159] S102, when the permanent magnet synchronous motor is running at medium and high speed, an adaptive sliding mode observer is used to estimate the rotor position and speed of the permanent magnet synchronous motor.

[0160] S103, using a weighted average switching method to design the optimal switching intervals of the two algorithms of step S101 and step S102, so as to realize position sensorless control of the permanent magnet synchronous motor at full speed.

[0161] like Fig.11As shown in the figure, in order to realize the full-speed composite control, the speed interval is divided into a low-speed section, a medium-high speed section and a switching section [ω1, ω2]. The pulse high-frequency voltage injection method is used in the low-speed stage, and the adaptive SMO is used in the medium-high speed stage. In the switching section, the estimated rotor position and speed values ​​are weighted averaged by a linear function. The advantage of this method is that the switching process is smooth, and there will be no large signal jumps in the speed and rotor position estimation, and its estimation accuracy is higher.

[0162] (1) Define the switching interval [ω1, ω2], where ω1 is the lower speed limit of the transition zone switching, and ω2 is the upper speed limit of the transition zone switching.

[0163] (2) The switching function is defined as:

[0164]

[0165]

[0166] They are respectively the motor speed and rotor position estimated by the pulse high frequency voltage injection method in step S101; They are respectively the motor speed and the rotor position estimated by the adaptive sliding mode observer in step 102.

[0167] The above-mentioned implementation modes are preferred cases of the present invention and are not used to limit the protection scope of the present invention. Various deformations or modifications that can be made by ordinary technicians in this field without creative work within the scope of the attached claims are still within the protection scope of this patent.

Claims

1. An adaptive sliding mode observer control method based on a novel composite approach rate, characterized in that: A composite reaching law based on exponential function and power function is constructed. When the absolute value of the error between the estimated state variables and the actual state variables is greater than the switching parameter, that is, the system state is far away from the sliding surface, the composite reaching law converges to the sliding surface with the exponential reaching term and the power tangent saturation reaching term; conversely, when the system state is close to the sliding surface, the composite reaching law converges to the sliding surface with the exponential reaching term and the power hyperbolic tangent reaching term.

2. The adaptive sliding mode observer control method based on a novel composite reaching rate according to claim 1 is characterized in that: The composite reaching law based on exponential function and power function is as follows: Where: c is the linear gain and c > 0; K is the switching gain and K > 0; h is the adjustment parameter, 0 < h < 1; s is the error between the estimated state variable and the actual state variable; cs is the exponential reaching term, K|s h M(s) is the power reaching term; M(s) is the switching function; The switch function is: Where: a is the switching parameter, and a>0; H(s) is the tangent saturation function; tanh(s) is the hyperbolic tangent function; therefore: When the system state is far away from the sliding surface, the reaching law is The exponential approach term cs plays a key role in accelerating the convergence of the system state to the sliding surface through its rapid response characteristics. When the system state is close to the sliding surface, that is, s→0, the approach law is The power approach term plays a key role, slowing down the convergence speed to the sliding surface through the hyperbolic tangent function.

3. The adaptive sliding mode observer control method based on a novel composite reaching rate according to claim 2 is characterized in that: The segment switch function is specifically: Where: e is the boundary layer thickness, and e>0; λ=arctan(1).

4. The adaptive sliding mode observer control method based on a novel composite reaching rate according to claim 2 is characterized in that: The mathematical model of the permanent magnet synchronous motor is: Where: R is the phase resistance; L s is the phase inductance; i α 、i β are the stator currents in the α-β coordinate system; E α 、E β are the stator back electromotive force on the α-axis and β-axis respectively; u α 、u β are the stator voltages on the α-axis and β-axis, ω e is the speed information; θ e is the rotor position; ψ f is the permanent magnetic flux linkage; The sliding surface expression is: in: i x , are the observed value, actual value and estimated error value of the stator current of the permanent magnet synchronous motor respectively; wherein the value of x is α, β; For the sliding surface Based on the composite reaching law based on the exponential function and the power function, we get: in, When the system state variable approaches the sliding surface, that is, s→0, According to the equivalent control principle of sliding mode control, The expanded back electromotive force is: According to the mathematical model of the permanent magnet synchronous motor and the extended back electromotive force, the rotor position and speed information of the permanent magnet synchronous motor can be obtained.

5. The adaptive sliding mode observer control method based on a novel composite reaching rate according to claim 1 is characterized in that: The adaptive sliding film mode observer uses a phase-locked loop (PLL) instead of the original arctan(x) to obtain the rotor position and speed information of the permanent magnet synchronous motor.

6. The full-speed position sensorless composite control method according to claim 5 is characterized in that: The phase-locked loop comprises a phase detector, a loop filter and a voltage-controlled oscillator connected in sequence, and the loop filter is composed of a PI controller; The output of the voltage-controlled oscillator is the speed information of the motor. The speed information of the motor After passing through the integrator, the rotor position of the motor is obtained The phase-locked loop transfer function is: in: G is the gain coefficient of the phase detector; K i , K p is the proportional coefficient of the PI controller.

7. A full-speed position sensorless composite control method based on a novel composite reaching law, based on the novel composite reaching rate-based adaptive sliding mode observer control method according to any one of claims 1 to 6, characterized in that: The steps include: (1) When the permanent magnet synchronous motor is running at low speed, pulse high-frequency voltage injection control is used to estimate the rotor position and speed information of the permanent magnet synchronous motor; (2) When the permanent magnet synchronous motor is running at medium and high speed, the adaptive sliding mode observer control based on the new composite reaching rate is used to complete the estimation of the permanent magnet synchronous motor rotor position and speed information; (3) The weighted average switching method is used to design the optimal switching intervals of the two algorithms in step (1) and step (2) to realize position sensorless control of the permanent magnet synchronous motor in the full speed range.

8. The full-speed position sensorless composite control method based on the novel composite reaching law according to claim 7 is characterized in that: The step (1) is specifically: (1.1) A high-frequency voltage signal is injected into the direct axis of the estimated two-phase rotating coordinate system. Under the excitation of the high-frequency voltage signal, a high-frequency current component is generated, and the high-frequency current component of the quadrature axis (q axis) is extracted. (1.2) High-frequency current component With the same frequency sinusoidal high frequency signal sin(ω h t) multiply to perform signal modulation; (1.3) The modulated high-frequency current component passes through a frequency of ω h A band-pass filter (BPF) is used to filter out the carrier frequency signal and the fundamental frequency signal; (1.4) The signal after bandpass filtering is low-pass filtered (LPF) to extract the low-frequency signal related to the rotor position and speed, and output the error function f v (△θ e ), after the rotor position is tracked, the rotor position and speed information of the permanent magnet synchronous motor are estimated.

9. The full-speed position sensorless composite control method according to claim 8 is characterized in that: The rotor position tracking includes a PI regulator and an integrator; The error function f v (△θ e ) The motor speed information is obtained through the PI regulator The speed information of the motor After the integrator, the rotor position of the motor is obtained 10. The full-speed position sensorless composite control method according to claim 1, characterized in that: The step (3) is specifically: (3.1) Define the switching interval [ω1, ω2], where: ω1 is the lower speed limit of the transition zone switching, ω2 is the upper speed limit of the transition zone switching; (3.2) The switching function is defined as: They are respectively the permanent magnet synchronous motor speed information and rotor position estimated by the pulse high frequency voltage injection method in step (1); They are respectively the permanent magnet synchronous motor speed information and rotor position estimated by the adaptive sliding mode observer in step (2).

Images