An adaptive quasi-proportional resonant sensorless control method with parameter identification
By adopting the adaptive quasi-proportional resonant sensorless control method in permanent magnet synchronous motor, using the AQPR_PI control model and sliding mode observation function for parameter identification, the problem of low signal and noise at zero speed or extremely low speed is solved, and the control is improved Accuracy and robustness enhance control performance at different carrier ratios.
Patent Information
- Application Number
- CN202210837956.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-07-16
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2042-07-16
AI Technical Summary
When permanent magnet synchronous motors are at zero speed or extremely low speed, the signal-to-noise of the effective signal is relatively low, making it difficult to extract the speed and position information of the rotor, resulting in a decrease in control accuracy. The existing methods have poor control performance under different carrier ratios.
A self-adaptive quasi-proportional resonant sensorless control method with parameter identification is proposed. By establishing the current state equation of the permanent magnet synchronous motor on the stationary coordinate system, the extended back electromotive force estimate is obtained using the AQPR_PI control model, combining the sliding mode observation function and Popov ultra-stability theory performs online identification of stator resistance and inductance parameters, substitutes it into the AQPR_PI control model to calculate the estimated value of the extended back electromotive force to reduce system output error and jitter phenomenon.
It improves the robustness of system parameters and control accuracy, enhances the tracking accuracy and anti-interference ability under different carrier ratios and operating conditions, reduces the system output error and jitter phenomenon, and improves the accuracy of speed and position estimation.
Smart Images

Figure CN115208251B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of permanent magnet synchronous motor control, and in particular relates to an adaptive quasi-proportional resonance sensorless control method with parameter identification. Background Art
[0002] Permanent magnet synchronous motors are widely used in daily life due to their simple structure, small size, and low manufacturing cost. In order to achieve excellent control performance of permanent magnet synchronous motors (PMSM), it is necessary to accurately obtain the speed and position information of the motor. The traditional position detection method is to directly measure using mechanical sensors such as electronic or electromechanical sensors, such as Hall effect devices, resolvers, and optical encoders.
[0003] However, the installation of these mechanical sensors will increase the size, volume and weight of the system, which seriously limits the development of motor miniaturization. At the same time, some sensors have strict requirements on the working environment, the measurement accuracy is greatly affected by the environment, and the number of electrical component connections is increased, which reduces the anti-interference performance. Therefore, the research on sensorless control of PMSM has received widespread attention, and many control methods have been proposed, such as the extended back electromotive force method (EMF), flux observation method, extended Kalman filter method and high-frequency signal injection method. Among them, the EMF method is simple and reliable, and is widely used in medium and high-speed occasions. The control methods based on EMF mainly include sliding mode control, state observer and model reference adaptive control. Since sliding mode control does not require high system model accuracy and is insensitive to parameter changes and external interference, it is a special nonlinear control system with variable structure. Therefore, it can be used to estimate EMF and extract speed and position information from EMF. The prior art analyzes that at medium and high speeds, the low switching frequency cycle is close to the LC resonance point, which causes resonance phenomenon and leads to poor performance of motor speed and position estimation. In order to solve this problem, an accurate discrete sliding mode control based on an adaptive filter is proposed. The results show that this method can improve the speed and position estimation accuracy and enhance the robustness of the system to uncertainty interference. There is also an existing technology that linearizes the PMSM mathematical model using Jacobi linearization for the nonlinearity, uncertainty and disturbance of the PMSM mathematical model, designs an integral state feedback controller, and cascades the sliding mode observer (SMO) and the disturbance observer. Experimental results show that this method improves the sensorless control performance within a wide speed range. Another prior art analyzes the rotor position estimation error and concludes that the harmonics in the estimation signal are difficult to eliminate. Therefore, an adaptive notch filter is proposed to extract the fundamental wave of the position estimation signal from the output of the SMO system and apply it to the phase-locked loop (PLL), which can effectively adaptively compensate for the EMF harmonic error. In view of the nonlinearity of the inverter and the spatial harmonics of the flux linkage, the EMF estimation is inaccurate, which causes the problem of harmonic fluctuation of the rotor position.
[0004] However, the sign function introduced by the sensorless control method based on SMO will deteriorate the system output and aggravate the high-frequency jitter phenomenon. In addition, LPF will produce phase shift while filtering high-frequency signals, resulting in a larger system output error and reduced control accuracy. In order to solve the above problems and improve the control performance of PMSM, there is an existing technology that combines SMO with a phase locked loop (PLL). On the basis of the traditional SMO principle, in order to reduce high-frequency jitter, a switching function with a double boundary layer structure is used to replace the traditional sign function, and the position information is extracted using PLL. The experimental results show that this method has good tracking performance. There is also an existing technology that establishes an adaptive frequency second-order disturbance observer to estimate magnetic flux and EMF, eliminating the requirement of traditional SMO for phase delay compensation for estimating EMF, and reducing the position estimation error. Experimental results verify that the proposed method can provide excellent sensorless control performance. Due to the presence of low-order harmonic components, magnetic field saturation, DC bias and other factors, the accuracy of the EMF information extracted by the traditional low-pass filter (LPF) is not high, resulting in low position estimation accuracy and significant phase delay. In order to solve this problem, the prior art also proposes a sliding mode control of a dual second-order generalized integrator phase-locked loop to eliminate the harmonic components in the EMF estimate, and performs delay compensation to improve the dynamic response speed and control accuracy. In response to the high-frequency jitter and the lack of time-limited convergence of the traditional SMO, another prior art proposes a high-order sliding mode control law to suppress high-frequency jitter and design a terminal sliding mode surface to achieve finite time convergence. Experiments have shown that this method has high estimation accuracy while ensuring robustness. However, although the above methods have good control accuracy, there is still a problem that when the permanent magnet synchronous motor runs at zero speed or very low speed, the signal-to-noise ratio of the effective signal is low and difficult to extract. Under different carrier ratios, the motor control performance cannot be guaranteed, resulting in the above method extracting the rotor speed and position information inaccurately when the motor runs at zero speed or very low speed, and the motor control accuracy is reduced. Summary of the invention
[0005] In view of the above problems, the present invention proposes an adaptive quasi-proportional resonant sensorless control method with parameter identification.
[0006] The technical solution to achieve the purpose of the present invention is:
[0007] An adaptive quasi-proportional resonant sensorless control method with parameter identification, characterized in that it comprises the following steps:
[0008] Step 1: Establish the current state equation of the permanent magnet synchronous motor in the stationary coordinate system:
[0009]
[0010] Where R0 represents the stator resistance, [iα ,i β ] T represents the current along the α-β axis, [E α ,E β ] T represents the extended back EMF;
[0011] Step 2: Establish the AQPR_PI control model according to formula (3);
[0012] Step 3: Obtain the extended back-EMF estimate using the AQPR_PI control model;
[0013] Step 4: Obtain the estimated value of the permanent magnet synchronous motor speed and the rotor angle according to the estimated value of the extended back electromotive force.
[0014] Furthermore, the step of establishing the AQPR_PI control model in step 2 includes:
[0015] Step 21: According to formula (3), a sliding mode observation function as shown in formula (10) is established to obtain the current observation value:
[0016]
[0017] in, and represents the observed value of stator current, Indicates the stator resistance identification value, Indicates the inductance identification value;
[0018] Step 22: Subtract the current detection value from the current observation value to obtain the stator current error:
[0019]
[0020] in:
[0021]
[0022]
[0023] Step 23: According to Popov's hyperstability theory, we get formula (17):
[0024]
[0025] Among them, γ1 2 represents any finite positive number;
[0026] Step 24: Reversely solve formula (17) to obtain formula (18):
[0027]
[0028] in:
[0029]
[0030] Step 25: Substituting formula (19) into formula (18), the identification values of the stator resistance and the inductance parameters are obtained as follows:
[0031] Identification value of stator resistance:
[0032]
[0033] Identification values of inductance parameters:
[0034]
[0035] Among them, k p , k i is the gain parameter.
[0036] Furthermore, the specific operation steps of step 3 include:
[0037] Step 31: Convert the time domain state equation of formula (10) into the S domain state equation of formula (11):
[0038]
[0039] Step 32: Calculate the estimated value of the extended back electromotive force using a transfer function, wherein the transfer function is:
[0040]
[0041] Among them, k p , k r represents the gain parameter, ω c represents bandwidth, ω0 represents resonant frequency;
[0042] Step 33: Substituting equation (12) into equation (11), the extended back electromotive force estimation formula is:
[0043]
[0044] Step 34: Substitute the identified value of the stator resistance and the identified value of the inductance parameter obtained in step 25 into formula (13), and finally calculate the estimated value of the extended back electromotive force.
[0045] Compared with the prior art, the present invention has the following beneficial effects:
[0046] First, the model structure of the present invention is simple, with fewer control parameters and easy adjustment; at the same time, the online calculation speed is fast, the model running time is short, and the stability is high; it can reduce the system output error, weaken the chattering phenomenon, and improve the system parameter robustness and control accuracy.
[0047] Second, the present invention proposes an adaptive quasi-proportional resonant sensorless control method with parameter identification, which utilizes the frequency domain selection characteristics and adaptive filtering performance of adaptive quasi-proportional resonant control to replace the sign function and LPF model, reduce the system output error, and improve the control accuracy; at the same time, the parameters in the controller are identified online and fed back to the controller to enhance the system parameter robustness and anti-interference ability. Finally, the experimental results show that the present invention has higher tracking accuracy and stronger parameter robustness under different carrier ratios and working conditions. BRIEF DESCRIPTION OF THE DRAWINGS
[0048] Figure 1 It is a sensorless control structure of PLL-based SMO;
[0049] Figure 2 It is the structure diagram of AQPR_PI control model;
[0050] Figure 3a-3c Bode diagram analysis of the impact of changes in the three parameters kp, kr, and ωc on the controller;
[0051] Figure 4a-4b The simulation results for parameter identification;
[0052] Figure 5a-5c is the root locus of the control system in discrete state;
[0053] Figure 6 It is a PMSM experimental platform;
[0054] Figure 7 The control model structure proposed by the present invention;
[0055] Figure 8a-8b The experimental results of SMO method under high carrier ratio;
[0056] Figure 9a-9b The experimental results of SMO method under low carrier ratio;
[0057] Figure 10a-Figure 10b The experimental results of AQPR_PI method under high carrier ratio conditions; Fig.10a Experimental results for estimating and measuring current; Fig.10b Experimental results of the estimated value of EMF, speed error and position error;
[0058] Figure 11a-Figure 11b The experimental results of AQPR_PI method under low carrier ratio conditions; Fig.11a Experimental results for estimating and measuring current; Fig.11b Experimental results of the estimated value of EMF, speed error and position error;
[0059] Fig.12 The experimental results of parameter identification of resistance and inductance under dynamic conditions;
[0060] Figure 13a-13d Comparison results of control performance before and after identification; Fig.13a The experimental results of the measured speed, estimated speed, and speed error before identification are given. Fig.13b The experimental results of the measured position, estimated position, and position error before identification are shown in Figure 2. Fig.13c The experimental results of the measured speed, estimated speed and speed error after identification are given. Fig.13d Experimental results of measured position, estimated position, and position error after identification. DETAILED DESCRIPTION
[0061] In order to enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.
[0062] 1. Based on the traditional SMO sensorless control
[0063] The mathematical model of PMSM in the stationary coordinate system can be expressed as:
[0064]
[0065]
[0066] Among them, [u α ,u β ] T represents the voltage under the α-β axis of the stationary coordinate system, [i α ,i β ] T represents the current under the α-β axis, R0 represents the stator resistance, ω e represents the electrical angular velocity, θ e Indicates the rotor position angle, L d and L q Represents the stator inductance. For surface-mount PMSM, L d =L q =L0. p represents the differential operator, Ψ f represents the permanent magnet flux, [E α ,E β ] T It stands for extended back-electromotive force (EMF).
[0067] From formula (2), it can be seen that EMF contains the speed and rotor position information, so as long as the EMF is accurately observed, the speed and rotor position can be solved. Therefore, the SMO method is used to obtain EMF. In order to facilitate the establishment of the SMO model, formula (2) is converted into the current state equation shown in formula (3) to obtain the current detection value:
[0068]
[0069] In formula (3), [i α ,i β ] T As the variable of the state equation, [u α ,u β ] T As the system input, a sliding mode observation function is established as shown in formula (4) to obtain the current observation value:
[0070]
[0071] in, and represents the stator current observation value, [Z α ,Z β ] T represents the observed value of EMF;
[0072] Subtracting formula (4) from formula (3), we get the stator current error equation (5), which is:
[0073]
[0074] Among them, [Δi α ,Δi β ] T Indicates the current error;
[0075] Design the sliding mode control rate, as shown in formula (6):
[0076]
[0077] Where k represents the sliding mode controller gain parameter, and sgn(·) represents the sign function. In order to make the current error converge in the end, according to the sliding surface reaching function condition Right now:
[0078]
[0079] Therefore, the gain k needs to satisfy:
[0080] k>max{-R0|Δi α |+E α sgn(Δiα ), -R0|Δi β |+E β sgn(Δi β )} (8)
[0081] Attached Figure 1 The sensorless structure of the PLL-based SMO is shown in Figure 2. The traditional SMO model uses the sign function as the sliding surface switching function, which will result in a large number of high-order harmonics being output, causing chattering. To reduce the chattering phenomenon, a low-pass filter is connected after the sign function. When the state variable converges to the sliding surface, that is, Δi α =0,Δi β =0, we can get formula (9):
[0082]
[0083] in, represents the estimated value of EMF, ω p is the cutoff frequency of the LPF.
[0084] 2. Sensorless control method of adaptive quasi-proportional resonance with parameter identification proposed in the present invention
[0085] In order to make the system converge quickly to the sliding mode surface, the traditional SMO method uses a sign function. However, the introduction of the sign function will lead to the generation of a large number of high-order harmonics, resulting in obvious system output jitter. At the same time, the LPF added by the traditional SMO will produce control delays and phase shifts. When the fundamental frequency is close to the switching frequency, the filtering effect is reduced, affecting the system control accuracy. Therefore, the present invention uses AQPR (adaptive quasi-proportional resonance) to replace the sign function and LPF. At the same time, the stator resistance and inductance parameters are identified online (PI) and fed back to the AQPR to establish an AQPR_PI control model, which ultimately improves the robustness of the system parameters, reduces the system output jitter, and improves the tracking accuracy. Appendix Figure 2 The AQPR_PI control model structure diagram shows the AQPR in the upper part and PI (parameter identification) in the lower part. The specific process includes:
[0086] According to formula (3), we can get formula (10):
[0087]
[0088] in, Indicates the stator resistance identification value, Indicates the inductance identification value;
[0089] The time domain state equation of formula (10) is transformed into the S domain state equation of formula (11):
[0090]
[0091]
[0092] Among them, k p , k r represents the gain parameter, ω c represents bandwidth, ω0 represents resonant frequency;
[0093] Substituting formula (12) into formula (11) yields an estimated value of EMF:
[0094]
[0095] Subtracting formula (10) from formula (3) yields the current error equation shown in formula (14), and the variable expressions involved in formulas (15), (16) and (14).
[0096]
[0097]
[0098]
[0099] According to Popov's hyperstability theory, we get formula (17):
[0100]
[0101] Among them, γ1 2 represents any finite positive number. From formula (17), we can see that the parameter identification model is stable.
[0102] Formula (17) is solved in reverse to obtain formula (18). Formula (19) is the expression of variable R1 in formula (18).
[0103]
[0104]
[0105] Substituting formula (19) into formula (18), we can obtain the identification value of the stator resistor, as shown in formula (20):
[0106]
[0107] Similarly, the identification value of the inductance parameter can be obtained, as shown in formula (21):
[0108]
[0109] where k p , k i is the gain parameter;
[0110] Since the stator resistance and inductance parameters in formula (13) are generally fixed values determined by the manufacturer when the motor is produced, but during the use of the motor, these two parameters will change with factors such as temperature rise and magnetic field saturation, so the actual values of these two parameters are measured by formulas (20) and (21), and finally the estimated values of the stator resistance and inductance parameters identified by formulas (20) and (21) are fed back to formula (13), thereby improving the parameter robustness of the system.
[0111] 3. AQPR controller and parameter identification analysis
[0112] The control model proposed in this invention includes two parts: AQPR controller and parameter identification. From formula (12), it can be seen that AQPR contains three parameters: k p , k r and ω c Figure 3 shows the Bode diagram analysis of the impact of these three parameter changes on the controller. Figure 3a Indicates that the resonant frequency ω0 is 251rad / s and the bandwidth ω c is 2.51rad / s, k r is 10, k p The Bode plots are 1.5, 1, and 0.5. p As the amplitude increases, the amplitude outside the mid-frequency band also increases, but the amplitude at the resonant frequency is almost the same, indicating that k p The effect on resonance is small. Usually, in order to suppress harmonics, k p The value should be less than 1. Figure 3b Indicates that the resonant frequency ω0 is 251rad / s and the bandwidth ω c is 2.51rad / s, k p is 0.5, k r The Bode plots are 20, 10, and 5. r increases, and the amplitude increases at the resonant frequency, indicating that k r It is to reduce the steady-state error, but k r The increase of makes the frequency band wider, increases the resonance effect, and amplifies the useless signal, which is not conducive to system stability. Figure 3c Indicates that the resonant frequency ω0 is 251rad / s, k p is 0.5, k r is 10, bandwidth ω c The Bode diagrams are 2.51rad / s, 5.02rad / s and 7.53rad / s respectively. cbecomes smaller, the amplitude at the resonant frequency remains almost unchanged, the bandwidth decreases in turn, the frequency domain selectivity becomes better, but the system dynamic response speed slows down. Therefore, in order to achieve a good control effect of AQPR, adjust k r Reduce the steady-state error and adjust ω c Suppress the influence of frequency fluctuation.
[0113] Figure 4 shows the simulation results of parameter identification. In the simulation, the PMSM speed is set to 1000rpm, the initial resistance R0 is 2.84Ω, and the inductance L0 is 8.5mH. In order to verify the convergence speed and tracking accuracy of the parameter identification method, the resistance and inductance are doubled at 0.1s, and restored to the initial values at 0.3s. Figure 4a , 4b It can be seen that in the steady-state process, the parameter identification tracking accuracy is high and can stably track the given value. In the dynamic process of parameter change, the parameter identification can quickly track the change and converge to the actual value, with a faster dynamic response speed.
[0114] 4. Stability Analysis of AQPR-PI Model
[0115] According to formula (13), the closed-loop transfer function of EMF is obtained as:
[0116]
[0117] Combining formula (12) and formula (22), according to the control principle, the EMF open-loop transfer function is obtained as:
[0118]
[0119] in, represents the gain parameter, T e represents the electrical time constant of PMSM, T r Represents the equivalent time constant.
[0120] Considering the one-beat delay of the control system, and using bilinear transformation, the transfer function in the continuous time domain is converted into a discrete time system, and the S plane is mapped to the unit circle of the Z plane, formula (24) is obtained:
[0121]
[0122] Where z represents a discrete operator.
[0123] Figure 5 shows the root locus of the control system in discrete states. Figure 5a Adopt T e =T r =15ms, as the speed increases, the gain parameter k aIt also gradually increases. By selecting appropriate gain parameters, the characteristic roots of the controller are located inside the unit circle, and the system is stable. Figure 5b By changing the size of the inductor and resistor, different electrical time constants are obtained. The change of the electrical constant has little effect on the gain parameter. Figure 5c At 1000 rpm, as T r The increase of gain parameter k a According to the root locus and the focus of the unit circle in Figure 5, the critical gain parameter is obtained, and the appropriate gain parameter is selected so that the root locus is within the unit circle. At this time, the system is in a stable state.
[0124] Example
[0125] In order to further illustrate the effectiveness of the control method proposed in the present invention, Figure 6 The experiment was carried out on the PMSM experimental platform shown in the figure. Figure 7 This is the AQPR_PI control model structure. The experimental platform is based on MATLAB / Simulink to design the servo system algorithm model. When the control algorithm is built, the target code is compiled and downloaded to the target machine through the developed software. The target machine controls the servo PMSM through a dedicated control card and controls the load PMSM through a motion control card. Table 1 shows the PMSM parameters.
[0126] Table 1 Parameters of PMSM model
[0127]
[0128] 1. Comparative test of control performance under different carrier ratios
[0129] In order to verify the control performance of the AQPR-PI method under low carrier ratio conditions, the method proposed in the present invention is compared with the SMO method based on PLL. The PMSM speed is set to 500rpm, and the switching frequency is set to 10kHz and 1kHz respectively, then the corresponding carrier ratios are 300 and 30. Among them, Figures 8 and 9 respectively show the experimental results of the SMO method under high carrier ratio and low carrier ratio. Figures 10 and 11 respectively show the experimental results of the AQPR_PI method under high carrier ratio and low carrier ratio.
[0130] Under high carrier ratio conditions, Figure 8a represents the measured and estimated values of α-β current by the SMO method, Figure 8b represents the EMF estimation, speed error and position error of the SMO method. Figure 7 It means that the estimated current can stably track the sampled current, and the estimated EMF is not distorted. The standard deviations of the speed error and position error are 7.1483 and 0.1802 respectively, and the error fluctuation is small, which means that the estimated values of the speed and position are close to the measured values.
[0131] Under low carrier ratio conditions, Figure 9a represents the measured and estimated values of α-β current by the SMO method, Figure 9b represents the EMF estimation, speed error and position error of the SMO method. Figure 9a It can be seen that the estimated current has been greatly distorted and cannot stably track the measured current. A large number of harmonics appear in the EMF estimation, resulting in serious distortion and jitter. The standard deviations of the speed error and position error are 12.9091 and 0.205 respectively, and the error fluctuations are large, which is not conducive to the smooth control of the PMSM.
[0132] Figure 10 shows the experimental results of the AQPR_PI method under high carrier ratio conditions. As shown in Figure 10, under high carrier ratio, the estimated current of the proposed method can stably track the measured current, the EMF estimation is stable, the jitter phenomenon is not obvious, the standard deviations of the speed error and position error are 5.8579 and 0.1316 respectively, the error fluctuation is small, and the tracking accuracy is high.
[0133] Figure 11 shows the experimental results of the AQPR_PI method under low carrier ratio conditions. As shown in Figure 11, the estimated current fluctuation becomes larger, but it can still stably track the measured current, the EMF jitter intensifies, and the standard deviations of the speed error and position error are 6.3405 and 0.1539 respectively. Compared with the AQPR_PI method under high carrier ratio, the fluctuation increases, but it can still be stable. Compared with the SMO method, under different carrier ratios, the estimated current and EMF of the method proposed in the present invention have smaller fluctuations and can stably track the measured value. Although the performance of the AQPI_PI method decreases under low carrier ratios, there is no serious distortion and jitter, and it still has high speed and position estimation accuracy.
[0134] 2. Parameter robustness verification of AQPR_PI method under dynamic conditions
[0135] In order to verify the parameter robustness of the method proposed in the present invention, it is observed whether the parameter identification can converge to the actual value under the complex dynamic conditions of variable speed and variable load. Fig.12 The figure shows the experimental results of resistance and inductance parameter identification under dynamic conditions, where a load of 6 N.m is added at t = 1s, the speed suddenly changes from 600 rpm to 1000 rpm at t = 2s, and the load is removed at t = 3s. Fig.12It can be seen that in the motor startup phase, since the parameters are in an alternating state, the identification results show a convergence process, but can quickly return to a stable state. After adopting the parameter identification method proposed in the present invention, in a stable state, the average value of resistance identification is 1.8511, and the standard deviation is about 0.0195, which can effectively identify resistance. In a stable state, the average value of inductance identification is 6.6721, and the standard deviation is about 0.0275. When a load is added, the current increases, causing the magnetic field saturation to intensify, resulting in a decrease in inductance, and the characteristics are just the opposite when the load is removed. The identification results show that this method can effectively identify parameters under dynamic conditions, and the identification error is less than 1%.
[0136] Figure 13 compares the control performance before and after identification. Fig.13a , Fig.13b It can be seen that the speed pulsation is large before identification, the speed error fluctuates significantly, and the standard deviation is 20.8062. At the same time, the position error increases, the jitter intensifies, and the standard deviation is 0.1523. Moreover, as the speed increases, the position error also tends to increase. Fig.13c , Fig.13d It can be seen that by feeding the parameter identification results back to the control model, the speed error and position error become smaller, the fluctuation is reduced, and the standard deviations are 13.2467 and 0.0791 respectively. The results show that this method can enhance parameter robustness, reduce speed and position errors, and improve estimation accuracy.
[0137] The contents not described in detail in this specification belong to the prior art known to the professional and technical personnel in this field. Although the present invention patent is described in detail with reference to the above embodiments, for the technical personnel in this field, it is still possible to modify the technical solutions recorded in the above embodiments, or to replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. An adaptive quasi-proportional resonant sensorless control method with parameter identification, characterized in that: The following steps are involved: Step 1: Establish the current state equation of the permanent magnet synchronous motor in the stationary coordinate system: Where R0 represents the stator resistance, [i α ,i β ] T represents the current along the α-β axis, [E α ,E β ] T represents the extended back EMF; Step 2: Establish the AQPR_PI control model; it includes the following steps: Step 21: According to formula (3), a sliding mode observation function as shown in formula (10) is established to obtain the current observation value: in, and represents the observed value of stator current, Indicates the stator resistance identification value, Indicates the inductance identification value; Step 22: Subtract the current detection value from the current observation value to obtain the stator current error: in: Step 23: According to Popov's hyperstability theory, we get formula (17): Among them, γ1 2 represents any finite positive number; Step 24: Reversely solve formula (17) to obtain formula (18): in: Step 25: Substituting formula (19) into formula (18), the identification values of the stator resistance and the inductance parameters are obtained as follows: Identification value of stator resistance: Identification values of inductance parameters: Among them, k p , k i is the gain parameter; Step 3: Obtain the extended back-EMF estimate using the AQPR_PI control model; Step 4: Obtain the estimated value of the permanent magnet synchronous motor speed and the rotor angle according to the estimated value of the extended back electromotive force.
2. The adaptive quasi-proportional resonant sensorless control method with parameter identification according to claim 1, characterized in that: The specific steps of step 3 include: Step 31: Convert the time domain state equation of formula (10) into the S domain state equation of formula (11): Step 32: Calculate the estimated value of the extended back electromotive force using a transfer function, wherein the transfer function is: Among them, k p , k r represents the gain parameter, ω c represents bandwidth, ω0 represents resonant frequency; Step 33: Substituting equation (12) into equation (11), the extended back electromotive force estimation formula is: Step 34: Substitute the identified value of the stator resistance and the identified value of the inductance parameter obtained in step 25 into formula (13), and finally calculate the estimated value of the extended back electromotive force.
