PMSM rotor position estimation method based on parameter adaptive PLL
By using parameter adaptive PLL in the PMSM system, combining adaptive notch filter and sliding average filter, the problem of the traditional method decreasing accuracy at variable speed is solved, and a high-precision and low-cost rotor position estimation is achieved. Suitable for a wide range of working conditions.
Patent Information
- Application Number
- CN202410980228.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-22
- Publication Date
- 2024-11-12
AI Technical Summary
The traditional PMSM rotor position estimation method, under the influence of parameter changes and noise at variable speed, leads to a decrease in estimation accuracy, and sensor installation increases system complexity and cost, limiting the application range.
A parameter adaptive PLL based on the QT1-PLL structure is adopted, and a frequency adaptive hybrid filter is formed through an adaptive notch filter and a sliding average filter in series, and the parameters are adjusted in real time to improve estimation accuracy and robustness.
It significantly improves the accuracy and robustness of rotor position estimation, is suitable for low-speed and dynamic operating conditions, reduces hardware dependence, reduces system complexity and cost, and is suitable for various special operating conditions.
Smart Images

Figure CN118944514A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of rotor position estimation, and in particular to a PMSM rotor position estimation method based on parameter adaptive PLL. Background Art
[0002] Permanent magnetic synchronous motor (PMSM) is widely used in many fields such as industry and agriculture due to its advantages such as high power density, wide speed regulation range and simple structure. Usually, the permanent magnetic synchronous motor drive system uses mechanical encoder or rotary transformer to provide rotor position information, but the traditional position sensor is restricted by temperature, humidity and vibration, which limits the application field of permanent magnetic synchronous motor drive system; on the other hand, the installation of position sensor will increase the volume and space size of the motor, and the installation process and maintenance requirements of position sensor are also high, which further increases the cost and complexity of the drive system and reduces the reliability of the drive system. In order to solve the drawbacks of traditional position sensors, researchers have proposed a sensorless rotor position estimation technology for permanent magnetic synchronous motors, which estimates the rotor position using relevant variable information such as motor current and stator voltage. It has the basic conditions to replace traditional position sensors and has the advantages of high reliability and low cost. Therefore, the sensorless rotor position estimation method of permanent magnetic synchronous motor has become one of the research hotspots in the field of motor control technology. Summary of the invention
[0003] In view of the shortcomings of the prior art, the present invention provides a PMSM rotor position estimation method based on parameter adaptive PLL; the present invention aims at the fact that the PMSM rotor position estimation using the traditional PLL with fixed PI parameters under variable speed will lead to a decrease in estimation accuracy, and proposes a parameter adaptive PLL based on QT1-PLL structure. Firstly, the working principle of the traditional QT1-PLL is analyzed and mathematical modeling is performed for it. Then, a parameter adaptive QT1-PLL topology structure is proposed, and the adaptive notch filter and MAF are connected in series to form a frequency adaptive hybrid filter structure, which is then combined into the QT1-PLL and parameter design is performed for it. Finally, the effectiveness of the proposed method is verified by experiments.
[0004] A PMSM rotor position estimation method based on parameter adaptive PLL comprises the following steps:
[0005] Step 1: Design a parameter adaptive quasi-type hybrid filter structure PLL;
[0006] The adaptive notch filter ANF and the sliding average filter MAF are connected in series as a hybrid filtering link to filter the harmonic components; the ANF is responsible for removing -2ω r Subharmonic components, ωr is the fundamental angular velocity of the electromotive force, and MAF is used to remove the remaining harmonics; since ANF replaces MAF to filter out 2ω r Subharmonic, MAF window length T ω Reduce to Estimate the angular frequency for the rotor;
[0007] The transfer function of the ANF after Laplace transformation is:
[0008]
[0009] Where ξ is the damping coefficient, and the hybrid filter structure PLL is expressed as:
[0010]
[0011] Among them, T ω is the time window length of MAF; MAF(s) is the MAF function, s is the complex variable in Laplace transform, and ξ is the damping coefficient;
[0012] Step 2: Design the phase-locked loop adaptive parameters;
[0013] A small signal model of the adaptive QT1-PLL is established, and the small signal model is transformed to obtain a closed-loop structure;
[0014] The open-loop transfer function is:
[0015]
[0016] in, To estimate the rotor position components, is the rotor position estimation component, k is the adaptive law that changes with the speed, that is,
[0017] The transfer function of the position estimation error is expressed as:
[0018]
[0019] Among them, θ e is the position estimation error. When the rotor position changes stepwise, θ e The response is expressed as:
[0020]
[0021] It can be deduced from formula (12) that when the speed changes, θ e The response is:
[0022]
[0023] Among them, Δθ is the change in rotor position, and Δω is the change in rotor speed.
[0024] The beneficial effects of adopting the above technical solution are:
[0025] The present invention provides a PMSM rotor position estimation method based on parameter adaptive PLL, which improves the accuracy and robustness of estimation by real-time adaptive adjustment of parameters, and is particularly suitable for low-speed and dynamic working conditions. Compared with the prior art, the beneficial effects are as follows:
[0026] 1. Improve estimation accuracy: In the prior art, the rotor position estimation accuracy is low due to parameter changes and noise when running at low speed. However, the present invention can adjust the parameters of the estimation model in real time by introducing a parameter adaptive mechanism, which significantly improves the estimation accuracy of the rotor position.
[0027] 2. Enhance system robustness: The traditional position sensorless method is easily affected by the change of motor parameters under dynamic conditions, resulting in increased estimation errors. The present invention adopts an adaptive PLL method, which can effectively cope with parameter changes, improve the robustness of the system, and ensure stable operation under various conditions.
[0028] 3. Reduce hardware dependence: The present invention does not need to rely on additional position sensors, and uses adaptive algorithms and PLL technology to achieve high-precision position estimation, which reduces the hardware cost and complexity of the system, and also reduces the workload of sensor installation and maintenance.
[0029] 4. Wide scope of application: The present invention is not only applicable to ordinary PMSM motors, but can also be applied to PMSM systems under various special working conditions, such as high speed, low speed, large load changes, etc., and has wide applicability.
[0030] 5. Improve dynamic response: The existing technology often has unsatisfactory response speed and accuracy under dynamic conditions. The present invention uses an adaptive PLL method to quickly respond to system dynamic changes, provide more accurate rotor position estimation, and help improve the overall performance of the motor control system.
[0031] 6. Strong adaptability: The present invention is applicable to different types of PMSM and has strong adaptability.
[0032] Through the above technical improvements, the present invention significantly improves the performance of PMSM rotor position estimation, overcomes many deficiencies in the prior art, and provides an efficient and reliable solution for the motor control field. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1It is a structural diagram of the PMSM rotor position estimation system based on QT1-PLL of the present invention;
[0034] Figure 2 It is a system structure evolution diagram of the QT1-PLL of the present invention;
[0035] Among them, (a) - type I PLL with low-pass filter, (b) - type I PLL with low-pass filter, (c) - type I PLL with inverse tangent function, (d) - quasi-type I PLL;
[0036] Figure 3 Three PLL system structure diagrams in the embodiments of the present invention;
[0037] Among them, (a)-SRF-PLL, (b)-MAF-PLL, (c)-QT1-PLL;
[0038] Figure 4 The three PLL Bode diagrams at a fixed speed of the present invention are shown;
[0039] Figure 5 The phase-locked loop structure diagram proposed by the present invention;
[0040] Figure 6 This is the H(s) frequency characteristic diagram of the present invention;
[0041] Figure 7 This is a structural diagram of the ANF frequency adaptation implementation of the present invention;
[0042] Figure 8 This is a small signal model diagram of the PLL proposed in the present invention;
[0043] Fig. 9 It is a closed-loop model diagram of the PLL proposed by the present invention;
[0044] Fig.10 The relationship curve diagram of the estimated setting time and k of the 2% position (solid line) and the speed (dashed line) when the speed is 500r / min of the present invention;
[0045] Fig.11 The relationship curve diagram of the estimated setting time and k of the 2% position (solid line) and the speed (dashed line) when the speed is 1000r / min of the present invention;
[0046] Fig.12 The relationship curve diagram of the estimated setting time and k of the 2% position (solid line) and the speed (dashed line) when the speed is 2800 r / min of the present invention;
[0047] Fig.13 This is a diagram showing the change of k value under the change of rotation speed of the present invention;
[0048] Fig.14The Nyquist curve diagram of the present invention when k=50, k=130, k=290;
[0049] Fig.15 This is a diagram of the experimental platform of the PMSM drive system of the present invention;
[0050] Fig.16 This is a discretization effect diagram of the phase-locked loop proposed in the present invention under back-electromotive force disturbance;
[0051] Fig.17 The waveform diagram of the rotation speed measured by the encoder of the present invention;
[0052] Fig.18 The waveform diagram of the stator current under the change of speed of the present invention;
[0053] Fig.19 The rotor position waveform diagram under the speed change of the present invention;
[0054] Fig. 20 The waveform diagrams of three PLL estimation results when the rotation speed is reduced in the present invention;
[0055] Fig.21 A comparison waveform diagram of the estimated rotation speed and the actual rotation speed of the PLL proposed by the present invention;
[0056] Fig. 22 This is a waveform diagram of the speed estimation error of the PLL proposed by the present invention;
[0057] Fig.23 The MAF-PLL estimated speed and actual speed waveform diagram of the present invention;
[0058] Fig.24 It is a waveform diagram of the speed estimation error of the MAF-PLL of the present invention;
[0059] Fig.25 It is the waveform diagram of the estimated speed and actual speed of QT1-PLL of the present invention;
[0060] Fig.26 This is a waveform diagram of the QT1-PLL speed estimation error of the present invention;
[0061] Fig. 27 The waveform diagram of the estimated speed and the actual speed of the PLL under voltage distortion proposed by the present invention;
[0062] Fig.28 This is a speed error waveform diagram of the PLL proposed by the present invention under voltage distortion;
[0063] Fig.29 A comparison diagram of the estimated speed and the actual speed of the MAF-PLL of the present invention under voltage distortion;
[0064] Fig.30The rotor position estimation error waveform diagram of the MAF-PLL of the present invention under voltage distortion;
[0065] Fig.31 The torque waveform diagram of the present invention when the load changes;
[0066] Fig.32 This is a waveform diagram of the speed estimation of the PLL proposed by the present invention when the load changes;
[0067] Fig.33 This is a waveform diagram of the rotor position estimation error of the PLL proposed in the present invention when the load changes. DETAILED DESCRIPTION
[0068] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.
[0069] A PMSM rotor position estimation method based on parameter adaptive PLL comprises the following steps:
[0070] The conventional PMSM rotor position estimation structure based on QT1-PLL is as follows: Figure 1 As shown in the figure. First, the stator voltage and current are input to the SMO module, and then the back EMF estimation is passed to the QT1-PLL. In the QT1-PLL, after the input signal passes through the filtering link, the estimated rotor position and speed are obtained through the inverse tangent function and PI control.
[0071] Generally speaking, type 1 PLL has good stability margin and response speed, but it cannot accurately estimate the rotor position under frequency fluctuations. In order to overcome its shortcomings, a "quasi-type 1" PLL was proposed. The QT1-PLL has a structure similar to that of a type 1 phase-locked loop, which not only eliminates the estimation error under frequency fluctuations, but also ensures fast response and harmonic suppression capabilities.
[0072] Figure 2 This is the system structure evolution process of QT1-PLL. Figure 2 The structure shown in (a) is a type I PLL structure with a low-pass filter (LPF). Compared with type II PLL, this structure has only one control parameter k. p In order to obtain a larger bandwidth, MAF is usually selected as the low-pass filter in the structure. MAF can be approximately considered as LPF under certain conditions and can be expressed as:
[0073]
[0074] Among them, T ωis the time window length of MAF. Formula (1) shows that the response time of MAF depends on the time window length. Assume that the time window length of MAF contains N samples of its input signal, N = T ω / T s , where T s is the sampling time of the control system. The transfer function and pole-zero expression of MAF in z domain are obtained as follows:
[0075]
[0076] It can be seen that MAF has N zeros and one pole. MAF provides unity gain at zero frequency and a zero gain at the notch frequency m / T. ω There is zero gain at (m=1,2,3…). This means that MAF passes the DC component without loss and can completely block 1 / T ω Integer multiples of harmonic components.
[0077] The rotor position estimation error of a type of PLL at non-rated frequency also depends on the input voltage amplitude V. In order to eliminate this dependence, a normalization link is added to the PLL structure, and the structure is shown in 2(b). The PLL of this structure adds an LPF to the d-axis path, and the normalization link makes the PLL unaffected by the input voltage amplitude. The inverse tangent function can also be added to the inner loop as a normalization link, and its topology is as follows Figure 2 (c) The rotor position estimation error of the modified PLL under frequency drift is linearly related to Δω, that is:
[0078]
[0079] exist Figure 2 The structure shown in (d) adds a forward channel to Figure 2 The estimated error result calculated in (c) is directly fed forward to the PLL output, so that the modified PLL can compensate for the estimated error of the rotor position at the output, thus forming the QT1-PLL structure. The QT1-PLL structure is evolved from the SRF-PLL and MAF-PLL. These three PLL structures are as follows Figure 3 shown.
[0080] in, represents the back-electromotive force (Back-EMF) estimated by SMO, is the estimated component in the dq coordinate system, ω r,ref is the rated angular velocity, Δω is the deviation of the rotation speed, and is the estimated value of the speed and rotor position. The open-loop transfer function of the three PLLs can be written as:
[0081]
[0082]
[0083]
[0084]
[0085] In the formula, k p and k i are the parameters in the PI controller.
[0086] Figure 4 is the Bode plot of the three PLLs. It can be seen from formula (6) and formula (7) that the number of poles of QT1-PLL and SRF-PLL is smaller than that of MAF-PLL, that is, these two PLLs can provide a larger bandwidth. At the same time, QT1-PLL can also provide satisfactory filtering capability like MAF-PLL, which can completely remove specific order harmonic components. However, MAF contains a fixed time delay of 0.01s, which reduces the speed of transient response. However, the PI parameter ( Figure 3 The k value in (c) is a fixed value, which cannot be adjusted according to different speeds when the speed changes, which will inevitably affect the rotor position estimation performance under variable speed conditions.
[0087] Step 1: Design a parameter adaptive quasi-type-1 hybrid filter structure PLL, namely Quasi-type-1PLL, QT1-PLL; used for PMSM rotor position estimation;
[0088] Inspired by the advantages of QT1-PLL, a phase-locked loop (PLL) based on a hybrid filter structure is proposed. Figure 5 As shown. The adaptive notch filter ANF and the sliding average filter MAF are connected in series as a hybrid filtering link to filter the harmonic components; the ANF is responsible for removing -2ω r Subharmonic components, ω r is the fundamental angular velocity of the electromotive force, and MAF is used to remove the remaining harmonics; since ANF replaces MAF to filter out 2ω r Subharmonic, MAF window length T ω Reduce to Estimate the angular frequency for the rotor;
[0089] The transfer function of the ANF after Laplace transformation is:
[0090]
[0091] Wherein, ξ is the damping coefficient, which is set to 0.7 in this embodiment. Combining formula (5) and formula (7), the hybrid filter structure PLL is expressed as:
[0092]
[0093] Among them, T ω is the time window length of MAF; MAF(s) is the MAF function, s is the complex variable in Laplace transform, and ξ is the damping coefficient;
[0094] According to formula (9), the frequency characteristics of the hybrid filter are obtained, and the corresponding Bode diagram is as follows: Figure 6 As shown. It can be seen that this filtering link can filter out -2ω in the dq coordinate system r 、±6ω r 、±12ω r The implementation form of ANF frequency adaptive structure is as follows: Figure 7 shown.
[0095] Step 2: Design the adaptive parameters of the phase-locked loop and perform stability analysis;
[0096] Before designing the parameters of the proposed adaptive QT1-PLL, its mathematical model needs to be established. Figure 5 The small signal model of the adaptive QT1-PLL is established based on the phase-locked loop structure proposed in Figure 8 In order to facilitate parameter design, the small signal model is transformed to obtain a closed-loop structure; Fig. 9 As shown,
[0097] The open-loop transfer function is:
[0098]
[0099] in, To estimate the rotor position component; is the rotor position estimation component;
[0100] The transfer function of the position estimation error is expressed as:
[0101]
[0102] Among them, θ e is the position estimation error. When the rotor position changes stepwise, θ e The response is expressed as:
[0103]
[0104] It can be deduced from formula (12) that when the speed changes, θ e The response is:
[0105]
[0106] Among them, Δθ is the change in rotor position, and Δω is the change in rotor speed;
[0107] When the rotor position and speed change suddenly, in order to ensure the good dynamic performance of the system, the Laplace inverse transformation is performed on formula (12) and formula (13). The influence of k value on the transient setting time under two sudden changes is studied at different speeds (500r / min, 1000r / min and 2800r / min). Fig.10 , Fig.11 and Fig.12 The solid line is the relationship curve corresponding to the rotor position jump, and the dotted line is the relationship curve corresponding to the speed mutation.
[0108] from Figure 10 to Figure 12 It can be seen that the system cannot obtain the fastest speed estimation and rotor position estimation at the same time. When the speed is 500r / min, the best values of k are k=48 and k=54 respectively. When the speed changes, the value of k should be adjusted with the speed change. In order to obtain a fast response time at the same time, the response speed of the speed estimation and the position estimation should be considered at the same time.
[0109] In order to achieve good dynamic performance under both rotor position and speed changes, it is necessary to select k when both curves reach a lower position. The lowest intersection point of the two curves is taken as the optimal solution for k. Figure 10 to Figure 12 It can be observed that the optimal k value is different at different speeds, and the relationship between the back EMF voltage frequency and the optimal k value can be obtained as follows: Fig.13 According to this relationship, the adaptive law of the parameter k in the proposed PLL changes with the speed is:
[0110] In order to verify the stability of the proposed phase-locked loop under different parameters k, the Nyquist stability analysis was performed with k values at different speeds. The Nyquist curves under different k values are shown in Figure 2. Fig.14 As shown. By observing the drawn Nyquist curve, it can be found that all open-loop transfer curves do not surround (-1, j0), which proves that the system is stable under the corresponding k value.
[0111] The specific design simulation and experimental results analysis in this embodiment are as follows:
[0112] In order to verify the effectiveness of this method, an experimental study was carried out on a 550W sensorless PMSM control system. The pole pair number of the PMSM used in the experiment is 2, the stator resistance is 0.63Ω, the direct axis inductance and the orthogonal axis inductance are 119mH and 29mH respectively, and the encoder is installed at the end of the load. The PMSM is driven by a DSP-controlled inverter-based motor driver, and the inverter switching frequency is 10kHz. The sensor samples data at the beginning of each switching cycle. The estimated position and electromotive force waveform are monitored in the host computer Labview software environment. The relevant parameters of the drive system are shown in Table 1, and the system experimental platform is shown in Fig.15 shown.
[0113] Table 1 PMSM drive system related parameters
[0114]
[0115] Although the proposed PLL is designed in the continuous domain, it is implemented in the discrete domain on the DSP control board of the experimental platform. Therefore, in practical applications, the continuous time needs to be discretized. Using the backward Euler method, the discrete approximation of the integral operation is as follows:
[0116]
[0117] The discretization effect of the proposed phase-locked loop is verified through simulation experiments. The sampling frequency in the discretization process is set to 10kHz. Fig.16 The discretization effect of the proposed PLL input back EMF when phase jump and frequency step change occur is given in Fig. Fig.16 It can be seen that the transient behaviors of the continuous domain and the discrete domain are almost the same under this change.
[0118] (1) Multi-speed operation experiment
[0119] The estimated performance of various PLLs is compared under speed variation conditions. The comparative experiments will be carried out on the proposed PLL, QT1-PLL[ and MAF-PLL. The operation of PMSM includes three stages: 800r / min, 50r / min and reverse 500r / min.
[0120] Fig.17 The speed change curve measured by the encoder. The motor speed climbed to 800r / min at 0.3s to complete the startup process, then switched to low-speed operation at 1s and maintained for 5s, and finally switched to the reverse state. The entire process tested in detail the performance of the proposed phase-locked loop in the transient operation stage under motor acceleration and deceleration, and in the steady-state operation stage at different speeds.
[0121] Fig.18The stator current in different stages is shown in Figure 2. The stator current rises to 1.8A at the moment of starting, and gradually stabilizes at around 0.65A as the speed increases. The stator current is maintained at around ±0.2A at low speed, and finally stabilizes to 0.7A in the reversal state. Fig.19 is the rotor position measured by the encoder.
[0122] Fig. 20 The rotor position estimation results of the proposed adaptive PLL, QT1-PLL and MAF-PLL during the speed reduction from 800r / min to 50r / min. It can be seen from the figure that the rotor estimated position of the proposed adaptive PLL is closer to the actual rotor position than other PLLs when the speed decreases. Since the MAF included in QT1-PLL and MAF-PLL has a large time delay, it affects its transient performance and there is a larger position estimation error during the deceleration process.
[0123] Fig.21 The comparison between the estimated speed and the actual speed of the proposed adaptive PLL is shown. It can be seen from the figure that the proposed adaptive PLL can accurately estimate the speed regardless of the speed climbing stage, stable stage or deceleration stage.
[0124] The speed estimation error during the experiment is as follows: Fig. 22 As shown in the figure. The speed error peak occurs during the acceleration and deceleration of the rotor, and the error is within 15r / min. The speed estimation error of the proposed PLL is always stable within 1r / min under low-speed conditions. Since the proposed PLL has adaptively adjusted filtering links and control parameters, the system can have good estimation performance in the entire operating condition.
[0125] Fig.23 The figure shows the comparison between the estimated speed and the actual speed of MAF-PLL. As can be seen from the figure, there is a significant estimation error when the motor accelerates and decelerates, which indicates that the transient performance of MAF-PLL is poor. Fig.24 The speed estimation error of MAF-PLL in the experimental process, especially in the acceleration and deceleration process, the average peak value of the speed error is greater than 40r / min, which is much larger than the speed estimation error of the proposed PLL. This is because the bandwidth of MAF-PLL is small, and the PI controller parameters cannot change with the speed change, resulting in poor transient performance under the motor speed change condition, and the estimation error will have a large overshoot. However, since MAF-PLL evolved from type II PLL, MAF-PLL can still achieve accurate estimation under steady-state conditions of different speeds.
[0126] Fig.25 Comparison between the estimated speed and the actual speed of QT1-PLL. Fig.26is the estimated error between QT1-PLL and the actual speed. As can be seen from the figure, QT1-PLL can quickly track the actual speed changes without overshoot. However, since QT1-PLL belongs to a quasi-type structure and uses a fixed control parameter k, it always has an estimated steady-state error greater than 15r / min at non-rated speeds.
[0127] Table 2 Summary of the results of the proposed PLL, QT1-PLL, and MAF-PLL under variable speed
[0128]
[0129] Table 2 summarizes the results of the three methods. It can be seen that the proposed adaptive PLL has high estimation accuracy both in the transient operation stage of motor acceleration and deceleration and in the steady-state operation stage at different speeds. Although QT1-PLL can quickly track the actual speed change when the motor accelerates and decelerates, and there is no overshoot. However, QT1-PLL always has a large estimation error when the motor is running at a non-rated speed. Since MAF-PLL is evolved based on type II PLL, its steady-state estimation accuracy at different speeds is high. However, type II PLL has a smaller bandwidth than type I PLL, and MAF-PLL uses fixed PI parameters and cannot adjust the bandwidth according to the speed, which will affect its transient performance in the acceleration stage.
[0130] (2) Distortion voltage injection experiment
[0131] In order to verify the filtering performance of the proposed PLL, a distorted voltage injection experiment is carried out. As shown in Table 2, under the condition of stable speed, MAF-PLL has an estimation performance similar to that of the proposed PLL, so a comparative experiment will be carried out between MAF-PLL and the proposed PLL. The distorted voltage signal consists of sinusoidal signals of the 5th harmonic and the 7th harmonic. The period from 0s to 2s is the stage of no distorted voltage interference. At 2s, two harmonic distorted voltages of 0.3V are injected, and at 4s, the amplitude of the distorted voltage is reduced to 0.1V. The rotor speed is set to 1500r / min and remains stable. During this period, the comparison of the speed estimation of the two PLLs is observed.
[0132] Fig. 27 The figure shows the comparison between the estimated speed and the actual speed of the proposed PLL under voltage distortion. It can be seen from the figure that the proposed PLL can accurately estimate the PMSM speed throughout the experiment. The position estimation error is as follows Fig.28 It can be seen that the proposed PLL position estimation error is always controlled within 0.4rad, indicating that the proposed method can effectively suppress the harmonic components.
[0133] Fig.29is the comparison between the estimated speed of MAF-PLL and the actual speed. It can be seen that under the distorted voltage injection, the estimated speed of MAF-PLL has a large error. Its estimated error is as follows: Fig.30 As shown in the figure, during the 0.3V harmonic voltage injection process, the position estimation error value is greater than 1.4rad, which is much larger than the proposed PLL. Although MAF-PLL belongs to the second type PLL, its MAF does not adopt an adaptive structure, which makes its filtering performance unable to adjust with the speed, resulting in the filtering ability being affected at non-rated speed.
[0134] Table 3 Summary of experimental results of harmonic voltage injection of the proposed PLL and MAF-PLL
[0135]
[0136] Table 3 summarizes the experimental results of harmonic voltage injection. It can be seen from the table that the speed estimation errors of the two PLLs are similar when harmonics are not injected. When harmonics are injected, the speed estimation error of MAF-PLL increases, and the speed estimation error peak reaches 27.5r / min when 0.3V harmonics are injected. The speed estimation error of the proposed PLL is kept around 3.0r / min. Even under 0.3V harmonic voltage injection, its peak estimation error and mean estimation error are 4.3r / min and 2.8r / min, respectively, which are much smaller than 27.5r / min and 16.2r / min of MAF-PLL. The experimental results show that the hybrid filtering link of the proposed PLL has a good suppression effect on the distorted voltage.
[0137] (3) Load change experiment
[0138] In order to verify the response speed and estimation accuracy of the proposed method under load changes, this section conducts speed estimation experiments under load changes. The rotor speed is maintained at 1500r / min, and the initial load of the motor is set to 4N·m. The load is increased from 4N·m to 12N·m in the first stage, and the load is decayed from 12N·m to 8N·m in the second stage. During this period, the estimation performance of the proposed parameter adaptive QT1-PLL is observed. Fig.31 is the load torque curve.
[0139] Fig.32 The comparison results of the proposed parameter adaptive QT1-PLL estimation results and the actual speed in the entire load change stage. It can be seen that it can still track the speed change when the load changes, and has a low estimation error under different loads.
[0140] Fig.33is the rotor position estimation error value of the proposed parameter adaptive QT1-PLL under load changes. It can be seen from the figure that the proposed PLL has no large overshoot during load changes, and the peak-to-peak error does not exceed 0.5rad.
[0141] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but should also cover other technical solutions formed by any combination of the above-mentioned technical features or their equivalent features without departing from the above-mentioned inventive concept. For example, the above-mentioned features are replaced with the technical features with similar functions disclosed in the embodiments of the present disclosure (but not limited to) to form a technical solution.
Claims
1. A PMSM rotor position estimation method based on parameter adaptive PLL, characterized in that: The following steps are involved: Step 1: Design a parameter adaptive quasi-type hybrid filter structure PLL; The adaptive notch filter ANF and the sliding average filter MAF are connected in series as a hybrid filtering link to filter the harmonic components; the ANF is responsible for removing -2ω r Subharmonic components, ω r is the fundamental angular velocity of the electromotive force, and MAF is used to remove the remaining harmonics; since ANF replaces MAF to filter out 2ω r Subharmonic, MAF window length T ω Reduce to Estimate the angular frequency for the rotor; Step 2: Design the phase-locked loop adaptive parameters to achieve PMSM rotor position estimation.
2. A PMSM rotor position estimation method based on parameter adaptive PLL according to claim 1, characterized in that: Step 1 is as follows: The transfer function of the ANF after Laplace transformation is: Where ξ is the damping coefficient, and the hybrid filter structure PLL is expressed as: Among them, T ω is the time window length of MAF; MAF(s) is the MAF function, s is the complex variable in Laplace transform, and ξ is the damping coefficient.
3. A PMSM rotor position estimation method based on parameter adaptive PLL according to claim 1, characterized in that: The step 2 specifically includes: A small signal model of the adaptive QT1-PLL is established, and the small signal model is transformed to obtain a closed-loop structure; The open-loop transfer function is: in, To estimate the rotor position components, is the rotor position estimation component, k is the adaptive law that changes with the speed, that is, The transfer function of the position estimation error is expressed as: Among them, θ e is the position estimation error; when the rotor position changes stepwise, θ e The response is expressed as: It can be deduced from formula (12) that when the speed changes, θ e The response is: Among them, Δθ is the change in rotor position, and Δω is the change in rotor speed.
