A speed sensorless construction method for a permanent magnet assisted bearingless synchronous reluctance motor

Abstract

The invention discloses a speed sensorless construction method of a permanent magnet assisted bearingless synchronous reluctance motor. The motor speed sensor is composed of a speed controller, a signal conversion module, a back electromotive force observer, an ISRF phase-locked loop and an adaptive controller. The ISRF phase-locked loop is used to extract a motor speed signal from the back electromotive force to realize sensorless control. The phase-locked loop uses frequency or phase as a feedback error to adjust to form a closed-loop control, so as to ensure that the frequency of an input signal is equal to that of an output signal, and the phase difference is constant, so as to complete high-precision speed closed-loop control adjustment. The adaptive controller can dynamically adjust the gain of a speed estimation loop according to the size of the phase deviation, so as to solve the problem that a back electromotive force estimation value has a large transient state during startup and when a large phase angle jump occurs in a traditional SRF phase-locked loop. An adaptive gain PLL loop system is proposed to provide a loop gain for a PI controller in the PLL loop system, so as to enable the system to respond faster.

Description

Technical Field

[0001] The invention belongs to the technical field of electric drive control equipment, and is a speed sensorless construction method of a permanent magnet assisted bearingless synchronous reluctance motor, which is suitable for high-performance control of a permanent magnet assisted bearingless synchronous reluctance motor. Background Art

[0002] The bearingless synchronous reluctance motor is an AC synchronous reluctance motor with a novel and unique structure and principle. It embeds two sets of windings with a pole pair number difference of 1 in the stator slot: the torque winding and the suspension winding, and controls the current in the two sets of windings to simultaneously generate electromagnetic torque and radial suspension force to achieve stable suspension and rotation of the rotor. The permanent magnet assisted bearingless synchronous reluctance motor (PMa-BSynRM) is based on the bearingless synchronous reluctance motor. By adding permanent magnets to the rotor magnetic barrier, the power density and torque density of the motor are improved, and the weak magnetic speed regulation capability is enhanced. It has the advantages of high speed, high precision, high efficiency and maintenance-free, further broadening the application range of bearingless synchronous reluctance motors and is highly competitive in the field of electrical transmission.

[0003] The key to achieving stable operation of the PMa-BSynRM rotor lies in the accurate detection of the rotor position and radial displacement. At present, a photoelectric encoder is usually used for rotor position detection. The motor speed and rotation direction are detected through the photoelectric encoder through the speed signal conditioning circuit. However, the use of the photoelectric encoder will bring inevitable defects, mainly including that it is greatly affected by the environment, which will affect the accuracy of measuring the rotor position, and requires a more complex circuit to process the output signal, which increases the cost of motor operation control and limits the promotion and application of PMa-BSynRM.

[0004] In order to solve the above problems, a pulsating high-frequency voltage injection method is adopted. By injecting a pulsating high-frequency voltage signal into the torque winding of the bearingless synchronous reluctance motor and utilizing the motor spatial convexity effect, the rotor speed can be effectively predicted. However, the accuracy of the rotor speed prediction is greatly affected by the injected high-frequency pulsating voltage, and the detection and extraction of the rotor position signal is relatively complicated. A method for constructing a bearingless synchronous reluctance motor speed estimation system is disclosed in the document with Chinese patent application number CN202110796374.9. By constructing modules such as a motor speed estimator, a virtual controller, and a generalized inverter, the speed sensorless rotation operation of the motor is realized. However, the virtual controller further estimates the estimated speed value obtained by the speed estimator to obtain the current, which increases the current error. While enhancing the speed tracking performance, it affects the speed estimation accuracy.

[0005] Therefore, while using sensorless control technology, how to design a sensorless control system with simple structure, high precision and high accuracy has become a key issue that needs to be urgently solved for the further development of PMa-BSynRM sensorless technology. Summary of the invention

[0006] The purpose of the present invention is to propose a speed sensorless construction method for a permanent magnet assisted bearingless synchronous reluctance motor, so as to solve the problems of complex structure and low precision of the existing permanent magnet assisted bearingless synchronous reluctance motor sensorless control system, reduce the motor control cost, improve the motor working performance, and better apply it to the field of electrical transmission.

[0007] To achieve the above object, the technical solution adopted by the speed sensorless construction method of a permanent magnet assisted bearingless synchronous reluctance motor of the present invention comprises the following steps:

[0008] Step 1): The current hysteresis PWM module is connected before the torque subsystem of the permanent magnet assisted bearingless synchronous reluctance motor to form a composite controlled object. The input of the current hysteresis PWM module is the two current differences in the dq coordinate system.

[0009] Step 2): The three-phase winding current and voltage output by the torque subsystem are converted into two current and voltage i in the α-β coordinate system after the signal conversion module. Mα / u Mα ,i Mβ / u Mβ And the two currents i in the dq rotating coordinate system Mq / i Md , current i Mq The q-axis current component of the speed controller output The current difference is obtained by subtracting Current i Md With a given d-axis current component The current difference is obtained by subtracting

[0010] Step 3): The two current and voltage i Mα / u Mα ,i Mβ / u Mβ The back-EMF observer outputs two back-EMF estimates. Two back EMF estimates They are input into the ISRF phase-locked loop and the adaptive controller respectively, and the adaptive controller outputs the adaptive parameters Give it to the ISRF phase-locked loop, and the ISRF phase-locked loop gets the estimated angle of the motor Actual rotor position angle θ e Estimated angle with motor The error angle Δθ between the motor and the estimated angular velocity Among them, λ is a positive constant, and the estimated rotor speed is ω e is the rotor angular velocity, k1 is the filter bandwidth, k f is the loop gain proportional parameter, k p and k i are the proportional and integral parameters respectively;

[0011] Step 4): The motor estimated angle The error angle Δθ is fed back to the signal conversion module, and the motor estimates the angular velocity With the given motor angular velocity ω * After comparison, the result is input into the speed controller;

[0012] Step 5): The motor speed sensorless system is composed of the speed controller, the signal conversion module, the back-EMF observer, the ISRF phase-locked loop and the adaptive controller.

[0013] Furthermore, the signal conversion module is composed of Clark transformation, Park transformation and Park transformation. Clark transformation converts the three-phase winding current and voltage i MU / u MU 、i MV / u MV 、i MW / u MW Convert the current and voltage i into the α-β coordinate system Mα / u Mα and i Mβ / u Mβ , the current i in the α-β coordinate system Mα and i Mβ Input Park transformation, current and voltage i in α-β coordinate system Mα / u Mα and i Mβ / u Mβ Enter the Park transform, which is based on the angle and current i Mα ,i Mβ Converted to the current i in the dq rotating coordinate system Md and i Mq , Park transform is based on the estimated angle and current voltage i Mα / u Mα ,i Mβ / u Mβ , converted to the current and voltage i in the γ-δ rotating coordinate system Mγ / u Mγ and i Mδ / u Mδ , the γ-δ rotating coordinate system is a coordinate system that differs from the dq rotating coordinate system by Δθ.

[0014] The transformation formula of the Clark transformation is N3 is the effective number of turns of each phase of the three-phase winding, and N2 is the effective number of turns of the two-phase winding.

[0015] The transformation formula of Park transformation is: The transformation formula of Park transformation (53) is:

[0016] The estimated back EMF value is Estimated back EMF ω c is a positive constant; s is the transformation parameter in the Laplacian transformation; e Mα , e Mβ is the component of the actual value of the back EMF on the α-β axis.

[0017] Furthermore, the rotor angular velocity u δ is the component of the torque winding voltage on the δ axis; R M is the torque winding resistance; L Mγ and L Mδ is the self-inductance of the torque winding γ-δ axis; ψ0 is the permanent magnet flux; i Mγ and i Mδ is the component of the torque winding current on the γ-δ axis; s is the transformation parameter in the pull-type transformation.

[0018] The actual rotor position angle ω c is a positive constant; s is the transformation parameter in the Pull transformation.

[0019] The advantages of the present invention after adopting the above technical solution are:

[0020] 1. The present invention adopts an improved synchronous reference frame phase-locked loop (ISRF-PLL) technology to realize the speed sensorless control of the permanent magnet assisted bearingless synchronous reluctance motor. Compared with the sensor control system, the control system is simplified, the cost of the control system is reduced, and the shortcomings of unstable parameter measurement caused by mechanical sensors are overcome. The phase-locked loop technology uses frequency or phase as feedback error adjustment to form a closed-loop control, which can ensure that the frequency of the input signal is equal to that of the output signal, and the phase difference is constant, and can complete high-precision speed closed-loop control adjustment. Moreover, it has a simple structure, stable parameter measurement, and can achieve accurate measurement of the motor speed.

[0021] 2. The present invention proposes an adaptive controller that can dynamically adjust the gain of the speed estimation loop according to the size of the phase deviation, thereby solving the problem of large transient back-electromotive force estimation values ​​occurring in traditional SRF phase-locked loops during startup and when large phase angle jumps occur; and proposes an adaptive gain PLL loop system that provides loop gain for the PI controller in the PLL loop system, enabling the system to respond faster.

[0022] 3. The present invention utilizes an ISRF phase-locked loop to extract the motor speed signal from the back electromotive force to achieve sensorless control, and improves the traditional SRF phase-locked loop, overcoming many shortcomings of other sensorless control technologies, such as: the pulse high-frequency voltage injection method utilizes the spatial convexity effect of the motor to achieve effective prediction of the rotor speed, but the accuracy of the rotor speed prediction is greatly affected by the injected high-frequency pulsating voltage, and the detection and extraction of the rotor position signal is relatively complicated; the model reference adaptive method based on voltage and current has poor identification accuracy due to the existence of a pure integral link, and is affected by the stator resistance and is unstable at low speed; the extended Kalman filter method uses the minimum variance optimal prediction estimation method to weaken random interference and measurement noise, but the algorithm is complex, the parameter configuration lacks certain standards, and the amount of calculation is large. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The present invention is further described in detail below with reference to the accompanying drawings and specific implementation methods:

[0024] Figure 1 A structural block diagram of a speed sensorless permanent magnet assisted bearingless synchronous reluctance motor constructed by the present invention;

[0025] Figure 2 for Figure 1 The internal and external structure diagram of the current hysteresis PWM module in;

[0026] Figure 3 for Figure 1 The information collection and conversion unit in the

[0027] Figure 4 It is the space vector diagram of the synchronously rotating frame;

[0028] Figure 5 for Figure 1 Schematic diagram of the back EMF observer in ;

[0029] Figure 6 for Figure 1 Simplified schematic of the ISRF phase-locked loop in Figure 1.

[0030] In the figure: 1. Speed ​​controller; 2. Composite controlled object; 3. Current hysteresis PWM module; 4. PMa-BSynRM torque subsystem; 5. Signal conversion module; 6. Back EMF observer; 7. ISRF phase-locked loop; 8. Adaptive controller;

[0031] 31. Park inverse transform; 32. Clark inverse transform; 33. Current hysteresis PWM; 34. Inverter; 51. Clrak transform; 52, 53. Park transform. DETAILED DESCRIPTION

[0032] See also Figure 1 The current hysteresis PWM module 3 is connected before the PMa-BSynRM torque subsystem 4 to form a composite controlled object 2. The permanent magnet assisted bearingless synchronous reluctance motor speed sensor constructed by the present invention is composed of a speed controller 1, a signal conversion module 5, a back-EMF observer 6, an ISRF phase-locked loop 7 and an adaptive controller 8, which are connected to the composite controlled object 2 to realize the speed sensorless control of the PMa-BSynRM.

[0033] The output signal of the current hysteresis PWM module 3 is the three-phase winding current i MU 、i MV 、i MW , three-phase winding current i MU 、i MV 、i MW As the input signal of PMa-BSynRM torque subsystem 4, the three-phase winding current i MU 、i MV 、i MW And the three-phase winding voltage signal u output by PMa-BSynRM torque subsystem 4 MU 、u MV 、u MW As the input signal of the signal conversion module 5. The three-phase winding current and voltage i MU / u MU 、i MV / u MV 、i MW / u MW After conversion by the signal conversion module 5, two current and voltage signals in the α-β coordinate system are output: i Mα / u Mα and i Mβ / u Mβ , and the two currents i in the output dq rotating coordinate system Mq / i Md Among them, the current i Mq The q-axis current component of the speed controller 1 output Difference is obtained by Current iMd The d-axis current component given by the system Difference is obtained by The difference between the two currents in the dq coordinate system Together they serve as the input of the current hysteresis PWM module 3 to form an inner closed-loop control.

[0034] The two current and voltage i in the α-β coordinate system output by the signal conversion module 5 are Mα / u Mα and i Mβ / u Mβ As input signals of the back-EMF observer 6, the back-EMF observer 6 outputs two back-EMF estimation values: and signal, two back EMF estimates and They are input into the ISRF phase-locked loop 7 and the adaptive controller 8 respectively. The adaptive controller 8 outputs the adaptive parameter u to the ISRF phase-locked loop 7, u, As the input signal, it is input into ISRF phase-locked loop 7, and ISRF phase-locked loop 7 obtains the estimated angle of the motor. Estimated angular velocity of the motor The error Δθ between the actual rotor position angle and the estimated rotor position. Motor estimated angle The error Δθ between the actual rotor position angle and the estimated rotor position is input into the signal conversion module 5. The motor angular velocity ω given by the system * After comparison, the data is input into the speed controller 1, thus forming an outer closed loop control, thereby forming a PMa-BSynRM speed sensorless system.

[0035] See also Figure 2 The current hysteresis PWM module 3 is composed of Park inverse transformation 31, Clark inverse transformation 32, current hysteresis PWM 33, and inverter 34 connected in sequence. The current hysteresis PWM module 3 obtains two currents of the torque winding in the dq rotating coordinate system. and After that, the current in the α-β coordinate system is obtained by Park inverse transformation 31 and Then the three-phase current is output through Clark inverse transformation 32 Current hysteresis PWM33 obtains three-phase current Output three-phase DC current i Ma 、i Mb 、i Mc After receiving the three-phase DC current signal, the inverter 34 outputs the three-phase winding current i MU 、iMV 、i MW Finally, the three-phase winding current i MU 、i MV 、i MW The signal is input to the PMa-BSynRM torque subsystem 4. Among them, the transformation formula of the Park inverse transformation 31 is:

[0036]

[0037] Where: Estimate the angle for the motor.

[0038] The transformation formula of Clark inverse transform 32 is:

[0039]

[0040] Where: N3 is the effective number of turns per phase of the three-phase winding, N2 is the effective number of turns of the two-phase winding, assuming that the total power remains unchanged before and after the transformation, the turns ratio

[0041] See also Figure 3 The signal conversion module 5 is composed of Clark transformation 51, Park transformation 52 and Park transformation 53. Clark transformation 51 converts the three-phase winding current and voltage i MU / u MU 、i MV / u MV 、i MW / u MW Convert the current and voltage i into the α-β coordinate system Mα / u Mα and i Mβ / u Mβ , the current i in the α-β coordinate system Mα and i Mβ Input Park transformation 52, current and voltage i in α-β coordinate system Mα / u Mα and i Mβ / u Mβ Enter Park Transform 53. Park Transform 52 is based on angle and current i Mα ,i Mβ , converted to the current i in the dq rotating coordinate system Md and i Mq , Δθ is the actual rotor position angle θ e With the estimated rotor position The error between the two. Park transform 53 is based on the estimated angle and current voltage i Mα / u Mα and i Mβ / uMβ , converted to the current and voltage i in the γ-δ rotating coordinate system Mγ / u Mγ and i Mδ / u Mδ , the γ-δ rotating coordinate system is a coordinate system that differs from the dq rotating coordinate system by Δθ.

[0042] Among them, the transformation formula of Clark transformation 51 is:

[0043]

[0044] Where: N3 is the effective number of turns per phase of the three-phase winding, N2 is the effective number of turns of the two-phase winding, assuming that the total power remains unchanged before and after the transformation, the turns ratio

[0045]

[0046] The transformation formula of Park transformation 52 is:

[0047]

[0048] Where: Δθ is the actual rotor position angle θ e With the estimated rotor position The error between .

[0049] The transformation formula of Park transformation 53 is:

[0050]

[0051] See also Figure 4 , is the space vector diagram of the synchronous rotating frame, where the d axis is in the same direction as the permanent magnet flux ψ0. Figure 4 In, e s is the back electromotive force, e d and e q is the back electromotive force e on the d-axis and q-axis s The orthogonal component of e is the actual rotor position angle, Δθ is the actual rotor position angle θ e and the estimated rotor position angle When Δθ is small enough, we can assume that Δθ≈e d Therefore, e d Can be used as an indicator of the estimated rotor position accuracy.

[0052] See also Figure 5 ,according to Figure 4The conclusions drawn were used to construct the back-EMF observation principle diagram. When the PMa-BSynRM rotor is stably suspended, the rotor is maintained in a balanced position. Ignoring the influence of the suspension force winding current and voltage on the rotor winding current and voltage, the voltage equation in the α-β coordinate system can be obtained:

[0053]

[0054] Where: u Mα and u Mβ 、i Mα and i Mβ 、e Mα and e Mβ are the components of the actual values ​​of the torque winding voltage, current and back EMF on the α-β axis respectively; R M is the torque winding resistance; L Mα and L Mβ is the self-inductance of the torque winding α-β axis.

[0055] According to the back-EMF observation principle diagram, the Lassen transformation equation of voltage and back-EMF in the α-β coordinate system can be obtained:

[0056]

[0057] Where: and and are the components of the estimated current and back-EMF of the torque winding on the α-β axis respectively; s is the transformation parameter in the pull-type transformation.

[0058]

[0059] Where: k p and k i are the proportional and integral parameters respectively.

[0060] By performing a pull-type transformation on equation (6) and substituting it into equation (8) together with equation (7), we can obtain:

[0061]

[0062] In formula (9), when i Mα and u Mα When k is the input signal, p =L Mα ω c , k i =R M ω c ,ω c is a normal number; when i Mβ and u Mβ When k is the input signal, p =L Mβ ωc , k i =R M ω c ,ω c is a positive constant, and s is the transformation parameter in the Las Vegas transformation. Based on the zero-pole cancellation method, equation (9) can be simplified to the following expression:

[0063]

[0064] Where: c is a positive constant; s is the transformation parameter in the Laplacian transformation; e Mα , e Mβ is the component of the actual value of the back EMF on the α-β axis, which can be directly obtained in equation (6).

[0065] See also Figure 6 ,In order to overcome the problem that the position estimation accuracy is affected by noise and harmonics and that the estimated value of the back EMF will experience large transient problems during the ,startup of the phase-locked loop and when a large phase angle jump occurs, this paper adopts the ISRF ,phase-locked loop method to extract the position signal from the back EMF, and adds an adaptive ,controller 2 and PI gain loop on the basis of the ,traditional phase-locked loop. Figure 6 Simplified schematic for the ISRF phase-locked loop. To construct the ISRF phase-locked loop, the PMa-BSynRM model needs to be reconstructed in the synchronous reference frame:

[0066]

[0067] Where: u γ and u δ 、i Mγ and i Mδ are the components of the torque winding voltage and current on the γ-δ axis respectively; L Mγ and L Mδ is the self-inductance of the torque winding γ-δ axis; ω e is the rotor angular velocity of the motor; considering the addition of permanent magnets in the PMa-BSynRM rotor, ψ0 represents the permanent magnet flux; Δθ is the actual rotor position angle θ e With the estimated rotor position The error between .

[0068] According to formula (11), the rotor angular velocity equation of the motor can be obtained:

[0069]

[0070] Figure 6 The adaptive parameter u is:

[0071]

[0072] Where: λ is a positive constant. When the phase angle jumps, u becomes a small value, which reduces the gain of the speed estimation and prevents a large transient in the estimated speed.

[0073] From this we get Figure 6 The five input signals in: the components of the estimated back EMF on the α-β axis and Adaptive parameters u, rotor angular velocity ω e , the expected error Δθ between the estimated reference frame and the true reference frame ref =0.

[0074] After obtaining the estimated value of the back EMF in the α-β coordinate system, the actual rotor position angle can be directly obtained through the inverse angle function:

[0075]

[0076] Where: c is a positive constant; s is the transformation parameter in the Lagrange transformation; the inverse tangent sign is Figure 6 It is represented by arctan.

[0077] The error angle Δθ between the actual rotor position angle and the estimated rotor position is:

[0078]

[0079] The error angle Δθ between the actual rotor position and the estimated rotor position is fed back to the signal conversion module 5 and enters the PI gain loop at the same time, so that the estimated error between the actual rotor position and the estimated rotor position can be obtained.

[0080]

[0081] Where: k f is the loop gain proportional parameter; Figure 6 abs in the equation represents the absolute value symbol; Δω e (t) represents Δω at the current time t e The value of .

[0082] When the position error signal is amplified by the proportional controller, Δω e As a compensation term for speed evaluation:

[0083]

[0084] Where: k p and k i are proportional and integral parameters, Δω e (t+1) represents Δω at time (t+1) e The value of .

[0085] ω e It is directly obtained from equation (12) and is regarded as a feedforward term. Therefore, the estimated rotor speed obtained before the filter can be expressed as:

[0086]

[0087] Where: Estimate the speed for a rotor with harmonics.

[0088] In order to reduce the impact of high-frequency noise, a first-order filter is used in the PLL structure:

[0089]

[0090] Where: k1 is the bandwidth of the filter structure; Estimate the angular velocity for the motor.

[0091] After calculating the rotor speed, the position is obtained by integration:

[0092]

[0093] After obtaining the rotor position, the rotor position signal is used as feedback to adjust and correct the input signal to form a closed-loop control.

[0094] In summary, the speed sensorless control of the permanent magnet assisted bearingless synchronous reluctance motor of the present invention can be achieved.

Claims

1. A speed sensorless construction method of a permanent magnet assisted bearingless synchronous reluctance motor, characterized in that The following steps are involved: Step 1): A current hysteresis PWM module (3) is connected before the torque subsystem (4) of the permanent magnet assisted bearingless synchronous reluctance motor to form a composite controlled object (2), and the input of the current hysteresis PWM module (3) is the two current differences in the dq coordinate system. Step 2): The three-phase winding current and voltage output by the torque subsystem (4) are output through the signal conversion module (5) as two current and voltage i in the α-β coordinate system. Mα / u Mα ,i Mβ / u Mβ And the two currents i in the dq rotating coordinate system Mq / i Md , current i Mq The q-axis current component output by the speed controller (1) The current difference is obtained by subtracting Current i Md With a given d-axis current component The current difference is obtained by subtracting Step 3): The two current and voltage i Mα / u Mα ,i Mβ / u Mβ The back-EMF observer (6) outputs two back-EMF estimates: Two back EMF estimates are input into the ISRF phase-locked loop (7) and the adaptive controller (8), and the adaptive controller (8) outputs the adaptive parameters Give it to ISRF phase-locked loop (7), ISRF phase-locked loop (7) gets the estimated angle of the motor Actual rotor position angle θ e Estimated angle with motor The error angle Δθ between the motor and the estimated angular velocity Where λ is a positive constant and the estimated rotor speed is ω e is the rotor angular velocity, k1 is the filter bandwidth, kf is the loop gain proportional parameter, kp and ki are the proportional and integral parameters respectively, Δω e (t+1) is the value of the compensation item of the speed evaluation at time (t+1); Δω e (t) is the value of the compensation term at time t in the speed evaluation; Step 4): The motor estimated angle The error angle Δθ is fed back to the signal conversion module (5), and the motor estimates the angular velocity With the given motor angular velocity ω * After comparison, the result is input into the speed controller (1); Step 5): The motor speed sensorless device is composed of the speed controller (1), the signal conversion module (5), the back electromotive force observer (6), the ISRF phase-locked loop (7) and the adaptive controller (8).

2. The construction method according to claim 1, characterized in that: the current The hysteresis PWM module (3) is composed of Park inverse transformation (31), Clark inverse transformation (32), current hysteresis PWM (33), and inverter (34) connected in sequence. The current hysteresis PWM module (3) obtains two currents of the torque winding in the dq rotating coordinate system. and Then, the current in the α-β coordinate system is obtained by Park inverse transformation (31): and Then the three-phase current is output through Clark inverse transformation (32) The current hysteresis PWM (33) obtains the three-phase current Output three-phase DC current i Ma 、i Mb 、i Mc After receiving the three-phase DC current signal, the inverter (34) outputs the three-phase winding current i MU 、i MV 、i MW , the three-phase winding current i MU 、i MV 、i MW The signal is input to the torque subsystem (4).

3. The construction method according to claim 2, characterized in that: The transformation formula of the Park inverse transformation (31) is The transformation formula of the Clark inverse transform (32) is N3 is the effective number of turns of each phase of the three-phase winding, and N2 is the effective number of turns of the two-phase winding.

4. The construction method according to claim 1, characterized in that: The signal conversion module (5) is composed of Clark transformation (51), Park transformation (52) and Park transformation (53). Clark transformation (51) transforms the three-phase winding current and voltage i MU / u MU 、i MV / u MV 、i MW / u MW Convert the current and voltage i into the α-β coordinate system Mα / u Mα and i Mβ / u Mβ , the current i in the α-β coordinate system Mα and i Mβ Input Park transformation (52), the current and voltage i in the α-β coordinate system Mα / u Mα and i Mβ / u Mβ Input Park Transform (53), Park Transform (52) is based on angle and current i Mα ,i Mβ Converted to the current i in the dq rotating coordinate system Md and i Mq , Park transform (53) is based on the estimated angle and current voltage i Mα / u Mα and i Mβ / u Mβ , converted to the current and voltage i in the γ-δ rotating coordinate system Mγ / u Mγ and i Mδ / u Mδ , the γ-δ rotating coordinate system is a coordinate system that differs from the dq rotating coordinate system by Δθ.

5. The construction method according to claim 4, characterized in that: The transformation formula of the Clark transformation (51) is N3 is the effective number of turns of each phase of the three-phase winding, and N2 is the effective number of turns of the two-phase winding.

6. The construction method according to claim 5, characterized in that: The transformation formula of ark transformation (52) is: The transformation formula of Park transformation (53) is:

7. The construction method according to claim 1, characterized in that: Estimated back EMF Estimated back EMF ω c is a positive constant; s is the transformation parameter in the Laplacian transformation; e Mα , e Mβ is the component of the actual value of the back EMF on the α-β axis.

8. The construction method according to claim 1, characterized in that: The rotor angular velocity u δ is the component of the torque winding voltage on the δ axis; R M is the torque winding resistance; L Mγ and L Mδ is the self-inductance of the torque winding γ-δ axis; ψ0 is the permanent magnet flux; i Mγ and i Mδ is the component of the torque winding voltage on the γ-δ axis; s is the transformation parameter in the pull-type transformation.

9. The construction method according to claim 1, characterized in that: The actual rotor position angle ω c is a positive constant; s is the transformation parameter in the Pull transformation.

Images