Method for constructing flux linkage controller of permanent magnet synchronous motor based on sensorless control

Abstract

The invention discloses a method for constructing a flux linkage controller of a permanent magnet synchronous motor based on sensorless control, which is composed of a first PI module, a stator flux linkage vector calculation module, a gating circulation unit, an inverter, a signal conversion module, a sensorless predictive control module and a speed conversion module, and constructs a novel second-order generalized integral system. Electric signals can be estimated only by measuring current and voltage signals of the B phase and the C phase, and the number of sensors is reduced; a stator flux linkage vector calculation module and a gating circulation unit are constructed, optimization iteration of a cost function in conventional model prediction control is replaced, then an inverter is directly controlled to drive a motor, redundant calculation links are reduced, and the calculation amount is reduced; information in the control process is fully utilized, accurate acquisition of angular velocity, flux linkage and torque information is achieved, the information is combined with model prediction control to form an angular velocity-torque-flux linkage closed-loop control structure, and the method has excellent static and dynamic characteristics and strong robustness.

Description

Technical Field

[0001] The invention belongs to the field of motor control and relates to a flux controller of a permanent magnet synchronous motor (PMSM), in particular to a construction method of the flux controller. Background Art

[0002] Permanent magnet synchronous motor sensorless control technology is a high-performance control method achieved by omitting position and speed sensors in the motor. This technology uses advanced control algorithms and observers to accurately estimate the motor rotor position and speed using the motor's own back electromotive force or flux information. Sensorless control technology has been widely used in motor systems, especially in permanent magnet synchronous motors, providing a feasible solution for improving system performance, reducing costs and improving reliability.

[0003] In permanent magnet synchronous motors, flux is a key parameter that describes the strength of the motor's magnetic field. By combining flux information with sensorless control, the motor's operating status can be estimated more accurately. This combination enables the system to achieve efficient and precise control without external sensors, reducing system complexity and cost. In addition, the use of flux information for sensorless control also improves the robustness of the system, making it more adaptable to different working conditions and environments. Therefore, the development of sensorless control technology for permanent magnet synchronous motors and the rational use of flux have jointly promoted the advancement of motor control technology, providing more reliable and efficient motor control solutions for various application scenarios.

[0004] The invention patent with the patent publication number CN116961513A provides a flux control method based on model predictive control. By injecting high-frequency signals into the motor stator, the rotor's flux information is reconstructed according to the response signal and corrected to form a flux observer to obtain the rotor's position information, thereby realizing sensorless control. The invention patent with the patent publication number CN111431460A provides a flux control method based on sensorless model prediction. The rotor's position information is obtained based on a sliding mode observer and a phase-locked loop, and the flux is controlled through a sliding mode controller in combination with model prediction. The problems with the above two flux control methods are that they are both simple combinations of sensorless control and flux control, and the information in the control process is not fully utilized. The integrated design and cost of the actual control system are not considered. As a result, the above two methods require multiple electrical signal sensors and redundant calculation links to achieve the ideal control effect. Summary of the invention

[0005] The purpose of the present invention is to provide a method for constructing a permanent magnet synchronous motor flux controller based on sensorless control. The constructed flux controller forms a closed-loop control structure of "angular velocity-torque-flux", and performs sensorless control only through two sensor electrical signals, which reduces the amount of calculation and makes the system have stronger robustness and good dynamic characteristics.

[0006] To achieve the above object, the technical solution adopted by the construction method of the permanent magnet synchronous motor flux controller based on sensorless control of the present invention includes the following steps:

[0007] Step 1): A second-order generalized integral estimation module is composed of a double second-order generalized integral system, a unit phase-locked loop, a flux calculation module, a discretization processing module, a torque calculation module and a discrete voltage and current calculation module; the voltage u of the motor β and current i β As the input of the biquad generalized integrator system, the estimated voltage u′ of the α-β axis is output α , u′ β and estimated current i′ α , i′ β ; A unit phase-locked loop is formed by a normalized amplitude estimation module and a phase-locked loop, and the estimated voltage u′ α , u′ β The normalized unit voltage is obtained by the amplitude estimation module The estimated angular velocity is then processed by the phase-locked loop Estimated voltage u′ α , u′ β and estimated current i′ α , i′ β As the input of the flux calculation module, the output is the estimated stator flux The torque calculation module is based on the current i β and the estimated stator flux Calculate the estimated torque The discretization processing module estimates the stator flux based on the motor load angle δ and the Calculate the discrete stator flux The voltage u′ β , u′ α and current i′ β , i′ α As the input of the discrete voltage and current calculation module, the output is the estimated discrete voltage u′ s (k) and discrete current i′ s (k);

[0008] Step 2): The sensorless predictive control module is formed by connecting the second-order generalized integral estimation module and the flux linkage prediction module. The control voltage u of the permanent magnet synchronous motor b ,,u cand control current i b ,i c The signal conversion module converts it into the voltage u of the α-β axis β and current i β , motor load angle δ, voltage u β and current i β Input into the second-order generalized integral estimation module to obtain the estimated angular velocity and estimated torque Discrete voltage u′ s (k), discrete current i′ s (k) and discrete stator flux The discrete voltage u′ s (k), discrete current i′ s (k) and discrete stator flux Input to the flux prediction module to predict the discrete stator flux at the next moment

[0009] Step 3): Estimated angular velocity The speed conversion module converts it into an estimated speed. Set the reference speed N * With the estimated speed The difference is input into the first PI module, and the reference torque is obtained through calculation and processing. The reference torque The set reference stator flux amplitude |ψ s | * and estimated torque The two are input into the stator flux vector calculation module and the reference stator flux is output.

[0010] Step 4): Reference stator flux and discrete stator flux The gated cyclic unit module is inputted together, and the gated cyclic unit module outputs six switch signals to the inverter;

[0011] Step 5): The first PI module, the stator flux vector calculation module, the gated circulation unit, the inverter, the signal conversion module, the sensorless predictive control module and the speed conversion module together constitute a flux controller.

[0012] The six-way switch signal output by the gated cycle unit module Where Z t =σ(W z X t +U z H t-1 +B z ), R t=σ(W r X t +U r H t-1 +B r ), t is the time step of each execution, is the input, is the final hidden state, i.e., output, and R is a set of real numbers; is the hidden state of the previous moment, R t ∈(0, 1) is the reset gate, Z t ∈(0, 1) is the update gate, is a candidate hidden state, ⊙ represents the vector element multiplication operation, σ(x) is the sigmoid function, and tanh is the hyperbolic tangent function; W r , U r , B r , W z , U z , B z , W, U, B are the parameters for learning and adjustment, m and n are the number of rows and columns of the output variables, m=2, n=6.

[0013] The beneficial effects of the present invention are:

[0014] 1. The present invention makes full use of the signal output by the motor by constructing a new second-order generalized integral system. The α-β axis electrical signal can be estimated by measuring only the current signal and the voltage signal of the B phase and the C phase, thereby reducing the number of sensors and effectively reducing the system cost.

[0015] 2. The present invention replaces the optimization iteration of the cost function in the conventional model predictive control by constructing a stator flux vector calculation module and a gated cyclic unit module, and then directly controls the inverter to drive the motor, reducing redundant calculation links, reducing the amount of calculation, and making the system more robust and have good dynamic characteristics.

[0016] 3. The present invention makes full use of the information in the control process to achieve accurate acquisition of angular velocity, flux and torque information and combines it with model predictive control to form a closed-loop control structure of angular velocity-torque-flux. It has excellent static and dynamic characteristics and strong robustness, achieves ideal control effects, and can be widely used in engineering practice. It has academic and practical significance for the research and application in the field of permanent magnet synchronous motor control. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a structural block diagram of a permanent magnet synchronous motor flux controller based on sensorless control constructed by the construction method of the present invention;

[0018] Figure 2 yes Figure 1The construction block diagram of the sensorless predictive control module in FIG.

[0019] Figure 3 yes Figure 2 The construction block diagram of the second-order generalized integral estimation module in ;

[0020] Figure 4 yes Figure 3 The structural block diagram of the unit phase-locked loop in;

[0021] Figure 5 yes Figure 1 The structural block diagram of the stator flux vector calculation module in .

[0022] In the figure: 1. First PI module; 2. Second PI module; 11. Stator flux vector calculation module; 12. Gated loop unit; 13. Inverter; 14. Speed ​​conversion module; 15. Signal conversion module; 16. Sensorless predictive control module; 23. Unit phase-locked loop; 51. First improved second-order generalized integrator; 52. Second improved second-order generalized integrator; 54. Flux calculation module; 55. Torque calculation module; 56. Discretization processing module; 57. Discrete voltage and current calculation module. DETAILED DESCRIPTION

[0023] The present invention is further described in detail below in conjunction with the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention:

[0024] like Figure 1 As shown, the permanent magnet synchronous motor flux controller based on sensorless control constructed by the present invention is composed of a first PI module 1, a stator flux vector calculation module 11, a gated circulation unit 12, an inverter 13, a signal conversion module 15, a sensorless predictive control module 16 and a speed conversion module 14, wherein the first PI module 1, the stator flux vector calculation module 11, the gated circulation unit 12, and the inverter 13 are connected in sequence, and the output end of the inverter 13 is connected to the PMSM 17 (i.e., the permanent magnet synchronous motor 17) to control the permanent magnet synchronous motor. The permanent magnet synchronous motor is connected to the signal conversion module 15, the sensorless predictive control module 16 and the speed conversion module 14 in sequence. The control current and control voltage signals of the permanent magnet synchronous motor are converted by the signal conversion module 15 and input into the sensorless predictive control module 16. The angular velocity output by the sensorless predictive control module 16 is converted by the speed conversion module 14 and used as the feedback speed. The output torque is fed back to the stator flux vector calculation module 11, and the output flux is fed back to the gated circulation unit 12 to form a closed-loop control.

[0025] like Figure 1-5As shown, the second-order generalized integral estimation module is composed of a double second-order generalized integral system, a unit phase-locked loop, a flux calculation module, a discretization processing module, a torque calculation module 55 and a discrete voltage and current calculation module 57; the voltage u of the motor β and current i β As the input of the biquad generalized integrator system, the estimated voltage u′ of the α-β axis is output α , u′ β and estimated current i′ α , i′ β The normalized amplitude estimation module and the phase-locked loop constitute a unit phase-locked loop 23, the estimated voltage u ' α , u′ β The normalized unit voltage is obtained by the amplitude estimation module The estimated angular velocity is then processed by the phase-locked loop Estimated voltage u′ α , u′ β and estimated current i′ α , i′ β As the input of the flux calculation module 54, the output is the estimated stator flux The torque calculation module 55 is based on the current i β and the estimated stator flux Calculate the estimated torque The discretization processing module (56) is based on the motor load angle δ and the estimated stator flux Calculate the discrete stator flux The voltage u′ β , u′ α and current i′ β , i′ α As the input of the discrete voltage and current calculation module 57, the output is the estimated discrete voltage u′ s (k) and discrete current i′ s (k); The sensorless predictive control module 16 is connected to form a second-order generalized integral estimation module and a flux prediction module, and the control voltage u of the permanent magnet synchronous motor b ,,u c and control current i b ,i c The signal conversion module 15 converts it into the voltage u of the α-β axis. β and current i β , motor load angle δ, voltage u β and current i β Input into the second-order generalized integral estimation module to obtain the estimated angular velocity and estimated torque Discrete voltage u′ s (k), discrete current i′ s(k) and discrete stator flux The discrete voltage u′ s (k), discrete current i′ s (k) and discrete stator flux Input to the flux prediction module to predict the discrete stator flux at the next moment Estimated angular velocity The speed conversion module 14 converts the estimated speed into Set the reference speed N * With the estimated speed The difference is input into the first PI module 1, and the reference torque is obtained after calculation processing. The reference torque The set reference stator flux amplitude |ψ s | * and estimated torque are input into the stator flux vector calculation module (11) and output as reference stator flux Reference stator flux and discrete stator flux The gated loop unit module 12 is inputted together, and the gated loop unit module 12 outputs six switch signals to the inverter 13. The first PI module 1, the stator flux vector calculation module 11, the gated loop unit 12, the inverter 13, the signal conversion module 15, the sensorless predictive control module 16 and the speed conversion module 14 together constitute a flux controller. The specific construction method of the flux controller is as follows:

[0026] When the PMSM 17 is running, the motor load angle δ of the PMSM 17 can be estimated by a conventional sensorless control method, and the motor load angle δ is used as the first input signal of the sensorless predictive control module 16. The control voltage u of the B phase and the C phase of the PMSM 17 is extracted by the voltage sensor and the current sensor. b ,,u c and control current i b ,i c Signal, control voltage u b ,,u c and control current i b ,i c The signal is converted into the voltage u of the α-β axis by the signal conversion module 15 β and current i β , which is expressed as follows:

[0027]

[0028] The transformed voltage u β and current i βThe signals are used as the second and third input signals of the sensorless predictive control module 16 , and are input into the sensorless predictive control module 16 together with the motor load angle δ.

[0029] Combination Figure 2 As shown, the sensorless predictive control module 16 is composed of a second-order generalized integral estimation module and a flux linkage prediction module. The input of the sensorless predictive control module 16 is the motor load angle δ, the voltage u β and current i β signal, the output is the estimated angular velocity Estimated torque and the discrete stator flux at the next moment Motor load angle δ, voltage u β and current i β The signal is input into the second-order generalized integral estimation module to obtain the estimated angular velocity Estimated torque The discrete voltage u′ at the current moment s (k), discrete current i′ s (k) and discrete stator flux The current discrete voltage u′ s (k), discrete current i′ s (k) and discrete stator flux Input into the flux prediction module, which predicts the discrete stator flux at the next moment. And output.

[0030] Combination Figure 2 As shown, the second-order generalized integral estimation module is composed of a double second-order generalized integral system, a unit phase-locked loop 23, a flux calculation module 54, a discretization processing module 56, a torque calculation module 55 and a discrete voltage and current calculation module 57, wherein the double second-order generalized integral system is composed of a first improved second-order generalized integrator 51 and a second improved second-order generalized integrator 52 in parallel.

[0031] Combination Figure 3 As shown, the voltage u β The signal is used as the input of the first improved second-order generalized integrator 51, and the current i β The signal is used as the input of the second improved second-order generalized integrator 52, and the first improved second-order generalized integrator 51 outputs the estimated voltage u′ of the α-β axis. α , u′ β The second improved second-order generalized integrator 52 outputs the estimated current i′ of the α-β axis α , i′ β , the estimated voltage u′ α , u′ β and estimated current i′ α , i′β As the output signal of the double second-order generalized integral system. The expression of the double second-order generalized integral system is:

[0032]

[0033] Among them, K is the gain of the motor system. In different engineering fields, there is no unified standard for the specific value of the gain. It is determined according to the specific needs of the motor's conventional control. is the estimated angular velocity.

[0034] Combination Figure 4 As shown, according to the estimated voltage u′ output by the biquad generalized integrator α , u′ β Construct a unit phase-locked loop 23, which is composed of a normalized amplitude estimation module and a phase-locked loop, and estimates the voltage u′ α , u′ β After the amplitude estimation module and normalization processing, the normalized unit voltage is obtained The expression is as follows:

[0035]

[0036] in, and is the unit voltage signal, K α and K β is the motor voltage amplitude, which can be obtained by querying the motor parameters.

[0037] The normalized unit voltage Input into the phase-locked loop, and the estimated angular velocity is obtained through the phase-locked loop transformation processing. And as the output signal of the unit phase-locked loop 23, the estimated angular velocity Returning to the double generalized second-order integral system, its expression is:

[0038]

[0039] Among them, K Gp and K Gi The general G PI The proportional and integral coefficients of the module need to be determined according to the different characteristics of the motor. s is the system sampling time, △e is the error signal. Figure 4 In the above formula, 1 / s represents the integral. Here, the angular velocity in formula (6) is is the angular velocity of the feedback.

[0040] like Figure 3 As shown, the estimated voltage u′ output by the biquadric generalized integrator system isα , u′ β and estimated current i′ α , i′ β As input to the flux calculation module 54, the flux calculation module 54 outputs the estimated stator flux Its expression is as follows:

[0041]

[0042] Among them, R s is the stator resistance.

[0043] The estimated stator flux output by the flux calculation module 54 is and The current i is input into the torque calculation module 55 and the discretization processing module 56 respectively. The torque calculation module 55 then calculates the current i based on the input current i β Signal to calculate estimated torque Its expression is as follows:

[0044]

[0045] Where Pn is the number of pole pairs of the motor.

[0046] The discretization processing module 56 then calculates the discrete stator flux at the current moment based on the motor load angle δ Its expression is:

[0047]

[0048] Where k represents the number of control cycles, T d The preset discrete sampling time is usually less than 10 -4 s, δ is the motor load angle of the motor operation, which is estimated by the ordinary sensorless control method.

[0049] The voltage u′ output by the biquadric generalized integrator system β , u′ α and current i′ β , i′ α As the input signal of the discrete voltage and current calculation module 27, the output is the estimated discrete voltage u′ s (k) and discrete current i′ s (k), the expression is as follows:

[0050]

[0051] Wherein, k represents the number of control cycles that have passed, and j is the sign of the complex number.

[0052] like Figure 2, the discrete stator voltage u′ output by the second-order generalized integral estimation module s (k), discrete stator current i′ s (k) and the estimated discrete stator flux Input to the flux prediction module, the flux prediction module predicts the discrete stator flux at the next moment The expression is as follows:

[0053]

[0054] For voltage and current signals, T d Discretize for a preset discrete sampling time.

[0055] like Figure 1 , the estimated angular velocity output by the sensorless predictive control module 16 Input to the speed conversion module 14, converted into an estimated speed The output-input relationship of the speed conversion module 14 can be expressed as follows:

[0056]

[0057] Where Pn is the number of pole pairs of the motor.

[0058] Set the reference speed N * With the estimated speed The difference is input into the first PI module 1, and the reference torque is obtained by the proportional-integral operation of the first PI module 1. As the output signal of the first PI module 1, its expression is:

[0059]

[0060] Among them, K p1 , K i1 They are respectively the proportional gain and integral gain coefficients of the first PI module 1 , and their values ​​need to be specifically determined according to different types of motors.

[0061] The reference torque The reference stator flux amplitude |ψ is set s | * and the estimated torque output by the sensorless predictive control module 16 They are input into the stator flux vector calculation module 11 together.

[0062] Recombination Figure 5 The stator flux vector calculation module 11 is composed of the second PI module 2, the integral module and the coordinate transformation module. The output of the second PI module 2 is the slip angular velocity ω s , the reference torque and estimated torque The second PI module 2 outputs the slip angular velocity ω s , whose expression is:

[0063]

[0064] K p2 , K i2 are the proportional gain and integral gain coefficients of the second PI module 2 respectively.

[0065] The slip angular velocity ω is calculated by the integral module s By integrating, we can get the slip angle θ s :

[0066] θ s =∫ω s dt (15)

[0067] The reference stator flux amplitude |ψ s | * and slip angle θ s Input into the coordinate transformation module, the output is the reference stator flux

[0068]

[0069] Where T is the matrix transpose.

[0070] like Figure 1 As shown, the reference stator flux and the predicted discrete stator flux at the next moment output by the sensorless prediction control module 16 are input into the gated recurrent unit module 12, As the input signal of the gated cycle unit 12, the gated cycle unit module 12 outputs six switch signals s1, s2, s3, s4, s5, s6. The input-output relationship can be expressed as:

[0071]

[0072] H t ={s1,s2,s3,s4,s5,s6}

[0073] Among them, t is the time step of each execution, is the input, is the final hidden state, i.e., output, and R is a set of real numbers; is the hidden state of the previous moment, R t ∈(0, 1) is the reset gate, Z t ∈(0, 1) is the update gate, is a candidate hidden state, ⊙ represents the vector element multiplication operation, σ(x) is the sigmoid function, and tanh is the hyperbolic tangent function; W r , U r , B r , W z , U z , B z , W, U, B are learnable and adjustable parameters, which are obtained through the optimization algorithm of gated recurrent unit (GRU) training; m and n are the number of rows and columns of the output variables, which are m=2 and n=6 respectively in the present invention.

[0074] The six-way switch signal s 1～6 As the output signal of the gated cycle unit 12, it is input into the inverter 13, and after the inversion process, the DC voltage is converted into a three-phase AC voltage. a ,s b ,s c The output of the inverter 13 is input to the permanent magnet synchronous motor to control the permanent magnet synchronous motor.

[0075] like Figure 1 As shown, the first PI module 1, the stator flux vector calculation module 11, the gated circulation unit 12, the inverter 13, the speed conversion module 14, the signal conversion module 15, and the sensorless predictive control module 16 together constitute a flux controller. When the flux controller is working, after the PMSM 17 is running, the detected motor current i b ,i c and voltage u b ,u c Input to signal conversion 15 and get voltage u after conversion β and current i β The signal is input into the sensorless predictive control module 16 to estimate the speed, torque and flux; the sensorless predictive control module 16 outputs the electrical angular velocity To the speed conversion module 14, the estimated torque is output To the stator flux vector calculation module 11, output the predicted discrete stator flux The speed conversion module 14 converts the input electrical angular velocity into After transformation, the estimated speed is With reference speed N * After the difference is made, it is input into the first PI module 1 to obtain the reference torque The reference stator flux amplitude |ψ s | * , reference torque and the predicted discrete stator flux The switching signals s1-s6 are input into the gated cycle unit 12, and then input into the inverter 13. The inverter 13 processes the switching signals and outputs a voltage signal S a , S b and S c The motor is driven to PMSM17 to run, and the motor system is controlled according to the reference flux and the reference mechanical angular velocity. The method adopted by the present invention optimizes sensorless control and model predictive control, further reduces the number of required sensors, and ensures the accuracy of the obtained signals. On the basis of obtaining good control performance, the proposed method makes full use of electrical signals, and only two electrical sensors are required to obtain all necessary electrical signals, thereby realizing the estimation of motor speed and torque and the prediction of flux, and further realizing the precise control of the motor; in addition, a stator flux vector calculation module and a gated loop unit are designed to replace the traditional cost function calculation module, thereby improving the reliability and dynamic performance of the system. The controller designed based on this method shows excellent control accuracy and robustness; compared with traditional methods, the demand for sensors is lower, thereby reducing the cost and space occupied by the system, and has more significant economic benefits.

[0076] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for constructing a permanent magnet synchronous motor flux controller based on sensorless control, characterized in that The following steps are involved: Step 1): A second-order generalized integral estimation module is formed by a dual second-order generalized integral system, a unit phase-locked loop (23), a flux calculation module (54), a discretization processing module (56), a torque calculation module (55) and a discrete voltage and current calculation module (57); the voltage u of the motor β and current i β As the input of the biquad generalized integrator system, the estimated voltage u′ of the α-β axis is output α , μ′ β and estimated current i′ α , i′ β ; The normalized amplitude estimation module and the phase-locked loop form a unit phase-locked loop (23), and the estimated voltage u′ α , u′ β The normalized unit voltage is obtained by the amplitude estimation module The estimated angular velocity is then processed by the phase-locked loop Estimated voltage u′ α , u′ β and estimated current i′ α , i′ β As the input of the flux calculation module (54), the output is the estimated stator flux The torque calculation module (55) is based on the current i β and the estimated stator flux Calculate the estimated torque The discretization processing module (56) is based on the motor load angle δ and the estimated stator flux Calculate the discrete stator flux The voltage u′ β , u′ α and current i′ β , i′ α As the input of the discrete voltage and current calculation module (57), the output is the estimated discrete voltage u′ s (k) and discrete current i s ′(k); Step 2): The sensorless predictive control module (16) is formed by connecting the second-order generalized integral estimation module and the flux linkage prediction module. The control voltage u of the permanent magnet synchronous motor b ,,u c and control current i b ,i c The signal is converted into the voltage u of the α-β axis by the signal conversion module (15). β and current i β , motor load angle δ, voltage u β and current i β Input into the second-order generalized integral estimation module to obtain the estimated angular velocity and estimated torque Discrete voltage u′ s (k), discrete current i s ′(k) and discrete stator flux The discrete voltage u′ s (k), discrete current i s ′(k) and discrete stator flux Input to the flux prediction module to predict the discrete stator flux at the next moment Step 3): Estimated angular velocity The speed conversion module (14) converts the speed into an estimated speed. Set the reference speed N * With the estimated speed The difference is input into the first PI module (1), and the reference torque is obtained through calculation processing. The reference torque The reference stator flux amplitude |ψ is set s | * and estimated torque are input into the stator flux vector calculation module (11) and output as reference stator flux Step 4): Reference stator flux and discrete stator flux The six switching signals are inputted into a gated cycle unit module (12) and the gated cycle unit module (12) outputs six switching signals to an inverter (13); Step 5): The first PI module (1), the stator flux vector calculation module (11), the gated loop unit (12), the inverter (13), the signal conversion module (15), the sensorless predictive control module (16) and the speed conversion module (14) together constitute a flux controller.

2. The method for constructing a permanent magnet synchronous motor flux controller based on sensorless control according to claim 1, characterized in that: In step 4), the six-way switch signal output by the gated cycle unit module (12) Where Z t =σ(W z X t +U z H t-1 +B z ), R t =σ(W r X t +U r H t-1 +B r ), t is the time step of each execution, is the input, is the final hidden state, i.e., output, and R is a set of real numbers; is the hidden state of the previous moment, R t ∈(0, 1) is the reset gate, Z t ∈(0, 1) is the update gate, is a candidate hidden state, ⊙ represents the vector element multiplication operation, σ(x) is the sigmoid function, and tanh is the hyperbolic tangent function; W r , U r , B r , W z , U z , B z , W, U, B are the parameters for learning and adjustment, m and n are the number of rows and columns of the output variables, m=2, n=6.

3. The method for constructing a permanent magnet synchronous motor flux controller based on sensorless control according to claim 1, characterized in that: In step 1), the estimated voltage output by the biquad generalized integrator system is Estimated current K is the gain of the motor system.

4. The method for constructing a permanent magnet synchronous motor flux controller based on sensorless control according to claim 1 is characterized in that: In step 1), the phase-locked loop estimates the angular velocity in, K Gp and K Gi G PI The module's proportional and integral coefficients, T s is the system sampling time.

5. The method for constructing a permanent magnet synchronous motor flux controller based on sensorless control according to claim 1, characterized in that: The stator flux vector calculation module (11) is composed of a second PI module (2), an integration module and a coordinate transformation module. The second PI module (2) outputs a slip angular velocity K p2 , K i2 are the proportional gain and integral gain coefficients of the second PI module (2) respectively; the integral module then calculates the slip angular velocity ω s Get the slip angle θ s =∫ω s dt, the reference stator flux amplitude |ψ s | * and slip angle θ s The reference stator flux is output through the coordinate transformation module T is the matrix transpose.

6. The method for constructing a permanent magnet synchronous motor flux controller based on sensorless control according to claim 1, characterized in that: In step 1), the unit voltage K α and K β is the motor voltage amplitude.

7. The method for constructing a permanent magnet synchronous motor flux controller based on sensorless control according to claim 1, characterized in that: In step 1), the estimated stator flux is output by the flux calculation module (54). R s is the stator resistance; the torque calculation module 55 is based on the formula Calculate the estimated torque Pn is the number of pole pairs of the motor; The discretization processing module 56 is based on the formula Calculate the discrete stator flux k is the number of control cycles, T d is the preset discrete sampling time; The discrete voltage u′ output by the discrete voltage and current calculation module 27 s (k) = u′ α (k)+ju′ β (k), discrete current i′ s (k) = i′ α (k)+ji′ β (k),u′ s (k) and discrete current i′ s (k), j is a complex number.

8. The method for constructing a permanent magnet synchronous motor flux controller based on sensorless control according to claim 1, characterized in that: In step 2), the flux prediction module predicts the discrete stator flux at the next moment R s is the stator resistance, T d is the preset discrete sampling time. For voltage and current signals, T d Discretize for a preset discrete sampling time.

9. The method for constructing a permanent magnet synchronous motor flux controller based on sensorless control according to claim 1, characterized in that: The reference torque output by the first PI module (1) K p1 , K i1 are the proportional gain and integral gain coefficients respectively.

10. The method for constructing a permanent magnet synchronous motor flux controller based on sensorless control according to any one of claims 1 to 9, characterized in that: In step 1), the first improved second-order generalized integrator (51) and the second improved second-order generalized integrator (52) are connected in parallel to form the double second-order generalized integrator system, and the voltage u β As the input of the first improved second-order generalized integrator (51), the estimated voltage u′ of the α-β axis is output α , u′ β ; Current i β As the input of the second improved second-order generalized integrator (52), the estimated current i' of the α-β axis is output α , i′ β .

Images