Design method of magnetic flux leakage adjustable permanent magnet motor sliding mode prediction position sensorless controller

Abstract

The invention discloses a magnetic flux leakage adjustable permanent magnet motor sliding mode prediction sensorless controller design method, which is characterized in that a permanent magnet motor double-closed-loop controller is connected to the front end of a magnetic flux leakage adjustable permanent magnet motor system, and a ClarkPark coordinate transformation module is connected to the rear end of the magnetic flux leakage adjustable permanent magnet motor system; the ClarkPark coordinate transformation module, the sliding-mode prediction observer and the improved phase-locked loop are connected in sequence, the sliding-mode prediction observer cancels selection of an approaching rate function and determination of a sliding-mode gain coefficient while ensuring quick response and minimum buffeting of sliding-mode motion, the phase-locked loop is redesigned through a mathematical iteration method, and the phase-locked loop is obtained. The parameter tuning process of a phase-locked loop PI regulator is canceled, the dichotomy iterative algorithm is utilized to realize observation of the position and the rotating speed of a motor rotor, the controllability of the observation precision is realized, and the bandwidth of a sensorless controller is effectively expanded; the effective flux linkage is introduced into a motor mathematical model, motor electromagnetic parameter coupling terms in the model are eliminated, and motor back electromotive force can be automatically tracked in the observation process.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of motor control, and relates to an adjustable leakage flux permanent magnet (ALF-PM) motor, in particular to a design method of a sliding mode predictive position sensorless controller thereof. BACKGROUND

[0002] Electric vehicles are popularized and promoted due to their advantages of environmental protection and energy saving. The operating conditions of electric vehicles are complex and changeable, and higher requirements are put forward for the performance of the driving motor of the electric vehicles, such as the speed regulation range, efficiency and output torque. The ALF-PM motor has the functions of controllable leakage flux, such as light load and more leakage flux and heavy load and less leakage flux, due to its special magnetic circuit design, and has the characteristics of wider speed regulation range and greater load carrying capacity, and is very suitable for application in the field of electric vehicles.

[0003] In order to realize high-performance driving control of the motor system of the electric vehicle, the rotor position information needs to be accurately obtained. The position sensorless control has the advantages of cost saving, space saving and stability, and is applied to the motor driving control of the electric vehicle. Among them, the position sensorless control technology based on the sliding mode observer has the advantages of strong robustness, simple structure and easy implementation, and has been widely applied.

[0004] For the ALF-PM motor, due to its special magnetic circuit design, the electromagnetic parameters change under complex operating conditions while realizing the controllable leakage flux. Therefore, under complex operating conditions, the ALF-PM motor sliding mode observation position sensorless driving control has the following problems: the fixed position controller parameters do not have universality due to the complex operating conditions and the change of electromagnetic parameters of the ALF-PM motor, especially the determination of the sliding mode gain coefficient, which cannot guarantee the low chattering and rapidity of the sliding mode motion, and will reduce the control performance of the position controller; at the same time, before the use of the sliding mode observer and the quadrature phase-locked loop, there is also a parameter adjustment process.

[0005] In order to solve the parameter adjustment problem of the position controller based on the sliding mode observer, ensure the low chattering and rapidity of the sliding mode motion, the control method disclosed in the document with the Chinese patent publication number CN109067289B and the name "a deep learning optimized position sensorless BLDC sliding mode observer control method" optimizes the hyperbolic tangent coefficient and the sliding mode gain parameter of the sliding mode observer through deep learning self-adaption, effectively reduces the chattering phenomenon caused by the traditional observer, and can accurately estimate the line back electromotive force, so as to better estimate the rotor position information; but this method has complex structure, large calculation burden and many parameters that need to be adjusted. The control system and method disclosed in the document with the Chinese patent publication number CN115189616A and the name "a permanent magnet synchronous motor position sensorless control system and control method" use a fuzzy PID control module to suppress the chattering in the control system; but this method does not improve the position controller, but achieves the effect of suppressing the chattering through other ways.

[0006] In general, the existing position controller based on the sliding mode observer often has complex structure, heavy calculation burden or limited application range, and cannot realize high-performance control of the ALF-PM motor. Therefore, it is necessary to design a new type of position sensorless controller based on the sliding mode observer of the ALF-PM motor, to solve the difficulties caused by the complex electromagnetic parameter changes and complex operating conditions of the leakage flux adjustable permanent magnet motor to the observation accuracy and parameter adjustment of the position sensorless controller, and realize high-performance driving control of the leakage flux adjustable permanent magnet motor. SUMMARY

[0007] The present application aims to solve the above problems by providing a design method of a sliding mode predictive position sensorless controller for a leakage flux adjustable permanent magnet motor, which combines the predictive control principle with the sliding mode control, ensures the low chattering and rapidity of the sliding mode motion, improves the control performance of the position controller, and thus realizes accurate rotor position self-detection of the ALF-PM motor.

[0008] Technical scheme: The design method of the sliding mode predictive position sensorless controller for the leakage flux adjustable permanent magnet motor adopts the following technical scheme, which includes the following steps:

[0009] Step A: connect the pulse width modulation module, the inverter module, the leakage flux adjustable permanent magnet motor and the current sensor module in series to form a leakage flux adjustable permanent magnet motor system, and the input signal of the leakage flux adjustable permanent magnet motor system is voltage The output signal is current

[0010] Step B: the speed loop PI controller module, the current loop PI controller module and the two inverse Park coordinate transformation modules are combined to form a permanent magnet motor double closed loop controller, the input signals of the permanent magnet motor double closed loop controller are the given reference speed the feedback motor estimated speed and the feedback motor estimated rotor position angle and the feedback current the output signals are the reference current and the voltage

[0011] Step C: the permanent magnet motor double closed loop controller is connected at the front end of the flux leakage adjustable permanent magnet motor system, the Clark\Park coordinate transformation module is connected at the rear end of the flux leakage adjustable permanent magnet motor system, the Clark\Park coordinate transformation module, the sliding mode prediction observer and the improved phase-locked loop are sequentially connected to form a sliding mode prediction sensorless controller;

[0012] the current and the motor estimated rotor position angle fed back by the improved phase-locked loop together as the input of the Clark\Park coordinate transformation module, the Clark\Park coordinate transformation module outputs the current

[0013] the given reference speed and the voltage and the reference current and the current together as the input signals of the sliding mode prediction observer, the sliding mode prediction observer outputs the estimated back electromotive force as the input of the improved phase-locked loop, the improved phase-locked loop outputs the motor estimated speed and the motor estimated rotor position angle

[0014] Further, in step C, the specific design process of the improved phase-locked loop is as follows:

[0015] Step 1): the initial observation angles are θ1=0 rad, θ2=2π / 3 rad and θ3=4π / 3 rad;

[0016] Step 2): the initial observation angles θ1, θ2 and θ3 and the estimated back electromotive force are sequentially substituted into the evaluation function to calculate, and θ1, θ2 and θ3 are used to replace the unknown at this time to obtain the evaluation functions

[0017] Step 3): After each evaluation function calculation, check whether the real-time small loop number g is 3, if g < 3, return to calculate a new round of evaluation function, if g ≥ 3, this loop ends and the observation angles with the maximum and the second maximum evaluation function results are redefined as new observation angles θ1, θ2, and the new observation angle θ3 is calculated by the formula θ3 = (θ1 + θ2) / 2;

[0018] Step 4): After obtaining the new observation angles θ1, θ2, θ3, enter the new loop iteration calculation, and the real-time loop iteration number flag is j. In the large loop with j as the flag, the same operation and processing as the last round of small loop with g as the flag are first performed, and then after the small loop is run, only the angle corresponding to the maximum evaluation function result is selected as the iteration angle θ2, and the other two iteration angles are obtained by the formula , wherein

[0019] Step 5): Determine whether the loop flag j reaches the set iteration number N, when the set iteration number N is reached, output the iteration angle θ2 with the maximum evaluation function, which is the rotor position angle output by the iteration-based phase-locked loop

[0020] Step 6): Calculate the estimated rotor position angle of the motor τ represents the time constant, and the estimated rotor position angle of the motor is differentiated to obtain the estimated speed of the motor

[0021] Further, in step C, the sliding mode predictor calculates the predicted back-EMF amplitude under the current working condition according to the reference speed ; f The predicted back-EMF amplitude E is amplified and reduced to determine the predicted back-EMF finite set interval, and the predicted back-EMF amplitude d E of each predicted back-EMF discrete point is obtained through the finite set interval.

[0022] The current is delayed by one step to obtain the predicted current at the current time k Based on the amplitude d E , the predicted current at time k , and the current , a series of predicted back-EMFs and are obtained through alternating change processing.

[0023] The predicted back-EMFs and are obtained.Substitute into the prediction model to obtain a series of next time prediction currents Substitute into the prediction model to obtain a series of next time prediction currents Substitute into the designed cost function J for calculation, and the counter electromotive force value corresponding to the discrete point corresponding to the minimum cost function value is the optimal estimated counter electromotive force value;

[0024] The optimal estimated counter electromotive force value is filtered by a low-pass filter to output an estimated counter electromotive force

[0025] The application has the following beneficial effects after adopting the above technical solutions:

[0026] 1) The new leakage magnetic adjustable permanent magnet motor sliding mode prediction position sensorless controller designed in the application has the advantages of weak dependence on motor electromagnetic parameters, no control parameter adjustment, fast response speed, high observation accuracy and controllability.

[0027] 2) The application cancels the selection of the approaching rate of the position sensorless controller and the sliding mode gain coefficient, introduces the concept of effective flux into the mathematical model of the leakage magnetic adjustable permanent magnet motor, eliminates the motor electromagnetic parameter coupling term in the model, and weakens the dependence of the model on the motor electromagnetic parameters. Based on the mathematical model of the leakage magnetic adjustable permanent magnet motor, the prediction control is integrated into the sliding mode control, a sliding mode prediction observer is designed, the approaching rate function is not selected, and the sliding mode gain coefficient is canceled, the motor counter electromotive force can be automatically tracked in the observation process, not only ensuring the simplicity and easy operation of the observer, but also ensuring the balance between the rapidity and low chattering of the sliding mode motion, and improving the performance of the observer. At the same time, the designed sliding mode prediction observer can realize multi-objective optimization observation. By modifying the structure of the cost function in the sliding mode prediction observer, the optimization target is introduced into the cost function, and the multi-objective optimization of the output of the sliding mode prediction observer can be realized.

[0028] 3) The application cancels the parameter tuning process of the phase-locked loop PI regulator, and the improved phase-locked loop is used to realize the observation of the rotor position and speed of the leakage magnetic adjustable permanent magnet motor by using a bisection iterative algorithm. The improved phase-locked loop does not need the parameter adjustment process and can be directly applied. The position observation accuracy of the improved phase-locked loop is controllable and can be applied to different accuracy requirements, which can effectively save the calculation resources of the controller. At the same time, since the improved phase-locked loop cancels the use of the PI regulator in the traditional quadrature phase-locked loop, the bandwidth is greatly improved, the overall stability and response speed are effectively improved, and the position sensorless controller can adapt to more complex working conditions. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1A structural block diagram of a flux-adjustable permanent magnet motor system;

[0030] Figure 2 A structural block diagram of a double closed-loop controller of a permanent magnet motor;

[0031] Figure 3 A structural block diagram of a flux-adjustable permanent magnet motor position sensorless controller based on a sliding mode predictive observer;

[0032] Figure 4 A structural block diagram of a sliding mode predictive observer and an improved phase-locked loop in Figure 3 ;

[0033] Figure 5 A design flowchart of the improved phase-locked loop in Figure 4 ;

[0034] Figure 6 A comparison chart of rotor position observation errors of a flux-adjustable permanent magnet motor;

[0035] Figure 7 A comparison chart of rotor speed observation errors of a flux-adjustable permanent magnet motor. DETAILED DESCRIPTION

[0036] In order to make the above-mentioned purposes, technical solutions and advantages of the present application clearer, the specific implementation method of the present application will be described in detail below with reference to the accompanying drawings.

[0037] As shown in Figure 1 , a flux-adjustable permanent magnet motor system includes a flux-adjustable permanent magnet motor. A pulse width modulation module, an inverter module, the flux-adjustable permanent magnet motor, and a current sensor module are connected in series to form the flux-adjustable permanent magnet motor system. The input signal of the flux-adjustable permanent magnet motor system is a voltage , which is also the input signal of the pulse width modulation module. The output signal is a current , which is the output signal of the current sensor module. The input voltage is converted into a series of switching pulse signals S a , S b , and S c through the PWM modulation of the pulse width modulation module. The switching pulse signals S a , S b , S c are input signals of the inverter module. The switching pulse signals are used to control the on-off of the switching tubes in the inverter to effectively adjust the three-phase driving voltage of the flux-adjustable permanent magnet motor, thereby achieving driving control of the flux-adjustable permanent magnet motor. The three-phase current of the motor is sampled in real time by the current sensor module and outputted.

[0038] like Figure 2 As shown, a dual-loop controller for a permanent magnet motor is designed. The dual-loop controller for the permanent magnet motor is constructed by combining a speed loop PI controller module, a current loop PI controller module, and two inverse Park coordinate transformation modules. The input signal of this dual-loop controller is a given reference speed. Feedback on estimated motor speed And feedback motor estimates rotor position angle and the feedback current The output signal is the reference current. and voltage Wherein, the given reference speed And feedback motor estimated speed As the input signal to the speed loop PI controller, the speed loop PI controller module outputs a two-phase reference current along the dq axis. Two-phase reference current And feedback motor estimates rotor position angle As the input signal to the first inverse Park coordinate transformation module, the first inverse Park coordinate transformation module outputs the reference current in the αβ coordinate system. and Two-phase reference current and feedback current As the input signal to the current loop PI controller module, the current loop PI controller module outputs the voltage in the dq coordinate system. and voltage As the input signal to the second inverse Park coordinate transformation module, the feedback motor estimated rotor position angle It is also input into the second inverse Park coordinate transformation module, which then outputs the voltage in the αβ coordinate system. The discrete forms of the equations constituting the speed loop PI controller module and the current loop PI controller module are as follows:

[0039]

[0040] Where u represents the output of the PI controller, e represents the input of the PI controller, and T s K represents the discrete time step. P and K I Let k and kT represent the proportional and integral coefficients of the PI controller, respectively, where k is a non-negative integer representing the sampling time of the discrete signal in the discrete system, i.e., kT. s Z represents the sampling time of the discrete signal in the discrete system, and its value remains a non-negative integer not greater than k. Specifically, in the speed loop PI controller module, u is... and e is In the current loop PI controller module, u is and e is and

[0041] The matrix equation of the two inverse Park coordinate transformation modules is:

[0042]

[0043] wherein is the feedback motor estimated rotor position angle. The input signal is the reference current The specific working process of the first inverse Park coordinate transformation module of

[0044]

[0045] The specific working process of the second inverse Park coordinate transformation module of the input signal voltage and

[0046]

[0047] As shown in Figure 3 , a sliding mode predictive sensorless controller is designed. The sliding mode predictive sensorless controller is composed of a permanent magnet motor double closed loop controller, a Clark\Park coordinate transformation module, a sliding mode predictive observer and an improved phase-locked loop. Among them, the permanent magnet motor double closed loop controller is connected at the front end of the leakage adjustable permanent magnet motor system, the Clark\Park coordinate transformation module is connected at the rear end of the leakage adjustable permanent magnet motor system, and the Clark\Park coordinate transformation module, the sliding mode predictive observer and the improved phase-locked loop are connected in sequence. The voltage in αβ coordinate system output by the permanent magnet motor double closed loop controller is input into the leakage adjustable permanent magnet motor system and the sliding mode predictive observer respectively, and the reference current in αβ coordinate system output by the permanent magnet motor double closed loop controller is input into the sliding mode predictive observer.

[0048] The three-phase current and output by the leakage adjustable permanent magnet motor system and the motor estimated rotor position angle fed back by the improved phase-locked loop are collectively used as the input of the Clark\Park coordinate transformation module. The matrix equation of the Clark\Park coordinate transformation module includes the general Clark transformation equation and the Park transformation equation, and the Clark\Park coordinate transformation module outputs the current in αβ coordinate system and the current in dq axis after general transformation. Among them, the current​ The current is fed back to the permanent magnet motor double closed-loop controller. The current and the current are calculated as follows:

[0049]

[0050] The given reference speed and the current and output by the permanent magnet motor double closed-loop controller, and the current output by the Clark\Park coordinate transformation module are taken as input signals of the sliding mode prediction observer, and the estimated back electromotive force is output by the sliding mode prediction observer. The estimated back electromotive force is input into the improved phase-locked loop, and the improved phase-locked loop outputs the motor estimated speed and the motor estimated rotor position angle . The motor estimated speed and the motor estimated rotor position angle are fed back to the permanent magnet motor double closed-loop controller.

[0051] Figure 4 The specific design of the sliding mode prediction observer is shown in the following table: the sliding mode prediction observer calculates the approximate predicted back electromotive force amplitude E under the current working condition according to the input given reference speed , and the calculation expression is as follows:

[0052]

[0053] In the formula, ψ f is the motor permanent magnet flux linkage.

[0054] The E value representing the predicted back electromotive force amplitude is amplified and reduced to determine that the predicted back electromotive force finite set interval is [k1E, k2E], wherein k1 and k2 are gain coefficients, k1 needs to ensure the finite time convergence of the sliding mode motion, the gain coefficient values are k1=0.8 and k2=3.0. Through the predicted back electromotive force finite set interval [k1E, k2E], each predicted back electromotive force discrete point (1, 2, …, n) in the predicted back electromotive force finite set interval [k1E, k2E] can be obtained. Then the predicted back electromotive force finite set interval [k1E, k2E] is discretized through mathematical iteration to obtain the amplitude d E of the predicted back electromotive force reflected by each discrete point.

[0055] The current After a delay step, the predicted current on the αβ axis at time k is obtained. Since the predicted back EMF finite set reflects the amplitude d of the back EMF, E Therefore, these discrete points need to be processed to reflect the AC changes in the back electromotive force. Specifically, based on the amplitude d... E Predicting the current at time k With current The alternating current is processed, and the current at time k is predicted using the feedback in sliding mode control. Compared with the actual input current A sliding structure design, using subtraction to represent the change in back electromotive force (EMF), is employed to handle discrete points. This allows for the generation of a series of predicted back EMFs based on effective flux linkage, utilizing discrete points within a finite set of predicted back EMFs. The AC variation processing of the predicted back EMF is represented as follows:

[0056]

[0057] In the formula, and It is the predicted back electromotive force corresponding to the discrete points (1, 2, ..., n) obtained by discretization. It is the predicted current at time k based on the feedback of the prediction model of effective magnetic flux, where b is a set boundary, which can be 1.

[0058] Current Predicting the current at time k The voltage output of the dual closed-loop controller for the permanent magnet motor Together with the finite set interval [k1E, k2E] of the predicted back electromotive force, it serves as the input signal for the prediction model module based on effective magnetic flux linkage. The discretized predicted back electromotive force... and Substituting each value into the following new prediction model based on effective magnetic flux, a series of different predicted currents for the next time step are obtained. and These can be used to predict the current at the next moment. and Unified representation as Among them, the predicted back electromotive force will be calculated for the finite set interval [k1E, k2E]. Substituting into the prediction model based on effective magnetic flux linkage, as shown in the following equation:

[0059]

[0060] In the formula, L q T is the q-axis inductance of a permanent magnet motor with adjustable leakage flux. sR represents the resistance of the motor stator winding.

[0061] The obtained series of predicted currents and are substituted into the designed cost function for calculation, and the calculation results are compared, wherein the discrete point value corresponding to the minimum cost function value is the optimal estimated back EMF value amplitude, at which time the optimal estimated back EMF value corresponding to the discrete point can be obtained according to the above-mentioned back EMF AC processing formula and output. The design of the cost function is improved according to the sliding surface of the sliding mode motion. When the traditional sliding mode motion reaches stability, the sliding surface S satisfies the following equation:

[0062]

[0063] Therefore, the traditional sliding surface S is improved and perfected in the design process of the sliding mode predictive observer, and a new sliding surface S is designed as follows: The discrete form is represented as follows: Based on the design of the new discrete sliding surface S n , a discrete form of the cost function can be designed by integrating the zero-error control principle as follows:

[0064]

[0065] The back EMF corresponding to the minimum cost function value is the optimal estimated back EMF, which is output.

[0066] In addition, the optimal estimated back EMF value obtained through the cost function needs to be filtered by a low-pass filter LPF, and the general form of the filter transfer function is represented as follows:

[0067]

[0068] In the formula, τ represents the time constant, reflecting the speed of response change of the control system, and s represents the input independent variable complex frequency of the frequency domain transfer function. Specifically, in the overall structure of the sliding mode predictive observer and the improved phase-locked loop, the input independent variable is and

[0069] The estimated back EMF obtained after the sliding mode predictive observer output is filtered by the LPF and serves as the input signal of the improved phase-locked loop, and the output signal of the phase-locked loop is the estimated motor rotor position angle and the estimated motor speed ​Meanwhile, since the LPF has a phase lag effect during low-pass filtering, it is also necessary to adjust the rotor position angle output by the iterative phase-locked loop. Phase angle compensation is performed to obtain an accurate estimate of the motor rotor position angle. The specific phase angle compensation operation is shown in the following formula:

[0070]

[0071] like Figure 5 The improved phase-locked loop's workflow is demonstrated, where g and j represent the real-time loop counts for different cycles, and N is the set final iteration count. The specific workflow steps are as follows:

[0072] Step 1): Since the range of change of the electric angle of the motor rotor is [0, 2π] rad, the initial observation angles selected during iteration are θ1 = 0 rad, θ2 = 2π / 3 rad, and θ3 = 4π / 3 rad.

[0073] Step 2): Initialize the observation angles θ1, θ2, θ3 and the estimated back electromotive force output by the sliding mode predictive observer. Substitute them sequentially into the evaluation function of the improved phase-locked loop. The calculations are performed using θ1, θ2, and θ3 to represent the unknown values ​​at this point. The evaluation functions were obtained respectively.

[0074] Step 3): After each evaluation function calculation is completed, check if the real-time loop count g is 3. If g < 3, return to the previous step and perform a new round of evaluation function calculation. If g ≥ 3, the current loop ends and the observation angles with the largest and second largest evaluation function results are selected and redefined as new observation angles θ1 and θ2. The new observation angle θ3 is calculated using the formula θ3 = (θ1 + θ2) / 2.

[0075] Step 4): After obtaining the new observation angles θ1, θ2, and θ3, the improved phase-locked loop begins a new iterative calculation, with the real-time iteration count flag being j. In the large loop with j as the flag, the same operations as the previous small loop with g as the flag are performed first, i.e., steps 2)-3). Then, after the small loop completes, only the angle corresponding to the maximum value of the evaluation function result is selected as the iteration angle θ2, while the other two iteration angles are determined by formula... Received, among which

[0076] Step 5): judging whether the cycle flag j reaches the set iteration number N, when the set iteration number N is reached, outputting the iteration angle θ2 with the maximum evaluation function as the rotor position angle output by the iteration-based phase-locked loop

[0077] Step 6): then performing LPF phase angle compensation operation to obtain the final motor estimated rotor position angle and the motor estimated rotor position angle is subjected to differential processing to obtain the motor estimated speed Referring to Figure 4 .

[0078] In the iteration process, the greater the iteration number N, the more the effective bits of the motor observed position angle, and the higher the position observation accuracy. After the speed is obtained through differential processing, low-pass filtering processing is further performed to filter out the high-frequency noise caused by the differential operation.

[0079] To realize the closed-loop control, the motor estimated rotor position angle is fed back to the Clark\Park coordinate transformation module as input to perform relevant coordinate transformation operation. The motor estimated rotor position angle and the motor estimated speed are subjected to the Clark\Park coordinate transformation module processing, and the output signal of the flux-regulated permanent magnet motor system is transformed to obtain the current and the given reference speed are input to the permanent magnet motor double-closed-loop controller, and the controller outputs and also outputs the voltage as the control voltage of the flux-regulated permanent magnet motor to realize the driving control of the motor.

[0080] Figure 6 Simulation comparison results of the rotor position observation error of the flux-regulated permanent magnet motor are given. The motor operating conditions are as follows: the initial speed is 600 rpm, the initial load torque is 10 N·m, the position sensor control is switched to the position sensorless control at 2 s, the speed suddenly changes to 800 rpm at 4 s, and the load torque suddenly changes to 15 N·m at 7 s. It can be seen from the simulation results that the position estimation error of the position sensorless controller designed in the application is smaller, and the response speed is faster and the control accuracy is higher under variable speed and variable load operating conditions.

[0081] Figure 7 Simulation comparison results of the rotor speed observation error of the flux-regulated permanent magnet motor are given, and the simulation operating conditions of the motor are the same as those of the position observation error Figure 6The simulation results show that the position sensorless rotational speed estimation error of the designed position sensorless controller is smaller, and the position sensorless controller can ensure stable operation and high control precision under variable speed and variable load conditions.

[0082] In conclusion, the application designs a novel position sensorless controller based on a sliding mode predictive observer for position sensorless driving control of the flux-adjustable permanent magnet motor. In order to suppress the influence of motor parameter variation on the sensorless control, the application introduces the concept of effective flux to design a new motor mathematical model. The application also refers to predictive control and proposes a sliding mode predictive observer, which ensures fast response and minimum chattering of the sliding mode motion while eliminating the selection of the reaching rate function and the determination of the sliding mode gain coefficient. In addition, the phase-locked loop is redesigned by a mathematical iterative method, which realizes controllable observation accuracy and effectively expands the bandwidth of the sensorless controller. The position sensorless controller designed by the application has the advantages of weak dependence on motor electromagnetic parameters, no control parameter adjustment process, fast response, high observation accuracy and controllability. Therefore, it has significant control advantages in the face of the complex operating conditions and electromagnetic parameter variations of the flux-adjustable permanent magnet motor.

[0083] The above examples are only used to illustrate the design idea and characteristics of the application, and the purpose is to enable those skilled in the art to understand the content of the application and implement it, and the protection scope of the application is not limited to the above examples. Therefore, any equivalent changes or modifications made according to the principles and design ideas disclosed by the application are within the protection scope of the application.

Claims

1. A method for designing a sensorless controller for sliding mode prediction of a permanent magnet motor with adjustable leakage flux, characterized in that... Includes the following steps: Step A: Connect the pulse width modulation module, inverter module, leakage flux adjustable permanent magnet motor, and current sensor module in series to form a leakage flux adjustable permanent magnet motor system. The input signal of the leakage flux adjustable permanent magnet motor system is voltage. The output signal is current. Step B: Combine the speed loop PI controller module, the current loop PI controller module, and two inverse Park coordinate transformation modules to form a dual closed-loop controller for the permanent magnet motor. The input signal of this dual closed-loop controller is the given reference speed. Feedback on estimated motor speed And feedback motor estimates rotor position angle and the feedback current The output signal is the reference current. and voltage Step C: The permanent magnet motor dual closed-loop controller is connected to the front end of the leakage flux adjustable permanent magnet motor system, and the Clark / Park coordinate transformation module is connected to the rear end of the leakage flux adjustable permanent magnet motor system. The Clark / Park coordinate transformation module, the sliding mode prediction observer, and the improved phase-locked loop are connected in sequence to form a sliding mode prediction sensorless controller. Current And improved phase-locked loop feedback for estimating the rotor position angle of the motor. Both serve as inputs to the Clark / Park coordinate transformation module, which outputs current. Given reference speed and voltage and reference current and current Together, they serve as the input signals to the sliding mode predictive observer, which outputs an estimated back electromotive force. As the input to the improved phase-locked loop, the improved phase-locked loop output estimates the motor speed. And the estimated rotor position angle of the motor 2. The design method according to claim 1, characterized in that: In step C, the specific design process of the improved phase-locked loop is as follows: Step 1): Initialize the observation angles as θ1 = 0 rad, θ2 = 2π / 3 rad, θ3 = 4π / 3 rad; Step 2): Initialize the observation angles θ1, θ2, θ3 and the estimated back electromotive force. Substitute into the evaluation function in sequence In the calculation, θ1, θ2, and θ3 are used to replace the unknown values ​​at this time. The evaluation functions were obtained respectively. Step 3): After each evaluation function calculation is completed, check if the real-time small loop count g is 3. If g < 3, return to perform a new round of evaluation function calculation. If g ≥ 3, the current loop ends and the observation angles with the largest and second largest evaluation function results are selected and redefined as new observation angles θ1 and θ2. The new observation angle θ3 is calculated using the formula θ3 = (θ1 + θ2) / 2. Step 4): After obtaining the new observation angles θ1, θ2, and θ3, a new iterative calculation is initiated. The real-time iteration count is marked with j. In the large loop marked with j, the same operation as the previous small loop marked with g is performed first. After the small loop completes, only the angle corresponding to the maximum value of the evaluation function result is selected as the iteration angle θ2. The other two iteration angles are determined by the formula... Received, among which Step 5): Determine if the loop flag j has reached the set number of iterations N. When the set number of iterations N is reached, output the iteration angle θ2 with the maximum value of the evaluation function, which is the rotor position angle output by the iteration-based phase-locked loop. Step 6): Calculate the estimated rotor position angle of the motor. τ represents the time constant and is used to estimate the rotor position angle of the motor. The estimated motor speed is obtained by differentiation.

3. The design method according to claim 1, characterized in that: In step C, the sliding mode predictive observer is based on the reference rotational speed. Calculate the predicted back electromotive force amplitude under the current operating conditions. ψ f The permanent magnet flux linkage of the motor is used; then, the value representing the predicted back EMF amplitude E is magnified and reduced to determine the finite set interval of the predicted back EMF. The amplitude d of the predicted back EMF at each discrete point is obtained through this finite set interval. E ; Current The predicted current at time k is obtained after a one-step delay. Based on amplitude d E Predicting the current at time k With current A series of predicted back electromotive forces were obtained by performing alternating current transformations. and n is the number of discrete points; Predicting back electromotive force and Substituting each value into the prediction model yields a series of predicted currents for the next time step. Predict the current at the next moment Substitute the cost function J of the design into the calculation, and the back electromotive force value corresponding to the number of discrete points with the minimum cost function value is the best estimated back electromotive force value. The optimal estimated back EMF value is filtered by a low-pass filter to output the estimated back EMF.

4. The design method according to claim 3, characterized in that: Predicting back electromotive force for: b is the defined boundary value, which is set to 1.

5. The design method according to claim 4, characterized in that: The prediction model is: L q T is the q-axis inductance of a permanent magnet motor with adjustable leakage flux. s R represents the discrete time step, and R represents the resistance of the motor stator winding.

6. The design method according to claim 5, characterized in that: The cost function is:

7. The design method according to claim 5, characterized in that: The low-pass filter transfer function is: τ represents the time constant, and s represents the complex frequency of the input independent variable of the frequency domain transfer function.

8. The design method according to any one of claims 1-7, characterized in that: The finite set interval is [k1E, k2E], where k1 and k2 are gain coefficients, k1 = 0.8 and k2 = 3.

0.

9. The design method according to claim 1, characterized in that: In step B, the discrete form of the equations constituting the controller module and the current loop PI controller module is as follows: u represents the output of the PI controller, e represents the input of the PI controller, and T s K represents the discrete time step. P and K I These represent the proportional and integral coefficients of the PI controller, respectively, where k is a non-negative integer.

10. The design method according to claim 9, characterized in that: The matrix equations of the two inverse Park coordinate transformation modules are as follows

Images