A method for suppressing torque ripple of a brushless doubly-fed machine

Abstract

The application discloses a torque ripple suppression method for a brushless doubly-fed motor, which comprises the following steps: firstly, a current loop and a rotating speed loop model of the brushless doubly-fed motor are established; then, an active disturbance rejection controller is established according to the current loop and the rotating speed loop model, wherein the active disturbance rejection controller comprises a d-axis improved current controller, a q-axis current loop active disturbance rejection controller and a rotating speed loop active disturbance rejection controller; the d-axis improved current controller comprises a d-axis current loop active disturbance rejection controller and an improved repetitive controller connected in parallel with the d-axis current loop active disturbance rejection controller; and then, the brushless doubly-fed motor is controlled by using the established active disturbance rejection controller. The active disturbance rejection controller is combined with the improved repetitive controller, the extended state observer is used for disturbance estimation, and the disturbance contains a coupling term, so that the current dynamic response is faster, and the parameter robustness is stronger; meanwhile, the improved repetitive controller connected in parallel can overcome the high-frequency current harmonics introduced by uncontrolled rectification, and can also limit the dynamic response to not produce excessive overshoot when the working condition is suddenly changed.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of motor control, and particularly relates to a torque ripple suppression method for a brushless doubly-fed motor. BACKGROUND

[0002] As a new type of alternating current motor, the brushless doubly-fed motor (BDFM) has attracted more and more attention and research in recent years because of its many advantages in structure. Two sets of windings are arranged on the stator of the motor, and the two sets of windings are independent of each other. Among them, the set of stator windings directly connected to the power grid is called the power winding (PW), and the other set of stator windings is called the control winding (CW). There is only one set of specially designed windings (Rotor Winding, RW) on the rotor of the motor. Through the modulation of the rotor, the indirect coupling of the magnetic fields of the two sets of stator windings is realized, thereby completing the electromechanical energy conversion of the motor.

[0003] The brushless doubly-fed motor can operate in various modes and can realize asynchronous or synchronous operation. Its fault tolerance stability is greatly improved compared to traditional motors, and as long as the power winding and the control winding of the motor are designed according to different voltage levels, the purpose of controlling a high-voltage motor using a low-voltage frequency converter can be achieved, which greatly reduces the cost of the speed regulation system. Moreover, there is a reactive term in the control target, which can realize unit power factor control. However, due to cost considerations, the brushless doubly-fed motor speed regulation system generally uses grid uncontrolled rectification as the DC source of the control winding inverter, so it contains many harmonics, which will directly cause motor torque ripple. To suppress motor torque ripple, the traditional method is to parallel a controller based on the internal model principle on the current controller, but the brushless doubly-fed motor has variable working conditions, and its speed needs to be adapted to various working condition requirements in real time. This suppression method will have overshoot in the dynamic process and take a long time to recover to stability. SUMMARY

[0004] The purpose of the present application is to provide a parameter robustness brushless doubly-fed motor decoupling disturbance control method that can effectively improve the dynamic response speed to solve the above problems in the prior art.

[0005] To achieve the above purpose, the technical solution of the present application is as follows:

[0006] A brushless doubly-fed motor torque ripple suppression method, comprising the following steps in sequence:

[0007] Step A, building a current loop and a speed loop model of the brushless doubly-fed motor;

[0008] Step B, establishing an active disturbance rejection controller according to the current loop and the speed loop model, the active disturbance rejection controller comprising a d-axis improved current controller, a q-axis current loop active disturbance rejection controller and a speed loop active disturbance rejection controller, wherein the d-axis improved current controller comprises a d-axis current loop active disturbance rejection controller and an improved repetitive controller connected in parallel with the d-axis current loop active disturbance rejection controller:

[0009] Step C, controlling the brushless doubly-fed motor by using the established active disturbance rejection controller.

[0010] In step A, the current loop model is:

[0011]

[0012]

[0013]

[0014]

[0015]

[0016]

[0017]

[0018]

[0019]

[0020] In the above formula, i cd , i cq are d-axis and q-axis currents of the control winding in the power winding-oriented synchronous coordinate system, u cd , u cq are d-axis and q-axis voltages of the control winding in the power winding-oriented synchronous coordinate system, are d-axis and q-axis ideal voltages of the control winding, t is time, σ, K d , K q , L M are intermediate parameters, b0, f d , f q are current loop control gain, d-axis total disturbance and q-axis total disturbance, L sc , L sp , L r are self-inductances of the control winding, the power winding and the rotor, r c is the control winding resistance, M cr , M pr are mutual inductances between the control winding and the rotor and between the power winding and the rotor, ω p , ωc grid frequency, control winding electrical frequency, r r is rotor resistance, i pd pq are d, q axis currents of the power winding in the power winding oriented synchronous coordinate system, ψ p is the power winding flux linkage in the power winding oriented synchronous coordinate system;

[0021] The speed loop model is:

[0022]

[0023]

[0024]

[0025] In the above formula, ω r is rotor angular velocity, J is moment of inertia, p p c are pole pair numbers of the power winding and the control winding, T L is torque, B is friction coefficient, b speed w are control gain and total disturbance of the speed loop, respectively, is the d axis ideal output voltage of the control winding.

[0026] In step B, the method for establishing the d axis improved current controller comprises: designing an improved repetitive controller based on the uncontrolled rectification harmonic frequency, and resonating the improved repetitive controller at the selected frequency to feed forward compensate the harmonics; based on the current loop model, regarding all items other than the current state variable as disturbance, observing and feeding forward using an extended state observer; and using the feed forward value and the command value to design a control law.

[0027] The control law of the d axis improved current controller is:

[0028]

[0029] In the above formula, u is the output of the improved current controller, b0 is the current loop control gain, K p is the control law proportional parameter, is the d axis ideal current of the control winding, z 1_d 2_d are the d axis observed current and the d axis disturbance estimation value of the control winding, G falrc (z) is the improved repetitive controller.

[0030] The expression of the improved repetitive controller is as follows:

[0031] ​​​​

[0032]

[0033]

[0034] In the above formula, fal(ε,α,δ) is a fal function, ε is an input, α is a constant between 0 and 1, δ is a constant determining the length of the linear section of the function, z -N is a delay parameter, N is 1 indicating a unit delay operator, Q(z) is a stability coefficient, K rc is an open-loop gain for adjusting the output size, z K is a phase adjustment parameter, K is a delay compensation coefficient for adjusting the output phase, f s is a switching frequency of PWM, f dbase is the frequency of 6th harmonic interference.

[0035] The establishment method of the q-axis current loop active disturbance rejection controller comprises the following steps: based on a current loop model, regarding all terms other than a current state variable as disturbance, observing and feeding forward by using a linear extended state observer, and designing a control law by using a current feedforward value and an instruction value;

[0036] The control law of the q-axis current loop active disturbance rejection controller is as follows:

[0037]

[0038] In the above formula, u is the output of the q-axis current loop active disturbance rejection controller, b0 is a current loop control gain, K p is a control law proportional parameter, is a d-axis ideal voltage of a control winding, z 1_q , and z 2_q are a q-axis observed current and a q-axis disturbance estimation value of the control winding respectively.

[0039] In step B, the establishment method of the speed loop active disturbance rejection controller comprises the following steps: based on a speed loop model, regarding all terms other than a d-axis current state variable of a control winding as disturbance, observing by using a linear extended state observer, and then feeding forward and compensating.

[0040] The control law of the speed loop active disturbance rejection controller is as follows:

[0041]

[0042] In the above formula, u * is the output of the speed loop active disturbance rejection controller, b speed is a speed loop control gain, K p is a control law proportional parameter, is an ideal angular velocity of a rotor, z 1_speed , and z 2_speedThese are the observed rotor angular velocity and the estimated speed loop disturbance, respectively.

[0043] The active disturbance rejection controller also includes a reactive power PI controller, which is used to control the q-axis current of the control winding to achieve zero reactive power. The reactive power PI controller is designed based on the following formula and according to the q-axis power current feedback:

[0044]

[0045] Step C includes the following steps in sequence:

[0046] S1. For the improved d-axis current controller, the d-axis current i of the control winding will be... cd The controller's output u is input to the linear expansion state observer to obtain the d-axis observed current z of the control winding. 1_d With d-axis disturbance estimate z 2_d For the q-axis current loop active disturbance rejection controller, the q-axis current i of the control winding will be... cq The controller's output u is input to the linear expansion state observer to obtain the q-axis observed current z of the control winding. 1_q With q-axis disturbance estimate z 2_q For the speed loop active disturbance rejection controller, the rotor angular velocity ω is... r and the output u of the controller * The rotor angular velocity observation value z is obtained by inputting it into the linear expansion state observer. 1_speed With the speed loop disturbance estimate z 2_speed ;

[0047] S2, the ideal angular velocity of the rotor Rotor angular velocity observation z 1_speed Speed ​​loop disturbance estimate z 2_speed Substituting this into the control law of the speed loop active disturbance rejection controller, we obtain the ideal d-axis current of the control winding.

[0048] S3, control the ideal d-axis current of the winding. d-axis observed current z 1_d With d-axis disturbance estimate z 2_d Substituting this into the control law of the improved d-axis current controller, we obtain the ideal d-axis voltage of the control winding. The q-axis reference current of the control winding q-axis observed current z of the control winding 1_q With q-axis disturbance estimate z 2_q Substituting this into the control law of the q-axis current loop active disturbance rejection controller, we obtain the ideal q-axis voltage of the control winding.

[0049] S4, subtracting the feedback value of the q-axis ideal current of the power winding from the q-axis ideal current value 0 of the power winding to obtain the q-axis ideal current of the control winding q-axis ideal voltage The ideal voltage is input to the vector modulation module through coordinate transformation, and 6 switching signals are generated to control the inverter switches.

[0050] The active disturbance rejection controller further comprises a reactive PI controller;

[0051] The step S2 further comprises:

[0052] The q-axis ideal current of the control winding is obtained by subtracting the feedback value of the q-axis ideal current of the power winding from the q-axis ideal current value 0 of the power winding and inputting it into the reactive PI controller

[0053] Compared with the prior art, the present application has the following advantages:

[0054] 1. The brushless doubly-fed motor torque ripple suppression method of the present application first establishes the current loop and speed loop models of the brushless doubly-fed motor, then establishes an active disturbance rejection controller according to the current loop and speed loop models, and the active disturbance rejection controller comprises a d-axis improved current controller, a q-axis current loop active disturbance rejection controller and a speed loop active disturbance rejection controller, the d-axis improved current controller comprises a d-axis current loop active disturbance rejection controller and an improved repetitive controller connected in parallel thereto, and then the established active disturbance rejection controller is used to control the brushless doubly-fed motor. By connecting the improved repetitive controller in parallel to the d-axis current loop active disturbance rejection controller, on the one hand, the improved repetitive controller provides the ability of high-frequency harmonic suppression; on the other hand, in the dynamic process, the improved repetitive controller can suppress the rapidly changing input to avoid overshoot.

[0055] 2. The brushless doubly-fed motor torque ripple suppression method of the present application regards all items other than the current state variables as disturbances for the design of the d-axis current loop active disturbance rejection controller and the q-axis current loop active disturbance rejection controller, and uses a linear extended state observer to estimate the disturbances. The disturbances contain the coupling terms, and the design of the controller only needs b0, which saves various motor parameters involved in the traditional coupling terms. Compared with the traditional calculation of decoupling terms and feedforward compensation, the current dynamic response is faster, and the parameter robustness is stronger.

[0056] 3. For the speed loop active disturbance rejection controller, the brushless doubly-fed motor torque ripple suppression method of the present application regards all items other than the control winding d-axis current state variables as disturbances, and uses a linear extended state observer to observe and feedforward compensate, so that the internal and external disturbances of the motor system can be compensated by the designed linear extended state observer, and the speed dynamic response speed is improved. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 It is the structure diagram of the d-axis improved current controller.

[0058] Figure 2 Structure diagram of q-axis current loop active disturbance rejection controller.

[0059] Figure 3 Structure diagram of speed loop active disturbance rejection controller.

[0060] Figure 4 Structure diagram of reactive power PI controller.

[0061] Figure 5 Overall structure diagram of brushless doubly-fed motor vector control.

[0062] Figure 6 Current steady-state waveform diagram obtained by traditional PI control method.

[0063] Figure 7 Fourier analysis result of current steady-state waveform obtained by traditional PI control method.

[0064] Figure 8 Current steady-state waveform diagram controlled by improved current controller of the application.

[0065] Figure 9 Fourier analysis result of current steady-state waveform controlled by improved current loop control of the application.

[0066] Figure 10 Current command tracking waveform diagram of traditional control PI method;

[0067] Figure 11 Current command tracking waveform diagram controlled by traditional repetitive controller;

[0068] Figure 12 Current command tracking waveform diagram controlled by improved current controller of the application. DETAILED DESCRIPTION

[0069] The application will be further described below in conjunction with specific embodiments and drawings.

[0070] Active disturbance rejection control is a new control method with broad application prospects in the field of motor control in recent years. The extended state observer in the active disturbance rejection controller can observe the internal and external disturbances of the system. The system can be converted into an integral series form after the feedforward, which greatly reduces the dependence of the controller parameters, and greatly improves the anti-disturbance ability and dynamic response ability. Repetitive control is a multi-frequency harmonic suppression technology based on the internal model principle, but it is sensitive to input mutations and suitable for steady-state conditions. The application combines the active disturbance rejection controller with the repetitive controller and improves the repetitive controller by introducing the fal function based on the traditional repetitive controller. To solve the problem of weak anti-interference ability of the traditional current loop, the extended state observer is used for disturbance estimation, and the disturbance contains its coupling term. Compared with the traditional calculation of decoupling items and feedforward compensation, the current dynamic response is faster, and the parameter robustness is stronger. The improved repetitive controller in parallel can overcome the high-frequency current harmonics introduced by uncontrolled rectification, and can also limit the dynamic response to produce excessive overshoot when the working condition changes.

[0071] Embodiment 1:

[0072] A torque ripple suppression method for a brushless doubly-fed motor is performed in the following steps:

[0073] 1. Build a current loop and speed loop model for the brushless doubly-fed motor.

[0074] The voltage equation is as follows:

[0075]

[0076] The flux linkage equation is as follows:

[0077]

[0078] In the above formula, u pd , u pd are the d, q axis voltages of the power winding in the power winding oriented synchronous coordinate system, u cd , u cq are the d, q axis voltages of the control winding in the power winding oriented synchronous coordinate system, i pd , i pq are the d, q axis currents of the power winding in the power winding oriented synchronous coordinate system, i cd , i cq are the d, q axis currents of the control winding in the power winding oriented synchronous coordinate system, i rd , i rq are the d, q axis currents of the rotor in the power winding oriented synchronous coordinate system, r p , r c , r rare the resistances of the control winding, the control winding, the rotor, respectively, ψ pd , ψ pq are the d, q axis flux linkage of the power winding in the power winding oriented synchronous coordinate system, respectively, ψ cd , ψ cq are the d, q axis flux linkage of the control winding in the power winding oriented synchronous coordinate system, respectively, ψ rd , ψ rq are the d, q axis flux linkage of the rotor in the power winding oriented synchronous coordinate system, ω p , ω r are the grid frequency, the rotor angular velocity, respectively, p p , p c are the pole pairs of the power winding, the control winding, respectively, L sc , L sp , L r are the self-inductance of the control winding, the power winding, the rotor, respectively, M cr , M pr are the mutual inductance between the control winding and the rotor, the mutual inductance between the power winding and the rotor, respectively, t is time.

[0079] i pd , i pq are obtained by coordinate transformation of the collected three-phase current of the power winding, i cd , i cq are obtained by coordinate transformation of the collected three-phase current of the control winding, that is:

[0080]

[0081] In the above formula, i pa , i pb , i pc are the a, b, c phase currents of the motor power winding, i ca , i cb , i cc are the a, b, c phase currents of the motor control winding, θ p , θ c are the electrical angles of the motor power winding, the control winding, respectively, indicating the angle between the d axis of the motor and the a phase of the power winding or the control winding, which can be directly obtained by the grid phase-locked loop PLL.

[0082] When the power winding is oriented, θ c has the following relationship (assuming that the initial d axis coincides with the a phase of the power winding and the control winding):

[0083] θ c = ω p t-(p p +p c )ω r t Formula 4

[0084] When the d-axis of the synchronous rotating coordinate system is oriented to the power winding voltage vector U p , the transient process of the power winding flux linkage is ignored (i.e., dψ pd / dt = 0, dψ pd / dt = 0), and the resistance r p of the power winding is ignored, there is:

[0085]

[0086] Substituting equation 2 into equation 1, we have:

[0087]

[0088] Substituting equation 6 into equation 1, we have:

[0089]

[0090] Substituting equation 7 into equation 1, we have:

[0091]

[0092] In the above equation, ψ p is the power winding flux linkage in the power winding oriented synchronous coordinate system, and ω c is the control winding electric frequency.

[0093] The above equation reveals the relationship between the control winding voltage and the control winding current, i.e., the appropriate control winding current can be obtained by controlling the control winding voltage. In order to control the control winding, the derivative of the power winding current needs to be eliminated.

[0094] Substituting equation 6 into the rotor flux linkage equation, we have:

[0095]

[0096] Let Substituting the above equation and equation 7 into the rotor voltage equation, we have:

[0097]

[0098] Substitute the derivative of the power winding current into the control winding voltage equation:

[0099]

[0100]

[0101] The purpose of power winding voltage vector orientation or flux linkage vector orientation is to obtain the decoupling of d-axis component and q-axis component of control winding current. For power winding voltage vector orientation, the ideal case is that the d-axis component of control winding current corresponds to the electromagnetic torque of motor and the q-axis component corresponds to the reactive power of power winding. Thus, by controlling the d-axis current and q-axis current of control winding respectively, the decoupling control of motor torque and reactive power can be realized, so that the ideal system dynamic response characteristic is obtained.

[0102] According to formula 10, at steady state, the differential term can be ignored, and the following formula can be obtained:

[0103]

[0104] In the above formula, since the rotor resistance is small and ω p is large, their ratio is small, so they can be ignored, and the above formula can be simplified as:

[0105]

[0106] The above formula shows the coupling relationship between control winding and power winding. The d-axis current of power winding can be controlled through the d-axis current of control winding, and the q-axis current of power winding can be controlled through the q-axis current of control winding.

[0107] The active power and reactive power of power winding can be calculated by the following formula respectively:

[0108]

[0109] The relationship between motor electromagnetic torque and control winding current is:

[0110]

[0111] In summary, the current loop model can be written as:

[0112]

[0113]

[0114]

[0115]

[0116]

[0117]

[0118] In the above formula, are the ideal d-axis voltage and q-axis voltage of control winding respectively, b0, f d , fq are current loop control gain, d-axis total disturbance, q-axis total disturbance, respectively.

[0119] The speed loop model can be written as:

[0120]

[0121]

[0122]

[0123] In the above formula, b speed , f w are speed loop control gain, speed loop total disturbance, respectively, J is the moment of inertia, T L is torque, B is friction coefficient, is the d-axis ideal current of control winding.

[0124] 2. Based on the uncontrolled rectification harmonic frequency, the improved current controller is designed, all items except the current state variable are regarded as disturbance, the extended state observer is used for observation, the decoupling effect of few parameter dependence is realized after feedforward, the fast dynamic response is realized, the appropriate parameters are selected, the improved repetitive controller resonates at the selected frequency, and the harmonic is feedforward compensated.

[0125] Taking the d-axis as an example, let x 1_d = i cd , x 2_d = f d , The state equation can be written as follows:

[0126]

[0127] y = Cx

[0128] Wherein, x = [x 1_d x 2_d ] T , C = [1 0],

[0129] The linear extended state observer LESO is designed:

[0130]

[0131]

[0132] Wherein, z = [z 1_d z 2_d ] T , L = [β1 β2] T , z 1_d , z 2_drespectively 1_d , x 2_d The estimated value of the parameter selection is obtained by using the bandwidth method:

[0133] β1=2ω0

[0134]

[0135] K p =ω0 / 5

[0136] Wherein, β1, β2 are LESO parameters, ω0 is the bandwidth, K p is the control law proportional parameter.

[0137] The DC bus of the frequency converter is established by using the diode uncontrolled rectification mode, and the harmonic in the power winding directional synchronous coordinate system is expressed as 6k (k = 0, 1, 2…) harmonic, and the fundamental frequency is 50 Hz of the power grid frequency.

[0138] The traditional repetitive controller not only has a large amplitude near the resonance frequency, but also presents a proportional amplification characteristic to the low frequency signal in the low frequency region. The amplitude changes sharply at the resonance, which easily causes the Nyquist curve to approach the critical point, increases the sensitivity function of the system, especially for the load mutation type disturbance, the amplification characteristic in the low frequency band causes the dynamic response to be too violent, the stable time is long, and a large overshoot appears. Therefore, the following fal function is introduced in the embodiment:

[0139]

[0140] In the above formula, ε is the input, α is a constant between 0 and 1, the smaller the nonlinearity is greater, and δ is a constant that determines the length of the linear interval of the function. In order to improve the dynamic response problem caused by the additional repetitive controller, the embodiment selects δ = 0.4 and α = 0.6.

[0141] The expression of the improved repetitive controller is as follows:

[0142]

[0143]

[0144] In the above formula, z -N is a delay parameter, N is 1 indicating a unit delay operator, Q(z) is a stability coefficient, which is generally designed as 0.95, K rc is an open-loop gain for adjusting the output size, z K is a phase adjustment parameter, K is a delay compensation coefficient for adjusting the phase of the output, by adjusting K, the phase of the output can be adjusted, f s is the switching frequency of PWM, f dbaseis the frequency of 6th harmonic disturbance. For convenience, the embodiment takes f s = 15 kHz, f dbase = 300 Hz.

[0145] The control law of the d-axis improved current controller is designed as follows:

[0146]

[0147] The structure of the d-axis improved current controller is shown in Figure 1 The decoupling method eliminates various motor parameters involved in the traditional coupling term, achieving higher parameter robustness and faster dynamic response. The parallel improved repetitive controller provides the ability of high-frequency harmonic suppression, and the improved controller can suppress rapidly changing inputs during the dynamic process to avoid overshoot.

[0148] Since torque is only related to the d-axis control winding current, only the q-axis current loop active disturbance rejection controller needs to be designed, and the design principle is the same as the d-axis. The structure is shown in Figure 2 The control law is:

[0149]

[0150] 3. Based on the speed loop model, all terms except the d-axis current state variable of the control winding are regarded as disturbance, and linear extended state observer is used for observation and feedforward compensation to realize fast dynamic response after torque mutation.

[0151] Let x 1_speed = ω r , x 2_speed = f w , The state equation can be written as follows:

[0152]

[0153] y = Cx

[0154] Where x = [x 1_speed x 2_speed ] T , C = [1 0],

[0155] Design linear extended state observer LESO:

[0156]

[0157]

[0158] Where z = [z 1_speed z2_speed T , L = [β1 β2] T , z 1_speed , z 2_speed are the estimated values of x 1_speed , x 2_speed , the parameter selection uses the bandwidth method:

[0159] β1 = 2ω0

[0160]

[0161] K p = ω0 / 5

[0162] Therefore, the control rate of the speed loop active disturbance rejection controller is designed as follows:

[0163]

[0164] The structure of the speed loop active disturbance rejection controller is shown in Figure 3 , and the disturbances inside and outside the motor system can be compensated by LESO, improving the dynamic response speed.

[0165] 4、Based on the following reactive power expression, the PI controller shown in Figure 4 is designed to control the q-axis current of the control winding to achieve reactive power control of 0:

[0166]

[0167] The PI controller expression is as follows:

[0168]

[0169] Where S is the Laplace operator.

[0170] Finally, the brushless doubly-fed motor vector control structure diagram is constructed as shown in Figure 5 .

[0171] 5、Through the microcontroller to set the rotor mechanical reference speed of brushless doubly-fed motor and load torque, the brushless doubly-fed motor is controlled with reactive power of 0, i.e. the q-axis current of the power winding is set to 0.

[0172] 6、The rotor angular velocity ω r of the motor is obtained through the position encoder installed on the brushless doubly-fed motor; the three-phase currents i pabc of the power winding and the three-phase currents i cabc of the control winding are obtained through the current sensor, and they are transformed into the d-axis current i pd of the power winding and the q-axis current i​pq , control the d-axis current i of the winding cd , control the q-axis current i of the winding cq .

[0173] 7. For the improved d-axis current controller, the d-axis current i of the control winding will be... cd The controller's output u is input to the linear expansion state observer to obtain the d-axis observed current z of the control winding. 1_d With d-axis disturbance estimate z 2_d For the q-axis current loop active disturbance rejection controller, the q-axis current i of the control winding will be... cq The controller's output u is input to the linear expansion state observer to obtain the q-axis observed current z of the control winding. 1_q With q-axis disturbance estimate z 2_q For the speed loop active disturbance rejection controller, the rotor angular velocity ω is... r and the output u of the controller * The rotor angular velocity observation value z is obtained by inputting it into the linear expansion state observer. 1_speed With the speed loop disturbance estimate z 2_speed .

[0174] 8. The ideal angular velocity of the rotor Rotor angular velocity observation z 1_speed Speed ​​loop disturbance estimate z 2_speed Substituting this into the control law of the speed loop active disturbance rejection controller, we obtain the ideal d-axis current of the control winding. The ideal q-axis current of the power winding (0) is subtracted from its feedback value and then substituted into the reactive PI controller to obtain the ideal q-axis current of the control winding.

[0175] 9. The ideal d-axis current of the control winding... d-axis observed current z 1_d With d-axis disturbance estimate z 2_d Substituting this into the control law of the improved d-axis current controller, we obtain the ideal d-axis voltage of the control winding. The q-axis reference current of the control winding q-axis observed current z of the control winding 1_q With q-axis disturbance estimate z 2_q Substituting this into the control law of the q-axis current loop active disturbance rejection controller, we obtain the ideal q-axis voltage of the control winding.

[0176] 10. The ideal d-axis current of the control winding q-axis ideal voltage The 6-way switch signals are generated by the coordinate transformation and input to the vector modulation module to control the switches of the inverter, the inverter output is connected to the brushless doubly-fed motor control winding to drive the motor to work normally.

[0177] To verify the effectiveness and superiority of the control method, the following experiments are carried out:

[0178] (1) disconnect the speed loop and the power loop, set the control winding current command to 2A, and respectively use the traditional PI control method and the improved current controller to carry out current tracking experiment, and investigate the current loop response performance. The current steady waveform and Fourier analysis results obtained by the two methods are shown in Figures 6-9

[0179] By comparison Figures 6-9 It can be seen that the current of the traditional PI control method contains a large amount of harmonics, and the Fourier analysis determines that it is 6k (k is a positive integer) harmonic; while the current harmonics of the present application can be greatly reduced.

[0180] (2) disconnect the speed loop and the power loop, set the control winding current command to respectively use the traditional PI control method, the ordinary repetitive controller and the improved current controller to carry out current tracking experiment, and investigate the current loop response performance of each. The current command tracking waveform diagrams of the three methods are shown in Figures 10-12

[0181] It can be seen from Figures 10-12 that the current controlled by the traditional PI control method contains a large amount of harmonics; the ordinary repetitive controller can reduce certain harmonics, but it will oscillate for sudden change of current command; the improved current controller used in the present application can not only suppress harmonics, but also accurately track the current command without overshoot. Therefore, it can be seen that the present application greatly improves the current loop control accuracy.​​

Claims

1.A method for suppressing torque ripple of a brushless doubly-fed machine, characterized in that: the method comprises the following steps in sequence: Step A, establishing a current loop model and a speed loop model of the brushless doubly-fed machine; Step B, establishing an active disturbance rejection controller (ADRC) based on the current loop model and the speed loop model, wherein the ADRC comprises a d-axis improved current controller, a q-axis current loop ADRC, and a speed loop ADRC, the d-axis improved current controller comprises a d-axis current loop ADRC and an improved repetitive controller connected in parallel with the d-axis current loop ADRC, and the d-axis improved current controller is established by: designing the improved repetitive controller based on a non-controlled rectification harmonic frequency and making the improved repetitive controller resonate at a selected frequency to feed forward compensate the harmonic; regarding all items other than the current state variables as disturbances based on the current loop model, observing and feeding forward using an extended state observer; and designing a control law using a feed forward value and an instruction value; the control law of the d-axis improved current controller is: Step C, controlling the brushless doubly-fed machine using the established ADRC. 2.The method according to claim 1, characterized in that: in Step A, the current loop model is: and the speed loop model is: 3.The method according to claim 1 or 2, characterized in that: the establishment method of the q-axis current loop ADRC comprises: regarding all items other than the current state variables as disturbances based on the current loop model, observing and feeding forward using a linear extended state observer, and designing a control law using a current feed forward value and an instruction value; and the control law of the q-axis current loop ADRC is: 4.The method according to claim 1 or 2, characterized in that: in Step B, the establishment method of the speed loop ADRC comprises: regarding all items other than the d-axis current state variables of the control winding as disturbances based on the speed loop model, observing and feeding forward compensating after observing using a linear extended state observer; and the control law of the speed loop ADRC is: 5.The method according to claim 2, characterized in that: the ADRC further comprises a reactive power PI controller, the reactive power PI controller is used to control the q-axis current of the control winding to achieve 0 reactive power, and the reactive power PI controller is designed based on the following formula according to q-axis power current feedback: 6.The method according to claim 1, characterized in that: Step C comprises the following steps in sequence: ； ； ； ； In the above formula, is the output of the current controller, is the current loop control gain, is the control law proportional parameter, is the d-axis ideal current of the control winding, , is the d-axis observed current of the control winding, and is the d-axis disturbance estimate value of the control winding, is the improved repetitive controller, is the fal function, is the input, is a constant between 0 and 1, is a constant that determines the length of the linear segment interval of the function, is a delay parameter, and N is a unit delay operator when N is 1, is a stability coefficient, is an open-loop gain for adjusting the output size, is a phase adjustment parameter, and K is a delay compensation coefficient for adjusting the phase of the output, is the switching frequency of the PWM, is the frequency of the 6th harmonic disturbance; 7.The method according to claim 6, characterized in that: the ADRC further comprises a reactive power PI controller; and Step S2 further comprises: ​ ​ ； ； ； ； ； ； ； ； in the above formula, , are d, q axis currents of the control winding in the power winding oriented synchronous coordinate system, , are d, q axis voltages of the control winding in the power winding oriented synchronous coordinate system, , are d, q axis ideal voltages of the control winding, t is time, , , , are intermediate parameters, , , are current loop control gains, d axis total disturbance, q axis total disturbance, , , are self-inductances of the control winding, power winding, rotor, is the control winding resistance, , are mutual inductances between the control winding and the rotor, between the power winding and the rotor, , are grid electrical frequency, control winding electrical frequency, is the rotor resistance, , are d, q axis currents of the power winding in the power winding oriented synchronous coordinate system, is the power winding flux linkage in the power winding oriented synchronous coordinate system; ​ ； ； ； In the above formula, is the rotor angular velocity, is the moment of inertia, , are the pole pair numbers of the power winding and the control winding, respectively, is the torque, is the friction coefficient, , are the control gain and the total disturbance of the speed loop, respectively, is the d-axis ideal output voltage of the control winding. ​ ​ ​ ； In the above formula, is the output of the q-axis current loop active disturbance rejection controller, is the current loop control gain, is the control law proportional parameter, is the d-axis ideal voltage of the control winding, , is the q-axis observed current of the control winding, and is the q-axis disturbance estimation value of the control winding, respectively. ​ ​ ​ ； In the above formula, is the output of the speed loop active disturbance rejection controller, is the speed loop control gain, is the control law proportional parameter, is the ideal angular velocity of the rotor, , is the rotor angular velocity observation value, and is the speed loop disturbance estimation value, respectively. ​ ​ 。 ​ ​ S1, for the d-axis improved current controller, the d-axis current of the control winding and the output of this controller is input into a linear extended state observer to obtain the d-axis observed current of the control winding and the d-axis disturbance estimation value ; for the q-axis current loop active disturbance rejection controller, the q-axis current of the control winding and the output of this controller is input into a linear extended state observer to obtain the q-axis observed current of the control winding and the q-axis disturbance estimation value ; for the speed loop active disturbance rejection controller, the rotor angular velocity and the output of this controller is input into a linear extended state observer to obtain the rotor angular velocity observation value and the speed loop disturbance estimation value ; S2, ideal angular velocity of the rotor , rotor angular velocity observation value , rotor speed loop disturbance estimation value , d-axis ideal current of the control winding obtained by bringing the rotor speed loop disturbance estimation value into the control law of the rotor speed loop active disturbance rejection controller ; S3, the d-axis ideal current of the control winding , the d-axis observed current and the d-axis disturbance estimation value into the control law of the d-axis improved current controller, to obtain the d-axis ideal voltage of the control winding ; the q-axis reference current of the control winding , the q-axis observed current of the control winding and the q-axis disturbance estimation value into the control law of the q-axis current loop active disturbance rejection controller, to obtain the q-axis ideal voltage of the control winding ; S4, the d-axis ideal current of the control winding , the q-axis ideal voltage Through the coordinate transformation, input to the vector modulation module, generate 6-way switching signal to control the inverter switch. ​ ​ ​ The q-axis ideal current value of the power winding is subtracted from the feedback value and input into the reactive power PI controller to obtain the q-axis ideal current of the control winding .

Images