CN110212838A - A kind of method for controlling position-less sensor extracting optimal rotor-position - Google Patents

Abstract

The invention discloses a position sensorless control method for extracting the optimal rotor position. The control method includes: S1: adopting the traditional of the vector control method, the reference input current Inject the high-frequency pulse vibration signal on the basis of the high-frequency voltage injection to obtain the superimposed signal of the reference input current after the high-frequency voltage injection; S2: Drive the motor according to the superimposed signal to obtain the coupling signal containing the rotor position of the motor; S3: Determine the parallel design according to the coupling signal S4: According to the control strategy, the coupling signal is processed in parallel with the phase-locked loop and the extended Kalman filter to extract the optimal solution of the rotor position reference value. The control strategy of the invention can effectively solve the transition problem between zero low speed and high speed of the permanent magnet synchronous motor and extract the optimal solution of the reference value of the rotor position in the full speed range.

Description

A Position Sensorless Control Method for Extracting Optimal Rotor Position

technical field

The invention belongs to the field of motor control, in particular to a position sensorless control method for extracting an optimal rotor position.

Background technique

Due to high efficiency, high power density and high power factor, permanent magnet synchronous motors are used more and more, such as electric vehicles or wind power generation pitch control. In order to realize a high-performance three-phase permanent magnet synchronous motor control system, it is necessary to obtain the rotor position information of the motor. There are two ways to obtain the rotor position information:

One is to install the position sensor. However, because the position sensor is prone to failure in a harsh environment or is subject to external disturbances, the sensor is damaged, and the maintenance cost increases.

The second is that there is no position sensor, and the rotor position information is estimated from other physical quantities during the design of the vector control system. Based on the position sensorless control of the permanent magnet synchronous motor, the existing technical solutions are mainly to obtain the rotor position information by controlling respectively at low speed and high speed.

The control principle at high speed is the control method based on the fundamental wave mathematical model, using the quantity related to the rotational speed in the fundamental wave excitation mathematical model to refer to the rotor position and rotational speed, such as the back EMF generated by the motor at high speed, in the back EMF Extract the rotor position information. But this method is only suitable for the motor at a high speed, and the method will fail when the motor speed is zero or at a very low speed.

The control principle at low speed is a control method based on high-frequency signal injection. When the motor speed is too low to generate back electromotive force, the high-frequency voltage or current signal is superimposed on the fundamental wave input signal of the control system, and the combined signal is combined. Applied to the three-phase winding of the permanent magnet synchronous motor, so that the high-frequency signal will have the corresponding rotor position information after coupling, and the high-frequency signal will be extracted through the band-pass filter to obtain the rotor position reference value.

This method is suitable for the case where the back electromotive force of the motor is extremely small when the motor speed is zero or low speed. When the motor is running at zero speed and low speed, if the method of extracting the back EMF is used to obtain the motor rotor position information, because the back EMF is extremely small and the signal-to-noise ratio of the useful signal is very low, it is usually difficult to extract the rotor position information.

The above two methods can obtain the rotor position information in two specific situations, but there is no guarantee whether the rotor position information can be effectively obtained during the transition stage between low speed and high speed, so these two methods cannot make the motor operate within each speed range. Effective sensorless control.

Contents of the invention

Purpose of the invention: In view of the above problems, the present invention proposes a position sensorless control method based on phase-locked loop and extended Kalman filter parallel design to extract the optimal rotor position, to solve the problem of permanent magnet synchronous motor between zero low speed and high speed. The optimal solution of the rotor position reference value is extracted in the range of the full speed range.

Technical solution: In order to achieve the purpose of the present invention, the technical solution adopted in the present invention is: a position sensorless control method for extracting the optimal rotor position, the method includes the following steps:

S1, using traditional The vector control method, the reference current command under the reference coordinate system d ^* -q ^* After being processed by the current controller, the reference voltage command in the d ^* -q ^* coordinate system is obtained In the reference voltage command Inject high-frequency pulse vibration signals on the basis of high-frequency voltage injection to obtain superimposed signals of reference voltage commands after high-frequency voltage injection;

S2, after coordinate transformation and SVPWM processing of the superimposed signal, output a drive coupling signal containing rotor position information, which is used to drive the motor PMSM;

S3, according to the coupling signal, design a parallel control strategy to extract the optimal rotor position and angular velocity ω ^* ;

S4, Feedback the optimal angular velocity ω ^* to the current controller, and the optimal rotor position The parameters that replace the coordinate transformation are circulated to form a closed-loop system.

Further, the specific method of S1 is as follows:

Step S1.1: The current equation in the d-q coordinate system of the known three-phase permanent magnet synchronous motor is:

Among them, i _d , i _q are dq axis current; u _d , u _q are dq axis voltage; L _d , L _q are dq axis inductance components; R is stator resistance; ψ _f is flux linkage of permanent magnet; ω _e is electric current angular velocity;

If i _d and i _q are completely decoupled, the above formula can be changed to:

Among them, u _d0 and u _q0 are dq-axis voltage output after current decoupling; _id and i _q are dq-axis currents; L _d and L _q are dq-axis inductance components; R is stator resistance; ψ _f is permanent magnet magnetism chain; ω _e is the electrical angular velocity;

After fully decoupling the voltage equation of i _d and i _q , the Laplace transform can be obtained: Y _(s) = G _(s) U _(s)

in,

Among them, u _d0 (s), u _q0 (s) are complex voltage values after Laplace transform; i _d (s), i _q (s) are complex current values after Laplace transform;

Using a PI regulator and performing feed-forward decoupling, the output voltage in the reference coordinate system d ^* -q ^* can be obtained as:

in, is the reference current command under the synchronous rotating coordinate system d ^* -q ^* , and i _d and i _q are the current commands under the synchronous rotating coordinate system dq coordinate system; is the voltage under the reference coordinate system d ^* -q ^* obtained after being processed by the current controller in the step S1; L _d and L _q are the inductance components of the dq axis; ψ _f is the flux linkage of the permanent magnet; ω _e is the electrical angular velocity; K _pd and K _pq are proportional gains, K _id and K _iq are integral gains;

Step S1.2: According to step S1.1, after obtaining the output voltage of the d ^* -q ^* coordinate axis On the basis of , a high-frequency sinusoidal voltage signal is injected in the d ^* axis

in: is the injected high-frequency sinusoidal voltage signal, u _in is the amplitude of the high-frequency sinusoidal voltage signal, ω _in is the frequency of the high-frequency sinusoidal voltage signal, and thus the superimposed signal is obtained.

Further, the specific method of S2 is as follows:

Step S2.1: Obtain the superimposed signal after 2r/2s coordinate transformation θ _e is the actual rotor position angle; ω is the angular velocity of the dq coordinate axis; the principle of 2r/2s coordinate conversion is as follows:

Transform the synchronous rotating coordinate system dq to the stationary coordinate system α-β, T _2r/2s is the coordinate transformation matrix, which can be expressed as:

Step S2.2: After processing by SVPWM algorithm, output AC side phase voltage u _a , _ub , _uc and sinusoidal phase current _ia , _ib , _ic with mutual difference of 120° electric angle to drive motor PMSM;

In the natural coordinate system ABC, the phase voltages are u _a , u _b , uc , and the sinusoidal phase currents with _a mutual difference of 120° electrical angle are _{ia , i b , ic} . After 3s/2r coordinate transformation, the phase current The equivalent of the phase voltage on the d ^* -q ^* coordinate axis is expressed as Therefore: That is, the coupling signal containing the rotor position of the motor under the d ^* -q ^* coordinate axis;

in:

solve The specific process is as follows:

Establish the relationship between the reference rotor synchronous coordinate system d ^* -q ^* and the actual rotor synchronous coordinate system dq;

Among them: α-β is the stationary coordinate system, ω is the angular velocity of the dq axis, ω ^* is the angular velocity of the d ^* -q ^* axis, is the reference rotor position angle, _θe is the actual rotor position angle, and _Δθe is the rotor reference error angle, namely:

Treat the permanent magnet synchronous motor as an RL circuit, and inject a high-frequency voltage signal into the d ^* axis of the d ^* -q ^* coordinate axis:

Among them: u _in is the amplitude of the injected high-frequency voltage signal, ω _in is the frequency of the injected high-frequency voltage signal;

At this time, the voltage equation of the three-phase permanent magnet synchronous motor under high-frequency excitation in the synchronous rotating coordinate system, that is, the coupled voltage equation, can be simplified as:

Among them: i _din and i _qin are the current response of the motor under high-frequency excitation; u _din and u _qin are the voltage response of the motor under high-frequency excitation; L _d and L _q are the dq axis inductance components;

In the synchronous rotating coordinate system d-q, the stator inductance of the motor can be expressed as:

In the stationary coordinate system α-β, the stator inductance of the motor can be expressed as:

where: average inductance half differential inductance

Then, under the reference rotor synchronously rotating coordinate system d ^* -q ^* , the relationship between high-frequency voltage and current is:

in: It is the reference value of the current component under d ^* -q ^* ; It is the reference value of the voltage component under d ^* -q ^* .

Further, the specific method of S3 is as follows:

Step S3.1, convert i _a , i _b , and i _c into currents i _α and i _β in the α-β coordinate system after 3s/2s coordinate transformation, and transform i _α and i _β under 2s/2r coordinate transformation is the actual current i _d ′, i _q ′ in the dq coordinate system, input i _d ′, i _q ′ to the LPF for low-pass filtering, and then input it to the current controller as a negative feedback signal;

Step S3.2, in step S3.1: i _a and i _β are known, and now they are used as the input of the following two observer parallel processing links, the specific process is as follows:

Step S3.2.1, phase-locked loop PLL link

Filter i _α and i _β , extract the high-frequency part signals containing rotor position information in i _α and i _β through the band-pass filter BPF, and convert the high-frequency part signals into the reference rotor after 2s/2r coordinate transformation Under the synchronous rotating coordinate system d ^* -q ^*

In order to obtain the rotor position, the q ^* axis high frequency current Amplitude modulation is performed, and the modulation current sinω _in t is injected, where ω _in is the angular frequency of the modulation current;

The modulation function can be expressed as:

Among them, u _in is the amplitude of the high-frequency sinusoidal voltage signal; ω _in is the angular frequency of the modulation current; L is the average inductance, ΔL is the half-difference inductance; Δθ _e is the rotor reference error angle;

When the rotor reference error angle Δθ _e tends to zero, Δθ _e = 0, cosΔθ _e = 1, sin2Δθ _e = 2sinΔθ _e cosΔθ _e = 2sinΔθ _e , the error signal can be linearized at this time:

Among them, k _ε =u _in (L _q -L _d )/4ω _in L _d L _q ;

Input the modulation signal to the phase-locked loop PLL composed of low-pass filter LPF and PI controller, and obtain the reference angular velocity through the processing of the phase-locked loop PLL Angle with reference rotor position By adjusting f(ΔΔ _e )=0, that is Then the rotor position reference value can converge to the rotor position actual value;

Step S3.2.2, the extended Kalman filter EKF link;

The voltage responses u _a , u _b , u _c under high-frequency voltage excitation are known, and the voltage signals u _α , u _β are obtained after 3s/2s coordinate transformation, and u _α , u _β are compared with i _α in step S3.1 , i _β are jointly input to the extended Kalman filter EKF, and the reference angular velocity is obtained after being processed by the extended Kalman filter EKF and the reference rotor position angle

Further, the specific method of S4 is as follows:

Step S4.1, the observer PLL output and the observer EKF output of Send it to the least square solver to find the optimal solution Optimal solution Feedback to the 2r/2s coordinate transformation of step S2.1, the 2s/2r coordinate transformation of step S3.1 and the 2s/2r coordinate transformation of step S3.2, replace θ _e as a control parameter to participate in the calculation of the coordinate transformation process;

Step S4.2, by converting the observer PLL output and the observer EKF output of Feed into the least squares solver, in and Find the best parameter match ω ^* between the data;

Step S4.3, send ω ^* into the speed controller and output the reference current command after processing exist Under the control method, the reference current command in the current controller The error value is obtained by performing difference operation with the actual current command i _d ′, i _q ′ fed back in step S3.1, and the current controller forms a closed-loop system by controlling the error value.

Beneficial effects: Compared with the prior art, the technical solution of the present invention has the following beneficial technical effects:

(1) The phase-locked loop PLL and the extended Kalman filter EKF are designed in parallel, and the high-frequency pulse vibration voltage input signal can accurately obtain the rotor position information of the motor when the motor is at zero or low speed; the extended Kalman filter is An adaptive system that can work in a large speed range such as high speed; therefore, the parallel design can obtain rotor position information in the full speed range, avoiding the separate design of the low speed and high speed systems;

(2) Since the rotor position angle θ _e is obtained by the serial integration of the angular velocity ω, a small change in the angular velocity ω will cause the error of the rotor position angle θ _e to be amplified. The parallel design calculates the rotor position angle θ _e and the angular velocity ω separately, It can effectively avoid the closed-loop control error from being amplified;

(3) Within a sampling period of a certain width, the reference input signal and Carry out least squares parameter identification, select the optimal solution ω ^* and Minimize the impact of large error margins;

(4) The optimal solution ω ^* and It can adapt to the position sensorless control of the motor at each speed stage, so that the motor can effectively obtain the rotor position information during the acceleration or deceleration process, solve the good transition between the zero low speed and high speed section of the permanent magnet synchronous motor and extract the optimal rotor The problem of sensorless control of position angle.

Description of drawings

Figure 1 is a structural schematic diagram of conversion between coordinate systems;

Figure 2 is a relationship diagram between the reference rotor synchronous coordinate system d ^* -q ^* and the actual rotor synchronous coordinate system dq;

Figure 3 is the improved PLL and EKF in parallel to obtain the optimal solution ω ^* and The block diagram of the permanent magnet synchronous motor position sensorless control system.

Detailed ways

The technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

The technical scheme of the present invention is based on the parallel design of the phase-locked loop and the extended Kalman filter to reduce the amplification of the closed-loop feedback control error, and at the same time, it can obtain the best motor rotor position angle in the full speed range. The control scheme can be divided into for the following steps:

Step S1: As shown in Figure 3, adopt the traditional The vector control method, the reference current command under the reference coordinate system d ^* -q ^* After being processed by the current controller, the reference voltage command in the d ^* -q ^* coordinate system is obtained In the reference voltage command The high-frequency pulse vibration signal is injected on the basis of the high-frequency voltage to obtain the superimposed signal of the reference voltage command after the high-frequency voltage injection. Specifically include the following:

Step S1.1: The current equation in the d-q coordinate system of the known three-phase permanent magnet synchronous motor is:

Among them, i _d , i _q are dq axis current; u _d , u _q are dq axis voltage; L _d , L _q are dq axis inductance components; R is stator resistance; ψ _f is flux linkage of permanent magnet; ω _e is electric current angular velocity.

If i _d and i _q are completely decoupled, the above formula can be changed to:

Among them, u _d0 and u _q0 are dq-axis voltage output after current decoupling; _id and i _q are dq-axis currents; L _d and L _q are dq-axis inductance components; R is stator resistance; ψ _f is permanent magnet magnetism chain; ω _e is the electrical angular velocity.

After fully decoupling the voltage equation of i _d and i _q , the Laplace transform can be obtained: Y _(s) = G _(s) U _(s)

in:

Among them: u _d0 (s), u _q0 (s) are complex voltage values after Laplace transform; i _d (s), i _q (s) are complex current values after Laplace transform.

Using a PI regulator and performing feed-forward decoupling, the output voltage in the reference coordinate system d ^* -q ^* can be obtained as:

in, is the reference current command under the synchronous rotating coordinate system d ^* -q ^* , and i _d and i _q are the current commands under the synchronous rotating coordinate system dq coordinate system; is the voltage under the reference coordinate system d ^* -q ^* obtained after being processed by the current controller in the step S1; L _d and L _q are the inductance components of the dq axis; ψ _f is the flux linkage of the permanent magnet; ω _e is the electrical angular velocity; K _pd and K _pq are proportional gains, and K _id and K _iq are integral gains.

Step S1.2: According to step S1.1, after obtaining the output voltage of the d ^* -q ^* coordinate axis On the basis of , a high-frequency sinusoidal voltage signal is injected in the d ^* axis

in: is the injected high-frequency sinusoidal voltage signal, u _in is the amplitude of the high-frequency sinusoidal voltage signal, ω _in is the frequency of the high-frequency sinusoidal voltage signal, and thus the superimposed signal is obtained.

Step S2: After the superimposed signal is processed by 2r/2s coordinate transformation and SVPWM, a driving coupling signal containing rotor position information is output, and this signal is used to drive the motor PMSM. Specifically include the following steps:

Step S2.1: Obtain the superimposed signal after 2r/2s coordinate transformation

As shown in Figure 1: _θe is the actual rotor position angle; ω is the angular velocity of the dq coordinate axis; the principle of 2r/2s coordinate conversion is as follows:

Transform the synchronous rotating coordinate system dq to the stationary coordinate system α-β, T _2r/2s is the coordinate transformation matrix, which can be expressed as:

Step S2.2: After processing by SVPWM algorithm, it outputs AC side phase voltages u _a , _ub , uc and sinusoidal phase currents _ia , _ib , _ic with _a mutual difference of 120° electrical angle to drive the motor PMSM.

In the natural coordinate system ABC, the phase voltages are u _a , u _b , uc , and the sinusoidal phase currents with _a mutual difference of 120° electrical angle are _{ia , i b , ic} . After 3s/2r coordinate transformation, the phase current The equivalent of the phase voltage on the d ^* -q ^* coordinate axis is expressed as Therefore: That is, it is the coupling signal containing the rotor position of the motor under the d ^* -q ^* coordinate axis.

in:

solve The specific process is as follows:

As shown in Figure 2: Establish the relationship between the reference rotor synchronous coordinate system d ^* -q ^* and the actual rotor synchronous coordinate system dq.

Among them: α-β is the stationary coordinate system, ω is the angular velocity of the dq axis, ω ^* is the angular velocity of the d ^* -q ^* axis, is the reference rotor position angle, _θe is the actual rotor position angle, and _Δθe is the rotor reference error angle, namely:

Usually, the frequency of the high-frequency injection signal is 0.5~2kHz, which is much higher than the fundamental frequency of the motor. At this time, the permanent magnet synchronous motor can be regarded as an RL circuit, in the d ^* axis of the d ^* -q ^* coordinate axis Inject high frequency voltage signal:

Among them: u _in is the amplitude of the injected high-frequency voltage signal, ω _in is the frequency of the injected high-frequency voltage signal.

At this time, the voltage equation of the three-phase permanent magnet synchronous motor under high-frequency excitation in the synchronous rotating coordinate system, that is, the coupled voltage equation, can be simplified as:

Among them: i _din and i _qin are the current response of the motor under high-frequency excitation; u _din and u _qin are the voltage response of the motor under high-frequency excitation; L _d and L _q are the dq-axis inductance components.

In the synchronous rotating coordinate system d-q, the stator inductance of the motor can be expressed as:

In the stationary coordinate system α-β, the stator inductance of the motor can be expressed as:

where: average inductance half differential inductance

Then, under the reference rotor synchronously rotating coordinate system d ^* -q ^* , the relationship between high-frequency voltage and current is:

in: It is the reference value of the current component under d ^* -q ^* ; It is the reference value of the voltage component under d ^* -q ^* .

Step S3: According to the coupling signal, a parallel control strategy is designed to extract the optimal rotor position.

According to step S2.2, refer to the reference value of the current component in the rotor synchronously rotating coordinate system d ^* -q ^* Can be expressed as:

in: is the reference value of the current component under d ^* -q ^* ; u _in is the amplitude of the injected high-frequency voltage signal, ω _in is the frequency of the injected high-frequency voltage signal; L is the average inductance, ΔL is the half-difference inductance; Δθ _e is the rotor reference error horn.

It can be seen that when there is a difference in the inductance of the d-axis and the q-axis, that is, ΔL≠0, the reference value of the current component in the reference rotor synchronously rotating coordinate system d ^* -q ^* The amplitude of is related to the rotor position reference error angle Δθ _e . When Δθ _e = 0, the high-frequency current component of q ^* reference is equal to zero.

The position of the rotor is obtained by processing the high-frequency current of the q ^* axis, which specifically includes the following steps:

Step _S3.1 : In step _S2.2 , the current responses ia, ib, and _ic under high-frequency voltage excitation are known.

i _a , i _b , i _c are converted into currents i _α , i _β in the α-β coordinate system after 3s/2s coordinate transformation, and i _α , i _β are converted into actual currents in the dq coordinate system under 2s/2r coordinate transformation The current i _d ′, i _q ′, input i _d ′, i _q ′ to the LPF for low-pass filtering, and then input it to the current controller as a negative feedback signal.

Step S3.2: In step S3.1: i _α and i _β are known, which are now used as the input of the following two observer parallel processing links. The specific process is as follows:

Step S3.2.1: phase-locked loop PLL link.

Now filter i _α and i _β , and extract the high-frequency part signals containing rotor position information in i _α and i _β through the band-pass filter BPF, and convert the high-frequency part signals into Reference to rotor synchronously rotating coordinate system d ^* -q ^*

In order to obtain the rotor position, the q ^* axis high frequency current Amplitude modulation is performed, and the modulation current sinω _in t is injected, where ω _in is the angular frequency of the modulation current.

The modulation function can be expressed as:

Among them, u _in is the amplitude of the high-frequency sinusoidal voltage signal; ω _in is the angular frequency of the modulation current; L is the average inductance, ΔL is the half-difference inductance; Δθ _e is the rotor reference error angle.

When the rotor reference error angle Δθ _e tends to zero, Δθ _e = 0, cosΔθ _e = 1, sin2Δθ _e = 2sinΔθ _e cosΔθ _e = 2sinΔθ _e , the error signal can be linearized at this time:

Among them, k _ε =u _in (L _q -L _d )/4ω _in L _d L _q ;

Input the modulation signal to the phase-locked loop PLL composed of low-pass filter LPF and PI controller, and obtain the reference angular velocity through the processing of the phase-locked loop PLL Angle with reference rotor position

By adjusting f(ΔΔ _e )=0, that is The rotor position reference value can then converge to the rotor position actual value.

Step S3.2.2: Extended Kalman filter EKF link.

Known voltage responses u _a , u _b , u _c under high-frequency voltage excitation, after 3s/2s coordinate transformation, voltage signals u _α , u _β are obtained, and u _α , u _β are compared with i _α in step S3.2 , i _β are jointly input to the extended Kalman filter EKF, and the reference angular velocity is obtained after being processed by the extended Kalman filter EKF and the reference rotor position angle

Step S3.3: Extract the optimal rotor position to form a closed-loop control system.

Step S3.3.1: In the step S3.2.1 PLL link, The reference angular velocity is obtained after being processed by the phase-locked loop PLL Angle with reference rotor position In the step S3.2.2EKF link, u _α , u _β and i _α , i _β are processed by EKF to obtain the reference angular velocity and the reference rotor position angle

Step S3.3.2: The least squares method is used within a sampling period of a certain width. The least squares method refers to a mathematical optimization technique, which finds the best function matching of data by minimizing the sum of squares of errors.

will be the observer PLL output of the and the output of the observer EKF Send it to the least square solver to find the optimal solution

Assume that within a set of sampling periods and N sets of sampling values are obtained, for several sets of data input to the least squares solver It can be regarded as 2n discrete points in the plane Cartesian coordinate system, and 2n discrete data sequences can be obtained

The 2n discrete points are described by a best fitting curve L, so that the 2n discrete points are not far above or below the curve. The fitting curve L can not only reflect the overall distribution of the data, but also avoid large local fluctuations, and can better reflect the characteristics of the approximate discrete data points.

Assuming that the coordinates of 2n discrete points are ( _xi , y _i ), (i=1, 2, ... 2n), the function f( _xi ) is a function expression passing through these 2n discrete points, and the fitting curve It does not need to pass through all known points, but it can reflect the basic relationship of the data and obtain the most appropriate function model from a large number of discrete data.

When the fitting function does not require strict passage of all discrete data points ( _xi , y _i ), (i=1, 2, ... 2n), the fitting function and f( _xi ) have residuals at _xi (i=1, 2,...2n), and δ( _xi ) is not necessarily all zero, in order to make the fitting curve reflect the variation trend of all discrete data as much as possible, δ( _xi ) is required to take the minimum value. Note the vector e=[δ ₁ , δ ₂ ... δ _2n ], according to the 2-norm theory, When is the minimum, the least squares solution can be obtained.

Let the fitting function

By 2-norm theory should make the residual function:

Take the partial derivative of the function S and make it zero, that is

Write it in matrix form as:

when When linearly independent, the equation system has a unique solution, and this unique solution is the optimal solution, that is, the fitting function Then the fitting function That is, the function expression of the optimal solution ω ^* .

Since the observer PLL is mainly used to observe the state of the motor at low speed and output The extended Kalman filter EKF has a large speed observation range, and is usually used for state observation and output at medium and high speeds Acquire several groups within a certain sampling width Reference value and input to the least squares solver, in Find the best parameter match for the data between Make the motor complete a good transition between zero low speed and medium high speed.

Optimal solution Feedback to the 2r/2s coordinate transformation of step S2.1, the 2s/2r coordinate transformation of step S3.1 and the 2s/2r coordinate transformation of step S3.2, replace θ _e as a control parameter to participate in the calculation of the coordinate transformation process;

In the same speed context, by adding the observer PLL output and the observer EKF output of Feed into the least squares solver, in and Find the best parameter match ω ^* between the data, the greatest degree of attenuation due to and There is a large magnitude error between them, so that the motor can complete a good transition between zero and low speeds and medium and high speeds, and can effectively obtain the optimal rotor position information during acceleration or deceleration.

Send ω ^* into the speed controller and output the reference current command after processing exist Under the control method, the reference current command in the current controller Perform difference calculation with the actual current command i _d ′, i _q ′ fed back in step S3.1 to obtain the error value, and the current controller forms a closed-loop system by controlling the error value, and reduces the error value, so that the system gradually tends to be stable .

To sum up: the phase-locked loop PLL and the extended Kalman filter EKF are designed in parallel, and the method of high-frequency pulse vibration voltage input signal is used to calculate the rotor position angle θ _e and angular velocity ω separately, effectively avoiding the closed-loop control error from being enlarge. At the same time, within a sampling period of a certain width, the reference input signal and Carry out least squares parameter identification, select the optimal solution ω ^* and Minimize the impact of large error margins.

The optimal solution ω ^* and It can adapt to the position sensorless control of the motor at each speed stage, so that the motor can effectively obtain the rotor position information during the acceleration or deceleration process, solve the good transition between the zero low speed and high speed section of the permanent magnet synchronous motor and extract the optimal rotor The problem of sensorless control of position angle.

Claims (5)

1. A position-sensorless control method of extracting an optimal rotor position, the method comprising the steps of:

s1, using conventionalReference coordinate system d^*-q^*Lower reference current commandD is obtained after being processed by a current controller^*-q^*Reference voltage command under coordinate systemAt reference voltage commandInjecting a high-frequency pulse vibration signal on the basis of the reference voltage instruction, and acquiring a superposed signal of the reference voltage instruction after high-frequency voltage injection;

s2, outputting a drive coupling signal containing rotor position information after the superposed signal is subjected to coordinate transformation and SVPWM processing, wherein the signal is used for driving a motor PMSM;

s3, according to the coupling signal, designing a parallel control strategy and extracting the optimal rotor positionAnd angular velocity ω^*；

S4, optimizing the angular speed omega^*Feedback to the current controller to optimize the rotor positionAnd (4) replacing the parameters of coordinate transformation to carry out circulation to form a closed-loop system.

2. The position sensorless control method for extracting an optimal rotor position according to claim 1, wherein the specific method of S1 is as follows:

step S1.1: the current equation under the known d-q coordinate system of the three-phase permanent magnet synchronous motor is as follows:

wherein i_d、i_qIs d-q axis current; u. of_d、u_qIs the d-q axis voltage; l is_d、L_qIs the d-q axis inductance component; r is a stator resistor; psi_fIs a permanent magnet flux linkage; omega_eIs the electrical angular velocity;

if i_d、i_qFully decoupled, the above equation can be changed to:

wherein u is_d0、u_q0D-q axis voltage output after current decoupling; i.e. i_d、i_qIs d-q axis current; l is_d、L_qIs the d-q axis inductance component; r is a stator resistor; psi_fIs a permanent magnet flux linkage; omega_eIs the electrical angular velocity;

to i_d、i_qThe voltage equation after complete decoupling is subjected to Laplace transform to obtain: y is_(s)＝G_(s)U_(s)

Wherein:

wherein: u. of_d0(s)、u_q0(s) is a complex voltage value after laplace transformation; i.e. i_d(s)、i_q(s) is a complex current value after laplace transform;

adopting PI regulator and carrying out feedforward decoupling to obtain a reference coordinate system d^*-q^*The following output voltages are:

wherein,synchronizing the rotation of the coordinate system d for reference^*-q^*Lower reference current command, i_d、i_qThe current instruction is under a d-q coordinate system of a synchronous rotation coordinate system;is the reference coordinate system d obtained after the processing of the current controller in the step S1^*-q^*A lower voltage; l is_d、L_qIs the d-q axis inductance component; psi_fIs a permanent magnet flux linkage; omega_eIs the electrical angular velocity; k_pd、K_pqTo proportional gain, K_id、K_iqIs the integral gain;

step S1.2: according to step S1.1, d is obtained^*-q^*Coordinate axis output voltageOn the basis of d^*Injecting high frequency sinusoidal voltage signals into the shaft

Wherein:for an injected high-frequency sinusoidal voltage signal u_inIs the amplitude, omega, of a high-frequency sinusoidal voltage signal_inIs the frequency of the high frequency sinusoidal voltage signal, resulting in a superimposed signal.

3. The position sensorless control method for extracting the optimal rotor position according to claim 2, wherein the specific method of S2 is as follows:

step S2.1: the superposed signals are obtained after 2r/2s coordinate transformationθ_eIs the actual rotor position angle; omega is d-q coordinate axis angular velocity; the 2r/2s coordinate transformation principle is as follows:

transforming the synchronous rotating coordinate system d-q to a stationary coordinate system α - β, T_2r/2sIs a coordinate transformation matrix, which can be expressed as:

step S2.2:after being processed by SVPWM algorithm, the AC side phase voltage u is output_a、u_b、u_cSinusoidal phase current i with mutual difference of 120 DEG electrical angle_a、i_b、i_cDriving a motor PMSM;

in the natural coordinate system ABC, the phase voltage is u_a、u_b、u_cThe sinusoidal phase current with the difference of 120 degrees is i_a、i_b、i_cAfter 3s/2r coordinate transformation, the phase current and the phase voltage are at d^*-q^*Equivalent representation under the coordinate axis asTherefore, the method comprises the following steps:i.e. at d^*-q^*The coupling signal of the motor rotor position is contained under the coordinate axis;

wherein:

solving forThe specific process is as follows:

establishing a reference rotor synchronous coordinate system d^*-q^*Relation to the actual rotor synchronous coordinate system d-q;

wherein α - β is a static coordinate system, omega is d-q axis angular velocity, omega^*Is d^*-q^*The angular speed of the shaft is set to be,for reference rotor position angle, θ_eFor actual rotor position angle, Δ θ_eFor the rotor reference error angle, i.e.:

regarding the permanent magnet synchronous motor as RL circuit, at d^*-q^*D of the coordinate axis^*Injecting a high-frequency voltage signal into the shaft:

wherein: u. of_inFor injecting high-frequency voltage signal amplitude, omega_inIs the frequency of the injected high-frequency voltage signal;

at this time, the voltage equation of the three-phase permanent magnet synchronous motor under the high-frequency excitation in the synchronous rotating coordinate system, namely the voltage equation after coupling, can be simplified as follows:

wherein: i.e. i_din、i_qinIs the current response of the motor under high-frequency excitation; u. of_din、u_qinIs the voltage response of the motor under high-frequency excitation; l is_d、L_qIs the d-q axis inductance component;

in the synchronous rotating coordinate system d-q, the motor stator inductance can be expressed as:

under the stationary coordinate system α - β, the motor stator inductance can be expressed as:

wherein: average inductanceHalf-difference inductor

Then synchronously rotating the coordinate system d on the reference rotor^*-q^*The relationship between the high-frequency voltage and the current is:

wherein:is d^*-q^*A lower current component reference value;is d^*-q^*The lower voltage component reference value.

4. The position sensorless control method for extracting the optimal rotor position according to claim 3, wherein the specific method of S3 is as follows:

step S3.1, mixing i_a、i_b、i_cAfter 3s/2s coordinate transformation, the current is converted into current i under α - β coordinate system_α、i_βI is transformed under 2s/2r coordinate_α、i_βConverted into actual current i under d-q coordinate system_d′、i_q', will i_d′、i_qThe signal is input to an LPF for low-pass filtering and then is input to a current controller as a negative feedback signal;

step S3.2, in step S3.1: known as i_α、i_βThe method is used as the input of the following two observer parallel processing links, and the specific process is as follows:

step S3.2.1, PLL link of phase locked loop

Will i_α、i_βFiltering, and extracting i through a band-pass filter BPF_a、i_βHigh-frequency part signal containing rotor position information is converted into reference rotor synchronous rotation coordinate system d after 2s/2r coordinate transformation^*-q^*Is as follows

To obtain the rotor position, q is needed^*Shaft high frequency currentAmplitude modulation is carried out, and modulation current sin omega is injected_int, where ω_inIs the modulation current angular frequency;

the modulation function can be expressed as:

wherein u is_inIs the amplitude of the high frequency sinusoidal voltage signal; omega_inIs the modulation current angular frequency; l is the average inductance, and Delta L is the half-difference inductance; delta theta_eIs a rotor reference error angle;

when rotor reference error angle delta theta_eTowards zero, Δ θ_e＝0，cosΔθ_e＝1，sin2Δθ_e＝2sinΔθ_ecosΔθ_e＝2sinΔθ_eAt this point, the error signal may be linearized:

wherein k is_ε＝u_in(L_q-L_d)/4ω_inL_dL_q；

The modulation signal is input into a phase-locked loop PLL consisting of a low-pass filter LPF and a PI controller, and the reference angular velocity is obtained through the processing of the phase-locked loop PLLAngle with reference rotor positionBy adjusting f (Δ)_e) Is equal to 0, i.e The rotor position reference value can converge to the rotor position actual value;

step S3.2.2, expanding a Kalman filter EKF link;

knowing the voltage response u under high frequency voltage excitation_a、u_b、u_cThe voltage signal u is obtained after the coordinate transformation of 3s/2s_a、u_βWill u_α、u_βWith step S3.1 i_α、i_βThe reference angular velocity is obtained after the common input to the extended Kalman filter EKF and the processing of the extended Kalman filter EKFAnd a reference rotor position angle

5. The position sensorless control method for extracting the optimal rotor position according to claim 4, wherein the specific method of S4 is as follows:

step S4.1, outputting observer PLLAnd observer EKF outputSending the solution to a least square solver to obtain an optimal solutionOptimal solutionThe 2r/2S coordinate transformation fed back to step S2.1, the 2S/2r coordinate transformation of step S3.1 and the 2S/2r coordinate transformation of step S3.2, instead of θ_eParticipating in a coordinate transformation process as a control parameterThe calculation of (1);

step S4.2, by outputting observer PLLAnd observer EKF outputIs fed to a least squares solver atAndfind the best parameter match omega of the data between^*，

Step S4.3, mixing omega^*The reference current command is sent to a speed controller and is output after being processedIn thatUnder the control method of (2), the reference current instruction is set in the current controllerAnd the actual current command i fed back in step S3.1_d′、i_q' A difference operation is carried out to obtain an error value, and the current controller forms a closed loop system by controlling the error value.