阅读说明：本技术 一种高速永磁同步电机无位置传感器控制方法 (High-speed permanent magnet synchronous motor position-sensorless control method ) 是由 苑国锋 张宝利 于 2020-04-15 设计创作，主要内容包括：本发明公开了一种高速永磁同步电机无位置传感器控制方法,该控制方法能够在降低传统高频注入法产生的噪音的同时提高高速永磁同步电机静止或运行于低速区的动态性能,并且能够保障电机在突加、减载等动态操作过程中稳定运行。(The invention discloses a position sensorless control method of a high-speed permanent magnet synchronous motor, which can improve the dynamic performance of the high-speed permanent magnet synchronous motor in a static or low-speed area while reducing the noise generated by the traditional high-frequency injection method, and can ensure the stable operation of the motor in the dynamic operation processes of sudden loading, load shedding and the like.)

1. A control method of a high-speed permanent magnet synchronous motor without a position sensor comprises the following steps:

step (1): a high-frequency random square wave is injected,

step (2): detecting three-phase stator current i_A、i_B、i_C，

And (3): the carrier separation operation is carried out, and the carrier separation operation is carried out,

and (4): obtaining the envelope curve of the high-frequency current component,

and (5): calculating position estimation error

And (6): calculating estimated rotor estimated electrical angular velocityAnd estimating rotor position

And (7): obtaining fundamental current i by selecting proper current loop bandwidth_df、i_qf，

And (8): calculating the given value of the stator q-axis current under a rotating coordinate system

And (9): calculating reference values of d-axis and q-axis voltages of the stator in a rotating coordinate system

Step (10): the inverter is driven.

2. The control method of claim 1, wherein the random square wave comprises two sets of square wave signals with different frequencies and different amplitudes,

wherein the amplitude of the higher frequency signal is twice that of the lower frequency signal, while the frequency of the higher frequency signal is twice that of the lower frequency signal.

3. Control method according to claim 1, characterized in that the three-phase stator current i_A、i_B、i_CMay be obtained by a current sensor.

4. The control method according to claim 1,

the carrier separation operation requires that the three-phase stator currents i detected first_A、i_B、i_CStator current i under a two-phase static coordinate system is obtained through 3/2 transformation_α、i_β(ii) a For stator current i under two-phase static coordinate system_α、i_βCarrying out carrier separation operation to obtain a high-frequency current component i under a two-phase static coordinate system_αh、i_βhObtaining fundamental current component i under two-phase static coordinate system through current loop filtering_αf、i_βf。

5. The control method according to claim 1,

the fundamental current component i_αf、i_βfCan be used for calculating and calculating fundamental current i_df、i_qf，

The high-frequency current component i_αh、i_βhCan be used to calculate the envelope of the high frequency current component.

6. The control method according to claim 5,

the envelope Δ i of the high-frequency current component_αh、Δi_βhCan be obtained by combining the high-frequency current component i in a two-phase static coordinate system_αh、i_βhAnd multiplying the signal by a demodulation signal, and then performing normalization processing to obtain the signal.

7. The control method according to claim 1,

the position estimation errorCan be obtained by dividing the envelope of the high-frequency current component Δ i_αh、Δi_βhRespectively with the estimated rotor positionThe sine value and the cosine value of (1) are multiplied, and the results are subtracted.

8. The control method according to claim 1,

by rotor position error valueAs input, the estimated rotor electrical angular velocity can be obtained by a rotor position robust observerAnd estimating rotor position

9. The control method according to claim 1,

setting the motor speedAnd the estimated electric angular velocity calculated in the step (6)After the difference value of the reference voltage is processed by a PI regulator, the given value of the stator q-axis current under a rotating coordinate system can be obtained

10. The control method according to claim 1,

the voltage reference valueCan pass the given value of the stator q-axis current under the rotating coordinate systemSet value of d-axis current of statorRespectively comparing the current with the d and q axis fundamental wave currents i of the stator under the rotating coordinate system obtained by calculation in the step (7)_df、i_qfSubtracting, then obtaining by a complex vector regulator,

will be provided withUsing the estimated rotor position calculated in step (6)Carrying out inverse Park conversion on the conversion angle to obtain the stator voltage u under the two-phase static coordinate system_α、u_β(ii) a And the SVPWM is used as an input to generate a driving signal to drive the inverter.

Technical Field

The invention belongs to the field of motor control, and particularly relates to a position sensorless control method of a high-speed permanent magnet synchronous motor.

Background

The high-speed alternating current motor is used as a power source of a high-speed transmission technology and is widely applied in the fields of aerospace, flywheel energy storage, high-precision numerical control machines, distributed energy sources and the like. The permanent magnet synchronous motor has the advantages of high efficiency, high power density and the like, and is widely applied to high-speed systems. However, the rated frequency of the high-speed motor is generally over 500Hz, and the switching frequency of the inverter cannot be too high because the inverter is limited by the heat dissipation capacity of the device. The high-speed motor control method needs to meet the operating condition of low ratio of inverter switching frequency to motor fundamental wave operating frequency, namely low carrier ratio. Meanwhile, the high-speed motor is not suitable for installing a mechanical encoder due to the complex structure of the high-speed motor and the like, so that the research on the control method without the position sensor meeting the condition of low carrier ratio can promote the application of the high-speed permanent magnet synchronous motor.

The zero and low speed area position-sensorless control method mainly comprises an I/F algorithm and a signal injection method. The principle of the I/F control algorithm is that a known current is injected into a static motor to generate a magnetic field, and the rotor is attracted to rotate until the magnetic field generated by the rotor permanent magnet is in the same direction as the known magnetic field, so that the position of the rotor is obtained. The method is relatively simple and easy to implement, but the motor rotates and even possibly rotates reversely in the process of acquiring the position of the rotor, the accuracy of application is not satisfied, and the method is an open loop method, so that the stability is poor, and therefore an auxiliary signal injection method is mostly adopted in the field of high-accuracy control. The auxiliary signal injection method mostly adopts voltage signals, and common injection voltage forms include rotating voltage signals and high-frequency voltage signals, but the two implementation methods both need to use a plurality of filters, and the filters usually limit the bandwidth of a control system and are difficult to implement, so that the dynamic performance of the system is influenced.

Meanwhile, when the motor is controlled in a zero-speed area and a low-speed area, sharp and harsh noise can be caused by the use of high-frequency signals, and serious influence is caused on operators. And the non-linear effects of the inverter can affect the accuracy of the estimation of the rotor position. However, in practical control systems, it is usually required that the high-speed permanent magnet synchronous motor can drive the micro gas turbine or the electric motor to start and operate at a low speed, and the electric motor is switched to operate in a power generation mode after entering a middle-speed and high-speed region. Therefore, the high-speed permanent magnet synchronous motor is required to have good dynamic performance when being static and running in a low-speed region, and can quickly respond and stably run in the loading and unloading processes.

Therefore, a position-sensor-free control strategy which meets the zero-speed and low-speed zone loading capacity of the high-speed permanent magnet synchronous motor, is low in noise and can adjust the injection amplitude on line is needed to be designed.

Disclosure of Invention

In order to solve the above problems, the present inventors have made intensive studies to design a position sensorless control method for a high-speed permanent magnet synchronous motor, which can improve the dynamic performance of the high-speed permanent magnet synchronous motor in a static or low-speed region while reducing noise generated by a conventional high-frequency injection method, and can ensure stable operation of the motor in dynamic operation processes such as load addition and subtraction.

According to a first aspect of the present invention, there is provided a position sensorless control method for a high-speed permanent magnet synchronous motor, the control method comprising the steps of:

step (1): a high-frequency random square wave is injected,

step (2): detecting three-phase stator current i_A、i_B、i_C，

And (3): the carrier separation operation is carried out, and the carrier separation operation is carried out,

and (4): obtaining the envelope curve of the high-frequency current component,

and (5): calculating position estimation error

And (6): calculating estimated rotor estimated electrical angular velocityAnd estimating rotor position

And (7): obtaining fundamental current idf and iqf by selecting proper current loop bandwidth,

and (8): calculating the given value of the stator q-axis current under a rotating coordinate system

And (9): calculating reference values of d-axis and q-axis voltages of the stator in a rotating coordinate system

Step (10): the inverter is driven.

In the invention, the random square wave comprises two groups of square wave signals with different frequencies and different amplitudes,

wherein the amplitude of the higher frequency signal is twice that of the lower frequency signal, while the frequency of the higher frequency signal is twice that of the lower frequency signal.

In the present invention, the three-phase stator current i_A、i_B、i_CCan be obtained by a current sensor.

In the present invention, the carrier separation operation requires first detecting the three-phase stator current i_A、i_B、i_C3/2 transformation is carried out to obtain stator current i under a two-phase static coordinate system_α、i_β(ii) a For stator current i under two-phase static coordinate system_α、i_βCarrying out carrier wave separation operation to obtain a high-frequency current component i under a two-phase static coordinate system_αh、i_βhObtaining fundamental current component i under two-phase static coordinate system through current loop filtering_αf、 i_βf。

Further, the fundamental current component i_αf、i_βfCan be used for calculating fundamental current i_df、i_qf，

The high-frequency current component i_αh、i_βhCan be used to calculate the envelope of the high frequency current component.

The envelope Δ i of the high-frequency current component_αh、Δi_βhCan be obtained by converting a high-frequency current component i under a two-phase stationary coordinate system_αh、i_βhAnd multiplying the signal by a demodulation signal, and normalizing the signal to obtain the signal.

In the present invention, the position estimation error isBy applying a high frequencyEnvelope of current component Δ i_αh、Δi_βhRespectively with the estimated rotor positionThe sine and cosine values of (a) are multiplied and the results are subtracted.

According to the invention, the rotor position error value is usedAs input, the estimated rotor electrical angular velocity can be obtained by a rotor position robust observerAnd estimating rotor position

Setting the motor speedAnd the estimated electrical angular velocity calculated in the step (6)After the difference value of the reference voltage is processed by a PI regulator, the given value of the stator q-axis current under a rotating coordinate system can be obtained

The voltage reference valueCan pass the given value of the stator q-axis current under the rotating coordinate systemSet value of d-axis current of statorAre respectively provided withSubtracting the d and q axis fundamental wave currents idf and iqf of the stator in the rotating coordinate system obtained by calculation in the step (7), then obtaining the d and q axis fundamental wave currents idf and iqf through a complex vector regulator,

will be provided withUsing the estimated rotor position calculated in step (6)Carrying out inverse Park conversion on the conversion angle to obtain the stator voltage u under the two-phase static coordinate system_α、u_β(ii) a And the SVPWM is used as an input to generate a driving signal to drive the inverter.

The invention has the advantages that:

1) according to the invention, the square wave with fixed frequency used in the traditional high-frequency injection method is replaced by the square wave with random frequency, so that the power spectral density of the current is effectively widened, and the noise audible to human ears is reduced;

2) according to the invention, by improving the frequency of the injection voltage, the separation of high-frequency current and fundamental current can be realized by simple algebraic operation, so that time delay and system bandwidth limitation caused by using a band-pass filter and a low-pass filter are avoided, and meanwhile, the calculation can be simplified;

3) the invention can realize the real-time dynamic adjustment of the random square wave amplitude, and effectively reduce the nonlinear effect of the inverter caused by dead time;

4) the invention combines the random frequency square wave injection method with the position robust observer, thereby avoiding the problem of error convergence when the observer changes in a large dynamic range of load.

Drawings

FIG. 1 is a schematic diagram illustrating the control principle of the control method of the present invention;

figure 2 shows a schematic diagram of a permanent magnet synchronous machine coordinate system;

FIG. 3 shows a flow chart of a random square wave generator;

figure 4 shows a schematic diagram of a permanent magnet synchronous machine coordinate system;

FIG. 5 illustrates a signal processing schematic of the random square wave injection method of the present invention;

FIG. 6 shows a high frequency injection method schematic diagram with real-time amplitude adjustment;

FIG. 7 shows the relationship of the sampling current, the high frequency current and the fundamental current;

fig. 8(a) shows a separation method of a higher frequency high frequency current component;

FIG. 8(b) shows a method of separating the lower frequency high frequency current components;

FIG. 9 shows a robust observer structure diagram;

FIG. 10 illustrates a fixed frequency current signal;

FIG. 11 shows a random frequency current signal switched over a full period;

FIG. 12 shows an injected fixed frequency square wave signal;

FIG. 13 shows an injected random frequency square wave signal;

FIG. 14(a) shows a motor frequency spectrum plot for high frequency square wave injection;

FIG. 14(b) shows a plot of the motor frequency spectrum with random square wave injection;

fig. 15(a) shows the inverter non-linear high frequency voltage injection current response without consideration;

FIG. 15(b) shows a filtered inverter nonlinear high frequency voltage injection current response;

Detailed Description

The invention is explained in more detail below with reference to the figures and examples. The features and advantages of the present invention will become more apparent from the description.

The word "exemplary" is used exclusively herein to mean "serving as an example, embodiment, or illustration. Any embodiment described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments. While the various aspects of the embodiments are presented in drawings, the drawings are not necessarily drawn to scale unless specifically indicated.

The control method of the high-speed permanent magnet synchronous motor without the position sensor is characterized by comprising the following steps as shown in figure 1:

step (1): a high frequency random square wave is injected.

Unlike conventional high frequency square wave injection, the random frequency square wave generator generates two sets of square wave signals with different frequencies and amplitudes, and these two sets of signals have the relationship that the amplitude of the higher frequency signal is twice that of the lower frequency signal, while the frequency of the higher frequency signal is twice that of the lower frequency signal, as shown in fig. 2. A random frequency square wave is randomly generated every cycle, and a demodulation signal is generated by the random frequency signal generator, wherein the phase of the demodulation signal is orthogonal to the phase of the random frequency square wave signal.

Specifically, the random square wave injected in step (1) occurs on the following principle: a number between 0 and 1 is generated and compared to 0.5. If the generated number is greater than 0.5, a square wave signal with a higher frequency and a demodulation signal corresponding to the square wave signal are generated, otherwise, a square wave signal with a lower frequency and a demodulation signal corresponding to the square wave signal are generated, whether a square wave period is completed or not is judged, if the square wave period is completed, a next period is entered to generate a new square wave, otherwise, comparison is continued, and the flow chart is shown in fig. 3.

Step (2): detecting three-phase stator current i_A、i_B、i_C

In the permanent magnet synchronous motor coordinate system shown in fig. 4, a two-phase rotating coordinate system is estimatedThe included angle between the rotor position estimation error and the two-phase rotating coordinate system d-q is the rotor position estimation error

Wherein, theta_eIn order to be the actual rotor position,to estimate the rotorA location;

further, to estimating a two-phase rotating coordinate systemAxial injection random frequency high frequency square wave type voltageComparing the ideal d-axis high-frequency response current with an actual high-frequency response current sampling value, and adjusting the amplitude of the injection voltage according to the distortion condition of the actual high-frequency response current, wherein the amplitude adjustment principle is shown in fig. 6;

high-frequency current distortion caused by inverter nonlinearity is suppressed through different injection voltage amplitudes, and the influence of the inverter nonlinearity on rotor position estimation accuracy is reduced.

Obtaining three-phase stator current i through current sensor_A、i_B、i_C。

And (3): and carrying out carrier separation operation.

The three-phase stator current i obtained by sampling in the step (2)_A、i_B、i_CStator current i under a two-phase static coordinate system is obtained through 3/2 transformation_α、i_β(ii) a For stator current i under two-phase stationary coordinate system_α、i_βCarrying out carrier separation operation to respectively obtain fundamental current components i under two-phase static coordinate system_αf、i_βfAnd a high-frequency current component i in a two-phase stationary coordinate system_αh、i_βh；

Specifically, fig. 7 is a diagram illustrating the relationship between sampling, high frequency and fundamental current;

the fundamental current component i_αf、i_βfObtained by current loop filtering and used for calculating and calculating fundamental current i_df、i_qf。

Fig. 8(a) and 8(b) show a method of separating two square wave signals having different frequencies included in random frequencies.

In particular, the high-frequency current component i_αh、i_βhCan pass throughThe principle of separating the higher frequency signals is shown in (1) and (2), and the process of separating the lower frequency signals is shown in (3) and (4).

Wherein i_αhRepresenting high-frequency current component i in alpha axis_αs(n) stator current in α axis at present sampling time, i_αs(n-1) represents a stator current i in an alpha axis system at the last sampling timing_αs(n-2) represents a stator current in the α axis system at the last sampling timing, i_αf(n) represents the fundamental current component in the alpha axis system at the present sampling time, i_αf(n-1) represents the fundamental current component in the alpha axis system at the last sampling time, i_αfAnd (n-2) represents the lower fundamental current component of the alpha axis at the last sampling moment.

High frequency current component i_βhCan be obtained by the following formula,

in the formula i_βhRepresenting high-frequency current components in the beta axis system, i_βs(n) stator current in α axis at present sampling time, i_βs(n-1) stator current i in the beta axis system at the last sampling time_βs(n-2) stator current in a beta axis system at the time of up-sampling; i.e. i_βf(n) represents the fundamental current component in the beta axis system at the present sampling time, i_βf(n-1) represents the shafting lower fundamental current component, i, of the last sampling instant_βf(n-2) represents a β -axis lower fundamental current component at the last sampling timing.

The high-frequency current component i_αh、i_βhCan be used for calculating envelope curve of high-frequency current component and further calculating position estimation error

And (4): obtaining envelope curve of high-frequency current component

Obtaining the high-frequency current component i under the two-phase static coordinate system calculated in the step (3)_αh、i_βhMultiplying the envelope curve by a demodulation signal, and obtaining a high-frequency current component envelope curve delta i under a two-phase static coordinate system after normalization processing_αh、Δi_βhEnvelope Δ i of high-frequency current component in two-phase stationary coordinate system_αh、Δi_βhRespectively containing the actual rotor position theta_eCosine and sine quantities. The use of a filter is avoided, and the system performance is improved;

in which the envelope Δ i of the current component_αh、Δi_βhCan be obtained by the following formula:

in the formula, L_dRepresenting the direct-axis inductance of the motor, u_dThe high frequency voltage injected by the d-axis is shown.

And (5): calculating position estimation error

Dividing the envelope line delta i of the high-frequency current component in the two-phase static coordinate system obtained by calculation in the step (4)_αh、Δi_βhSeparately connecting the envelope of the high-frequency current component with the estimated rotor positionThe sine value and the cosine value are multiplied, and the results are subtracted to obtain the position estimation error

Preferably, the envelope Δ i is extracted from the high-frequency current component containing the rotor position information_αh、Δi_βhAnd estimating rotor position angleThe approximate position estimation error is shown as follows:

and (6): calculating estimated rotor estimated electrical angular velocityAnd estimating rotor position

Obtaining the rotor position error value in the step (5)Obtaining as input an estimated rotor estimated electrical angular velocity by a rotor position robust observerAnd estimating rotor positionThe structure diagram of the robust observer is shown in FIG. 9;

and (7): obtaining fundamental current i by selecting proper current loop bandwidth_df、 i_qf

Dividing the fundamental wave current component i in the two-phase static coordinate system obtained by calculation in the step (3)_αf、i_βfUsing the estimated rotor position calculated in step (5)Performing Park conversion on the conversion angle to obtain fundamental current i under a rotating coordinate system_df、i_qf(ii) a The estimated rotor position calculated in step (6)Can be used for step (5) to obtain the estimated rotor positionThe position error of the rotor can be extracted through a sum and difference formula of a trigonometric function according to the sine value and the cosine value;

and (8): calculating the given value of the stator q-axis current under a rotating coordinate system

Given value of motor speedAnd the estimated electric angular velocity calculated in the step (6)The difference value of the reference voltage is processed by a PI regulator to obtain the given value of the stator q-axis current under the rotating coordinate system

And (9): calculating reference values of d-axis and q-axis voltages of the stator in a rotating coordinate system

Setting the given value of the stator q-axis current in the rotating coordinate system obtained by calculation in the step (8)Set value of d-axis current of statorRespectively comparing the current with the d and q axis fundamental wave currents i of the stator under the rotating coordinate system obtained by calculation in the step (7)_df、i_qfSubtracting, and then obtaining reference values of d-axis and q-axis voltages of the stator under a rotating coordinate system through a complex vector regulator

Step (10): drive inverter

The stator voltage reference value under the rotating coordinate system obtained by calculation in the step (9) is used for calculatingUsing the estimated rotor position calculated in step (6)Carrying out inverse Park conversion on the conversion angle to obtain the stator voltage u under the two-phase static coordinate system_α、u_β(ii) a The SVPWM is used as an input of the SVPWM to generate a driving signal to drive the inverter.

According to the control method of the high-speed permanent magnet synchronous motor without the position sensor, which is provided by the invention, the control method has the following beneficial effects:

the random square wave injection method can reduce the noise of the motor.

The power spectral density function is used for calculating the average power of the signal and can be used for noise analysis of the permanent magnet synchronous motor. Further, when the power density at a certain point is large, noise is generated. The power density spectra of the permanent magnet synchronous motor by the fixed frequency square wave injection method and the random square wave injection method are respectively as follows:

1) fixed frequency square wave injection

The current Fourier transform of one period of the injected fixed frequency square wave signal shown in FIG. 10 can be expressed by

According to the Euler formula there is e^jωtCos ω t + jssin ω t; ω 2 pi f; j represents an imaginary symbol.

According to the formula Fourier transform (12) have

In the formula, T₁＝T₂T/2, where T is the square wave period,

f is square wave frequency; n is a non-negative integer (0, 1, 2.).

By substituting f ═ n/T into the above formula, the following can be concluded:

wherein T represents a square wave period, n is a non-negative integer (0, 1, 2.)

From equation (15), it can be concluded that the discrete spectrum will appear at odd multiples of the injected fixed frequency signal.

2) Random frequency square wave injection

Injecting a random square wave signal as shown in fig. 11, the random square wave signal comprising both a higher frequency signal and a lower frequency signal, and the power density of which can be expressed by the following formula:

when E [ E ]^j2πfT]≠1

Formula (16) is a continuum;

when E [ E ]^j2πfT]＝1

Formula (17) is a discrete spectrum;

where S (f) is the current power density, E [ ] is the desired operator, and I (f) is the Fourier transform over a period of the current signal.

When f satisfies the following formula (18), and E [ I (f) ] ≠ 0, scattering spectrum is generated, which generates noise:

[p₁,p₂]is a probability matrix generated by two random signals; in the formula, p₁At a higher frequencyProbability of rate signal, p₂Probability of lower frequency signal, T₃For higher frequency signal periods, T₄A lower frequency signal period.

Preferably, let p₁＝p₂＝0.5；

Preferably, T₃＝0.5T₄。

Since there is a 2-fold relationship between the two injection frequencies, f is also₃₄Is the frequency f of a higher frequency signal₃With the frequency f of the lower frequency signal₄The least common multiple of. The following relationships coexistn is a positive integer.

According to

When n is an even number, the number of n,at this time have

f₃₄Is f₃And f₄Of (a) is a least common multiple of (b), wherein f₃Representing the frequency of the higher frequency signal, f₄Representing the frequency of the lower frequency signal.

In the formula, E [ I (nf)₃₄)]Representing the current Fourier transform I (nf)₃₄) A periodic weighted average. The specific calculation process is shown as formulas (17) and (18).

At this time, the result is E [ I (nf)₃₄)]Equal to zero, so there is no discrete spectrum.

When n is an odd number, the number of the transition metal atoms,at this time have

f₃₄Is f₃And f₄Of (a) is a least common multiple of (b), wherein f₃Representing the frequency of the higher frequency signal, f₄Representing the frequency of the lower frequency signal.

In the formula, E [ I (nf)₃₄)]Represents the current I (nf)₃₄) A weighted average of one period of the fourier transform. Due to E [ I (nf)₃₄)]Not equal to zero, when there is a discrete spectrum.

As can be seen from the above equations (19) and (20), the current signal induced by the full-period switching random frequency square wave includes both a discrete spectrum and a continuous spectrum, and both the discrete spectrum and the continuous spectrum are at frequency points that are odd multiples of the higher injection frequency.

Nonlinear effect compensation principle analysis of (II) inverter

And analyzing the principle of inverter nonlinear effect compensation. Due to the existence of the nonlinearity of the inverter, the actual high-frequency voltage signal is distorted, and the d-axis shows a value delta u_dhThe voltage error of (1) is generated in the q-axis by a value of δ u_qhVoltage error of (2). Make the injection voltageIn a step shape, resulting in an induced currentCreating a non-linear effect.

Since random high-frequency voltage is injected on the estimated d-axis, the error caused by the nonlinearity of the inverter is considered, and the random high-frequency voltage is injected on the estimated d-axis

Representing an estimated rotational coordinate system; v_injRandom frequency voltage amplitude for preamble injection

It is transformed into an actual rotating coordinate system to obtain,

r denotes the actual rotational coordinate system.

According to the motor voltage equation, have

By coordinate transformation

The voltage and current of d-axis high-frequency injection method have the following relationship

Can be obtained from the above formula

Substitution of (26) into (24) gives

The error voltage and the error current have the following relationship

Substituting (27) into (28)

In the formula (I), the compound is shown in the specification,representing the actual angle and the estimated angle error; Δ i_qhRepresenting a q-axis current variation value; Δ i_dhRepresenting a d-axis current variation value; l is_qRepresenting motor quadrature axis inductance; l is_dRepresents the direct-axis inductance of the motor; delta i_qhRepresenting the q-axis current magnitude error.

Constructing a PI regulator as shown in FIG. 6, in which the latter term L contained is eliminated (29)_qδi_qh/Δi_dh(L_q-L_d) The nonlinear effect of the inverter can be eliminated, and the purpose of real-time compensation is achieved.

Example 1

Compared with the injection of a fixed frequency signal, the injection of the random square wave signal has a more remarkable effect on reducing the noise of the permanent magnet synchronous motor, and the injection of the random square wave signal has the following specific effects:

1) motor current frequency spectrum diagram injected by fixed-frequency square wave signal

A square wave signal with a fixed frequency, an amplitude of 60V, a frequency of 2.5kHz and a period T of 0.004s is injected as shown in fig. 12.

The discrete spectral density can be obtained from equation (15),

as shown in fig. 14(a), the motor current spectrogram obtained by the above formula calculation has more obvious peaks, and the amplitudes thereof are: 7.2dB, -0.1dB, -6.2 dB; the spikes of the motor current spectrogram represent noise. The larger the amplitude of the spike, the greater the noise accordingly.

2) Motor current frequency spectrogram injected by random frequency square wave signal

Fixed frequency square wave signals are injected as shown in fig. 13, the amplitude of each of the signals is Vinj 1-60V and Vinj 2-30V, and the frequency of each of the signals corresponds to f₃2.5kHz andf₄1.25 kHz. T3 ═ 0.004s is the higher frequency signal period, and T4 ═ 0.008 is the lower frequency signal period.

The motor discrete spectral density can be obtained from equation (20):

in the formula, E [ I (nf)₃₄)]Represents the current I (nf)₃₄) A weighted average of one period of the fourier transform. At this time E [ I (nf)₃₄)]Not equal to zero, so a discrete spectrum exists.

f₃₄Is f₃And f₄Of (a) is a least common multiple of (b), wherein f₃Representing the frequency of the higher frequency signal, f₄Representing the frequency of the lower frequency signal.

As shown in fig. 14(b), the motor current spectrogram obtained by calculating the above equation (20) has a relatively obvious peak, and the amplitudes thereof are: 3.8dB, -4.2dB, -10.3dB, and the comparison with the graph 14(a) shows that the value of each discrete point on the motor current spectrogram is reduced compared with the amplitude of each peak in the graph 14(a), so the control method provided by the invention achieves the purpose of noise reduction.

Example 2

In another preferred embodiment of the present invention, the actual high frequency voltage signal is distorted due to the non-linearity of the inverter, and δ u appears on the d-axis, for example, on the d-axis_dhVoltage error of (2) so that the injection voltage isStepped, resulting in induced currentA nonlinear effect is produced as shown in fig. 15 (a).

Therefore, based on the principle of the high-frequency injection method with real-time amplitude adjustment shown in FIG. 6, K in the figure_phThe proportional coefficient of the PI regulator; k_ihIs an integral coefficient; tsamp is adoptedSample time; z is a discretization operator; the error is eliminated by constructing the regulator as shown in equation (29) below:

in the formula (I), the compound is shown in the specification,representing the actual angle and the estimated angle error; Δ i_qhRepresenting a q-axis current variation value; Δ i_dhRepresenting a d-axis current variation value; l is_qRepresenting motor quadrature axis inductance; l is_dRepresents the direct-axis inductance of the motor; delta i_qhRepresenting the q-axis current magnitude error.

Construction of regulator Elimination (29) of the latter term L involved_qδi_qh/Δi_dh(L_q-L_d) The inverter non-linear effects can be eliminated.

By adjusting the cancellation, the voltage can be avoided by dynamically adjusting fig. 15(a) to fig. 15(b)Step phenomenon and induced currentAnd nonlinearity is adopted to improve the estimation accuracy of the rotor position angle.

Therefore, the injection amplitude self-adaptive adjustment can inhibit the nonlinearity of the inverter, and the signal distortion caused by the nonlinearity of the inverter is inhibited by changing the injection voltage amplitude at different moments.

The present invention has been described above in connection with preferred embodiments, which are merely exemplary and illustrative. On the basis of the above, the invention can be subjected to various substitutions and modifications, and the substitutions and the modifications are all within the protection scope of the invention.