KR19990015353A - Magnetic flux reference vector control method of induction motor by high frequency current injection - Google Patents

Abstract

The present invention relates to a method of controlling a magnetic flux reference vector of an induction motor by high frequency current injection, and in particular, to control the torque ripple caused by high frequency current injection. Since the method according to the present invention does not require accurate magnetic flux information, vector control and speed control are smoothly performed even at a low speed with an inverter output frequency of about 0.5 Hz, and control is performed regardless of any motor constant other than the stator resistance. It has the advantage of being very robust against fluctuations in motor constants.

Description

Magnetic flux reference vector control method of induction motor by high frequency current injection

The present invention relates to a method of controlling a magnetic flux reference vector of an induction motor by high frequency current injection, and in particular, to control the torque ripple caused by high frequency current injection.

Induction motors are very advantageous in terms of maintenance and repair because they are simpler and more robust than DC motors.However, from the viewpoint of variable speed driving, induction motors are motors for constant speed operation, which makes speed control extremely difficult without changing the input frequency. As a result, its application range has been limited.

Therefore, induction motors have been used for industrial purposes for a long time, but until the 50's, they were mainly used for constant speed operation due to control difficulties, and almost all of the places where variable speed operation is required or where adaptive control is required such as servo operation. DC motors have been used. In the late sixties, vector control theory was announced, and the development of high-speed power semiconductor devices and the development of high-performance microprocessors enabled not only variable speed operation of induction motors, but also the characteristics of induction motors above DC motors. Induction motors replace the application range of DC motors.

In order to perform vector control of an induction motor, speed information or magnetic flux information is essential. Generally, to obtain speed information, a sensor such as a tacho generator, a resolver, or a pulse encoder, In order to obtain the magnetic flux information, a sensor such as a searching coil is required. However, these sensors cause various problems in the operation of the induction motor, and limit the application. First of all, the coupling with the motor is difficult, and it is sensitive to the installation environment, which reduces the reliability and economically causes the price to rise. In other words, these sensors eliminate environmental conditions, reliability and economic advantages, which are advantages of induction motors. In addition, when the motor is far away from the inverter, difficulty in signal transmission occurs, and a problem such as a weak noise occurs, which affects stable operation.

For this reason, many studies have been made on sensorless vector control to estimate the speed information or the magnetic flux information using only voltage and current information without attaching a separate sensor to the induction motor. Typical methods include estimating magnetic flux to obtain direct angle information, vector control, MRAS (Model Reference Adaptive System), and EKF (Extended Kalman Filter).

However, the above-mentioned conventional methods are very vulnerable to fluctuations in the motor constants because of the limited information that can be obtained in the low-speed operation, because they do not work well due to the small size of the voltage and current, and the basic formula of the motor must be used. . Due to these problems, sensorless vector controllers are used only in limited ranges.

An object of the present invention is to solve the disadvantages of the conventional methods as described above, the present invention is independent of the size error when estimating the magnetic flux information by using the torque ripple caused by the injection of high-frequency current, only phase To provide a direct vector control method of an induction motor that depends only on the accuracy of.

Still another object of the present invention is to provide a direct vector control method of an induction motor capable of performing stable control insensitive to constant fluctuations of an induction motor.

It is still another object of the present invention to provide a direct vector control method of an induction motor that is resistant to noise by injecting high frequency current and increasing voltage information at low speed to reduce distortion of the output voltage.

1 is a view showing the error between the actual angle and the control angle in the dq coordinate system,

2 is an equivalent circuit diagram of an d-axis of an induction motor;

3 is a block diagram of an experimental apparatus for performing a direct vector control method of an induction motor according to the present invention;

4 is a diagram of the forward and reverse rotation waveform at -30 rpm at 30 rpm no load,

5 is a diagram of a forward and backward rotation waveform at -30 rpm at 30 rpm no load,

6 is a diagram of the forward and reverse waveforms at -30 rpm and 30 rpm half load,

7 is a diagram of the forward and reverse waveforms at -50 rpm to 50 rpm half load,

8 is a diagram of a forward and backward rotation waveform at -30 rpm to 30 rpm full load,

9 is a diagram of the forward and backward rotation waveform at -50 rpm to 50 rpm full load,

10 is a view showing a correlation between a torque current and torque;

11 is a diagram of the forward and reverse rotational waveform at -18 rpm no load at 18 rpm;

12 is a diagram of the forward and reverse waveforms at -18 rpm half load at 18 rpm;

13 is a diagram of the forward and reverse rotation waveform at -30 rpm no load at 30 rpm;

14 is a diagram of the forward and reverse waveforms at -30 rpm half load at 30 rpm;

15 is a diagram of the forward and reverse rotation waveform at -500 rpm no load at 500 rpm;

Fig. 16 is a diagram of the forward and backward rotation waveform at -500 rpm full load at 500 rpm;

Fig. 17 is a diagram showing the 3/4 load characteristic of the rating in the 30 rpm constant speed operation.

In order to achieve the above object, the magnetic flux reference vector control method of the induction motor according to the present invention comprises the steps of: applying a high frequency sinusoidal current to the d-axis of the synchronous coordinate system; Obtaining a torque ripple generated by the current applied in the step; And changing the control angle in a direction of reducing the torque ripple obtained in the step.

In the magnetic flux reference vector control method of the induction motor according to the present invention described above, the step of changing the control angle in the direction of reducing the torque ripple reduces the error Δθ which is an error between the control angle obtained from the torque ripple and the magnetic flux angle of the actual motor. It can be performed by.

In the method of controlling the magnetic flux reference vector of the induction motor according to the present invention, the step of obtaining the torque ripple may be performed by estimating the magnetic flux of the induction motor to obtain the torque and then obtaining the torque ripple through the high pass filter. .

In the above-described method of controlling the magnetic flux reference vector of the induction motor according to the present invention, the step of estimating the magnetic flux of the induction motor may be obtained using a first delay filter equal to the time constant and the rotor time constant of the induction motor.

In the magnetic flux reference vector control method of the induction motor according to the present invention described above, the step of estimating the magnetic flux of the induction motor includes adding a magnetic flux command value through a low pass filter to a measured value in order to improve the characteristics of the magnetic flux estimation at a low speed. You can also use the method.

Hereinafter, the magnetic flux reference control method of the induction motor according to the present invention will be described in detail with reference to the accompanying drawings.

First, the torque ripple that appears during high frequency current injection will be described.

The torque equation of a general induction motor is shown in Equation 1 below.

In Equation 1, P is the number of poles of the induction motor, L _m represents the motor magnetization inductance, L _r represents the rotor inductance, respectively λ _dr ^e , λ _qr ^e represents the dq axis rotor magnetic flux (synchronous coordinate system), λ _dr ^s and λ _qr ^s denote the dq-axis rotor flux (stop coordinate system), respectively, i _ds ^e and i _qs ^e denote the dq-axis stator current (synchronous coordinate system), i _ds ^s and i _qs ^s denote the dq-axis stator current (Static coordinate system) is shown respectively.

In the case of performing the rotor flux reference vector control, the q-axis rotor flux λ _qr ^e becomes 0, so that Equation 1 is expressed as Equation 2 below.

There is a first-order delay relationship shown in Equation 3 below between the d-axis rotor flux and the d-axis stator current in the synchronous coordinate system.

In Equation 3, T _r is the rotor time constant.

As shown in Equation 3, in the general case in which a DC current is applied, only a delay occurs, but in the case where an alternating current is applied, not only a delay but also a decay of magnitude appears. Equation 4 below shows frequency characteristics for the magnitude of Equation 3.

If the high frequency sinusoidal current is injected to the d-axis as shown in Equation 5 below, as shown in Equation 4, the effect of the high frequency current appearing on the rotor flux is almost negligible, and the frequency of the injected current is high. The smaller the influence on the magnetic flux.

Also, as can be seen from Equation 2, when the current is decoupled control and perfect vector control is performed, the influence of the high frequency current does not appear in the torque.

Figure 1 shows the error between the actual angle and the control angle in the dq coordinate system.

As shown in Fig. 1, if the angle being controlled and the magnetic flux angle of the actual motor have an error of Δθ, λ _qr ^e is not 0 but λsinΔθ, which causes ripple in torque due to the injected current.

Therefore, if the control angle can be changed in the direction of reducing the torque ripple, accurate vector control can be performed. The torque ripple may be expressed as in Equation 6 below.

[Lambda] in Equation 6 Is the error between the actual angle and the control angle. In order to easily obtain the angle error Δθ from Equation 6, a high frequency current component is provided at both sides of Equation 6 Multiply by to obtain the following equation (7).

Since Equation 7 is composed of a DC component and a component having a frequency twice the high frequency, the angle error Δθ can be obtained as shown in Equation 8 below by extracting only the DC component using a low pass filter.

In Equation 8, LP represents the output of the low pass filter.

If Δθ is small, Δθ may be almost equal to sinΔθ, and thus an angle error may be obtained without calculating a trigonometric function. In this case, the control angle is updated as in Equation 9 below.

θ [n] = θ [n-1] + Δθ

However, it is unreasonable to apply the calculated angular error directly to vector control due to various factors such as fluctuations in motor constants, filter characteristics, measurement errors, etc. Reduce the impact to ensure smooth control.

Now, estimating torque ripple will be described.

In the present invention, torque is estimated, and torque ripple is obtained by using a high pass filter for the estimated torque. Since the torque obtained by the equation of the still coordinate system from Equation 1 is the same as that obtained by the equation of the synchronous coordinate system, the equation of the static coordinate system that is easy to calculate is used in the present specification.

In order to find the torque, the flux must first be estimated. The magnetic flux is obtained by integrating the value obtained by subtracting the voltage drop caused by the stator resistance from the power supply voltage as shown in Equation 10 below.

In Equation 10, L _σ is a stator instantaneous inductance and is a value determined according to the model of the motor.

However, since the pure integral cannot be implemented because there is a problem of diverging when there is an offset of the measured value, it is obtained using a first order delay filter as shown in Equation 11 below.

In general, in Equation 11, the time constant of the first order delay filter is set equal to the rotor time constant of the motor. In this process, the estimate of the magnetic flux results in the actual magnetic flux, phase delay, and magnitude, and the difference becomes more severe at lower speeds. Therefore, in order to obtain the magnetic flux closer to the actual value, the present invention uses a method of adding the magnetic flux command value to the measured value through a low pass filter. The magnetic flux command value is used where the inverter output frequency is smaller than the frequency of the primary delay filter, and the measured value is used where the inverter output frequency is larger than the frequency of the primary delay filter. The magnetic flux estimate by this method is shown in Equation 12 below.

In Equation 12, λ _qdr ^{s *} is a command value of the rotor flux, and the time constant of the low pass filter used in Equation 12 is equal to the time constant of the first-order delay filter.

By using the magnetic flux estimate value shown in Equation 12, the estimated torque and torque ripple can be expressed as Equation 13 and Equation 14 below.

On the other hand, in the equation 10 Lσ may further increase the accuracy of the flux estimate calculated accurately L _σ by injection as a value, or a high-frequency current that is determined by the motor model.

2 is an d-axis equivalent circuit diagram of an induction motor.

As shown in Fig. 2, in the d-axis equivalent circuit of the induction motor, the high frequency component of the stator current hardly flows into the excitation inductance, and most of it flows toward the rotor. Under this assumption, a voltage equation such as the following equation (15) holds.

In Equation 15, L _ls denotes a stator leakage inductance, and L _lr denotes a rotor leakage inductance.

In general, the following equation (16) holds.

In Equation 15, since the frequency of the injected high frequency current is very large, the voltage drop of the leakage inductance is much larger than that of the resistor. Therefore, even if the resistance value changes to some extent, it does not significantly affect the estimation of L _σ .

If both sides of Equation 15 are summed up and squared, the following Equation 17 is obtained.

By extracting the DC distribution using the lowpass filter, L _σ can be estimated as shown in Equation 18 below.

There was no significant difference when comparing L _σ estimated by the method as described above with the measurement of the motor constant. In addition, this method uses only basic voltage and current values, and since the measured value itself is large, there is an advantage that it can be estimated online relatively accurately.

It is now described to estimate speed information for speed control.

The estimation of the velocity can be conceptually obtained as the derivative of the angle, as shown in Equation 19 below. In general, the implementation of the derivative is difficult due to the sensitivity to noise. However, since the speed of the motor cannot change instantaneously, stable speed information can be obtained by using an appropriate lowpass filter.

In Equation 19, ω _e represents a synchronous rotational angular velocity.

When the control period for calculating the angle is small, Equation 19 may be converted into a discrete model, as shown in Equation 20 below.

In Equation 20, T _samp is a sampling period of the current controller.

The speed of the rotor is given by the output frequency minus the slip frequency.

On the other hand, the slip in the vector control is given by the following equation (21).

In Equation 21, ω _sl represents a slip angular velocity.

Although the d-axis rotor flux is estimated in the present invention, the magnitude of this estimate does not exactly match the actual magnetic flux at low speeds of 1/10 or less of the rated speed, so it is inappropriate to calculate slip using this flux estimate. . Therefore, the slip estimate was used as the slip frequency directly on the assumption that the current control is precisely performed without converting the rated exciting current. The slip command value is shown in the following equation (22), and the estimated speed is shown in the following equation (23).

Example

3 is a block diagram of an experimental apparatus for performing a direct vector control method of an induction motor according to the present invention. As shown in FIG. 3, an experimental apparatus for performing a direct vector control method of an induction motor according to the present invention is largely composed of a control board, an inverter, a motor, and a load generator. The controller basically consists of current controller, flux estimator, torque ripple estimator and vector controller, and additionally includes speed controller and L _σ estimator.

A high performance current controller is essential to accurately control high frequency currents of hundreds of Hz. As a current controller, a PI controller of a synchronous coordinate system including forward electromotive force compensation is used. The controller has superior stability and control performance and is easier to implement than a hysteresis controller or a predictive controller. The control period of the current controller was PWM with a switching frequency of 5 kHz with 100 μsec. The space vector method was used as the PWM method. The space vector method has the advantages of greater usable voltage, better transient characteristics, and smaller current ripple than the triangular wave comparison method. The vector controller is also composed of a simple PI controller, and the speed controller is also a PI controller. All these controllers have the advantage of being easy to modify and simple to implement in software.

The control microprocessor uses TI's TMS320C30 to configure all the controllers in software. The specifications of this processor are as follows.

Ultra fast operation (60㎱ single cycle instruction) 16.7MIPS, 33.3MFLPOS

Built-in 4K × 32-bit high speed ROM

Built-in 2K x 32-bit high speed RAM

32-bit data bus, 24-bit address bus

Built-in 40 / 32-bit floating point / integer multiplier

Built-in DMA serial I / O, 32-bit timer

The control board is equipped with 32K × 32bit EPROM and 256K × 32bit high speed SRAM, 12bit A / D converter for reading phase current and DC link voltage, D / A converter for outputting internal variables and user's There is an RS232C serial port that communicates with the IBM PC to receive setpoint changes. All digital circuits use EPLD (Erasable Programmable Logic Device) to simplify the structure and increase the reliability.

As the inverter, voltage type IGBT inverter using 600V 75A IPM module was used. In addition to the three pairs of switching elements, the IPM module provides a regenerative braking switch, an internal gating board, and even a fault signal in case of an accident.

The motor used was a 5-horsepower 4-pole three-phase squirrel cage induction motor, and a direct current motor was used as a load. Two induction motors were used. The first induction motor was a test set equipped with a dynamo motor and used in the forward and reverse rotation experiments. The second induction motor was a set with a direct current generator connected as a frequency device. It was used for the control characteristic experiment for.

Specifications of the two motors used in the embodiment of the present invention are shown in Table 1 below.

1st induction motor 2nd induction motor 5hp, 220V, 4poles 50Hz 5hp, 220V, 4poles 60Hz R _s : 0.55Ω R _r : 0.3787ΩL _ls : 2.1mH L _lr : 2.1mH L _m : 59mH R _s : 0.27Ω R _r : 0.223Ω L _ls : 1.3mH L _lr : 1.3mH L _m : 33mH

First embodiment

Forward and reverse rotation experiments were performed at no load, half load, and full load at 30 rpm and 50 rpm for the first induction motor, and the correlation between torque and torque component current at the stationary state and various speed ranges was also investigated. Experimental results are shown in FIGS. 4 to 9.

Figure 4 shows the forward and reverse waveform at 30 rpm no load at -30 rpm, Figure 5 is the forward and reverse waveform at 30 rpm no load at -30 rpm, Figure 6 is the forward and reverse waveform at 30 rpm half load at -30 rpm, Figure 7 at -50 rpm A forward rotation waveform at 50 rpm half load, FIG. 8 shows a forward rotation waveform at 30 rpm full load at -30 rpm, and FIG. 9 shows a forward rotation waveform at 50 rpm full load at -50 rpm, respectively. In each of Figures 4-9, Is the waveform of the estimated angle, Is the waveform of estimated velocity, ω _rpm is the waveform of actual velocity, Denotes the waveform of the control variable, respectively. From the results shown in Figures 4 to 9, it can be seen that the characteristics controlled by the method according to the present invention are very excellent, and that the speed control and the vector control are stably performed under any load conditions.

Fig. 10 is a diagram showing a correlation between torque current and torque. From the correlation between torque component current and torque, we can see the linearity of the output torque with respect to torque component current, which is the basic characteristic of vector control. The closer the torque and torque component current is to the straight line passing through the origin, the more perfect the vector control is. It is indicated. 10, it can be seen that the control method according to the present invention operates relatively well at various speed ranges.

Second embodiment

The operating characteristics at the low speed of 1/100 of the rated speed and the operating characteristics at the general speed were investigated for the second induction motor. In addition, the control characteristics of the load variation during normal operation were tested.

Fig. 11 shows the results of the forward and backward waveforms at -18 rpm no load at 18 rpm, and Fig. 12 shows the results of the forward and backward waveforms at -18 rpm half load at 18 rpm.

Fig. 13 shows the results of the forward and backward waveforms at -30 rpm no load at 30 rpm, and Fig. 14 shows the results of the forward and backward waveforms at -30 rpm half load at 30 rpm.

Fig. 15 shows the results of the forward and backward waveforms at -500 rpm no load at 500 rpm, and Fig. 16 shows the results of the forward and backward waveforms at -500 rpm full load at 500 rpm.

In each of Figs. 11-16, Is the waveform of the estimated angle, Denotes a waveform of estimated velocity, ω _rpm denotes a waveform of actual velocity, and i _as denotes a waveform of phase current, respectively.

It can be seen from the results shown in FIGS. 11 to 16 that the control method according to the present invention is very robust to load variation and has good characteristics in various speed ranges.

Fig. 17 shows the 3/4 load characteristic of the rating at 30 rpm constant speed operation. Denotes a waveform of estimated speed, ω _rpm denotes a waveform of actual velocity, I _qs ^{e *} denotes a waveform of torque-minute current, and I _qs ^e denotes a waveform of torque-minute current command value.

As shown in Fig. 17, the characteristic of the load fluctuation during the constant speed operation was about 10 rpm for the 3/4 load of the rating, and returned to the original speed after about 1 second.

In the control method according to the present invention, it is impossible to control 0 rpm at no load, but since the control is not divergent, the original control performance may be exhibited when the control speed is brought back into the controllable speed range.

As described above, the method according to the present invention enables the vector control by controlling in the direction of minimizing the ripple of the torque generated by the high-frequency current injection, from which it is possible to estimate the speed. Since the method according to the present invention does not require accurate magnetic flux information, vector control and speed control are smoothly performed even at a low speed with an inverter output frequency of about 0.5 Hz. It has the advantage of being very robust against fluctuations in motor constants.

Claims (5)

In the magnetic flux reference vector control method of an induction motor, Applying a high frequency sinusoidal current to the d-axis of the synchronous coordinate system; Obtaining a torque ripple generated by the current applied in the step; And And changing the control angle in a direction of reducing the torque ripple obtained in the step. The induction motor according to claim 1, wherein the step of changing the control angle in the direction of reducing the torque ripple is performed by reducing Δθ, which is an error between the control angle obtained from the torque ripple and the magnetic flux angle of the actual motor. Magnetic flux reference vector control method. The method of controlling a magnetic flux reference vector of an induction motor according to claim 1, wherein the step of obtaining the torque ripple is performed by estimating the magnetic flux of the induction motor to obtain the torque and then obtaining the torque ripple through a high pass filter. . The method of controlling a magnetic flux reference vector of an induction motor according to claim 3, wherein the step of estimating the magnetic flux of the induction motor is performed using a first delay filter equal to the time constant and the rotor time constant of the induction motor. . The induction motor according to claim 3, wherein the step of estimating the magnetic flux of the induction motor uses a method of adding a magnetic flux command value through a low pass filter and adding the measured value to improve the characteristics of the magnetic flux estimation at a low speed. Control method of magnetic flux reference vector.