AC motor - torque control
I. Introduction
Several relatively common direct torque control strategies, for small and medium capacity, the control program focuses on the torque, flux linkage control and improve the carrier frequency. For large capacity, the difference is that the low speed using indirect torque control, so as to achieve low speed torque pulsation purposes.
Second, the direct torque control technology overview
Compared with the DC motor in the simple structure, easy maintenance, low environmental requirements and energy saving and improve productivity and so has sufficient advantages, making the exchange speed has been widely used in industrial and agricultural production, transportation, defense and daily life. With the power electronics technology, microelectronics technology, the rapid development of control theory, AC speed control technology has also been developed by leaps and bounds. In the field of high-performance AC speed control, there are two kinds of vector control and direct torque control. In 1968, Dr Hasse, of Darmstader University of Engineering, initially proposed Field orientation theory, and then gradually improved in 1971 and formed various vector control methods.
Third, the characteristics
For direct torque control, the general literature that it by the German Ruhr University Professor M.Depenbrock and Japan's I. Takahashi in 1985 were first proposed. For the direct torque control of the circular circle, the basic idea is to accurately observe the spatial position and size of the stator flux and maintain its amplitude constant and accurately calculate the load torque conditions, by controlling the motor instantaneous Input voltage to control the motor stator flux instantaneous rotation speed, to change its instantaneous slip of the rotor, to achieve direct control of the motor output purposes. In the control of thinking and vector control is different from the direct torque control by direct control of torque and flux to indirectly control the current, does not require complex coordinate transformation, it has a simple structure, fast torque response and good robustness to the parameters Etc.
Fourth, control
In fact, in 1977, a direct flux linkage and torque adjustment method similar to that of direct torque control was proposed in which the difference between the torque reference and the feedback is adjusted by the PI to obtain the slip frequency. Slip frequency plus the motor rotor mechanical speed to get the inverter should output the voltage stator frequency; stator flux reference and feedback difference through the integral operation to get a voltage and frequency ratio, and make it with the stator frequency multiplication Get the inverter should output the voltage, and finally through the SPWM method to control the motor.
Direct torque control has been proposed for nearly 20 years. On this basis, a variety of control strategies and their digital implementation schemes, flux observation and velocity identification have been developed. They are classified and analyzed. Comparison.
Five, direct torque control strategy
Direct torque control is based on the still coordinate system to control, as shown in Figure 1, in the traditional direct torque control, by detecting the stator two-phase current, DC bus voltage and motor speed (no speed sensor DTC is not Need to speed) to calculate the stator flux and torque calculation, so that the two are given with the stator flux and torque given subtraction, the difference and respectively through the respective hysteresis compared to the output torque and flux Of the increase and decrease the signal, the two signals into the optimization vector switch table, coupled with the stator flux where the sector has been to meet the flux for the circular, torque output followed by a given torque vector torque. Flux and torque hysteresis can be set multi-stage, and its width is variable, the hysteresis width is smaller, the higher the switching frequency, the more accurate control.
Six, direct torque control method
Direct torque control has the advantages of simple structure, fast torque response and good robustness to the parameters. But it is based on Bang-Bang control of single vector, torque and flux hysteresis. Inevitably caused by low-speed switching frequency is low, the switching frequency is not fixed and the torque ripple, limiting the direct torque control in the low-speed area applications. In view of this, many scholars at home and abroad put forward a variety of ways to improve the switching frequency, fixed switching frequency and reduce the torque ripple, this section will be listed one by one analysis and comparison.
Seven, space vector modulation method
T.G.Habetler's Space Vector Modulation Method
The application of the no-difference method to direct torque control is first proposed by American T.G.Habetler. The main idea of this method is to get the difference between the given value and the feedback value of the torque during this sampling period.
The magnitude and phase of the space voltage vector are arbitrary and can be synthesized by two adjacent fundamental voltage vectors. Using the calculated space voltage vector can achieve the purpose of torque and fluxless shooting.
Using Habetler's no-difference method, it is theoretically possible to completely make the flux linkage and torque error zero, thus eliminating the torque ripple, can make up the traditional DTC Bang-Bang control of the lack of the motor can run at very low speed. In addition, the spatial voltage vector obtained by the no-beat control can increase the switching frequency and make it relatively large with respect to a single vector, which is helpful for reducing the voltage harmonics and the motor noise.
However, the space voltage vector action time may be greater than the sampling period, which shows that the flux linkage and torque can not meet the beat control. So the author put forward three steps, first whether the torque to meet the no-beat, if not satisfied to see if the flux to meet the no-beat, if not satisfied in accordance with the original direct torque control vector table to select the next cycle Single voltage vector. Therefore, according to Habetler's worst-count method, there are four steps, which will cost a lot of computing resources, and it is not easy to achieve. In addition, the dependence on the motor parameters is relatively large throughout the calculation process, which will reduce the control Sticky.
8, torque or flux control method
In the direct torque control method of T · G · Habetler's no-beat, it is difficult to realize because of the large amount of computation. Therefore, a series of simplified
AC motor - South Korea SPG AC motor full range
AC motor - South Korea SPG AC motor full range
Moment control, the more typical is the torque tracking prediction method. In this method, the low-speed torque ripple is analyzed, and the conclusion that the torque ripple is asymmetric is obtained.
The nonzero voltage vector and the zero voltage vector have different effects on the torque change, and the former can cause the torque to rise or fall while the latter always causes the torque to drop. In addition, the rate of change of the torque effect is also varied in the different speed ranges. In the torque prediction control method, the voltage vector in the space position is fixed, synthesized in the middle of two single voltage vector, but the voltage vector is not the role of the entire sampling cycle, but a certain duty cycle, in a The sampling period can be divided into nonzero voltage vector and zero voltage vector. If the next sampling period, the torque variation caused by the nonzero voltage vector and the zero voltage vector is equal to the torque error calculated in this cycle.
Will eliminate the torque error, to achieve the purpose of torque no beat control. Even if the calculated voltage vector action time exceeds the sampling period, can also be replaced with full voltage vector, it is very easy to achieve, from the experimental results, the torque ripple is basically symmetrical, indicating that the pulse of the torque has been Greatly reduced. On the law that the flux chain is accurately controlled or slow changes, but did not consider the flux linkage of the beatless control, in the literature on the flux also carried out predictive control.
Nine, predictive control
In this method, the relationship between the spatial vector and the voltage vector of the flux can be approximated:
Where ΔΨS is the magnitude of the flux linkage amplitude under the action of the voltage vector, and θVΨ is the spatial angle of the two. Assuming that the flux error of the k-th sampling period is ΔΨSO, then the vector action time can be obtained according to the formula (5). With the principle of torque control priority, the vector time can be obtained by the vector action time calculated by the torque prediction control and the action time calculated by the flux prediction control. Considering the lossless beat control of the flux linkage, it is effective to eliminate the torque ripple without generating the flux linkage distortion, and the calculation amount is not too large. In addition to the above torqueless beat control method, a similar method is used in the literature, and the final voltage vector calculation time is basically the same, not described here. As with Habetler's no-beat method, the prediction method also uses more motor parameters. If the stator resistance and rotor time constant can be identified in real time online, the control accuracy will be greatly improved.
Ten, discrete time direct torque control
Discrete-time direct torque control The use of discrete-time methods for asynchronous motor control has been described in more detail in the literature. In the literature, this method was first used in direct torque control. The method is as follows: The voltage equation and the flux linkage equation obtained from the basic circuit model of the motor are discretized as follows:
The definition of a and b is also discretized for the torque equation, and the equation ⑺ is substituted into it, and the equation ⑺ is substituted into the amplitude square expression of the flux linkage. Using the discrete torque equation and the discrete flux linkage The amplitude-level method can solve the incremental ΔVSx and ΔVSy of the space voltage vector of the next cycle, and obtain the voltage vector of the torque and fluxless beat control and replace it with the following equation:
Discrete-time direct torque control can be through the difference equation, the k + 1 cycle should be reached by the torque and the magnetic chain recursive, so you can achieve both torque and flux linkage control, from the implementation of the way Very suitable for digital control, and this method is mainly based on the stator side of the control, the required motor parameters only stator resistance and inductance, the motor parameters of the robustness of the better, from the experimental results, the system's dynamic response Performance is better. However, in this method, the need to detect the motor phase voltage, which increases the complexity of the system hardware, in addition, the calculation is also relatively large.
Eleven, geometric graphics of the no-beat control
In the literature, the stator flux equation, the rotor flux equation and the torque equation expressed by the stator and rotor flux are discretized, and then the first two equations are brought into the torque equation. By analyzing the discrete torque equation, it can be seen that applying a voltage vector can make the torque error zero and the torque changes to a straight line on the plane, which is parallel to the direction of the rotor flux vector. The same method can be used to know that the applied voltage vector can make the flux error zero, the flux changes to a plane on the plane, which is concentric with the magnetic chain. So that the intersection of the straight line and the garden can be used to get the torque and flux linkage control of the voltage vector, of course, the voltage vector by the inverter can output the voltage limit.
The introduction of geometric graphics into the control of no-beat is a better idea, can get the optimal voltage control vector, but also contribute to the theoretical analysis. But on how to graphics and digital control from the realization of the way or there is a certain degree of difficulty.
Discrete space vector modulation method
Direct torque control of the no-beat can theoretically maximize the elimination of torque and flux errors, overcoming the weaknesses of Bang-Bang control inaccuracies, but requires relatively large computational costs, and these calculations are And the motor parameters, easy to cause the calculation of the error. Therefore, a discrete space vector modulation method is proposed in the literature which does not require much computation and can improve the accuracy of torque and flux control.
In the discrete space vector modulation method, the adjacent voltage vector and the zero voltage vector in the six basic voltage vectors output by the two-level inverter are regularly synthesized. As shown in Fig. 3, the adjacent single vector 2 And a single vector 3 and zero voltage vector synthesis of the space voltage vector. It can be seen from Figure 3 that the synthesis method is to divide the entire sampling period into three segments, each of which consists of nonzero voltage vector or zero voltage vector. The spatial voltage vector 23Z is composed of vector 2 and vector 3 and zero voltage vector The use of 1/3 sampling cycle, you can use 5 or 7-stage synthesis (not described), the use of this regular synthesis method can be synthesized a total of 10 voltage vector.
The refined voltage vector can be used to control the torque and flux more precisely. The traditional 2-stage hysteresis Bang-Bang control is used in the literature, and the torque is required to be fast 5-level hysteresis Bang-Bang control, as shown in Figure 4, different error bands using different voltage vector table. In addition, the author derives the influence of the voltage vector on the torque change as follows:
From the type ⑽ can be seen in the same voltage vector in the low speed and high speed on the torque changes are different. Thus, different voltage vectors are used in different speed ranges, as shown in Fig. On the other hand, low-speed use of small voltage vector and high-speed amplitude of the large voltage vector is also in line with the V / f = C this law. The traditional direct torque control at low speed continuous use of more zero voltage vector to make the switching frequency is very low, torque ripple. In the case of the discrete space vector modulation method, since the voltage vector with a small amplitude is used at a low speed, the zero voltage vector is used continuously, the switching frequency is high, and the torque ripple is small. In addition, due to the high voltage signal at high speed, you can divide 12 sectors, the use of two voltage vector table, which can be more precise control.
It can be seen from the above analysis that the discrete space vector modulation method is easy to implement, and it is not necessary to have the same calculation as the no-beat control, and keep the advantages of the traditional Bang-Bang control, so the robustness is good, but compared with the traditional direct Torque control can also improve torque and flux control accuracy, reduce low-speed torque ripple. However, the more accurate the control accuracy, the finer the vector division, the greater the voltage vector control table, which will increase the complexity of the control. Therefore, if you can make discrete space vector modulation and no difference beat control together, will help to overcome this shortcoming.
Thirteen, the output space voltage vector method
In the direct torque control, if the space voltage vector of arbitrary phase can be obtained, it will help to reduce the torque ripple at low speed and achieve steady-state performance of vector control at low speed. Section 3 in the no-beat control can be arbitrary phase of the space voltage vector, but the calculation is more complex and difficult to achieve. Another way to obtain a spatial voltage vector of arbitrary phase is to use a PI regulator. A · B · Plunkett direct torque and flux adjustment method is a PI adjustment method, but then there is no space voltage vector concept, only the use of SPWM method to output motor control voltage. In the literature, the proposed direct torque control uses the PI adjustment method, and the method for SVM outputs the spatial voltage vector.
By the torque reference and torque feedback to obtain the torque error input PI regulator, through the PI adjustment to get the q-axis voltage vector, given by the stator flux linkage and stator flux feedback to obtain the stator flux error input PI regulator, The d-axis voltage vector is obtained by PI adjustment, and then the voltage vector of the d-axis and q-axis is rotated to the α-axis and β under the stationary coordinate system for the output of the space voltage vector. It is clear that the space voltage vector is in the spatial position The phase is arbitrary. Structural control based on PI direct torque control similar to the stator flux orientation vector control