High-performance general-purpose inverter

1 Limitations of general frequency converters

When a general-purpose inverter is used to power an asynchronous motor, it can achieve stepless and smooth speed regulation, and starting and stopping are very convenient. However, there is a static error during speed regulation, the accuracy is not high, the speed regulation range is only about 1:10, and it cannot provide high dynamic performance like a DC speed regulation system.
2 Control strategy of high-performance general-purpose inverter

To achieve high dynamic performance, the physical model and dynamic mathematical model of the motor must be fully studied. Currently, the commonly used high-performance control strategies are vector control and direct torque control.

The characteristics of the vector control system are: using the dq synchronous rotating coordinate system whose d-axis direction is determined by the rotor flux, decomposing the stator current of the asynchronous motor into its excitation component and torque component, obtaining a torque model similar to that of a DC motor, and then taking measures to transform the nonlinear system into two independent speed and rotor flux subsystems, thereby imitating the DC motor to be controlled by PI regulators respectively. When a high-precision photoelectric encoder speed sensor is selected, the speed regulation range of the vector control system can reach 1:1000, and the dynamic performance is also very good. However, the orientation according to the rotor flux will be distorted by the change of motor parameters, thereby reducing the speed regulation performance of the system. The use of an intelligent regulator can overcome this shortcoming and improve the robustness of the system.

The direct torque control system abandons the more complex rotating coordinate transformation, and only forms the feedback signals of torque and stator flux in the two-phase stationary coordinate system, and uses two-position bang-bang control instead of linear regulator to control torque and stator flux, and selects the voltage space vector PWM (SVPWM) switching state according to the changes of the two to control the speed of the motor. This system has a simple control structure, fast torque response, and avoids the influence of rotor parameter changes. However, bang-bang control will cause pulsation of output torque, affecting the low-speed performance of the system.

From a theoretical perspective, both vector control systems and direct torque control systems are controlled based on the dynamic mathematical model of asynchronous motors. The asynchronous motor in the two-phase coordinate system has a 4th-order voltage equation and a 1st-order motion equation. Its state equation should be 5th-order, and 5 state variables must be selected. In the dynamic model of the system, the input variables are Usd, Usq, ω1, and TL. For cage rotor motors, the rotor is short-circuited, and Urd=Urq=0. Therefore, there are 9 state variables available for selection, namely, the speed ω, 4 current variables isd, isq, ird, irq, and 4 flux variables ψsd, ψsq, ψrd, and ψrq. The rotor currents ird and irq are unmeasurable and should not be used as state variables. Only the sub-currents isd, isq and the rotor flux ψrd, ψrq can be selected, or the sub-currents isd, isq and the stator flux can be selected. In other words, there can be two types of state equations: ω-ψr-is and ω-ψs-is. Vector control uses the ω-ψr-is equation, while direct torque control uses the ω-ψs-is equation.

From the overall control structure, both use torque and flux control separately, and the torque control loop (or the torque component loop of the current) is in the inner loop of the speed loop, which can suppress the influence of flux changes on the speed subsystem, thereby achieving approximate decoupling of the speed and flux subsystems. Therefore, both systems can achieve higher static and dynamic performance.

However, due to the difference in specific control schemes, the two have their own advantages in control performance. The vector control system adopts rotor flux orientation, thus realizing the decoupling of the stator current torque component and the flux component. The speed and flux regulators can be designed separately according to the linear system theory (generally using PI regulators) to implement continuous control, thereby obtaining a wider speed regulation range; but the orientation according to ψr is affected by the change of the motor rotor parameters, which reduces the robustness of the control system. The direct torque control system implements Te and ψs bang-bang control, avoiding the rotation coordinate transformation and simplifying the control structure; controlling the stator flux instead of the rotor flux, is not affected by the change of the rotor parameters; the bang-bang control itself belongs to P control, which can obtain a faster dynamic response than the PI regulator (due to the lack of a current inner loop, it is necessary to limit the maximum impact current); but it inevitably generates torque pulsation, and the flux voltage model with an integral link has poor accuracy at low speeds, which limits the low-speed performance of the system.

3 Application and development of vector control and direct torque control

Both the vector control system and the direct torque control system are high-performance AC speed control systems that have been widely used in practice. Due to their respective characteristics, they have different focuses in the application field. In addition to general speed regulation, vector control is more suitable for wide-range speed regulation systems and servo systems, while direct torque control is more suitable for large inertia motion control systems (such as electric locomotives) that require fast torque response. In view of the fact that both control strategies still have some shortcomings, the research and development of both systems are moving towards overcoming their shortcomings.

Further research on vector control systems is mainly aimed at improving the robustness of their control. For a long time, people have naturally thought of using adaptive control to solve the impact of rotor parameter changes on the accuracy of rotor flux orientation, but the research results have rarely been put into practical application. Most of the use is temperature compensation for rotor resistance changes. Modern intelligent control methods can make the controlled system independent of or less dependent on the mathematical model of the control object, so that the vector control system is not affected or less affected by changes in motor parameters. A more convenient way is to use an adaptive PID controller composed of a single neuron.

Research on direct torque control systems focuses on improving their low-speed performance. In the early 1990s, the EAEE research laboratory of Ruhr University in Germany, under the leadership of Professor Depenbrock and Professor Steimel, proposed an indirect self-control ISR (IndirektSelbstregelung in German) system as an improvement plan for direct torque control systems. Among them, the bang-bang controller is changed to a continuous regulator, and the PI regulator is used to perform closed-loop control on the stator flux amplitude to establish a circular stator flux trajectory. The angle Δθ corresponding to the flux vector increment is calculated based on the deviation of the electromagnetic torque. Finally, the stator voltage vector is calculated according to the output of the flux and torque regulators to obtain the corresponding inverter switch state. It can be seen that the ISR system is actually a fusion and compromise of the direct torque control system and the vector control system. In addition, there are many improvement plans proposed based on the bang-bang controller of direct torque control, such as the refinement of flux deviation and torque deviation, the beat-free modulation of the voltage space vector, the predictive control and intelligent control of the switch state, and the predictive tracking control of torque or flux alone.

Any improvement to the control strategy requires an increase in the burden on the control software, and therefore requires an improvement in the hardware capabilities. Nowadays, high-performance inverters often use high-end single-chip microcomputers or digital signal processors DSP.

4Speed ​​sensorless control

Whether it is a vector control system or a direct torque control system, speed closed-loop control is required. The required speed feedback signal comes from a speed sensor coaxial with the motor. For high-performance systems, photoelectric encoders are generally used, which have problems with cost, installation, and reliability. If the photoelectric encoder can be eliminated while maintaining good control performance, it will obviously be very popular. This is a high-performance speed control system without a speed sensor. As a high-performance general-purpose inverter, it is hoped that speed sensorless control will be adopted.

At this time, the speed can be indirectly obtained through the easily measured stator voltage and current signals. Common methods are:
(1) using the motor model to derive the speed equation to calculate the speed;

(2) Use the motor model to calculate the slip frequency and perform compensation;

(3) According to the model reference adaptive control theory, select the appropriate reference model and adjustable model, and identify the speed and rotor flux at the same time; Information from: Power Transmission and Distribution Equipment Network

(4) Use other identification or estimation methods to obtain the speed;
(5) Use the motor's tooth harmonic potential to calculate the speed; etc.
However, no matter which method is used, the calculation or identification accuracy is limited, and the accuracy of the dynamic speed is even more limited. Therefore, the current practical speed sensorless speed control system can only achieve general dynamic performance, and its high-precision speed control range of 10 is considered good. At present, several varieties of speed sensorless high-performance general-purpose inverters have been launched, but research work is still ongoing.