By Gianluigi Forte and Andrea Spampinato, STMicroelectronics
The three-phase induction motor is one of the most reliable types of electric machine. Induction motors are known to work for many years with very little maintenance effort. They also offer great operational flexibility.
Today, induction motors are the industrial sector’s most widely used electric machine, and so are responsible for a very high proportion of the industrial sector’s total electricity consumption. This means that improvements to the energy efficiency of induction motor systems resulting from a reduction in energy losses will have enormous benefits, both in cutting operating expenses and in supporting compliance with efficiency regulations.
This has led to a growth in the adoption of variable-speed drive technology in preference to fixed-speed drives. STMicroelectronics provides a complete solution for controlling a variable-speed induction motor using either scalar or vector controls. This Design Note describes how an efficient variable-speed drive design may be developed quickly on the basis of a combination of ST boards which implement control and power functions.
ST Boards for Induction Motor Control
The proposed solution can be evaluated by assembling a system composed of the following:
- A NUCLEO-F303RE control board based on the STM32F303RE, a 32-bit microcontroller which includes an Arm® Cortex®-M4F processor core.
- An STEVAL-IPM10B power board based on an STGIB10CH60TS-L second-generation SLLIMM™ Intelligent Power Module (IPM). It is an easy-to-use demonstration board for driving electric motors up to 1.2kW supplied by a 125V to 400V DC bus voltage. The board is provided with bootstrap and snubber capacitors, short-circuit protection, a fault event signal, and temperature monitoring.
- A motor control connector expansion board, the X-NUCLEO-IHM09M1.
With the architecture shown in Figure 1, it is possible to assemble a full inverter system which is simple, cheap and flexible, and fits the requirements of the chosen application in terms of computational and electrical power, using the appropriate STM32 microcontroller and IPM.
In addition, the NUCLEO board supports the STM32CubeMX system, which provides a full array of expansion elements for functions such as sensors and communication channels. It also provides a graphical configuration tool and project generator, and enables the user to set up peripherals in just a few steps.
Induction Motors: Control Techniques
An induction motor can provide torque only if the frequency of the three-phase stator voltages and currents, ωe, is higher than the electrical shaft rotation frequency, ωre. This difference is called the slip frequency, ωslip, and the value of its normalization with respect to ωe is the slip.
The motor itself can increase or decrease the torque in response to changes in the mechanical load, respectively decreasing or increasing the shaft speed in a slip range of around 20%. Using an inverter, it is possible to change the frequency of ωe, greatly reducing the start-up current.
It is possible to change the rotor speed by varying the synchronous frequency, ωe. If the voltage amplitude remained the same, the electromagnetic flux would change and saturation problems could occur. But by maintaining a constant ratio between the stator voltage amplitude and ωe, the electromagnetic flux remains constant, a technique known as V/f control.
Assuming the load is constant, this method allows the rotor speed to change at a constant slip, minimizing power losses. Since this is an open-loop control technique, for a fixed value of ωe an increase or a decrease in the mechanical load will cause a variation of the rotor speed.
By implementing a closed-loop version of this technique, the developer can add the ability to control the motor’s speed, while retaining the combination of low cost and low dynamic performance offered by the open-loop version. A speed sensor must be used to vary the slip frequency according to the actual rotor speed and to the mechanical load.
The only MCU peripherals required for these techniques are a timer for generating the six PWM signals, and a DAC for debugging. The closed-loop technique also requires a timer for decoding the speed/position sensor output.
An alternative to either of the above techniques, Field-Oriented Control (FOC) is an advanced control method which achieves high efficiency and excellent dynamic control by using a simple estimation of the rotor flux position. The indirect FOC method estimates the motor flux by using rotor speed information from a speed sensor, and the electrical rotor time constant, τr. It requires a speed sensor even if speed control is not needed.
Fig. 1: Block diagram of induction motor system based on ST power and control boards
In an induction motor, both the magnetization fi eld and the stator field are provided by the stator windings, so it is more difficult to control them independently than it is in a DC motor. FOC enables such control, but requires continuous information about the rotor flux position.
In fact, FOC is based on the co-ordinate transformation theory, which transforms vectors from a 120°-abc (uvw) reference frame to a 90°-qd; the angle of the rotor flux is used in these transformations. But unlike a permanent magnet synchronous motor, in an induction motor the rotor flux angle does not coincide with the shaft electrical angle because of the slip frequency. This means that when FOC is implemented in an induction motor, regardless of the speed control loop, the rotor flux angle must be known.
In practice, to save cost the rotor flux angle is normally estimated rather than measured. Once the angle is known, the currents in the stator, Iqd, will control respectively the electromagnetic torque and the magnetizing flux.
To perform this type of FOC, the STM32 MCU only requires two or three ADCs for measuring the motor phase currents, one ADC for measuring the DC bus voltage, one PWM timer to generate the gate commands, and if needed, a timer for decoding the shaft speed sensor output.
The indirect FOC technique estimates the rotor flux angle by using the rotor speed information from a speed sensor. The technique is ‘indirect’ since the flux vector is not directly estimated but only its momentary position. The speed sensor enables the indirect FOC technique to work at zero speed.
Sensor-less FOC off ers benefits in terms of both reliability and cost saving. There are several techniques for estimating rotor flux and shaft speed: the ST solution is a Model Reference Adaptive System (MRAS) observer which can work from low to very high speed, but not when the motor is stationary.
Demonstrated Performance of the ST System
The performance of an induction motor running ST’s control algorithms on an STM32F303 MCU board was tested on a three-phase induction motor with the following specifications:
- 1.9Arms nominal current
- 380Vrms nominal voltage
- 50Hz frequency
- 750W nominal power
- 2,650rpm maximum speed
Figure 2 shows the speed step response from standstill to 2,500rpm when the closed-loop V/f control technique is in action. Although no mechanical load is applied to the shaft, the graph shows that, as expected, the response time is longer than when using the FOC technique.
Fig. 2: Speed step response under no load from 0rpm to 2,500rpm using closed-loop V/f control
Nevertheless, this closed-loop control technique is suitable for any application that requires a cheap and simple implementation and that can tolerate relatively low dynamic performance.
If cost is the main factor, indirect FOC should be chosen instead of closed-loop V/f control. The indirect FOC technique is little more complex than closed-loop V/f control: both require a speed sensor, and the algorithm is ready to use once the STM32 peripherals are configured to match the topology of the hardware sensing network. What is more, the dynamic performance and efficiency of the indirect FOC technique are clearly better than that of a scalar method such as the closed-loop control scheme, as shown in Figure 3.
What is true for the indirect FOC method is even more true for sensor-less FOC, as shown in Figure 4. The main limitation of this method is that it does not operate when the motor is stationary.
Fig. 3: Speed step response under a 1Nm load from 1,000rpm to 2,500rpm using indirect FOC control
Fig. 4: Response time when reversing under sensor-less FOC control from 2,300rpm to -2,300rpm