Asynchronous machines - Direct torque control

Due to its low cost and robustness, the asynchronous machine is currently the most widely used machine for variable speed drives. The squirrel-cage asynchronous machine is structurally inferior to DC and synchronous machines. The fact that the machine is powered by a single armature means that the same current creates both flux and torque, so that variations in torque cause variations in flux. This type of coupling gives the asynchronous machine a completely non-linear model, which makes control of this machine very complex. A great deal of work has gone into developing efficient controls for the asynchronous cage machine.

Scalar control [ ], although well suited to certain types of drive, does not deliver very high performance, especially at low speeds and high torques. It is not at all suitable for positioning asynchronous machines.

Vector control by rotor flux orientation [ ] was developed to eliminate the machine's internal coupling, causing flux variations linked to torque variations. A great deal of work has been carried out on vector control, and numerous drives with this control are being produced and used for a wide range of applications, in a wide variety of power and speed ranges. However, although it gives high performance to the asynchronous machine, vector control by orientation of the rotor flux has a number of drawbacks:

• low robustness to parametric variations and in particular to those of the rotor time constant;

• the need for a modulator for close inverter control, which causes delays, especially at low modulation frequencies (high power). These delays are responsible for an increase in torque response time, which penalizes inverters used in traction applications;

• presence of coordinate transformations dependent on an estimated angle ;

• the speed of rotation is explicitly included in the control algorithm. When this speed is not measured (i.e. when the drive has no speed sensor), errors in the speed estimate degrade drive performance.

The use of flux observers reduced the control's sensitivity to variations in the rotor time constant, but did not cancel it out. In addition, rotor time constant estimation structures have been developed to compensate for the effects of rotor time constant variations. More and more drives are being designed without speed sensors, to reduce their negative effects: prohibitive cost, fragility... When the speed sensor is replaced by an estimator or observer, performance is degraded at very low speeds (cancellation of the stator pulsation), which corresponds to a critical zone due to the loss of speed observability....