Model reference adaptive control of servodrives

Interest is focused on model reference adaptive systems (MRAS) which represent an interesting group of adaptive systems for electric drives and their requirements. The basic adaptive model following algorithm is presented. In this case the adaptive mechanism has a PI-character with expressed product nonlinearities. It is necessary to simplify the adaptive mechanism in order to realize the adaptive control of plants with fast dynamics within the available means. For this purpose a modified adaptive algorithm called the unit vector adaptive law, found in the works of Balestrino et al. (1982) is presented. Finally the two presented modified adaptive algorithms are used in the position control of DC servosystems where great changes in the moment of inertia are common. A combined control structure will be used for the needs of positioning. The classical control loop is used for current and position control, and speed control is performed by means of model following control. The advantages of adaptive control over the classical fixed cascade servodrive control are shown.