This paper describes the application of the Fractional Order PIs (FOPI) in the speed loop of a high performance induction motor electrical drive. In particular the speed tracking and load rejection capability of FOPI controller has been investigated and compared with both an integer-order PI and an IP both in simulation and experimentally with constant settling time. Illustrative study proves the simplicity and efficiency of the presented design method over integer controllers.

No abstract

No abstract

This paper describes a new method for dynamical estimation of load disturbance in induction motors by using Nonlinear Unknown Input Observers (NUIO). This estimation is then used to compensate dynamically the load torque in a Field Oriented Control (FOC) induction motor drive to increase its load-rejection capability. The method has been verified both in simulation and experimentally on a experimental rig. I. INTRODUCTIONHigh-performance electrical drives represent the backbone of manufacturing and transportation industry. Consequently big attention has been paid over the best control strategies to obtain speed tracking and load rejection capability. With this respect, one of the key factors that can affect heavily the dynamical performance of an electrical drive is the presence of load disturbance on the shaft. A first way to make the system robust to their effects is to predict such disturbance in order to obtain better rejection capabilities, thus resulting in improved performance via proper dynamical compensation. This problem has been addressed in the case of induction motor drives in [1], where load torque is estimated by using the mechanical equation and assuming load torque has a slow dynamics. Other subsequent works [2][3][4] have followed the same approach. Another strategy that has been considred is to use dynamical filters or observers, in particular the Kalman Filters or the Unscented Kalman Filters [5][6][7], but at the cost of increasing complexity and requiring assumptions on the dynamical equation of the load.This work follows on the second kind of approach, but it addresses the load torque estimation problem in an innovative way, by using the theory of Nonlinear Unknown Input Observer (NUIO) [8],[9], whereby the system states are reconstructed without the knowledge of the unknown disturbance input, such as the load torque.The theory of observers originated from the work of Luenberger in 1964 [11]. According to Luenberger, any system driven by the output and the knowledge of the systems state space can serve as an observer for that particular system. Wang later used Luenberger approach to design UIO by estimating the both the states of the system and extracting the known inputs [12]. Later on, several approaches for designing linear UIO for linearized systems were proposed using different techniques [13]-[15].However from the last two decades, research has shifted from linear to nonlinear system observers. Earlier, class of non-linear Lipschitz design was based on the linear part of the system by imposing certain conditions on the nonlinearity. However, this approach inherited drawbacks that were related to observer convergence conditions which is considered difficult to satisfy for large value of Lipschitz constant [16]. One way to get around this is the use of a NUIO which makes use of the differential mean value theorem (DMVT) together with Linear Matrix Inequalities (LMI) [17], which however

The Institution of Engineering and Technology is registered as a Charity in England & Wales (no. 211014) and Scotland (no. SC038698).