Higher Order Sliding Mode Based Switched Reluctance Motor Control System Design
This thesis presents a novel scheme for speed regulation/tracking of Switched Reluctance (SR) motors based on Higher-Order Sliding-Mode technique. In particular, a Second-Order Sliding-Mode Controller (SOSMC) based on Super Twisting algorithm is developed. Owing to the peculiar structural properties of SRM, torque produced by each motor phase is a function of phase current as well as rotor position. More importantly, unlike many other motors the polarity of the phase torque in SR motors is solely determined by the rotor position and is independent of the polarity of the applied voltage or phase current.
SR motor needs an electronic commutation scheme for its operation. So design of commutation scheme plays an important role in motor efficiency and performance. This commutation scheme is embedded in its power supply as switching timers. The existing commutation schemes cause high power loss and based on those commutation schemes, the existing controller techniques for SR motor show low robustness especially when motor’s parameters change. Therefore a new commutation scheme is developed which optimizes power consumption in motor phases. On the bases of this commutation scheme, a new controller technique is used to design controller for SR motor which is highly efficient, simple to design and easy to implement and also provides sufficient robustness against parameter variations and unknown disturbances. The proposed controllers take advantage of this property and incorporate a commutation scheme which, at any time instant, selects to energize only those motor phases for the computation of control law, which can contribute torque of the desired polarity at that instant. This feature helps in achieving the desired speed regulation/tracking objective in a power efficient manner as control efforts are applied through selective phases and phases producing the torque of opposite polarity are kept switched off. This approach also minimizes the power loss in the motor windings reducing the heat generation within the motor.
The common techniques for designing the SR Motor controls are fuzzy logic control, Artificial Neural Network (ANN) and feedback linearization. Fuzzy logic control provides sufficient robustness against parameter variations but at a high computational cost. Artificial Neural Network (ANN) shows good dynamic response against unknown disturbances but problem in using this technique is requirement of a large training data set. In the feedback linearization technique, nonlinear control problem is transformed into linear control problem and then any one of the well established and mature linear controller techniques are applied on the resulting system. Feedback linearization cannot be applied to all types of nonlinear systems; and in case of parameters uncertainties, the robustness cannot be guaranteed. All the deficiencies in discussed techniques can be overcome by introducing sliding mode control which is simple, easy to implement and provides robustness. The inherent problem of chattering in classical FOSMC can further be improved by using higher order sliding mode control (HOSMC).
In order to highlight the advantages of Higher-Order Sliding-Mode controller, a classical First-Order Sliding-Mode controller (FOSMC) is also developed and applied to the same system. The comparison of the two schemes shows much reduced chattering and low power consumption in case of SOSMC. This feature is especially very important for SR motor control, due to reduced chattering; wear and tear problem of actuators is reduced. The responses of synthesized controllers are also investigated against changes in moment of inertia which could be due to engagement of load; stator phase resistance which could vary due to temperature variations in winding during operation and coefficient of viscous friction as a model uncertainty. The performance of the proposed SOSMC controller for speed regulation is also compared with that of another sliding mode speed controller published in the literature and also with dynamic sliding mode controller. The same technique is also applied on position control problem and, FOSMC and SOSMC are developed for position regulation problem; making it possible candidate for servo drive application.