Neural Network Integrated Adaptive Backstepping Control of DC-DC Boost Converter

Abstract

This paper deals with the output voltage regulation problem of dc-dc boost converter feeding a resistive load. A new control mechanism based on Chebyshev neural network embedded in an adaptive backstepping framework is proposed for the boost converter control. Since the converter is complex, time varying and non-linear in nature, it exhibits high sensitivity to unanticipated disturbances in the load current. Hence, designing a robust control mechanism to attain a satisfactory transient and steady state performance over a wide range of operating points is a challenging task. In this work, a control law is derived based on the systematic and recursive design strategy of adaptive backstepping method. A single layer functional link Chebyshev neural network is employed for a fast estimation of uncertain and time varying load profile of the boost converter. The stability of overall converter equipped with the proposed controller is proved using Lyapunov stability criterion. Further, in order to validate the proposed methodology, the boost converter is simulated in MATLAB/Simulink software and is subjected to different load perturbations. The efficacy of the proposed control is highlighted by evaluating it against the conventional adaptive backstepping control under identical conditions. The results obtained reveals that the proposed control is much faster in estimating the unknown load parameter and offers satisfactory output voltage tracking, yielding fast response and low peak overshoot/undershoot in the event of unknown load perturbations. Experimental investigation using dspace DS1103 controller is further carried out to validate the efficacy of proposed control scheme.