An intelligent adaptive control of DC–DC buck converters

Abstract

Buck DC–DC converter is used in many applications to supply a fixed amount of DC voltage. They are highly sensitive to the frequently changing loading conditions. Such a situation demands a robust control mechanism which can guarantee satisfactory performance of the buck converter over a widely changing load. This can be made possible by developing an adaptive control scheme which can estimate the true values of the uncertain load parameters in the least possible time. This paper proposes an adaptive Chebyshev neural network (CNN) based backstepping control technique for the output voltage regulation of a DC–DC buck converter. The proposed control strategy utilizes neural networks in approximating the unknown non-linear nature of load resistance by using orthogonal basis Chebyshev polynomials. CNN approximation tool in conjunction with the conventional backstepping procedure yields a robust control mechanism. The weights of neural network are tuned online using adaptive laws satisfying the overall closed loop stability criterion in the Lyapunov sense. The performance of the proposed control is demonstrated for wide range perturbations by subjecting the buck converter to changes in load resistance, input voltage and reference output voltage. Simulation studies are conducted to evaluate the performance of the proposed controller against radial basis function neural network based adaptive backstepping control and conventional adaptive backstepping. The results obtained are further verified from experimentation on a hardware setup using DSP based TM320F240 processor. Thus, the investigation confirms effectiveness of the proposed control scheme as the output voltage shows a fast and accurate response besides successfully rejecting the disturbances acting upon it.