PERFORMANCE ENHANCEMENT OF QUADRATIC BUCK– BOOST DC–DC CONVERTER USING ANN CONTROL

Abstract

This study presents the design and MATLAB/Simulink-based performance evaluation of a quadratic buck-boost DC–DC converter governed initially by a conventional Proportional– Integral (PI) regulator, and subsequently by an Artificial Neural Network (ANN) controller. The converter topology affords continuous input and output currents, rendering it well-suited to renewable-energy systems and electric-vehicle applications that demand stable voltage regulation amid varying input and load conditions. In the first instance, the system was simulated under the PI controller to govern the output voltage in both buck and boost regimes. Thereafter, the PI regulator was supplanted by an ANN controller, trained on input– output data derived from the converter’s dynamic response. The ANN controller adaptively captured the system’s nonlinear characteristics, yielding improved transient behaviour, diminished steady-state error, and enhanced voltage stability. Comparative simulation results demonstrate that the ANN-controlled converter attains a higher output voltage in boost mode and sustains a lower but stable voltage in buck mode, with smoother transitions and reduced overshoot. These outcomes indicate that the ANN-based strategy out-performs the conventional PI approach, offering superior robustness, more rapid convergence and improved efficiency under dynamic operating conditions. Consequently, the proposed ANN control technique enhances the overall performance of the quadratic buck-boost converter for advanced power-electronic applications