AI-based Intelligent Control for High-Efficiency DC-DC Converters in Advanced Applications

Abstract

Compact design, fast regulation, and low-loss operation at high efficiency make direct current-to-direct current (DC-DC) conversion increasingly important for renewable energy interfaces, electric vehicle (EV) subsystems, microgrids, storage units, and Internet of Things (IoT) devices. This work aims to bridge the gap between conventional converter operation and intelligent control under varying source and load conditions by implementing an artificial intelligence (AI)-based control structure with buck and boost converter topologies. The main purpose is to enhance the performance of voltage regulation, tracking accuracy, and conversion efficiency while decreasing ripple, dynamic error, and switching-induced performance limitations. Closed-loop converter operation includes voltage and current sensing, signal conditioning, intelligent duty-cycle regulation, pulse width modulation (PWM) generation, and gate driving all of which are implemented in a numerical simulation platform using the control architecture described. It shows that the voltage gain across the boost converter is 2 over an input of 24 V and a load current of 1 A, delivering an output power of 48 W at an efficiency squeeze of 96% along with maintaining a tracking error in voltage of approximately 0.8 V, duty cycle updates are common, which soon relates to a required updated value where constant effort remains on the duty cycle to lock it into the control stage, hence obtaining an updated duty cycle value as nearly around 0.538. This performance is evident in the buck converter waveforms, which maintain well-regulated output near 12 V, and the ripple response from the boost converter settles at around 48 V after transient fluctuation—ideal for energy management and smart power conversion applications.