Adaptive Machine Learning Framework for Harmonic Reduction in Cascaded H-Bridge Multilevel Inverters

The growing use of renewable energy systems, electric vehicles, and smart power applications has increased the demand for efficient and intelligent inverter control. Cascaded H-bridge multilevel inverters (CHB-MLIs) offer improved output quality, but conventional control methods often face challenges under varying load conditions. This paper presents an artificial intelligence (AI)-based control strategy for a CHB-MLI using convolutional neural networks (CNNs), k-nearest neighbors (KNN), and recurrent neural networks (RNNs). CNN is used for waveform feature analysis, KNN for load classification, and RNN for fault prediction and system behavior analysis. These AI techniques enable adaptive pulse-width modulation (PWM), improving inverter performance under different operating conditions. Simulation and experimental results demonstrate reduced total harmonic distortion (THD), improved voltage regulation, and enhanced system reliability compared with conventional control methods. The proposed AI-assisted approach provides an effective solution for multilevel inverter applications in renewable energy systems, smart grids, electric vehicles, and industrial power conversion.