Predictive Control With Parallelized Models Using Inductance Updating Mechanism for Split-Phase T-Type Inverters

The split-phase T-type (SPT) inverter features a symmetrical two-phase bridge structure that generates two sinusoidal waveforms with a 180${^\boldsymbol{\circ}}$ phase shift. However, this symmetry is often disrupted by time-varying circuit parameters, such as inductance variations, which degrade control performance. To address this challenge, this paper proposes a predictive current control with parallelized models using inductance updating mechanism (PCC-PM-IUM). By decoupling the inverter phases and constructing a parallel model, the proposed method enables independent and precise control of each phase leg, ensuring stable inverter operation under both balanced and unbalanced load conditions. Furthermore, the linear extended state observer (LESO) is improved by reconstructing the ultralocal model to estimate the lumped disturbance and inductance value. These estimates are fed back to the controller during each computation cycle, enhancing robustness against parameter mismatches and ensuring high-quality waveform output under complex operating conditions. The effectiveness of the proposed method is validated through experiments under various load conditions.