In this study, an enhanced hybrid architecture for the shunt active power filter (SAPF) based on the synchronous reference method (SRF), is proposed. The aim of this study is to develop a unified control strategy that integrates high-precision harmonic current extraction and robust DC-link voltage regulation within electrical grids characterised by highly non-linear loads and sudden variations in operating conditions. The approach’s originality lies in the integration of adaptive Laguerre functions coupled with a neural architecture to simultaneously manage reference current extraction and DC-link regulation. In contrast to the conventional approach, which utilises second-order low-pass filters and PI controllers, the developed method exhibits enhanced resilience to sudden fluctuations in load. The performance of the system was validated in the Matlab/Simulink numerical environment using two scenarios of sudden network variation. The findings of the study provide substantial evidence to support the hypothesis that the proposed strategy is indeed effective. In fact, it has reduced total harmonic distortion (THD) by 23.8%, representing a significant improvement on the conventional SRF-PI method. Additionally, a notable enhancement in DC bus stability has been observed, with an RMSE reduction of up to 60.6%. Analysis using the Hilbert transform and the root mean square (RMS) confirms the excellent quality of the energy produced, establishing this solution as a high-performance alternative for future smart grids.
Pesdjock et al. (Mon,) studied this question.