AI-driven power quality optimization in LCL-based D-Statcom supported distribution networks

Abstract In low and medium voltage distribution networks, the LCL integrated conventional converter based distributed static compensator (D-Statcom) has shown to be a practical solution for shunt compensation. Despite numerous efforts in this area, the traditional control approach still has a number of issues. This article describes the development of LCL integrated D-Statcom for shunt compensation utilizing a deep learning technique. The voltage source converter (VSC) and LCL filter are included in the new framework. The operation and control of the distribution network are directly impacted by the proposed system’s performance. MATLAB Simulink software and an experimental research based on d-SPACE are used to demonstrate the synchronization, which solves current related power quality (PQ) issues such as poor power factor (p.f.), current harmonics, unbalanced voltage at point of common coupling (PCC) and poor voltage regulation. In order to provide precise reference currents for control, the deep learning technique is used to monitor the essential active and reactive components of load currents. Furthermore, it precisely ascertains the remaining constituents, attaining a swifter transient reaction and superior system stability. Comparisons are then made between VSC and LCL integrated VSC using deep learning technique by considering the implementation procedure. With a lower DC-link voltage and a smaller converter power rating, the LCL integrated VSC system improves PQ of distribution network more than the VSC.