LDO Optimization with Evolutionary Neural Network

As grid-connected Photovoltaic (PV) based inverters are being used more, these systems play a more important role in the electricity generation by distributed power generators. Power injection to the grid needs to meet predefined standards.In order to meet the harmonics requirement of standards, they need an output filter. The connection through an LCL filter offers certain advantages, but it also brings the disadvantage of having a resonance frequency.LCL filter can easily help the system to satisfy these requirements but also introduce a resonance peak which makes the system control a challenging task. In this paper, a three-level Neutral Point Clamped (NPC) inverter is connected to the grid through an LCL filter. The injected current of the inverter is controlled using Proportional-Resonant (PR) controllers. The resonant peak of the filter is also damped using capacitor current feedback. A systematic mathematical design procedure for controller and filter capacitor current feedback coefficients is investigated in details. Simulations are carried out in MATLAB/Simulink environment and results depict suitable performance of the system with designed parameters

LDO Optimization with Evolutionary Neural Network Keywords:

LCL filter , active damping , PR Controller , NPC Inverter