Balanced Model Order Reduction Techniques Applied to Grid-tied Inverters In a Microgrid

Abstract

Multiple model order reduction strategies applied to grid-tied inverters in a microgrid setup are assessed in the proposed work. The full-order mathematical model of a grid-tied inverter is first derived for this purpose. The nonlinear equations of a grid-tied inverter system are derived in the synchronous reference frame and then linearized around an operating point to construct a 15 state small-signal model. The 15th order state full order model is then reduced using balanced truncation, balanced residualization, and balanced stochastic truncation. Each of these model reduction techniques results in a 6th order reduced model. Simulation results obtained using PLECS are used to compare the accuracy and equivalency of the reduced-order models to that of the one found using the full-order model. After comparing the results, the balanced residualization method is found to be the most suitable method for the given system.