State relevance and modal analysis in electrical microgrids with 100 grid-forming converters

In traditional power systems, dominated by syn-chronous generators, the separation between time scales is well-known due to the physical response of the generating units. However, the fast-growing deployment of renewable energy resources (RERs) is bringing an increasing number of generating units based on fast-acting electronic power converters to modern power systems. Nowadays, having an islanded portion of the grid with 100 electronic generation, at least temporarily, is not unthinkable and, in this case, time-scale separation among dynamics would depend on control design and would not be straightforward. Therefore, the classical model-reduction approach neglecting the dynamics of fast-varying state variables is, at least, questionable. This paper proposes a method to identify relevant states in a modern power system in order to have an accurate input-output response description in any reduced-order model. The method is based on the modal analysis of a balanced realisation of the system (i.e., a linear transformation in which the transformed states’ energies in the output response are known). The proposed method is illustrated on a microgrid with 100 grid-forming electronic power converters. Simulation results show that the proposed method can identify the system’s relevant states even in cases with unclear time-scale separation.