Dynamic equivalent of a real distribution grid hosting photovoltaic and synchronous generators

This paper presents a methodology to derive a reduced-order, “grey-box” model of an active distribution network. This dynamic equivalent is intended for dynamic simulations of large disturbances taking place in the transmission system. To deal with the uncertainty affecting dynamic model parameters, Monte-Carlo simulations are used, and the parameters of the equivalent are adjusted to match as close as possible the average randomized system responses. To avoid over-fitting, multiple disturbances are considered; they are automatically selected among a set of candidates. Moreover, to reduce the computational burden, only parameters with significant impact are adjusted in the identification procedure. Simulation results are reported on a real Australian distribution grid. The latter hosts synchronous generators and residential photovoltaic units. Its loads are modelled with a static and a motor part.