Dynamic equivalent of an active distribution network taking into account model uncertainties

A reduced-order model of a distribution network hosting inverter-based generators is derived. This model aims to be used in dynamic simulations of large disturbances in the transmission system. Uncertainties affecting the parameters of the component models are addressed using Monte-Carlo simulations. For a given disturbance, the average and the standard deviation of the randomized system evolutions are extracted. The dynamic equivalent is of the “grey-box” type. Its parameters are adjusted to match as closely as possible the average system response. This involves minimizing a weighted least-square objective function, in which the weights reflect the dispersion of the randomized Monte-Carlo responses. Multiple disturbances are considered to avoid over-fitting, and a recursive procedure is used to select a small subset of them from which the parameters of the ADN equivalent are identified. Simulations results are reported on the CIGRE MV distribution network benchmark. They show a good accuracy of the equivalent.