Reduced-order harmonic modeling and analysis of droop-controlled distributed generation networks

This work describes the development of a reduced-order nonlinear dynamic phasor modeling tool for distribution networks containing a large number of P-f and Q-v droop-controlled inverters. Attention is focused on constructing reduced-order models that accurately predict the nonlinear behavior of coherent clusters of droop-controlled inverters in response to a large disturbance, such as an asymmetric fault. In such a scenario, dynamic phasor models are shown to offer valuable insights into how the signal conditioning associated with local power measurement can have a significant effect on the large-signal stability of a droop-controlled inverter or inverter cluster. An algorithm conventionally used to identify coherent groups of synchronous generators is adopted for the investigation of coherency between large numbers of droop-controlled inverters within a distribution network. The coherency results help guide the aggregation of droop-controlled inverters and enable nonlinear reduced-order dynamic phasor modeling. For large disturbances, a reduced-order dynamic phasor simulation of a 50-inverter system is validated against full-order simulation results obtained using an OPAL-RT power system simulator.