Hammerstein Models and Real-Time System Identification of Load Dynamics for Voltage Management

Abstract

Distributed generators, controllable appliances, electric vehicle charging infrastructures, and energy storage systems introduce new technical challenges to the management of distribution networks, especially when there are large power fluctuations. Interactive dynamics between load and distributed generators in a distribution network carry significant impact on voltage variation and transient during load power disturbances. It is shown in this paper that the traditional static power flow analysis, in which load dynamic behavior is not counted, is not sufficient to model and predict voltage excursion after a power disturbance. To capture the behavior of load types and dynamics, this paper employs Hammerstein model structures to represent such behavior and explore their real-time identification. This is especially important for voltage quality management since the load dynamics depend on active and reactive load power, and hence change substantially due to load/generator power perturbations, electric vehicle charging activities, and subsystem load type varieties. Identification algorithms are introduced and their convergence properties are established. The algorithms are applied to a generic grid structure first, then evaluated on a 33-Bus system with multiple dynamic loads.