Measurement-based Modeling of Smart Grid Dynamics: A Digital Twin Approach

Abstract

The renewable energy resources have paved the way for distributed energy resources (DERs) integration in to the distribution grid. As a result, the load composition and their dynamics have become complex. The weather phenomena and new emerging consumer load patterns like electric vehicle contribute to time varying dynamics of these loads. In order to optimize the utilization of system assets and flexibility of DERs, the identification of time varying load dynamics is necessary. In this paper, the identification of time varying load dynamics is explored by combining system identification methods and nonlinear numerical optimization. The identified model parameters are then related to measurement data by means of artificial neural networks, which enables the identification of similar dynamics without opting to numerical optimization methods.