Distribution-level Optimal Power Flow with Fast DER Control

The large-scale integration of distributed energy resources (DERs) with controllable smart inverters, coupled with the need for improved efficiency, calls for mechanisms to optimize the electric distribution grid operations by managing its controllable assets. This has motivated the development of optimal power flow methods (OPF) for power distribution systems that aim at optimizing a pre-specified network-level objective-- subject to the distribution grid's operating constraints. The highly intermittent and variable nature of the DER power generation profile, for example, those observed from distributed PVs, make it challenging to achieve network-level optimization with the current optimal power flow (OPF) methods that are typically open-loop and are slow in response. Managing the rapidly varying phenomena calls for new approaches with a fast response that can simultaneously provide the voltage control while also optimizing for the network-level objectives. This panel will present recent advances in fast DER control methods that address the challenge of optimizing the grid operations while being mindful of the distribution grid constraints. Specifically, the talks will focus on innovations in modeling and analytical approaches that have led to faster control of DERs ranging from purely local algorithms to distributed real-time feedback-based online control algorithms and advanced machine learning techniques for OPF. The panel will conclude with the lessons learned, remaining challenges, and the directions for future research.