Design of Resource Agents with Guaranteed Tracking Properties for Real-Time Control of Electrical Grids

Abstract

We target the problem of controlling electrical microgrids with little inertia in real time. We consider a central controller and a number of resources, where each resource is either a load, a generator, or a combination thereof, like a battery. The controller periodically computes power setpoints for the resources based on the estimated state of the grid and an overall objective, and subject to safety constraints. Each resource is augmented with a resource agent that a) implements the setpoint requests sent by the controller on the resource, and b) translates device-specific information about the resource into a device-independent representation and transmits this to the controller. We focus on the resource agents and their impact on the overall system's behavior. Intuitively, for the system to converge to the objective, the resource agents should be obedient to the requests from the controller, in the sense that the actually implemented setpoint should be close to the requested setpoint, at least on average. This can be important especially when a controller that performs continuous optimization is used (for the sake of performance) to control discrete resources (which have a discrete set of implementable setpoints). We formalize obedience by defining the notion of $c$-bounded accumulated-error. We then demonstrate its usefulness, by presenting theoretical results (for a simple scenario) and simulation results (for a more realistic setting) that indicate that, if all resource agents in the system have bounded accumulated-error, the closed-loop system converges on average to the objective. Finally, we show how to design resource agents that provably have bounded accumulated-error for various types of resources, such as resources with uncertainty (e.g., PV panels) and resources with a discrete set of implementable setpoints (e.g., on-off heating systems).