Abstract

The use of Internet-of-Things (IoT) technology in electric power distribution is growing, specially with smart meters, automated controls, demand response (DR), and renewable distributed energy resource (DER) integration. These advancements have made it easier to monitor, control, and optimize DER-rich cyber-physical distribution systems (CPDS). However, as IoT devices become more widespread, they also increase the risk of cyber-attacks, which makes ensuring system resiliency critical. This work presents a novel resiliency-driven load restoration technique leveraging demand response through IoT devices to enhance the resiliency of a three-phase unbalanced power distribution systems. A hierarchical optimization framework is proposed, featuring a continuous linear program for primary optimization and a binary linear program for secondary optimization, aimed at achieving load-source energy balance and connection of secondary-level house/building loads based on their criticality. To address the limitations of traditional theoretical methods in analyzing distribution system losses across multiple devices and configurations, this paper proposes a novel deep neural network-based approach for system loss calculation. Validation of the proposed methodology is conducted using the IEEE 123-node system modeled in the HELICS co-simulation testbed, to demonstrate optimally configuring switch status, supplying critical loads, balancing load generation, and maintaining feeder voltages.