Testbedding of AI-Powered Resilient Dual-Wireless Network Solutions for Virtual Power Plant

Virtual Power Plants (VPPs) are pivotal in integrating distributed energy resources (DERs) into power grids, yet they demand ultra-reliable, low-latency communications (URLLC) to function effectively. This paper presents a testbed implementation of a resilient VPP communication architecture using dual wireless networks enhanced with Artificial Intelligent (AI)-powered redundancy management. A real-world setup was developed using a Battery Energy Storage System (BESS), commercial off-the-shelf (COTS) wireless equipment for building the dual private cellular networks based on Openairinterface (OAI) library, and a User Equipment (UE) emulator to create practical scenarios. The AI-powered resilient network solution features Machine Learning (ML)-based prediction and hedging mechanisms that ensure high reliability and low latency communication. We demonstrate that the proposed design significantly enhances communication reliability and latency performance, confirming the feasibility of AI-powered redundancy management in future wireless-enabled VPPs.