Data-Driven Approaches for Multi-Party Resources Management in Energy Internet Paradigm

The information and communication technology continue shifting the modern power system into the ìEnergy Internetî paradigm. Various multi-party resources are controlled by edge-cloud computational resources with information exchange among them via communication system. The architectures, functions, algorithms, and validation tools towards this future vision need to be further investigated. This panel session will discuss the data-driven approach for real-time control, operation, analysis, and testing of modern power systems with multi-party resources. This panel session has several presentations with recent advances in this topic. The data-driven stability assessment and fault diagnosis for power systems by using advanced machine learning will be introduced. The deep reinforcement learning as a data-driven control approach is applied to perform online approximate optimal control in inverter-interfaced grid environment. The co-simulation techniques and platform to realize data-driven control and analysis approaches will also be presented. Presentations in this panel session: - Consensus Multi-agent Reinforcement Learning for Decentralized Volt-VAR control in power distribution systems (23PESGM4202) - Emerging power systems controls and the challenge of trusted validation (23PESGM4360)