Cloud–Edge Collaborative Flexibility Assessment of Virtual Power Plants Considering Communication and Device Response Delays

Abstract

Virtual power plants (VPPs) aggregate geographically distributed flexible loads to provide balancing services for renewable-rich power systems. However, conventional flexibility assessments mainly consider physical regulation capability and may overestimate the power that can be delivered within a specified response-time budget when communication latency, route failures, and device response delays are neglected. This paper proposes a communication- and device-delay-aware flexibility assessment method for VPPs under a cloud–edge collaborative architecture. Wired and wireless transmission delays, cloud and edge computing latency, device response characteristics, and communication reliability are jointly modeled. The k-shortest paths (KSP) algorithm determines feasible primary and backup communication routes, while proximal policy optimization (PPO) coordinates edge-node and base-station selections to maximize the dispatchable flexibility subject to response-time constraints. Case studies on an IEEE 33-bus system with residential, commercial, and industrial flexible loads demonstrate stable convergence and accurate online assessment. Under the studied settings, neglecting communication effects overestimates the available flexibility by 20.3%, with the deviation being most pronounced under tight response-time budgets. The proposed method provides a communication-aware basis for reliable VPP aggregation and regulation.