A Profit-Oriented Attack by Virtual Power Plants for Economic Risk Transfer: Profitability Analysis and Defense Strategy

Abstract

Virtual power plants (VPP) may face the risk of economic losses arising from the deviation between day-ahead schedule and real-time energy delivery. This article introduces a novel profit-oriented attack by some greedy VPP (G-VPP) for the first time, which aims to transfer its risk of economic losses to other normal VPPs (N-VPPs) and obtain a risk-free profit. Particularly, the G-VPP launches the load-altering attack (LAA) on N-VPPs to fail their real-time (RT) generation schedule via modifying the generation instructions of the managed resources. The optimal LAA strategy considering stealthiness is proposed to avoid priori detection and the optimal bidding strategy using CVaR is proposed to quantify bidding risk and maximize profit. To defend against such profit-oriented attacks, this article constitutes the first attack traceability research. A defense scheme with attack detection and attacker identification is proposed, which is executed by VPPs and the independent system operator (ISO). VPPs employ an attack detection algorithm based on complex features, and if any, report it to the ISO. The ISO constructs a knowledge graph of electricity market transactions and embeds a rules-based identification algorithm to identify suspected attackers. Simulation results show the effectiveness of the defense scheme and Neo4j is used to visualize.