A Distributed, Model-Based Approach to Cyber-Security: Application to Wind Farms

In the U.S. Department of Energy’s report, ‘20% Wind Energy by 2030’, mitigating the risk associated with owning and operating wind turbines was identified as one of wind energy’s key challenges to overcome in order to promote industry growth. Similar to other large power infrastructure systems, one of the most important aspects of risk associated with wind power is its vulnerability to cyber-attacks. As wind power becomes increasingly integrated with the smart-grid, cyber-attacks pose a threat not only for the immediate physical damage they may cause to turbines, but also threaten to cause serious economic damage from power blackouts due to instability in the grid. This paper presents a distributed, model-based intrusion detection system (IDS) algorithm that has the ability to identify the presence of certain parameter manipulating cyber-attacks within a wind farm. The algorithm draws upon existing IDS schemes such as reputation scoring and collaborative nodes, but is unique in that leverages application-layer insight gained from understanding the interaction between wind speed dynamics within a wind farm and wind turbine supervisory control. Properties from Denmark’s Horns Rev wind farm were used to help develop a credible simulation environment for the algorithm’s testing.