The scale and complexity of modern power grids have increased significantly over the years. The once centralized control structure has paved way for a more distributed control framework with multiple operators. The adoption of standard Internet Protocol based communication and data access systems have increased the openness of the network. It has enabled enterprises to access the data information network for data sharing. Reliable operations of the power grid require grid operators to constantly monitor and control the state estimation at each bus. Problems such as load fluctuations, power supply disruptions, transmission line outages need to be detected quickly and accurately to take remedial actions. Hence, the integrity of the power system state estimation process is paramount. However, this integrity is coming under immense pressure due the explosion in the number of entry points into the grid metering system due to the openness of the grid. Each new entry point provides cyber-attackers with an opportunity to inject malicious data into the metering system and compromise the integrity of the state estimation. These attacks are very hard to detect as the state estimation system can no longer be relied upon. Thus, they can lead to power outages or electricity market manipulation causing huge losses to the grid operator.
The focus of this research is to protect the grid against data spoofing attacks which can manipulate the electricity market and cause economic losses to the grid operator. Due to the size of the grid, protecting each meter is not feasible. Hence, we research focuses on developing data-driven models to identify the buses whose state estimation is more critical than others in terms of their effect on the electricity market.