This paper is motivated by major needs for fast and accurate on-line data analysis tools in the emerging electric energy systems, due to the recent penetration of distributed green energy, distributed intelligence, and plug-in electric vehicles. Instead of taking the traditional complex physical model based approach, this paper proposes a data-driven method, leading to an effective early event detection approach for the smart grid. Our contributions are: (1) introducing the early event detection problem, (2) providing a novel method for power systems data analysis (PowerScope), i.e. finding hidden power flow features which are mutually independent, (3) proposing a learning approach for early event detection and identification based on PowerScope. Although a machine learning approach is adopted, our approach does account for physical constraints to enhance performance. By using the proposed early event detection method, we are able to obtain an event detector with high accuracy but much smaller detection time when comparing to physical model based approach. Such result shows the potential for sustainable grid services through real-time data analysis and control.
Keywords
- Power systems
- Smart grid
- Early event detection
- Data mining
- Nonparametric method
- Machine learning