Reliability modeling and data analytics in cyber-enabled power distribution systems

Abstract

The twenty-first century power system is witnessing major transformation. It is becoming smarter to meet the demand with higher efficiency and reliability. The smart grid calls for ubiquitous deployment of advanced monitoring and automated control equipment known as cyber-enabled devices which highly rely on an information and communication network. This transformation means the grid should be considered cyber-physical system (CPS), where secure interactions between the two networks (power and cyber) are critical for the optimal operation of the grid. This dissertation focuses on the interaction of cyber and power components and investigates the impact of communication malfunction on power grid outage management and energy management systems. It focuses on developing an analytical reliability model for fault detection, isolation and service restoration (FDISR) for smart distribution feeders. The impact of end-to-end outage probability of data communication along with sending, receiving, and relaying communication node failures is incorporated into the model. Vulnerability of system to cyber attack as an emerging cause of reliability degradation is also investigated. An optimal placement of fault detectors (FDs) and switching devices are determined in this work to improve reliability. Short-term prediction of multivariate dynamical processes evolving over time when data are partially observable is a challenging task. In smart grid, data are being collected from multiple sources. Short-term future is predicted based on real-time observations available up to the current time step. This dissertation also addresses the impact of partially observable measurements on short-term time-series prediction.