The project, led by Professor Gian Paolo Cimellaro and Professor Stephen Mahin, will develop a model to assess the vulnerability of power networks to cascading failures.
While greater interconnectedness of power networks has led to improvements in operational efficiency, it has also increased their vulnerability to cascading failure, a phenomenon where loss of a single component or piece of equipment can, through a complex web of interactions and dependencies, trigger a sequence of additional failures that ultimately leads to a large power outage. Several recent (2003-2006) major blackouts in the United States and Europe were caused by cascading failures, resulting in billions of dollars of losses. Therefore, a vital component in improving the reliability of power networks involves protecting against the threat of cascading failures.
Towards this end, a hybrid temporal network model will be developed that integrates several advanced modeling techniques (including dynamic fault and event trees, Bayesian networks, and fuzzy sets) to more fully characterize the vulnerability of power networks to cascading failures. The model will enable electrical operators and regulators to identify high-risk assets and critical failure sequences within a network, investigate the impact of different network configurations and/or mitigation strategies on reliability, and monitor network performance in near real-time to detect conditions that can trigger cascading failures.
Gian Paolo Cimellaro is Associate Professor at the Politecnico di Torino and currently Visiting Professor at UC Berkeley. Dr. Cimellaro’s current research interests address community resilience to natural disasters, risk mitigation of civil infrastructures failures including interdependencies, using advanced technologies such as smart phones.
Stephen Mahin is the Byron and Elvira Nishkian Professor of Structural Engineering at UC Berkeley, and is the Director of PEER. His research focuses on improving understanding of the seismic behavior of systems by integrating high performance numerical and experimental simulation methods to improve seismic performance of systems.
The Siebel Energy Institute is an international consortium dedicated to accelerating and sharing advancements in machine learning applied to power systems.