Multidimensional Risk Management for Underground Electricity Networks

• Main Authors: Garcez Thalles V., Szufel Przemyslaw
• Format: Article
• Language: English
• Published: Sciendo 2014-08-01
• Series: Studies in Logic, Grammar and Rhetoric
• Subjects: cascade failure; network decision making; underground electricity grids
• Online Access: https://doi.org/10.2478/slgr-2014-0017

In the paper we consider an electricity provider company that makes decision on allocating resources on electric network maintenance. The investments decrease malfunction rate of network nodes. An accidental event (explosion, fire, etc.) or a malfunctioning on underground system can have various consequences and in different perspectives, such as deaths and injuries of pedestrians, fires in nearby locations, disturbances in the flow of vehicular traffic, loss to the company image, operating and financial losses, etc. For this reason it is necessary to apply an approach of the risk management that considers the multidimensional view of the consequences. Furthermore an analysis of decision making should consider network dependencies between the nodes of the electricity distribution system. In the paper we propose the use of the simulation to assess the network effects (such as the increase of the probability of other accidental event and the occurrence of blackouts of the dependent nodes) in the multidimensional risk assessment in electricity grid. The analyzed effects include node overloading due to malfunction of adjacent nodes and blackouts that take place where there is temporarily no path in the grid between the power plant and a node. The simulation results show that network effects have crucial role for decisions in the network maintenance – outcomes of decisions to repair a particular node in the network can have significant influence on performance of other nodes. However, those dependencies are non-linear. The effects of network connectivity (number of connections between nodes) on its multidimensional performance assessment depend heavily on the overloading effect level. The simulation results do not depend on network type structure (random or small world) – however simulation outcomes for random networks have shown higher variance compared to small-world networks.