Electricity Network Reliability Optimization

Electricity distribution networks are subject to random faults. On occurrence of a fault in the distribution network, the circuit breaker on the faulty feeder trips, which disconnects supply to all customers connected to that feeder. Once the fault is located, the faulty component is isolated and then repaired. The cost of isolating and repairing faults depends on the way the network is configured, and the size and location of its components, since a major part of this cost comes from compensating customers if a fault in the network causes the interruption duration to be greater than a given threshold duration. We describe a model developed in collaboration with a local lines company for minimizing the cost of reliability of distribution networks by switch reconfiguration. Three local search heuristics to reconfigure radial distribution networks are investigated. For each component, the time between failures, the isolation duration, and the repair duration are modelled using Weibull distributions. The model allows the company to reconfigure their current radial distribution networks to minimize their total expected cost of reliability, as well as providing a tool for investigating capital investment scenarios.