Intelligent switching planning models for fault location and isolation in MV-feeders

The control center operator can be supported by an intelligent computer system to inference the switchings necessary to locate and isolate a fault in a distribution network. The paper proposes two different inference models combining the heuristic knowledge of operators and analytical methods (e.g. a decision tree) in detail for the problem considered. The focus is on the manually operated switchings coordinated by the operator. The main contribution of the models is the allowance for and effective use of the outage costs of customers, the likelihood of the fault location, and the moving times of the working crew in solving the switching planning problem. The inference models have been implemented in a support system for distribution network operation which is being tested in one distribution utility.