Generating fault rate surfaces using network fault statistics and geographic information

Abstract

This paper applies logistic regression to network fault statistics and geographic information. The logistic regression model is further processed in order to obtain raster formatted fault rate surfaces that aid network planning, e.g., in automated routines or as background maps. The fault rate surfaces are applied as a part of an automated network planning routine planning case based on real network data from the same distribution system operator as the fault data. The results can help in the planning of more reliable, yet economically feasible, power distribution networks.