Pipeline damage assessment using cluster analysis

Abstract

Clustering techniques are used commonly in a variety of engineering and scientific fields although it is rarely used in lifeline earthquake engineering. Cluster analysis deals primarily with the discovery of structures or groupings within data. This paper presents the application of cluster analysis in pipeline damage assessment and identification of high damage areas. Identification of sites where pipeline damage concentrates has special importance because these sites are generally problematic areas and/or pipelines in these sites have some weaknesses. Understanding why damage is high there may contribute and improve future works related to pipeline damage prevention and mitigation. The 1994 Northridge earthquake water distribution pipeline damage in the city of Los Angeles was used herein to illustrate the application of subtractive and fuzzy c-means cluster analysis. © 2009 ASCE.