Statistical Analysis of Monitoring Data for Concrete Arch Dams

Major dams in the world are often instrumented and monitored in order to detect anomalies, and to enable a timely response either in the form of repairs, reservoir management, or evacuation. Traditionally, the analysis of monitoring data has been done on an instrument-by-instrument basis, which is tedious to perform and difficult to interpret. The objective of this paper is to propose a statistical estimation procedure, principal component analysis, to identify the principal modes of deformation of the dam from a historical record of instruments. This approach reduces the number of analyses that need to be performed, provides a more representative model for the response of the dam, and a more robust model for the detection of the onset of anomalies in the response of the facility.