Ranking Models Used for Condition Assessment of Civil Infrastructure Systems

Abstract

Statistical ranking models are employed to extract information from an historical database on the prioritization of maintenance and repair operations for embankment dams within the U.S. Army Corps of Engineers. This information is used to develop weighting functions that express the relative importance of general classes of maintenance and repair operations in an effort to implement a Function-Based Condition Indexing methodology. In particular, the weighting functions are used to convert the condition index vector [a representation of the condition of the embankment dam under several repair, evaluation, maintenance, and rehabilitation (REMR) objectives, one element per objective] into a condition indexing scalar (a representation of the overall condition of the embankment dam). Goodness-of-fit statistics are used to evaluate the ranking criteria models. The procedure is demonstrated on embankment dams for the REMR objectives of Prevention of Surface Erosion and Collection of Performance Information using the Automated Budget System (ABS) database of the Corps of Engineers. The statistical analysis indicates that physical parameters such a dam age and height have had a tremendous influence on the historical prioritizations for individual operations under the two REMR objectives studied even though these may not have been explicitly considered during the ranking process. Other important parameters such as reservoir size, fetch, downstream hazard, and economic effect of the dam should be investigated in the future to determine the weighting functions for the final version of the condition indexing system.