Abstract
Assessing the safety of a civil engineering structure requires using a large quantity of data. These data are frequently “imperfect”, meaning that they comprise uncertainty, imprecision and incompleteness. It is important to develop systems for assessing and controlling these imperfections in order to both quantify the quality of the data used to evaluate the condition of structures and decide on the corrective actions to reduce these imperfections. Here, we propose a generic approach to controlling the quality of the data used when analysing a dam. Different methodologies can be considered to implement this approach and an illustration is given for two methods developed independently: one by Hydro-Québec, in Canada, and another, by Irstea, in France.