In engineering system control, human beings can play various key roles, in particular concerning measurement, global assessment and decision. This paper focuses on methods that allow the representation and aggregation of heterogeneous data (sensory evaluations, physical measurements, outputs of mathematical models, etc.) used in a global dam assessment process. It is acknowledged that in such complex systems many of the variables involved are evaluated with uncertainty. We propose a possibility theory-based approach to deal with all the different uncertain pieces of information and propagate them in aggregation models for global dam assessment. Finally, decision-making and communication applications relating to dam safety are presented.