1. Long-term behavior

Predicting behaviour is very complex because of the multitude of materials that exist, their combination and the conditions of use that control their initial quality and their evolution in service (see article ). Robust predictive models can only be built if the defects present after installation or generated during ageing are known and precisely identified, as well as their propagation modes.

The ageing of a material, whether composite or not, is a combination of a multitude of processes, interacting or independent, occurring at the same time or sequentially. The evolution of a COM is very often described as a set of elementary acts governed by a single mechanism. Knowledge of the interactions between all these parameters is a key factor in the accuracy of behavior prediction models (see article ).