This communication describes a case-study concerning a real hydraulic power-plant affected by AAR. The displacements imposed to the sensitive equipments (pumps, turbines, pipes) are difficult to predict due to the complexity of the structure and to strongly heterogeneous moisture distribution. Hence the prediction of serviceability and the managing of the plant require the complete chemo-mechanical modeling of the structure. Using a performant model designed to take into account influence of moisture and temperature on both kinetics and amplitude of AAR, the different phases conducting to the re-assessment of the structure are presented : monitoring of the structure, data collection from material samples, fitting of the numerical model, validation and exploitation. The paper focuses on the fitting of the model based on global and local strain monitoring. It aims to demonstrate the efficiency of a global method based on both field monitoring and numerical re-assessment to answer a critical structure-management problem.