Data driven maintenance cycle focusing on deterioration mechanism of road bridge RC decks

Effective maintenance of huge number of aged infrastructures with limited budget and resources has become a serious problem in many developed countries. A new concept called "Society 5.0", which integrates cyber space and physical space, provides a feasible solution strategy for infrastructures' maintenance. The authors focus on developing a rational maintenance system for existing reinforced concrete (RC) decks. Here, two predictive models are compared by analyzing in-service RC decks in Japan, i.e., survival analysis based on statistical-processing of site- information, and artificial intelligence (AI) models based upon data assimilation technology of site-inspected cracks. Although the two methods have different terminologies, the results show that they are highly correlated, which demonstrates their reliability for life assessment of RC decks, and their consistency for integration. Finally, a comprehensive maintenance system is introduced by setting priorities of inspection, as well as, updating inspection interval. Thus, rational countermeasure planning and maintenance can be successfully implemented.