Efficient masonry vault inspection by Monte Carlo simulations: Case of hidden defect

This paper presents a methodology for probabilistic assessment of masonry vaults bearing capacity with the consideration of existing defects. A comprehensive methodology and software package have been developed and adapted to inspection requirements. First, the mechanical analysis model is explained and validated by showing a good compromise between computation time and accuracy. This compromise is required when probabilistic approach is considered, as it requires a large number of mechanical analysis runs. To model the defect, an inspection case is simulated by considering a segmental vault. As the inspection data is often insufficient, the defect position and size are considered to be unknown. As the NDT results could not provide useful and reliable information, it is therefore decided to take samples with the obligation to minimize as much as possible their number. In this case the main difficulty is to know on which segment the coring would be mostly efficient. To find out, all possible positions are studied with the consideration of one single core. Using probabilistic approaches, the distribution function of the critical load has been determined for each segment. The results allow to identify the best segment for vault inspection.