Causal Probabilistic Networks in Assessing the Vulnerability of Existing Buildings

See related content

by F. Casciati, Univ of Pavia, Pavia, Italy,
L. Faravelli, Univ of Pavia, Pavia, Italy,

Document Type: Proceeding Paper

Part of: Structural Engineering in Natural Hazards Mitigation:

Abstract: A probabilistic assessment of the 'state of health' of the structure is pursued on the basis of pieces of evidence collected by inspection and by querying, in case, historical data bases. The diagnosis is reached by propagating evidence in a causal probabilistic network representing the available expertise. In the first implementation each piece of evidence to be collected by inspection is just the classification of each input variable into one of its possible states. The paper investigates the possibility of adopting a 'learning' strategy just by exploiting the same Bayesian updating scheme which governs the probabilistic propagation through the network.

Subject Headings: Probability | Inspection | Building systems | Case studies | Historic buildings | Databases | Diagnosis | Professional societies |

Services: Buy this book/Buy this article


Return to search