PSI - Issue 62

Available online at www.sciencedirect.com Structural Integrity Procedia 00 (2022) 000 – 000 Available online at www.sciencedirect.com ScienceDirect

www.elsevier.com/locate/procedia

ScienceDirect

Procedia Structural Integrity 62 (2024) 89–96

II Fabre Conference – Existing bridges, viaducts and tunnels: research, innovation and applications (FABRE24) An Artificial Neural Network for the prediction of the structural and foundational attention class of bridges according to the Italian Guidelines Lorenzo Principi 1 , Michele Morici 1* , Agnese Natali 2 , Walter Salvatore 2 , Andrea Dall’Asta 1 1 University of Camerino, School of Architecture and Design (SAAD), Ascoli Piceno (AP) 63100, Italy 2 University of Pisa, Department of Civil and Industrial Engineering, Pisa (PI) 56122, Italy Abstract The state of conservation, maintenance and monitoring of bridges has gained attention in the last decade over the world especially in Italy, where a large number of structures are present and, multiple cases of collapses of bridges have recently occurred. To address this problem and to provide a prioritization on bridges where detailed safety assessments are necessary, in 2020 the Italian Ministry of Transport and Infrastructure issues Guidelines, based on a multi-level and multi-risk approach. Six levels of assessment are foreseen: the first three must be applied to all bridges (Levels 0-2), while the last three (Levels 3-5) only for the bridges which are characterized by high risk deriving from the analyses of the previous levels. Focusing on the first three levels, Level 0 consists of a census of all the existing bridges, collecting registry data mainly deriving from the existing documentation. Level 1 consists of visual inspections, which are used to point out the conservation status of the bridge and the surrounding area. Level 2 provides for a risk-based classification starting from the data previously collected. Levels 0-2 must be applied indiscriminately to all bridges. Thus, to prioritize inspections, it could be helpful to have a tool capable to predict the state of conservation of bridge and to assess the associated risks, starting from data gathered with census. For this reason, this paper proposed an Artificial Neural Network (ANN) capable to assess the level of degradation and structural and foundational risk level of existing bridges, using a reduced set of information derived from Level 0 activities. This tool can be used to: I) rationally schedule the inspections, starting from structures that could have a higher probability to be heavily degraded, II) support managing and planning activities at territorial level, promptly furnishing information about the structural and foundational risk of the bridges. © 2024 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license ( https://creativecommons.org/licenses/by-nc-nd/4.0 ) Peer-review under responsibility of Scientific Board Members Keywords: Existing Bridges; Bridge Inspections; Defectiveness; Retrofit prioritization; Artificial Neural Network, Risk Assessment © 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of Scientific Board Members

2452-3216 © 2024 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license ( https://creativecommons.org/licenses/by-nc-nd/4. 0 ) Peer-review under responsibility of Scientific Board Member s

2452-3216 © 2024 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of Scientific Board Members 10.1016/j.prostr.2024.09.020