PSI - Issue 44

Agnese Natali et al. / Procedia Structural Integrity 44 (2023) 2020–2027 Agnese Natali et al./Structural Integrity Procedia 00 (2022) 000–000

2021

2

1. Introduction The Italian “Guidelines for classification, risk management, assessment of level of safety and monitoring of the existing bridges” (Ministero delle infrastrutture e della mobilità sostenibili, 2020) together with the “Operating Instructions” (Ansfisa, 2022) provide a new method for the management of the existing bridges. The method is structured in two big macro-phases: the first one is applied to all the bridges and aims at the classification of the structures considering the structural, seismic, and hydro-geological risks; the second one consists in further activities after classification (from periodic inspections, to monitoring or assessment of the safety level), based on the level of alarm obtained. To make the method easily and quickly applicable by the operators of the road managers companies, the evaluation of each source of risk is performed by applying a qualitative approach based on logical flows through which the so called “Warning Class” (WaC) of the structure is evaluated. Basically, by applying this procedure, a classification of the existing structures is performed, considering all the relevant sources of damage and the probability of the structures to be damaged by them. The WaC for each source of risk is a product of vulnerability, exposure and hazard classes, which are obtained through logical flows that consider the interaction of different parameters, which are determined or collected in the previous steps of the method (Level 0, to gather all the data for the classification step; Level 1, to perform visual inspections for the evaluation of the conservation status of the structure its context). These two steps and the quality of the data gathered during the inspections, are fundamental for a reliable evaluation of the WaC. Although the method is structured to be easily applied by the operators, the actions to the classification are many, and specifically those referring to the inspections may also be time consuming. Besides, they might be carried out for a large number of structures, elongating the time to obtain their final classification. As a consequence, without planning also the first activities of Levels 0 and 1, some critical situations may not be immediately analysed and might evolve in irreparable events. For this reason, a criterion of prioritization should be defined for a smart application of the method. In this regard, one of the possible defining parameters for the prioritization may be the conservation status of the structure, which also has a relevant role in determining the structural WaC, considering the evaluation of the vulnerability parameter. In this framework, this paper shows an ongoing activity about the use of Artificial Intelligence to develop a smart tool that recognizes the different elements that compose the bridge, the defects, their intensity, size and position. This tool could be applied to the images gathered, as an instance, from a drone, which independently flies by the structure and captures images of the structure. With the purpose of the definition of a prioritization criterion, drones could be sent to the structures, and the tool could analyze the images and return a primary evaluation of the level of defectiveness of the structure matching the detection of damage with the procedure from the Operating Inspections for the evaluation of the level of defectiveness. 2. Objectives of the smart tool and work phases The smart tool is conceived to be an instrument for a primary evaluation of the level of defectiveness of the structure, as defined in the Operating Instructions. For the correct evaluation of the level of defectiveness, the first step is the damage detection, that in the process of classification is made in the Level 1 during the visual inspections. The second step is the proper evaluation of the level of defectiveness, that is guided by the indications given by the Operating Instructions and depends on the types of damage detected in terms of severity, intensity and extension, their quantity and position/distribution in the component, and the possible individuation of critical situations. Starting from this, the smart tool should be able to accomplish the following tasks: A. Individuation of the global configuration of the structure (structural scheme). Individuation of all the components and their material and distinguishing the critical ones. B. For each component, detection of possible damages. C. For each damage, definition of its extension and intensity. D. Depending on the observed damage and its position and/or distribution in the component, being capable of recognizing the possible critical situations.