PSI - Issue 44

Mattia Zizi et al. / Procedia Structural Integrity 44 (2023) 673–680

674

2

Mattia Zizi et al. / Structural Integrity Procedia 00 (2022) 000–000

criteria and mainly accounting for vertical actions rather than horizontal ones. Since they often represent strategic infrastructures, the estimation of their safety against exceptional loads, such as the seismic one, is a topic of noteworthy importance for the engineering community. The evaluation of the structural response under external loads of existing masonry arch bridges is a complex issue. As a matter of fact, the behavior of such structures is notably affected by the uncertainties related to the mechanical parameters of the material and the interaction between them. In general, the safety assessment of masonry arch bridges can be performed at different detail levels. Whilst Limit Analysis (LA), based on Heyman’s assumption of null tensile strength of the material (Heyman, 1966), is recognized as a suitable yet generally conservative tool for estimating the bearing capacity under vertical and horizontal loads of such structures (Cavicchi & Gambarotta, 2005; Gilbert et al., 2020; Papa et al., 2021), approaches as FEM, DEM and DMEM are conversely able to simulate the overall response of the structure (both in the linear and nonlinear field) (Milani & Lourenço, 2012; Pantò et al., 2022; Portioli, 2020; Tubaldi et al., 2020; Zampieri et al., 2021). Existing masonry bridges are usually characterized by a set of masonry elements (i.e. arch ring, spandrel walls, piers and abutments) interacting with backfill material of soil nature filling the volume identified by the arch ring extrados and lateral spandrel walls (Hokelekli & Yilmaz, 2019). The capacity in withstanding external loads is mainly attributed to masonry elements, despite the backfill material could play a significant role in the overall response providing a restraint to the deformation of the arch ring and guaranteeing beneficial load redistribution effects (Callaway et al., 2015; Sarhosis et al., 2016; Zhang et al., 2018). Such an aspect has been widely addressed in the scientific literature, leading to the definition of several approaches for considering the backfill influence in numerical models (Gago et al., 2011). These differ on the level of complexity at which backfill effect is accounted for, such as: i. external load, ii. external load in combination with passive pressure, iii. elastic material, iv. nonlinear material. In general, it has been widely demonstrated that modelling assumptions can have a relevant influence on the final output of numerical analyses, also referring to other construction materials (Castaldo et al., 2022). Based on this premise, in the present study, the response under longitudinal seismic actions of a multi-span masonry arch bridge is analyzed. The numerical model is calibrated against an experimental test retrieved from the literature and includes the backfill considered as a nonlinear material. Then, the effects of different backfill material and interaction hypotheses on the seismic response of the bridge are studied and discussed. The experimental test considered in this study is “ Bridge no. 2 ” described in Melbourne et al. (1997). The reference test consisted of a three-span brickwork arch bridge vertically loaded until collapse at one-fourth length of the central span. The load was applied on the upper surface of the backfill by means of a concrete beam, whose imprint was 0.42 m wide. Ring arches had an intrados span of 3 m, an embrace angle of 106° and a thickness of 0.215 m. Piers had a height of 1.5 m and a rectangular cross-section of 0.44 m × 3.54 m, with the major dimension along the transversal direction of the model. Backfill over the arches had a height measured from the arch impost of 1.135 m, while its dimension in the transversal direction was 2.88 m. No spandrel was considered in the tested bridge, but the backfill was constrained in the transversal direction by the presence of additional detached walls. As far as the materials are concerned, the following information were available: • Brick masonry composing the ring arches and piers was characterized by a compressive strength f c =26.8 MPa, an Elastic Modulus E m =16200 MPa and a specific weight ρ =22.4 kN/m 3 ; • Backfill material consisted of incoherent soil ( c =0 MPa) having a friction angle φ =60° and a specific weight ρ =22.2 kN/m 3 . The test returned a capacity of about 320 kN of the bridge under the tested load condition. Several cracks at the intrados and extrados of the central and lateral arches, as well as of the right pier, were identified. In Fig. 1a the lateral 2. Model calibration 2.1. The reference test