PSI - Issue 84

Ivan Beltracchi et al. / Procedia Structural Integrity 84 (2026) 1015–1022

1020

compared with the response of the PRC beam obtained from a monotonic incremental loading analysis (pushover curve) performed in DIANA FEA . The model adopts a total-strain crack model, a non-linear constitutive approach for reinforced and prestressed concrete that captures crack initiation and propagation while accounting for the interaction of bending, shear, and compressive actions. The numerical model represents a single beam through a non-linear model that includes the transverse distribution beams and the associated deck segments, together with the overlying asphalt layers. This modelling strategy enables a detailed representation of the components governing the global response of the beam. Reinforced-concrete beams (C50/60) and the deck slab and transverse beams (C25/30) were modelled using Hordijk’s tension-softening law in tension (Hordijk, 1991) and parabolic stress–strain law in compression (DIANA user’s manual). Steel behaviour was described by a bilinear elastoplastic model for the ordinary reinforcement and a Ramberg-Osgood law for the prestressing tendons (Collins and D. Mitchell, 1991). For asphalt stiffening layers, the ongoing study adopts a dedicated procedure to define an appropriate mechanical characterisation of the mixture. The current implementation starts from literature-based average properties and models the asphalt as a viscoelastic material using a Maxwell-Kelvin representation (DIANA user’s manual, Fib 2013), which more realistically captures relaxation under static and dynamic loading and the influence of temperature (assumed constant in the present case).To improve numerical convergence and better capture the beam’s flexural response under a distributed load, the structural behaviour was evaluated using an analogue four-point bending test (4PBT). In this configuration, the load is increased under displacement control and applied at one-quarter of the span from each support, promoting bending-dominated, ductile failure mechanisms. The loading scheme is shown in the numerical model view with the complete mesh (Fig. 7). As shown in Fig. 8, accounting for the asphalt layers above the deck (light-blue dotted line) increases both the global capacity (ultimate load) and the overall stiffness compared with the model without asphalt. This inclusion also provides a closer match to the initial branch of the response observed in the field measurements.

Fig. 7 3D global view of nonlinear model of the beam in DIANA FEA and some local details.

(a) (b) Fig. 8 Equivalent distributed load of vehicle vs. mid-span deflection: (a) global response and (b) comparisons with experimental measures.