PSI - Issue 64

5

Author name / Structural Integrity Procedia 00 (2019) 000 – 000

M. Pedram et al. / Procedia Structural Integrity 64 (2024) 621–628

625

Fig. 2- FEM of a typical slab (D5)

3. Results 3.1. Validation of finite element models

Validation of the finite element models was the first step for the solution. For validation of the finite element (FE) models, the thermal properties of the material, heat flux on the sides and convection coefficient were updated. That is, the parameters were adjusted by sensitivity analysis so that the FEA conveyed the closest temperature estimations to the temperature records from IRT experiments. Initially, the thermal properties of the concrete and air were considered based on the hot disc measurements and assumed baseline values as formerly presented in the Table 1. In addition, the radiation heat transfer coefficient was set to a constant value of 53.865×10 -9 W/m 2 .K 4 . The other non-measured parameters of the model such as convection coefficient (hc) and heat flux on boundaries (f) were found by sensitivity analysis. As the convection coefficient depends on the temperature, for each initial temperature a different sensitivity study was performed on the control specimen. In the sensitivity study, the convection heat transfer coefficient and the boundary heat loss values that exhibit the lowest difference between the FEA surface temperature estimations and surface temperatures measured by IRT were selected. The heat flux on the boundaries was determined to boost the accuracy of the FEA results. The Euclidean norm of difference (ED) between the data sets was used to select the closest FE results to the experimental temperature records. Initially, for an assumed heat flux on the side and bottom faces, a first-stage sensitivity analysis was performed to find the convection coefficients. After fixing the convection coefficient, a second set of sensitivity studies was performed to find the heat flux on the boundaries. For the initial temperature of -20°C, 2.5 W/m 2 K, and 50W/m 2 were selected as the convection coefficient and boundary heat flux, respectively. For the initial temperature of 60°C, the convection coefficient and heat flux on the boundary were found as 1.5 W/m 2 K, and -40 W/m 2 , respectively. After finding the optimal values of the convection heat coefficient and heat flux on the boundaries, the conductivities of concrete and polystyrene were updated separately to enhance the precision of FEA estimations. The temperature estimations of FEA were compared for nodes on the defective concrete and the node on the intact

concrete to confirm the accuracy of prediction. 3.2. Variation of maximum thermal contrast

In this paper, the variation of the absolute maximum thermal contrast with the defect depth for a heat-up case from -20°C, and a cool-down case from 60°C are presented (Fig. 3). According to Fig. 3, the FEA estimations of