PSI - Issue 64

Abdalla Elhadi Alhashmi et al. / Procedia Structural Integrity 64 (2024) 1990–1996 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

1994

5

and bottom edges of the concrete block, about 50 mm away, to effectively replicate the impact of the steel frame as shown in Figure 2(b). 4. Results and discussion Figure 3 presents a comparison of ultimate stress under compression for four different DIP-FE models of a corroded concrete column. The observed variations in the results are based on the inclusion or exclusion of corrosion-induced crack mapping and the presence or absence of steel corrosion. The comparison between the DIP-FE models and the experimental data reveals that the ultimate load-bearing capacity of the column is predominantly influenced by the presence of corrosion-induced cracks rather than the decrease in the cross-sectional area of the reinforcement due to corrosion. This conclusion is supported by the nearly identical percentage difference observed between the models which included both cracks and corrosion (target model) and the model that only accounted for cracks. Additionally, incorporating the cracks into the FE model enhanced the precision of predicting the ultimate resistance by 22%, as evidenced by comparing the errors associated with the target model and the intact model (i.e., absence of corrosion and cracks). Hence, the efficacy of the DIP-FE analysis approach is confirmed. Figure 4 illustrates sample random field realizations, and the corresponding constructed resistance frequency histogram. A total of fifty RFE analyses were performed using the target model to build the resistance model. As shown in Figure 4, the resistance model followed a normal distribution with a mean, standard deviation, and COV of 32.27 MPa, 1.32 MPa, and 0.0410, respectively. The model accounted for spatial fluctuations of compressive strength, tensile strength, stiffness, and their associated damages. The yield of steel reinforcement was randomly sampled from a lognormal distribution with a bias of 1.145 and a COV of 0.05 (Nowak and Szerszen, 2003). The statistics of the resistance model (mean and COV) can then be used to estimate the probability of exceedance and reliability index for a given load using the limit state function presented in Equation (2).

Fig. 3. Comparison between different DIP-FE models and experimental results by Ma et al. (2022)

Made with FlippingBook Digital Proposal Maker