Model uncertainties of concrete cracking resistance models based on probabilistic simulations

Model uncertainties of concrete cracking resistance models based on probabilistic simulations

Paper in proceedings outside WoS and Scopus

In the paper, the probabilistic assessment of cracking resistance of concrete flexural members is presented. The aim of the performed analysis was to verify an alternative formula for cracking resistance calculation and to compare the proposed method with two standard methods. The experimental investigation performed at Lublin University of technology was used to verify the design models. The accuracy and reliability of the calculation methods was assessed by analyzing the model uncertainty θ. When model uncertainty θ > 1.0 the prediction model yields a lower value of cracking resistance and is thus conservative, while a value of θ < 1.0 implies that the prediction model yields higher cracking resistance than is actually available in the structure and is thus un conservative. A high model uncertainty θ = 1.53 was found when the cracking resistance was calculated by standard method assuming a linear distribution of normal stresses over the cross section and taking the maximum tensile stress as the concrete axial tensile strength. When applying the flexural tensile concrete strength defined in Eurocode 2 instead of axial tensile strength in a cracking resistance formula, the model uncertainty decreased to θ = 1.15 but the model was still conservative. The best prediction of cracking resistance was obtained for the proposed method in which the influence of a size effect and fracture properties of concrete on cracking moment were included. In this case the model uncertainty was close to 1.0 (θ = 0.94) with a relatively small scatter.

In the paper, the probabilistic assessment of cracking resistance of concrete flexural members is presented. The aim of the performed analysis was to verify an alternative formula for cracking resistance calculation and to compare the proposed method with two standard methods. The experimental investigation performed at Lublin University of technology was used to verify the design models. The accuracy and reliability of the calculation methods was assessed by analyzing the model uncertainty θ. When model uncertainty θ > 1.0 the prediction model yields a lower value of cracking resistance and is thus conservative, while a value of θ < 1.0 implies that the prediction model yields higher cracking resistance than is actually available in the structure and is thus un conservative. A high model uncertainty θ = 1.53 was found when the cracking resistance was calculated by standard method assuming a linear distribution of normal stresses over the cross section and taking the maximum tensile stress as the concrete axial tensile strength. When applying the flexural tensile concrete strength defined in Eurocode 2 instead of axial tensile strength in a cracking resistance formula, the model uncertainty decreased to θ = 1.15 but the model was still conservative. The best prediction of cracking resistance was obtained for the proposed method in which the influence of a size effect and fracture properties of concrete on cracking moment were included. In this case the model uncertainty was close to 1.0 (θ = 0.94) with a relatively small scatter.

Unreinforced concrete structures, Cracking resistance, Model uncertainty

Unreinforced concrete structures, Cracking resistance, Model uncertainty

SŁOWIK, M.; NOVÁK, D.; LEHKÝ, D.; SKRZYPCZAK, I.

23.04.2025

IA-FraMCoS

Vienna

978-3-903039-01-8

12th International Conference on Fracture Mechanics for Concrete and Concrete Structures

1

8

8

https://framcos.org/FraMCoS-12/Full-Papers/1378.pdf