PSI - Issue 64

Fadi Oudah et al. / Procedia Structural Integrity 64 (2024) 1983–1989 Fadi Oudah/ Structural Integrity Procedia 00 (2019) 000 – 000

1987

5

Fig. 1. Illustration for utilizing RFE models developed using RF-DYNA to develop a capacity histogram for reliability analysis.

6. Real-Life Case Studies of RFE for Reliability Assessment of RC structures The random finite element method (RFE) is seldom used in real life applications to assess the reliability of in service structures. Limited applications in the geotechnical field include assessing a real dyke failure (de Gast et al., 2018) and seismic analysis of gravity dam (Hariri-Ardebili et al., 2018). The authors have been involved in a few consulting jobs where RFE was utilized to assess the reliability of existing RC infrastructure in Canada at ULS by building representative resistance models. The scope of work and analysis approach of three real-life case studies are briefly reviewed. Specific project information and findings are not disclosed for confidentiality purposes. Case 1. AAR Affected Steel Reinforced Concrete Beam . Reliability analysis was conducted to determine the reliability index of a 50-year-old 9 span continuous reinforced concrete beam supporting moving live load. Field tests were conducted to estimate the mean and standard deviation of the concrete compressive strength. RFE was used to develop the capacity distribution at ULS – approximately 40 models were generated each with a different possible pattern. Three correlated random fields were considered including the concrete compressive strength, concrete tensile strength, and bulk modulus. The concrete tensile strength was reduced to account for cracking due to AAR. The reliability analysis was performed, and a reliability index was calculated. Case 2. Deteriorated Marine Concrete Mooring Wall . Reliability analysis was conducted to determine the minimum mooring load that can be applied to an existing 50-year-old mooring wall. RFE was utilized to develop the capacity distribution at ULS – approximately 40 models were generated each with a different possible pattern of concrete strength. A parametric analysis was conducted to determine the allowable mooring load. Case 3. Mechanically Damaged Concrete Slab . The probability of failure of a mechanically damaged in-service concrete slab in an industrial plant was assessed using RFE, where the mechanical damage was modelled based on three-dimensional point cloud data and the spatial variability of the concrete slab and the supporting soil (three layers of soil) were considered. Five random fields were considered including concrete compressive strength, concrete tensile strength, bulk modulus, soil angle of friction, and soil shear modulus. The following are common observations drawn from the above consulting projects: 1) the coefficient of variation of the capacity histogram at ULS was relatively low – ranges from 0.03 to 0.05; 2) the capacity histogram encompassed multiple possible failure modes of the assessed structure which indicates the usefulness of RFE based reliability analysis in assessing the global safety of the examined structures as discussed in Section 5; and 3) simulating the structural-soil interaction using RFE was critical for establishing a representative capacity histogram of the global response of Case Study 3.