Reliability assessment of RAC chloride concentration using Karhunen–Loève expansion with digital-image kernels. (10th June 2020)

• Record Type: Journal Article
• Authors: Wu, Yu-Ching; Zhi, Peng
• Is Part Of: Construction & building materials. Volume 245(2020)
• Publication Date: 2020-06-10
• Subjects: Reliability assessment -- Chloride diffusivity -- RAC -- Karhunen–Loève -- Digital-image kernel; Building materials -- Periodicals; 624.18
• DOI: 10.1016/j.conbuildmat.2020.118352
• Languages: English
• ISSNs: 0950-0618

Abstract

Highlights: The multiscale digital-image driven method is modified to explore stochastic effects. Karhunen–Loève spectral functions with digital-image covariance kernels are used. A novel numerical method is generated based on spectral functions for FEA. Reliability assessments of RAC Chloride diffusion influences are made. Abstract: Measurement of the durability of recycled aggregate concrete (RAC) is significant for steel reinforced concrete structures in chloride-rich environment. In literature, an ideal five-phase model is usually used as a stochastic volume element (SVE) of RAC. As an extension of previous researches, this paper improves the multiscale digital-image driven method developed previously to explore the influence of fine-scale randomness on the coarse-scale chloride diffusivity in SVE, which is used in the current stochastic analysis. In the novel numerical method, the Karhunen–Loève expansion with the digital-image covariance kernel is used to generate spectral functions for the finite element model. Results of the proposed method are in great agreement with ones in literature. After a couple of numerical experiments are made, it is demonstrated that the influence of fine-scale randomness is considerable on the global chloride diffusivity in RAC.