Structural safety evaluation using proof loads based on Bayesian inference

Jiali Tan

doi:10.1007/s13349-021-00523-7

2021-09-24

PRE

Jiali Tan

Sheng-En Fang *

DOI：10.1007/s13349-021-00523-7

原文链接

原文求助

摘要

En 中文

Structural safety evaluation is a critical issue in the realm of civil engineering. It involves factors such as uncertainties, external loads, and structural damage after long-term service. Presently, real-time safety evaluation of structural resistance remains challenging for real-world structures. This study proposes a resistance updating model for reinforced concrete beams based on Bayesian inference. The measured static loads of six experimental beams were taken as the proof loads for assessing real-time safety, and a stiffness degradation coefficient (SDC) was developed to embody resistance reduction. The SDC was continuously updated based on the deflection monitoring data of the beams under the proof loads. Specifically, the beam resistances were defined as the random distributions, and their initial distributions were calculated according to the material properties of the beams. Then, the corresponding SDCs of each proof load were estimated. Based on the initial resistance distributions and the real-time SDCs, the iterative updating of the resistance distributions was carried out using Bayesian inference. Subsequently, the future reliability indices of these beams were inferred using the JC method, which utilized the updated resistance distributions and the assumed future load distributions. Finally, a reliability index evolution model was established to evaluate the safety of the experimental beams. The analysis results showed that the proposed method can effectively estimate the resistance to provide an upper limit of future loading. The proposed method is sensitive to the proposed SDC but insensitive to the initial distributions of the resistance.

Keyword:

Safety evaluation

Resistance updating model

Stiffness degradation coefficient

Proof loads

Reinforced concrete structures