Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4965630 | Computers & Structures | 2018 | 12 Pages |
â¢A new method is proposed to model the temporal correlation in loads.â¢Copula-based approach is used to model the deterioration-load dependency.â¢Reliability analysis method is developed considering the load temporal correlation and deterioration-load dependency.
Aging of structural performance and significant external loads may impair structural safety and serviceability and cause potential economic losses. In the presence of uncertainties associated with both resistance deterioration and external loads, structural safety shall be estimated quantitatively under a probability-based framework. A stochastic load process is often auto-correlated on the temporal scale, with correlations arising from both the occurrence times and intensities. Moreover, a deterioration process is physically dependent on the load magnitudes. This paper investigates the impacts of load temporal correlation and deterioration-load dependency on time-variant structural reliability. The load occurrence process is modeled as a Poisson point process with correlated separation time between two load events. The correlation between the intensities of load events is described by the multi-variate Gaussian copula function. The resistance aging process is considered to be a combination of both gradual and shock deteriorations. Four candidate copula functions, namely Gaussian, Clayton, Gumbel and Frank, are considered to model the dependency of shock deterioration on load intensity. Two types of failure mechanisms are considered: the first is due to the load effect exceeding the resistance, and the second occurs when the cumulative damage within the considered service period reaches the permissible level. A simulation-based method is developed to estimate structural reliability considering the two failure modes. Illustrative examples are presented to demonstrate the applicability of the proposed method. Parametric studies are conducted to investigate the impacts of temporal correlation in loads and deterioration-load dependency on structural failure probability.