Prediction of limit states occurrence probability in curved bridges based on artificial intelligence and statistical analysis
Abstract
Purpose
It evaluated the seismic vulnerability based on fewer factors by presenting the effectiveness of seismic and structural parameters. The proposed method first demonstrated the effect of earthquake ground motion inputs on predicting the slight, moderate, extensive and collapse limit states and confirmed the method’s efficiency. The fragility curves illustrated with the approach of the present study are compared with the traditional techniques, such as analytical methods.
Design/methodology/approach
Based on the different macro- and micro-structural characteristics and the earthquake records, achieving a certain relation from regression analysis using artificial neural networks (ANNs) is difficult. With this background in mind, the present study aimed to compare the proposed model of the considered bridge with the analytical and ANN results. After statistical analysis and estimation of the most effective factors in predicting responses from the proposed approach, two-parameter two- and three-dimensional fragility curves are extracted.
Findings
Due to the structural differences between horizontally curved bridges, the methodology does not require any classification of bridge classes to predict responses. For a specific L/R of the bridge, the parameters cumulative absolute velocity (CAV) and Sa (T1) can provide a good estimate of the seismic fragility curves, and the proposed approach with less parameter assignment also leads to good results. With less computational effort, fragility curves can be illustrated.
Originality/value
The proposed method demonstrated the ability to accurately estimate the occurrence and non-occurrence limit states while maintaining a low computational cost and the derivation of a curved bridge’s seismic fragility curve.
Keywords
Citation
Karimi-Moridani, K. (2024), "Prediction of limit states occurrence probability in curved bridges based on artificial intelligence and statistical analysis", Engineering Computations, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/EC-03-2024-0237
Publisher
:Emerald Publishing Limited
Copyright © 2024, Emerald Publishing Limited