A neural network-based multivariate seismic classifier for simultaneous post-earthquake fragility estimation and damage classification. (15th March 2022)

• Record Type: Journal Article
• Authors: Yuan, Xinzhe; Chen, Genda; Jiao, Pu; Li, Liujun; Han, Jun; Zhang, Haibin
• Is Part Of: Engineering structures. Volume 255(2022)
• Publication Date: 2022-03-15
• Subjects: Artificial neural networks -- Seismic damage classification -- Fragility estimation -- Multivariate seismic classifier -- Intensity measures; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2022.113918
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: An ANN-based multivariate seismic classifier for simultaneous fragility estimation and damage classification. Due consideration of the primary record-to-record uncertainty using multiple IMs. Influence of IM selection to the performance of ANN seismic classifiers. System-level and element-level ANN classifiers for a reinforced concrete building. Satisfactory classification of all elements using a single unified ANN seismic classifier. Abstract: A scalar intensity measure ( IM ) could be insufficient to represent the earthquake intensity and variety in fragility estimation. Introducing multiple IM s to conventional regression of fragility functions can be computationally demanding and require priori assumptions of functional forms. In this study, multivariate seismic classifiers with multiple IM s as inputs are developed based on artificial neural networks ( ANN s) to address the above disadvantages of traditional regression approaches. Case studies of a four-story code-conforming benchmark building indicate that fragility estimates from multi- IM ANN classifiers lead to higher accuracy (5.0% to 7.7%) in system-level and element-level damage classification than the single- IM traditional fragility curves. Further studies of IM combinations show that the ANN performance can be improved by more IM s correlated with structural responses while compromised by redundant irrelevant IM s. The optimal IM set should be determined by correlation ranking and ANN predictive performance together. Moreover, the ANN configuration of the case-study building is optimized with five readily available IM s as inputs, which enable a near real-time (within 0.3 ms) prediction of future earthquake damage while maintain high predictive performance. Overall, the multivariate ANN seismic classifier can be a promising tool for simultaneous seismic fragility estimation and damage assessment.