A fast seismic assessment technique for reinforced concrete buildings: Machine learning-based Hassan Index

Fahri Baran Koroglu, Muhammet Fethi Gullu, Serdar Ciftci, Liam Pledger, Claudio Schill, Santiago Pujol,
A fast seismic assessment technique for reinforced concrete buildings: Machine learning-based Hassan Index,
Structures,
Volume 82,
2025,
110425,
ISSN 2352-0124,
https://doi.org/10.1016/j.istruc.2025.110425.
(https://www.sciencedirect.com/science/article/pii/S2352012425022404)
Abstract: Assessing large inventories of reinforced concrete structures in urban areas with high seismicity is a daunting task that requires tools that can be applied quickly to produce reliable results. The first goal should be to identify the most vulnerable structures that require rapid intervention. Existing assessment standards are often too complex for this purpose. In the literature, the index-based methods, for example the Hassan Index, provide more efficient assessment options based on simple geometric parameters. The question addressed here is whether machine learning (ML) algorithms trained to use the same parameters can better match field observations. The developed algorithm has been trained and tested on survey data from 1320 low- to mid-rise buildings, the model achieved 74 % accuracy on a held-out test set with 5 % “risk” (false negatives) and 21 % “cost” (false positives), improving over the simple index-threshold baseline (61 % accuracy, 8 % risk, 31 % cost). On an external dataset from the 2024 Taiwan earthquake, performance remained comparable (73 % accuracy, 3 % risk, 24 % cost). The approach is intended to prioritize structures for detailed assessment and early intervention; its applicability is limited to buildings whose attributes fall within the training data domain in terms of statistical properties.
Keywords: Reinforced concrete; Seismic vulnerability; Machine learning; Support point method; Rapid assessment