Multi-hazard assessment and mitigation for seismically-deficient RC building frames using artificial neural network models

Abstract

Non-ductile reinforced concrete building frames have seismic and blast vulnerabilities due to inadequate reinforcement detailing resulting in premature failure. One option to mitigate these vulnerabilities is the installation of a retrofit system on susceptible structures. However, differences in code-defined performance limits depending on loading type may result in a non-conservative retrofit design under multi-hazard loads. This paper presents a rapid tool for multi-hazard assessment and mitigation for the seismically-vulnerable building frames using artificial neural network models, which can rapidly generate large datasets. Using the models, energy-based performance limits for multi-hazard loading are derived, and a rapid decision-making approach for the retrofit design is developed under seismic and blast loads.