Economic loss assessment of precast parking structures based on seismic risk analysis and machine learning techniques

Abstract Seismic risk and the probability of seismic economic loss of precast parking structures are of great significance for earthquake resistance and disaster prevention before earthquakes, as well as emergency aid and repair after earthquakes. A probabilistic seismic economic loss assessment framework was established within the design reference period of precast parking structures, using a modified SPO2IDA method considering the spatial torsion effect of long‐span ramps to perform fragility and risk analysis. A precast‐specific Collapse Margin Ratio for Precast Parking Structures was proposed based on connector failure characteristics, and the probabilistic seismic economic loss including direct and indirect operational losses was evaluated. Furthermore, a multi‐model fusion prediction model was developed to improve the accuracy of economic loss prediction, and Bayesian inference was introduced to quantify the prediction uncertainty. Finally, a stiffness‐cost‐loss coupling optimization model for mechanical connectors was established to analyze the cost‐effectiveness of different stiffness connectors. The results showed that under the premise of fixed main structural configuration and connection layout, the use of rigid connectors can reduce the structural collapse probability by more than 70% and the expected total economic loss (including direct and indirect losses) by more than 50%. The additional cost of rigid connectors accounts for a limited proportion of the total structural cost (approximately 10%), presenting notable life‐cycle economic benefits within the study scope. The proposed modified analysis methods, machine learning models, and stiffness‐cost‐loss optimization model can provide a reliable reference for seismic economic loss assessment and connector selection of precast parking structures.