On a cost-effective integrated framework of earthquake-induced fragility assessment and GIS mapping concerning slopes near nuclear facilities

Seismic-triggered slope failures pose a potential threat to the critical infrastructure. This danger increases near nuclear power plants (NPPs) on steep slopes. However, uncertainties in seismic disasters and the nonlinear soil behavior make it challenging to predict slope failure vulnerability, and an accurate prediction also demands considerable numerical cost. Thus, this paper introduces a cost-effective framework for evaluating slope fragility near NPPs to address these issues. The proposed framework initiates a probabilistic sampling of soil parameters to cover wide-ranging spatial data, including statistical distribution. Geospatial analysis is utilized to deal with spatial variability. This framework identifies high-vulnerable areas by producing ensemble maps that depict slope stability and susceptibility for multiple return periods. Furthermore, fragility analysis establishes an explicit limit state based on field observation data and evaluates failure probabilities. Subsequently, the study employs a multiple linear regression (MLR) model to efficiently estimate probabilistic seismic performances of slopes under different slope conditions. Geographic Information System (GIS)-based visualization combined with MLR efficiently creates slope seismic fragility maps, enabling clear identification of the most likely vulnerable areas. Consequently, GIS map creation using the MLR model offers 99 % accuracy while achieving about 97 % computational cost efficiency compared to adopting detailed fragility evaluation.