Intelligent prediction and evaluation models for the seismic risk and vulnerability of reinforced concrete girder bridges in large-scale zones

The prediction of the seismic risk and vulnerability of bridge clusters can contribute positively to the development of large-scale regional earthquake resilience and loss models. Using artificial intelligence and machine learning techniques, a quantitative model that considers automated intelligence algorithms for predicting the seismic risk of reinforced concrete (RC) girder bridges was developed. Using earthquake risk theory and artificial intelligence methods, 1,198 RC girder bridges and 2428,407 seismic accelerations monitored by 15 typical seismic stations were processed during the Wenchuan earthquake in Sichuan Province, China, on May 12, 2008. Multiple intelligent models (support vector machine (SVM), K-nearest neighbor (KNN), decision tree (DT), and random forest (RF)) and nonlinear dynamic curves were developed on the basis of the influence of different intensity measures, and predictive parameter identification and comparative analysis were performed. Using the actual seismic risk theory of bridges and the developed automation model, an intelligent comparison confusion matrix and curve considering multidimensional prediction parameters were generated. The rationality of the developed intelligent prediction model for RC girder bridges was compared and validated via empirical seismic damage datasets.