A novel framework integrating downhole array data and site response analysis to extract dynamic soil behavior

Seismic site response analysis is commonly used to predict ground response due to local soil effects. An increasing number of downhole arrays are deployed to measure motions at the ground surface and within the soil profile and to provide a check on the accuracy of site response analysis models. Site response analysis models, however, cannot be readily calibrated to match field measurements. A novel inverse analysis framework, self-learning simulations (SelfSim), to integrate site response analysis and field measurements is introduced. This framework uses downhole array measurements to extract the underlying soil behavior and develops a neural network-based constitutive model of the soil. The resulting soil model, used in a site response analysis, provides correct ground response. The extracted cyclic soil behavior can be further enhanced using multiple earthquake events. The performance of the algorithm is successfully demonstrated using synthetically generated downhole array recordings.