BroadBand Ground Motion Synthesis using Generative Adversarial Neural Operator with Learnable Filtering Effect
Yaozhong Shi, Grigorios Lavrentiadis, Domniki Asimaki, Zachary E. Ross, & Kamyar AzizzadenesheliPublished September 10, 2023, SCEC Contribution #13284, 2023 SCEC Annual Meeting Poster #176
We develop a data-driven framework for 3-component ground motion synthesis with the ability of learning filtering effect of the ground motion. Leveraging the increase of ground-motion data from seismic networks and recent advancements in Machine Learning. We train a Generative Adversarial Neural Operator (GANO) to produce realistic three-component acceleration time histories conditioned on magnitude (M), rupture distance (Rrup), time-average shear-wave velocity at the top 30 m (Vs30), and faulting type based on a recent published California dataset. Such dataset contains ground motion time histories with two sampling frequencies, and each ground motion time history is filtered with a band pass filter. The validation results show our model can learn the filtering effect from time histories with arbitrary sampling frequency.
Citation
Shi, Y., Lavrentiadis, G., Asimaki, D., Ross, Z. E., & Azizzadenesheli, K. (2023, 09). BroadBand Ground Motion Synthesis using Generative Adversarial Neural Operator with Learnable Filtering Effect. Poster Presentation at 2023 SCEC Annual Meeting.
Related Projects & Working Groups
Ground Motions