Digital Appendix of "Surrogate Modeling of Structural Seismic Response Using Probabilistic Learning on Manifolds"

This is the digital appendix of "Surrogate Modeling of Structural Seismic Response Using Probabilistic Learning on Manifolds".

Probabilistic Learning on Manifolds (PLoM) is used as an alternative method of predicting seismic response of building structures. The PLoM algorithm proposed by Soize and Ghanem is implemented into an open-source Python
package, PLoM and SimCenter tools (quoFEM and EE-UQ). Two use cases are provided to validate and demonstrate the power of the PLoM software package applied to seismic response estimation. The first use case investigates the use of PLoM for predicting responses of buildings with different design parameters. The second use case investigates the use of a PLoM model for predicting building responses under site-specific seismic hazards.

The appendix contains two folders corresponding to the input data files used for the two use cases.
1. "Design&ModelingParameters" folder contains two subdirectories: (1) "EE-UQ_InputFiles" (input files and json configuration files that can be loaded by SimCenter EE-UQ application to automatically create the training and test building models), and (2) "AnalysisResults" (csv files of the response simulation data generated from EE-UQ and used as the input training data and test data sets for the PLoM).
2. "SiteSpecificGroundMotions" folder contains two subdirectories: (1) "OpenSeesModel" (contains the tcl scripts for creating the 12-story frame numerical model in OpenSees as well as the multi-stripe analysis ground motion sets and IDA ground motion set), and (2) "AnalysisResults" (contains three subdirectories for "IDA", "MSA_LosAngeles", and "MSA_SanFrancisco" results).