Cloud Accelerated Performance Based Seismic Design

Non-linear Time History Analysis (NLTHA) is a key enabler of Performance Based Seismic Design (PBSD). Arup Los Angeles office typically performs these simulations in LS-Dyna solver. In order to respond to the demands of concurrent design projects, the authors have adopted a cloud centric approach to accelerate our workflows and to enable the use of non-linear time history analysis as a design tool as opposed to a verification tool. This paper will present our custom workflow which enable a dramatic compression of the time required for these analysis. The workflow generates LS- Dyna models in parametric fashion via Rhino- Grasshopper. Since a single design iteration of analysis can result in 48 to 110 models from a range of ground motions and input parameters these models are typically executed on a compute cluster with a large number of compute cores. The resulting number of analyses generates a large amount of data (8-16TB) which we post process leveraging “Big Data” approaches typically used by other industries (financial or retail firms).