On a data-driven environment for multiphysics applications

The combination of the recent advances in computational and distributed sensor network technologies provide a unique opportunity for focused efforts on high confidence modelling and simulation of multiphysics systems. Responding to this opportunity, we present in this paper the architecture of a data-driven environment for multiphysics applications (DDEMA) as a multidisciplinary problem solving environment (MPSE). The goal of this environment is to support the automated identification and efficient prediction of the behavioral response of multiphysics continuous interacting systems. The design takes into consideration heterogeneous and distributed information technologies, coupled multiphysics sciences, and sensor originating data to drive and to steer adaptive modelling and simulation of the underlying systemic behavior. The design objectives and proposed software architecture are described in the context of two multidisciplinary applications related to material structure design of supersonic platforms and fire/material/environment interaction monitoring, assessment and management. These applications of DDEMA will be distributed over a highly heterogeneous networks that extend from light and ubiquitous resources (thin portable devices/clients) to heavy GRID-based computational infrastructure.