Project Title/ID Number | Integration of OpenSees and NEES—4192003 |
Start/End Dates | 10/1/03—9/30/04 |
Project Leader | Stephen Mahin (UCB/F) |
Team Members | Andreas Schellenberg (UCB/GS) |
F=faculty; GS=graduate student; US=undergraduate student; PD=post-doc; I=industrial collaborator; O=other
Click on images to enlarge in a new window
OpenSees provides a rich set of highly capable modules that support the nonlinear static and dynamic analysis of complex structural systems. Its advanced, object-oriented, open-source format makes it ideally suited for hybrid simulation, where some portions of a structure are modeled numerically, while others are physically tested. Many of the advanced hybrid simulation applications being considered as part of NSF’s George E. Brown, Jr. Network for Earthquake Engineering Simulation (NEES) and elsewhere require high performance analysis capabilities. Next-generation hybrid simulation techniques require an analysis framework able not only to carry out complex structural computations quickly, but also one that is able to operate transparently within a complex, real time computing and control environment. As such, OpenSees provides an ideal computing framework, and its capabilities will substantially advance the development of hybrid simulation as an effective tool for use by the earthquake engineering community.
The goal of this project is to develop, implement and demonstrate concepts for utilizing OpenSees where a structure is partitioned for the sake of hybrid simulation into multiple subassemblies. One or more of these subassemblies would represent regions of the structure to be physically tested, while other subassemblies would represent regions to be numerically simulated. The modular, object oriented format and information exchange protocols adopted by OpenSees allows its advanced algorithms for solution of the equations of motion and state-determination to be integrated into the real-time Matlab/Simulink control environment commonly used for hybrid simulation.
This project will have a tremendous impact on the range of simulations that can be contemplated using the hybrid simulation approach. It will thereby be of substantial benefit to PEER researchers as well as those involved in hybrid simulation and in the overall NSF NEES program. The project will also significantly expand the capabilities of OpenSees by contributing to its capabilities to analyze partitioned structures and to operate in a real-time operating environment.
The investigators will participate in PEER Thrust Area 4 activities.
The student has become fully aware of the capabilities of OpenSees. A series of algorithms for sub-structuring analyses were developed in Matlab and ported to OpenSees. A discrete event formulation of the hybrid simulation methodology was implemented in OpenSees in collaboration with a visiting scholar from Kyoto University (Yoshi Takanashi). This has been used on several small experiments, and is being currently being implemented using the nees@berkeley digital hybrid test control system. A series of simple validation tests to debug the software have been developed and will be executed.
It is assumed that concurrent work by researchers at PEER institutions and other NEES shared-used equipment sites will enhance general OpenSees capabilities. These separate efforts include a significant efforts at UC Berkeley to formulate real-time concepts for implementing hybrid simulation to partitioned structures, implementing communication strategies utilizing shared use memory, and developing features within OpenSees for partitioning, sub-structuring, parallel processing and so on. This project will collaborate as appropriate with investigators undertaking these other efforts and take advantage of the ensuing results where possible.
It is planned to extend the methodology to include more complex geometries of structures that can be handled and to fully integrate the OpenSees implementation into the NEESgrid architecture, including use of other experimental laboratories and computational resources. In addition, more complex validation tests will be performed in the laboratory.
This work has been adopted at Kyoto University, and is being incorporated as part of the NEES-MOST package.
Conduct a range of simple tests demonstrating the capabilities of the program to handle different classes of hybrid simulations.