Reduced-Order Wave-Propagation Modeling Using the Eigensystem Realization Algorithm
This paper presents a computationally efficient version of the Eigensystem Realization Algorithm (ERA) to model the dynamics of large-domain acoustic propagation from High Performance Computing (HPC) data. This adaptation of the ERA permits hundreds of thousands of output signals to be handled at a time. Once the ERA-derived reduced-order models are obtained, they can be used for future simulation of the propagation accurately without having to go back to the HPC model. Computations that take hours on a massively parallel high performance computer can now be carried out in minutes on a laptop computer.
KeywordsSingular Value Decomposition High Performance Computing Random Access Memory Hankel Matrice Markov Parameter
Unable to display preview. Download preview PDF.
- 1.Ketcham, S.A., Parker, M.W., Cudney, H.H., and Wilson, D.K.: Scattering of Urban Sound Energy from High-Performance Computations. DoD High Performance Computing Modernization Program Users Group Conference, IEEE Computer Society, pp. 341–348 (2008).Google Scholar
- 2.Cudney, H.H., Ketcham, S.A., and Parker, M.W.: Verification of Acoustic Propagation Over Natural and Synthetic Terrain. DoD High Performance Computing Modernization Program Users Group Conference, IEEE Computer Society, pp. 247–252 (2007).Google Scholar
- 5.Juang, J.-N., Cooper, J.E., Wright, J.R.: An Eigensystem Realization Algorithm Using Data Correlations (ERA/DC) for Modal Parameter Identification. Control Theory and Advanced Technology, 4, No. 1, 5–14 (1988).Google Scholar
- 6.Juang, J.-N.: Applied System Identification. Prentice-Hall, Upper Saddle River, NJ (2001).Google Scholar
- 7.Phan, M.Q., Ketcham, S.A., Darling, R.S., Cudney, H.H.: Superstable State-Space Representation for Large-Domain Wave Propagation. Proceedings of the 4th International Conference on High Performance Scientific Computing, Hanoi, Vietnam (2009).Google Scholar