Structured building model reduction toward parallel simulation
- Cornell University
Building energy model reduction exchanges accuracy for improved simulation speed by reducing the number of dynamical equations. Parallel computing aims to improve simulation times without loss of accuracy but is poorly utilized by contemporary simulators and is inherently limited by inter-processor communication. This paper bridges these disparate techniques to implement efficient parallel building thermal simulation. We begin with a survey of three structured reduction approaches that compares their performance to a leading unstructured method. We then use structured model reduction to find thermal clusters in the building energy model and allocate processing resources. Experimental results demonstrate faster simulation and low error without any interprocessor communication.
- Research Organization:
- Cornell Univ., Ithaca, NY (United States)
- Sponsoring Organization:
- USDOE National Energy Technology Lab; USDOE Office of Energy Efficiency and Renewable Energy (EERE), Energy Efficiency Office. Building Technologies Office
- Contributing Organization:
- Sibley School of Mechanical and Aerospace Engineering, Cornell University, Ithaca, New York, USA
- DOE Contract Number:
- EE0003921
- OSTI ID:
- 1089754
- Report Number(s):
- DOE/EE0003921-12
- Resource Relation:
- Conference: 13th International Conference of the International Building Performance Simulation Association
- Country of Publication:
- United States
- Language:
- English
Similar Records
Spatial parallelism of a 3D finite difference, velocity-stress elastic wave propagation code
An Integrated Approach to Locality-Conscious Processor Allocation and Scheduling of Mixed-Parallel Applications