We are aware of an issue with certificate availability and are working diligently with the vendor to resolve. The vendor has indicated that, while users are unable to directly access their certificates, results are still being stored. Certificates will be available once the issue is resolved. Thank you for your patience.

(32b) Energyplus Model-Based Predictive Control (EPMPC) Using Matlab/Simulink and Mle+

Zhao, J., Center for Building Performance and Diagnostics
Lam, K. P., Center for Building Performance and Diagnostics
Ydstie, B. E., Carnegie Mellon University

This paper proposes the EnergyPlus Model-based Predictive Control (EPMPC) system to control under floor air distribution (UFAD) in a multi-zone two stories open office building – The Center for Sustainable Landscapes – at Phipps Conservatory in the city of Pittsburgh. The building is equipped with a central air-handling unit with ground source heat pump as the cooling and heating source. The UFAD is used for the open office, conference rooms and other occupied spaces. Ceiling-based air distribution system is used for service spaces, such as restrooms, mechanical rooms, and storage rooms. The ASHRAE 90.1-2007 baseline and design phase whole building models are created using DesignBuilder and EnergyPlus program under the “Design-Build-Operate Energy Information Modeling (DBO-EIM)” infrastructure. The key of the DBO-EIM infrastructure is to use the detailed design phase energy model throughout the entire building life cycle by validating and adapting model assumptions during construction, commissioning and operation periods. Therefore, the model is expected to be “accurate” and easy-to-use for model-based building control systems in the daily operation to optimize energy and comfort. In this paper, the design phase model is applied in both the baseline and the EPMPC control. The EPMC controller includes both continuous and discrete variables and it is not possible to use optimization algorithms that rely on gradient search. In order to solve the optimization problem we developed an exhaustive search optimization algorithm within the Matlab/Simulink environment via maximum likelihood estimation+, co-simulation tool. In two weeks of winter and summer simulation using typical meteorological year 3 weather file, comparing to the baseline logic control, which is also implemented using the same co-simulation method, the EPMPC can save HVAC operation energy by 18.9%. The occupant predicted mean vote value is as good as the baseline if not better. The benefits of using the EnergyPlus model over the simplified first principle or identified linear model include: (1) it can model complex buildings and systems; (2) it can be continuously updated with little cost and effort; (3) it requires less mathematics skills to manage in daily operation once the system has been developed.