Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.

  

Scenario-Based Model Predictive Control for Energy-Efficient Building Climate Control

Author(s):

X. Zhang
Conference/Journal:

Master Thesis, FS12 (10172)
Abstract:

This thesis deals with the development of an advanced Model Predictive Control (MPC) strategy for energy-efficient building climate control. There are strong incentives to reduce energy consumption in buildings because this is where roughly 40% of the world's total energy is used. Heating, ventilation and air conditioning (HVAC) systems regulate the comfort ranges in buildings, but also consume approximately half of the energy, which makes them an attractive target for saving energy. In this project, a novel technique called Randomized MPC (RMPC) is investigated, which repeatedly optimizes the building energy consumption over a fi nite control horizon based on a given number of weather and occupancy scenarios. RMPC handles uncertainties even if they are non- Gaussian and non-additive because it is based on a sampling approach. Another advantage of RMPC is that it can directly tackle the bilinearity present in the system. RMPC is applied to eight representative scenarios and validated through extensive simulations based on empirically collected data. It is demonstrated that RMPC outperforms existing control methods in the regions of interest, in particular when adverse samples are removed iteratively after solving the initial MPC problem. Furthermore, simulations have shown that only a fraction of the theoretical number of samples are required for good performance. Hence, the linear programs (LP) resulting from the RMPC formulation can be solved very efficiently. Our findings suggest that RMPC and RMPC with sample removal are attractive alternatives to existing methods for control practitioners because they are easy to tune and tractable for large-scale systems.

Year:

2012
Type of Publication:

(12)Diploma/Master Thesis
Supervisor:

G. Schildbach

File Download:

Request a copy of this publication.
(Uses JavaScript)
% Autogenerated BibTeX entry
@PhdThesis { Xxx:2012:IFA_4215
}
Permanent link