Master's Thesis Alexander Kümpel

 

Exergy-based Control of a Building Energy System using Mixed Integer Linear Programming (MILP)

Due to the rising energy demand and resulting CO2 emissions, measures for the efficient handling
of energy are necessary. Heating, ventilation and air conditioning of buildings account for a large
amount of total primary energy consumption. In order to reduce the energy consumption of buildings,
new control strategies are required.
One concept for the assessment of complex thermodynamic systems is the exergy analysis. Exergy
indicates the maximal amount of energy convertible work. In contrast to energy, exergy can be destroyed.
Hence, the minimization of exergy destruction provides a means of increasing a system’s
efficiency.
In this work an exergy-based control method is developed. Furthermore, a model predictive control
strategy is employed. Model predictive control is a computer-based method which needs a mathematical
model for minimizing an objective function. In order to achieve short calculation periods,
mixed integer linear programming is used for the model predictive controller.
The exergy-based model predictive controller is developed for a generic building. Based on a thermodynamic
analysis of the generic building, a model of the building’s energy system is constructed.
The derived model is transformed into a mixed integer linear problemby linear approximation
of nonlinear equations. Furthermore, an objective function is defined considering both exergy destruction
and temperature deviations from the set points.
The developed controller is validated in a simulation and compared with a mode-based control
strategy. The exergy-based model predictive control saves up to 23% of energy compared to the
mode-based control. In addition, the developed controller is implemented in the E.ON ERC main
building for the test of a real-world application. The exergy-based controller is able to maintain
target temperatures and it makes reasonable decisions for reducing the exergy destruction.