| dc.description.abstract | Energy-usage in buildings is responsible for a large part of the total demands on energy production. Models are required to estimate the heating and cooling times of buildings, thus allowing a control system to accurately maintain comfort temperature only when strictly needed, lowering the demand for energy used for heating. Grey-box models based on Thermal Network Resistor-Capacitor equivalents are used to predict thermal behavior of buildings. Model structures are based on cognitive or intuitive understanding of thermodynamic behavior of buildings. Working with models and data sets requires software tools both for treatment of data, simulation of models and identification of parameters. In this project software is developed both in c# and MATLAB as and when applicable.
Grey-box models are shown to accurately predict temperature over the prediction horizon, leading to accurate estimation of heating time, when compared to measurement data. Further, the proposed control strategy is shown to overcome some of the shortcomings of standard heater control systems. | nb_NO |
