Felgner, F; Agustina, S.; Cladera Bohigas, R.; Merz, R.; Litz, L.:
Simulation of Thermal Building Behaviour in Modelica,
Proceedings of "2nd International Modelica Conference", pp. 147-154, Deutsches Zentrum für Luft- und Raumfahrt, Oberpfaffenhofen, Germany, March 2002.

During the past decades heating and air conditioning systems were usually designed and consequently oversized according to simplified, mostly static calculating procedures. The increase in primary energy costs, rising cost pressure felt by private and public clients as well as increased demands on comfort forced engineers to change the customary procedure. Thus the dynamic simulation of building and system behaviour plays an increasingly important role in planning and dimensioning heating and air conditioning systems. This change is supported by the growing performance of personal computers in use. This means that calculating methods which used to be too expensive and time-consuming became practicable and could even be improved.

Building and system simulation aims at emulating the thermal and energetic behaviour of an existing or a fictitious building and of its HVAC system as well as their interaction. For this purpose the external influences through the outdoor climate, user behaviour and internal loads are to be taken into account. The comprehensive building design requires the adequate description of real processes within a broad spectrum of mathematical, physical and engineering disciplines. The model of just an uncomplicated heating system includes various components from thermodynamics, fluid dynamics, mechanics, electrical and control engineering.

It is true there is a great variety of simulation tools - mostly conceived for architects and building engineers - varying according to the methods they use, the effects they consider as well as to their objectives. Such simulation tools pretend to offer a high transparency and flexibility through their menu-guided modelling but can often not be completely overlooked by the user as to their numeric methods, the effects considered and approximations applied. Operations going beyond what is provided by the menu are either not possible or can only be realized at great expense.

Therefore, we intended to take another way. Using an open simulation system, which provides the mathematical formalism, the model specification is done by the description of basic physical laws describing the relevant properties. An object-oriented, non calculation-causal simulation language like Modelica offers perfect conditions for this concept.
In the context of our work a model library for the simulation of thermal building behaviour has been developed in Modelica. Due to the interdisciplinary character of building simulation this domain is an ideal application of Dymola/Modelica. The new model library is divided into four sublibraries:

• Building (chapter 2),
• Weather (chapter 3),
• Heating (chapter 4),
• Controller (chapter 5).

Keywords: thermal building modelling, object-oriented modelling, Modelica.