In the first two chapters of this thesis, the novel method, developed by Mathews and Richards1, for predicting the thermal performance of buildings is introduced. The further enhancement and theoretical clarification of this method is the objective of this thesis. The method is based on a very simple thermo-flow network which models only the most important aspects of heat-flow in buildings. While Mathews and Richards based their network on analysis of the primary aspects of heat-flow in buildings, this thesis derives the simplified model by reduction from a comprehensive model. In this way, the assumptions and limitations is illuminated and the theoretical foundation of the method can be established. As a result of the investigation, a new simplified model with certain theoretical benefits is suggested. In later chapters, the method is extended and refined. Also, a new calculation procedure for finding solutions of the model is presented. In particular the method is extended to include multi-zone heat-flow, structural storage- and variable thermal systems. The new solution method is powerful, simple and efficient. This thesis contributes to the establishment of a viable tool for thermal analysis of buildings. / Dissertation (MEng)--University of Pretoria, 1990. / Mechanical and Aeronautical Engineering / unrestricted