Algorithmic and declarative modelling of a greenhouse. In 5th International Building Physics Conference, 2012. Kyoto University.
abstract   bibtex   
Greenhouses contain physical processes of heat and mass transfer that are not accounted for in typical building energy simulation programs, especially those involving the interaction between plants and the indoor environment. This includes convective, radiative, as well as latent heat exchange. The primary aim of this paper is to investigate modelling strategies for analyzing the historic greenhouses at the Kew Botanical Gardens in London for the purpose of evaluating appropriate retrofits for energy savings. Two modelling and simulation paradigms are explored in this paper: A general greenhouse model, adapted from the Gembloux Dynamic Greenhouse Climate Model (GDGCM) of Pieters and Deltour, is implemented in two independent platforms, one being algorithmic MATLAB code and the other being the acausal declarative modelling language Modelica. The paper describes the key differences between the two modelling paradigms in terms of following features: (a) adequate elucidation of the physical behaviour, (b) numerical behaviour of the model in terms of stability, (c) flexibility for model improvements and extensions, (d) relevance to retrofit objectives. Results from the two models are compared to measured data from the particular greenhouse. These comparisons help understand the relative accuracy of the two models and are also indicative of areas where model resolution needs improvements.

Downloads: 0