Author:

Keywords:

Concrete Core Activation, TABS, model based control, grey-box model, finite elements, Measurements

Abstract:

The quest towards low-energy buildings renews interest in Concrete Core Activation (CCA), because its low heating and high cooling temperatures harmonize with low-exergy production systems. However, CCA is characterized by a large thermal capacity and thus large time constant, which hampers adequate control. Model based control strategies can overcome this challenge if accurate CCA thermal models are available. But, computation time reduction in applied optimization routines compel simplified models. A sound trade-off between computation speed and accuracy needs thus to be found. In this paper a 1D controller model for an inhomogeneous CCA floor slab with air cavities is constructed. Initial model architecture is based on an available 1D-model for a homogeneous CCA slab and the physical properties of the inhomogeneous CCA slab determine the initial parameters. Firstly, experimental temperature and heat flow data from a full-scale test setup in a controlled environment are used to validate a detailed 3D finite element (FE) model of the CCA floor slab. Secondly, this 3D-FE-model generates multiple response data sets, that are used as training and validation data sets for the parameter identification of the modified 1D controller model. Compared to the 3D-FE-model, the 1D-model predicts steady state heat fluxes with an error smaller than 7.5%, while the time constants of surface heat flux, induced by a step in water supply temperature, are predicted with a maximal error of 6%. The adopted grey-box approach ensures that physics is adequately incorporated in the 1D-model creating model robustness against small dimensional variations.