Theoretical study

Simulation is a powerful tool to evaluate and optimise the system design. Accuracy of the performance of any model depends on the algorithm used and the accuracy of the data used.

The model was created using standard TRNSYS v. 15.0 components [1]. TRNSYS is an acronym for a ‘transient simulation program’. The model output is the free floating indoor air temperature. Model input variables are: the global and diffuse horizontal radiation, the outdoor air temperature and relative humidity and wind speed. The model includes the parameters describing the building which mainly concern the components geometry, materials thermophysical properties and surfaces optical properties. Thermophysical properties of the building materials and glazing optical properties are selected from [3]. Model simulation has been carried out using the measured input variables averaged and under-sampled at 1 h time step.

The floor has been treated as a wall with boundary conditions connected to the outdoor air temperature through an added insulation layer. The resistance of this added insulation was calculated as a function of the path length of the heat transfer through the soil layer using a technique recommended by [1].

Infiltration considered as air change rate has been averaged by a simple, single-zone approach based on the Lawrence Berkeley National Laboratory model (Sherman and Grimsrud 1980), indicated in [1]. This calculations are subject to high uncertainty, [1] indicates that: "The model has exhibited average errors on the order of 40% for many measurements on groups of houses and can be off by 100% in individual cases (Persily 1986)”.

An equivalent homogeneous multilayer wall is used to represent the ceiling. The soil reflectivity is supposed to be 0.2 and standard values for the convective heat transfer coefficients are adopted. The indoor air temperature is supposed to be homogeneous. When simulating a building in transient regime initial conditions must be specified. A estimation has indicated three days to reach the steady simulated thermal behaviour. Therefore the first three days has not been considered for analysis.

The time evolution of the simulated indoor air temperature is presented in Figure 8, its means is 20.1 °C and its standard deviation is 1.2 °C.

There are many sources of uncertainty when using modelling to assess the thermal performance of a proposed building. Sources of uncertainty can be categories as [10]: abstraction (concessions made to accommodate the design to the computer representation, e. g. building geometrical simplifications), database (the element to be modelled may not mach the information contained in the database and assumptions have to be made), modelled phenomena (simplification on the physical processes modelled, e. g. thermal contact with the ground), solution methods (e. g. in resorting to numerical discretisation techniques a discretisation error is introduced to the solution).

Studies have shown that the uncertainties can be quite substantial on model results. Uncertainties will be analyze in future works.