A. Thanachareonkit, M. Andersen and J.-L. Scartezzini

Solar Energy and Building Physics Laboratory (LESO-PB)
Swiss Federal Institute of Technology in Lausanne (EPFL)
CH — 1015 Lausanne (Switzerland)

ABSTRACT

Scale models represent a standard method for the assessment of daylighting performance in buildings. Recent studies however pointed out their general tendency to overestimate work plane illuminances and daylight factors. The cause of this inaccuracy between real buildings and scale models performance is due to several sources of experimental errors, such as modelling of building details, surfaces reflectance, glazing transmittance, as well as photometers features. To analyse the principal sources of errors, the comparison of a 1 : 1 daylighting test module and its 1 : 10 scale model, placed within identical outdoor conditions, were undertaken. Several scale model parameters were modified in order to determine their impact on the performance assessment, comprising the accurate mocking — up of surfaces reflectance, the model outdoor location, as well as the photometric sensors properties. This experimental study showed that large ranges of discrepancies between daylighting performance figures assessed in both ways can occur, some of them being even caused by slight differences of surfaces reflectance and photometers cosine responses. These discrepancies can be curbed down to a 20 % relative divergence, providing that enough time and effort is spent to mock-up the geometrical and photometrical features of the real building.

Keywords : Daylight factors, work plan illuminance, scale models, experimental errors.

INTRODUCTION

Daylighting contributes in essential way to buildings atmosphere and visual amenity of inhabitants. It is generally preferred to electric lighting, offering simultaneously a visual access to outdoor and contributing to sustainable development through displacement and saving of electricity into buildings (IEA, 2000). Scale models represent a standard method for the assessment of daylighting performance in buildings (Schiler, 1987), over passing computer modelling for certain practical daylighting design (Compagnon, 1993). They are usually used to mock-up real buildings and placed within sun and/or sky simulators to proceed to the assessment of their daylighting performance (Michel et al, 1995). Recent studies however pointed out their general tendency to overestimate work plane illuminances and daylight factors in comparison to the figures observed in real buildings.

The main causes of these inaccuracies were identified by former authors (Love and Navvab, 1991), as due to inappropriate model details construction, such as window frames and surfaces reflectance, leading even to parasitic light into the scale model; the calibration of photometric sensors, as well as their size, levelling and placement in the model, were also mentioned as a source of experimental errors. A more recent study

(Cannon Brookes, 1997), confirmed the former one, pointing out other physical factors, such as building maintenance and dirt, as contributors to these discrepancies.

In order to carry out a detailed analysis of the physical parameters responsible for the overestimation of daylighting performance of buildings in scale models, a comparison of illuminances and daylight factors, monitored within a 20 m2 single office room equipped with side lighting windows (a test module) and its 1:10 scale model placed within identical outdoor daylighting conditions, were undertaken. Moreover, the photometric sensors features, which were used in the test module and in the scale model, were compared. The principal causes of the performance overestimation were identified through this comprehensive experimental study : their understanding should lead to the elaboration of more appropriate roadmaps for architects and lighting designers in their assessment of daylighting performance in buildings.