The displacement of artificial lighting with daylighting in buildings has been shown to reduce energy consumption and provide preferable spaces for building occupants. The atrium is one means of achieving daylighting objectives in medium to large scale buildings. The transmittance of daylight through atria roofs is the least understood area of atrium design, with a particular dearth of information concerning roof transmittance in real buildings. This study aims to further the current knowledge base with regard to roof transmittance through the measurement of transmittance in real buildings, and comparison to a newly proposed photoanalysis technique. This has the potential advantage of being able to assess the transmittance of existing roofs far more efficiently than physical illuminance measurement approaches. The daylight factor at points immediately beneath the roof was measured in two case study buildings, one of which was a simple a-frame, the other a space frame monopitched roof with significant over-shading from the urban context. Hemispherical photographs were taken in these buildings, and the proportion of visible sky seen at the corresponding viewpoints found using the program HemiView. The effects of changing the photograph viewpoint, quality, and using the classifying tool within HemiView were investigated. The buildings were then modelled on a computer and a comparison made with hemispherical images of the roofs derived through a rendering process. Illuminance simulations were then run using the program Radiance. The effects of the well and external obstructions were isolated, and the relationship between illuminance at the photocell point and average roof plane illuminance found. The consequence of parametrically adjusting the structure reflectance and glazing transmittance was explored, and simple relationships relating these two factors to average roof plane transmittance derived. Further roof types were modelled on a computer and analysed in the same manner. A means of relating the photographic technique to transmittance is presented, and a methodology described for application of the procedure to existing roofs. The process was demonstrated on 15 roofs in Sheffield. The thesis concludes with a summary and suggestions for future work.

This thesis investigates the effect of, various vaulted and domed roof geometries on their solar behaviour under given summer and winter conditions. Roof is the building-envelope element that is most exposed to the sun as it receives a high amount of solar radiation, which is the main cause of summer overheating in hot-arid climates. In addition, to other climatic and physical factors, indoor thermal comfort in hot-arid climates is also influenced by the intensity of solar radiation received by roof surfaces. Therefore, roof form and geometry should be designed with careful consideration to insolation parameters. Domed, vaulted, and curved roofs have been used for a long time in hot-arid regions for historical, cultural, climatic, and structural reasons. The review of previous research work showed that different explanations have been given to the climatic effects of their forms and the environmental behaviour of their enclosed spaces. The research explores the previous attempts that discussed the relevant principles of solar radiation and solar geometry on horizontal and tilted surfaces with different orientations. The previous work that applied these principles and theories to evaluate the solar behaviour of architectural elements with arbitrary forms was also investigated. In order to evaluate the solar performance of flat and curved roofs geometrical configurations, a parametric study testing the received solar radiation intensity (W/m2) on flat, vaulted, and domed roofs with different span-to-height ratios and orientations was carried out using a published solar computer model. The results of this model were followed by validation tests using other two commercially available computer tools to carry out a brief solar and thermal analysis of selected curved-roof geometries. The evaluated curved-roofs solar performance and main findings of the present research have been compared with recently published independent research. It is believed that this research establishes a sound theoretical basis for the validity of various claims of the climatic advantages of different curved-roof forms in hot-arid regions. As part of this research outcome, solar and architectural design-guidelines for curved-roofs are introduced. The research concludes with a discussion of the architectural and solar potential of curved-roof forms, which is believed to be novel contribution to the knowledge and the understanding of curved-roofs solar behaviour and architectural applications in hot-arid climates.