Experimental and theoretical investigation of thermally insulated roof slabs

In this analysis, thermal performance of building roof elements subjected to periodic changes in temperature at a roof top was examined with and without insulating materials. The objective of this analysis is to reduce energy consumption and prevent heat flow through roofs by using various insulating materials on the roof slab. It aims to compare the thermal performance of roof slabs with and without insulation under different atmospheric temperature conditions. Further, the study seeks to accurately predict the roof bottom temperature using optimisation techniques in order to minimise experimental cost and time. Nine identical rooms were constructed and eight roofs were retrofitted with different insulating materials such as coconut shell, fly ash, Jhama brick bats, rice husks, glass fibre, rubber sheet, earthen pots and asbestos sheets. The roof of one room was kept untreated and was considered as a reference. The thermal behaviour of every innovative roof slab was analysed. Among the eight insulating materials considered in this study, the performance of the roof slab with fiberglass was found to be the best. Roof slab with fiberglass could provide temperature reduction in the room of about 6.17°C, whereas roof slab with earthen pot could provide temperature reduction in the room of 2.27°C. The time lag between the maximum roof top temperature and maximum room temperature was found to be maximum in the case of roof slab with fiberglass and minimum in the case of roof slab with earthen pot. The time lag between the maximum roof top temperature and maximum room temperature is 4 hours for roof slab with fiberglass. The variation of roof bottom temperature was predicted by five optimisation techniques such as cuckoo search, bacterial colony optimisation, group search optimisation, social spider optimisation and genetic algorithm. It was observed that the social spider optimisation technique predicted the temperature accurately and the error between the measured and predicted temperature values was minimum compared to the other optimisation techniques.