Development of an indoor air temperature predictive model for a naturally ventilated adobe mud hut; Botswana winter climatic conditions

Abstract

A transient heat transfer model to predict the indoor air temperature within a Botswana traditional hut (made of adobe/mud bricks) is developed from Fourier’s law of heat transfer, taking into consideration conduction, convection and irradiance. The predictive model was developed and analysed using MATLAB and correlated against in-situ experiments taken on July 5th 2018. They are then validated with a new data set as collected on July 7th2018. At 95% variance level the model demonstrates an excellent strength of association against actual indoor temperature at a correlation coefficient of 0.9, with an impressively low RMSE of 0.45◦C and an R-squared value of 65%, which is a fairly acceptable goodness of fit. Given the global need for energy saving, environment symbiosis and energy optimization, it is important to develop and improve natural ventilation predictive models to guide decision makers. With other environmental design methodologies, the model developed in this study should influence national housing policy towards sustainable indigenous construction; inform the development of natural building standards and indoor air quality strategies of the hot and dry climatic countries globally.