| dc.description.abstract |
The environmental impact of the building industry is significant. The construction industry constitutes a major part of the world's total energy consumption. As a result, building designers have constantly been urged to pay attention to the energy economics of buildings. Green building practices like passive solar building design, advanced construction, and building operation practices have been evolved over the years. Passive solar architecture basically refers the usage of structural and non-structural elements of the building for the comfort conditions with no additional operational costs. It makes the best possible use of the local geographic and climatic conditions. Such measures help in decreasing the operational costs of a building and increasing the thermal comfort of its occupants often leading to enhanced productivity. These energy efficiency measures also improve building marketability as the cost of operation is reduced. The temperature inside a building is one of the factors to identify the comfort level up to a great extent. Hence, it necessary to predict the internal air temperatures of any building. With this study, an attempt has been made at developing mathematical models that could predict the air temperature inside any building. The works attempts the development of a neural network regression tool which can predict the temperature inside a simple construction when fed with various building specifications as input. The results of this Artificial Neural Network were found to be in close agreement with the actual on-site recorded data. This is the first time that ANN is being used for such application. For the purpose of validation and testing, the data recorded for four rooms varying in architectural aspects over the past year have been taken. These recordings were taken on an hourly basis. |
en_US |