Artificial Neural Networks for predicting cooling load in buildings in tropical countries

Artificial neural networks have been used for prediction of cooling loads in buildings. Only cooling load simulations were run to evaluate the performance of the network for tropical countries like India where cooling is a critical issue. The input parameters taken for the simulation were thickness of the wall, presence of insulation on the western and the southern elevation, month of the year and day of the month as these parameter can provide quick comparison for different wall thicknesses and provision of insulation to the designer at the planning stage. The output taken is the cooling load in kWh. The building is simulated for four months from June to September when the ambient temperatures are high and cooling loads are demanding. The simulation produced satisfactory results in very less time as compared to commercially available softwares where the number of input parameter are quite large and comparative study of various options is not possible.