Artificial Neural Networks for Predicting Cooling Load Reduction using Roof Passive Cooling Techniques in Buildings

Main Article Content

Shrikant Pandey
D.A. Hindoliya, Ritu Pandey


Two identical prototype rooms having dimension 1m×1m×1m, were constructed using brick work. RCC roofs having
thickness of 100 mm were constructed on both rooms. Two roof passive cooling strategies (e.g., roof pond, insulation over the
roof) were applied on one of the test rooms one by one. The other room was kept with bare roof. The results show that the average
% reduction of roof cooling load was found to be 45 %., 30%, using roof pond, using insulation (thermocol) respectively.
The objective of this work is to train an artificial neural network (ANN) to learn to predict the reduction of cooling load
of buildings. Five training algorithms traincgb, traingdx, traingda, trainlm, and trainsc were used to create an ANN model. An
ANN has been trained based on number experimental data of cooling load. The network output is reduction in heat gain though
roof. The Intelligent model predicts reduction in cooling load with accuracy, more than 90%. The accuracy of the prediction could
be improved by more input data. The results show that the predicted data is in good harmony with the experimental data, which
indicates artificial neural network is a novel and reliable method to predict reduction in cooling load.



Keywords: Artificial neural network, Cooling-load reduction; Roof cooling; passive cooling, energy saving.


Download data is not yet available.

Article Details