Application of Artificial Neural Network for Predicting the Indoor Air Temperature in Modern Building in Humid Region

Abstract

This work was aimed to apply the artificial neural network (ANNs) for predicting indoor air temperature in modern building, seven hours in advance in humid region, using as inputs only the outdoor air temperature and the last six hourly values of indoor air temperature. The building experiment is built with cement hallow block in the town of Douala in Cameroon, and the experimentation was carried out for six months. Experimental data were used to determine the optimal ANN structure with Levenberg-Marquardt algorithm by using Matlab software. The optimal structure was the multilayer perceptron (MLP) with seven input variables, thirty hidden neurons and one neuron in the output layer. The activation functions were respectively the hyperbolic tangent in the hidden layer and the linear function in the output layer. Moreover, the indoor air temperature results simulated by using the developed ANN model were strongly correlated with the experimental data. These results testified that ANN can be valuable tool for hourly indoor air temperature prediction in particular and others indoor air parameters of building, such as relative humidity, cooling loads.