Artificial neural network for the control of the openings and cooling systems of the double skin envelope buildings

This study aimed at developing an artificial neural network (ANN)-based temperature control method for the double skin envelope buildings. For this, control logic for opening conditions of the inner and outer surfaces' openings as well as for cooling system's operation was developed based on the predictive and adaptive ANN model. The parametrical optimization process for the structure and learning methods of the ANN model was conducted in terms of the number of hidden layers, the number of neurons in the hidden layers, learning rate, and moment. Then, the performance of this optimized model was tested using the similarity analysis between the predicted values from the ANN model and the measured values from the actual double skin envelope building. Analysis revealed that the developed ANN model proved its prediction accuracy and adaptability in terms of stable Root Mean Square (RMS) and Mean Square Error (MSE) values. Based on this finding, it can be concluded that the developed ANN model showed potentials to be successfully applied to the temperature controls for the double skin envelope buildings.