innovative predictive modelling: artificial neural networks for chilled ceiling panel

In this research, a model using a machine learning algorithm based on artificial neural networks (ann) has been developed to predict the cooling capacity of the chilled ceiling panel as the cooling, ventilation and air conditioning system. the structure of the ann architecture and the accuracy of the model based on neural net fitting are evaluated in terms of input layers, which include the number of input variables, the proportion of training data, and the number of neurons. for this purpose, energy plus is used to generate simulation data for the chilled ceiling panel system. computational results on the real data of a company state that the developed model can predict the cooling capacity of the chilled ceiling panel with an average value of 97.1% accuracy based on seven input variables, 80% training data and 8 neurons. in addition, the results of the ann model compared to the mlr model show the superiority of the proposed ann model, which can be used to better design of the chilled ceiling panel systems.