Forecasting Summer and Winter Heat Loads of Buildings Using an Artificial Neural Network

The study employed a simplified approach using a one-dimensional and superpositional method to elucidate the contributions of different factors to the overall heat load. An artificial neural network (ANN), specifically a multi-layer perceptron, was chosen due to its ability to capture complex, nonlinear relationships among diverse datasets and its efficiency in model development. To assess the accuracy of the model, three evaluation parameters were employed: the coefficient of determination (R2), the Pearson coefficient (r), and the mean absolute error (MAE). The results indicated that the model achieved a remarkable accuracy of 99% for both summer and winter heat load predictions. Both the coefficient of determination and the Pearson coefficient approached the ideal value of 1, signifying high accuracy in predicting heat loads across seasons. This finding holds significant implications for effectively managing heat loads throughout the year, thereby contributing to improved energy efficiency in buildings. Furthermore, the developed model demonstrates potential scalability for incorporating additional parameters in future applications.