Machine Learning Methods for Estimating Energy Performance of Building Facade Systems

Building facade systems play a critical role in conserving building energy since they control heat transfer between outdoor and indoor environments. Energy inefficiency in the façade systems adversely affects building energy performance. One of the most effective methods of improving building energy efficiency is the energy-conscious design of building facade systems, which requires intensive energy performance evaluations. In this chapter, we investigate the applicability of various machine learning methods, such as support vector machines and gradient boosting machines, in estimating the energy savings of innovative facade systems as the substitute of conventional facade panels. Nine different machine learning methods are used to estimate the heating, cooling, and total energy savings of facade systems. These models are compared based on their accuracy measures and the time required for hyperparameter tuning and training. The results showed that the tree-based models, especially random forest and gradient boosting machines, outperformed linear regression models and support vector machines. However, tree-based models require more time for training compared to linear regression models and support vector machines.