Dynamic Statistical Learning with Engineered Features Outperforms Deep Neural Networks for Smart Building Cooling Load Predictions

Yiren Liu, S. Joe Qin, Xiangyu Zhao, Yixiao HUANG, Shenglong Yao, Guo Han

Keywords: Engineered Features, Dynamic Statistical Learning, Deep Neural Networks, Cooling Load Prediction.

Abstract:

Cooling load predictions for smart building operations play an important role in optimizing the operation of heating, ventilation, and air-conditioning (HVAC) systems.
In this paper, we report a cooling load prediction solution for real municipal buildings in Hong Kong set up in a recent global AI competition. We show that dynamic statistical learning models with engineered features from domain knowledge outperform deep learning alternatives with optimal efforts. The proposed solution for the global AI competition was conferred a Grand Prize and a Gold Award by the panel of internationally renowned experts. We report the results of data preprocessing based on cooling operation knowledge, feature engineering from HVAC system knowledge, and dynamic statistical learning algorithms to build the models. To search for the best model to predict the cooling load, deep learning models with LSTM and gated recurrent units are extensively studied and compared with our proposed solution.