Optimizing Building Heating Efficiency: A Data-driven Approach for Cost and Energy Savings

This study presents a data-driven modeling approach to enhance the heating control of buildings, aiming to address the rising energy costs and growing concerns about climate change. Two distinct models based on artificial neural networks (ANNs) are employed to estimate the cost of different control scenarios and their impact on the office space temperature, using operational data from a municipal building in Norway equipped with electric and thermal energy meters. The proposed method utilizes advanced tools built based on the data from the building and weather forecasts for predictive control, which can be readily accessed for practical implementation. The study demonstrates the potential of AI methods when combined with building data to lower costs and improve efficiency. A 7% reduction in power consumption and cost over working days of a week was achieved through the optimization of the building heating system.