BEEF-UP: Building Energy Efficiency Using Pre-cooling - overview
The United Nations estimates that buildings account for about 40% of global energy consumption. In commercial buildings in most developed nations, the heating, ventilation and air conditioning (HVAC) systems consume between 40% and 70% of the building’s total energy demand. As a result, energy bills are often one of three highest operating expenses of a building.
IBM Research – Australia has addressed this problem with an IoT-based solution called BEEF-UP (Building Energy Efficiency Using Pre-cooling). The solution relies on a simplified “grey-box” model of the building that has been trained using machine learning by data from historical building IoT sensor information combined with weather information. This results in a predictive model of how the building will respond to HVAC settings for the next day, assuming a given weather forecast.
BEEF-UP operates as a service delivered over Bluemix. Based on the weather forecast, the service provides an optimal plan for the building’s HVAC operation for the next day. This plan may involve pre-cooling the building before occupants arrive, when the cost of energy is low, and benefits from other predictive capabilities that arise from the grey-box model. The system is currently in testing with a client in Queensland, Australia, and preliminary results show a potential 30% savings in energy costs.