An occupancy-based model for building electricity consumption prediction: A case study of three campus buildings in Tianjin

The accurate prediction of a building's electricity consumption can provide baselines for energy management and indicate the building's energy-saving potential. However, electricity utilization indicators based on the building area are no longer applicable because of the overall increase in the building area per person and occupant energy demand of buildings. To tackle this challenge, the building electricity consumption was split into 'basic' and 'variable' forms in this study and a two-part building electricity consumption prediction model based on human behavior was established. The basic electricity consumption is related to the building area, while the variable electricity consumption is related to the building occupancy. The probability function and Markov model were used to describe the electricity consumption caused by the randomness of occupancy in buildings. The model was validated using three campus buildings. Based on the comparison of the actual electricity bills of the campus buildings with the model prediction results, the model accuracy error is less than 5%. The results show that the building electricity consumption of a building has a growth limit when multiple people share a room, which is related to a person's initiative or ability to control the electricity use. (c) 2019 Elsevier B.V. All rights reserved.