Time series prediction method based on sub-metering in building energy performance evaluation

Sub-metering of building energy consumption is a primary requirement for energy efficiency evaluation. However, the traditional method uses total energy consumption to evaluate the building energy performance, which cannot accurately analyze the energy saving of each sub-item. Therefore, this paper proposes an energy performance evaluation method based on time series prediction of sub-metering. First, a time series one-step ahead energy consumption prediction model is developed. Second, the energy consumption evaluation method is proposed based on the prediction error at each step and the overall prediction error. Third, five campus buildings in northern China are used as examples to validate the proposed method. The results show that the energy consumption sub-evaluation strategy is more conducive to improving energy use efficiency and exploring the energy-saving potential of the buildings. In addition, the evaluation method is based on a predictive model with a customized evaluation method for different sub-items of each building, which is very practical and versatile in practical use and provides an effective tool for the development of building energy efficiency work. •An energy consumption sub-metering model is proposed.•The proposed model is designed for energy performance evaluation.•The sub-item energy consumption prediction results are better than the total energy consumption prediction.•A generic energy consumption evaluation method is proposed.