Electrical consumption forecast using actual data of building end-use decomposition

• A high accuracy in the consumption forecast using end-uses decomposition is obtained.
• Number of selected days must be reduced and close to the day of prediction (DOP).
• End-use decomposition enables an optimal selection of similar days to the DOP.

The calculation of electricity consumption forecast a few days ahead is a complex issue and studies about this matter are continually being performed. Advances in this field enable obtaining consumption forecasts increasingly accurate. These consumption forecasts aim to improve the knowledge of the facilities, the planning and control of consumption and the measurement and verification of energy saving measures, among others. In this study the authors present several advances related to consumption forecast using end-use (EU) approach. In the disaggregation of the total consumption process, the correlation between energy and external variables, such as mean temperature, degree days or daylight, is studied. Additionally, an extrapolation method to obtain a total consumption forecast from forecasted EUs consumption that cover approximately 60% of total consumption is developed. With this procedure, total consumption forecasts with high accuracy can be obtained. The higher accuracy in each end-use, the better results are obtained in the total consumption forecast. For this reason, the study is focused in the end-uses disaggregation and its forecast calculation. The entire methodology is illustrated and contrasted using the consumption of the Universitat Politècnica de València (UPV).