Estimation models of heating energy consumption in schools for local authorities planning

• The annual heating energy consumptions of eighty school buildings are analysed.
• Two energy estimation models were developed to support public authorities planning.
• A multiple regression model was built using nine different influencing variables.
• CART enables also non-expert users to extract information for decision making.
• MAE, RMSE and MAPE were calculated to compare the performance of estimation models.

Large building stocks should be well managed, in terms of ordinary activities and formulating strategic plans, to achieve energy savings through increased efficiency. It is becoming extremely important to have the capability to quickly and reliably estimate buildings’ energy consumption, especially for public authorities and institutions that own and manage large building stocks. This paper analyses the heating energy consumption of eighty school buildings located in the north of Italy. Two estimation models are developed and compared to assess energy consumption: a Multiple Linear Regression (MLR) model and a Classification and Regression Tree (CART). The CART includes interpretable decision rules that enable non-expert users to quickly extract useful information to benefit their decision making. The output of MLR model is an equation that accounts for all of the major variables affecting heating energy consumption. Both models were compared in terms of Mean Absolute Error (MAE), Root Mean Square error (RMSE), and Mean Absolute Percentage error (MAPE). The analysis determined that the heating energy consumption of the considered school buildings was mostly influenced by the gross heated volume, heat transfer surfaces, boiler size, and thermal transmittance of windows.