Development of the heating load prediction model for the residential building of district heating based on model calibration

Abstract

Heating load prediction of district heating (DH) can give the heating demand, guide the strategies formulation and avoid excessive heat. However, existing prediction models from the supply side are to learn historical inefficient operations and have no energy-saving effect. Therefore, this study proposes the demand-side method, which predicts the heating load of terminal buildings considering influence of indoor temperature. Dynamic and steady-state white box model are taken as prediction techniques due to inapplicability of data-driven model, and model calibration is used to match the models accurately with the actual. Taking an actual residential building as the case, and the results show steady-state model has application advantages over dynamic model due to the greatly reduced calculation time under the situation of limited known information, although the accuracy of dynamic model is slightly better than that of steady-state model. The steady-state model can predict the heating load under different indoor temperature. The energy-saving rate reducing the actual indoor temperature to 18 °C is 11%–27% for different periods of a heating season. The method is a novel way to conduct load prediction for DH with energy-saving effect, and provides a meaningful basis for formulating heating strategies.