District Heating Network Demand Prediction Using a Physics-Based Energy Model with a Bayesian Approach for Parameter Calibration

Heat demand of a district heating network needs to be accurately predicted and managed to reduce consumption and emissions. Detailed thermal parameters are essential for predictions using physics-based energy models, but they are not always available or sufficiently accurate. To reduce the simulation time in calibration and the dependency of accurate data of buildings, this paper develops a prediction approach using a building energy model whose parameters are calibrated by Bayesian-based calibration method to match the recorded data of monthly heat demand. In the proposed calibration approach, an emulator is established to evaluate the untested values of thermal parameters using Bayesian method, and then use the evaluation results to search for the most suitable parameters value. The designed approach greatly accelerates the calibration speed. The method is used to calibrate a single parameter and multiple parameters of the building thermal energy models for a district heating network. After it has been verified with measured data, the developed calibration method is used to calibrate parameters of building energy models. The output of the calibrated model can predict the hourly building heat demand in district heating networks.