Study on the Operation Characteristics and Short-term Optimal Dispatch of Energy Supply System for Near-Zero Energy Buildings

Near-zero-energy buildings are characterized by high energy utilization and low carbon emission, which gradually become the development trend of the future building industry. In this paper, the energy supply system of a physical near-zero energy building in Lanzhou area is taken as the research object to analyze the energy transfer law of each subsystem and the year-round operation characteristics, analyze the dynamic relationship between the load and the system capacity, and put forward a prediction method to forecast the future energy demand and supply, so as to realize the efficient management and optimal scheduling of the energy system. The results show that: the MAPE of the LSTM model prediction result is 5.16%, and the RMSE is 6.65kW·h; the LSTM model can predict the PV power generation better; the load is predicted by using the SVM prediction model, and the MAPE of the electric load predicted by the SVM prediction model is 3.16%, and the RMSE is 5.65kW·h; the SVM prediction model The MAPE of thermal load predicted by the SVM prediction model is 5.25%, and the RMSE is 7.15kW·h. The SVM prediction model is able to predict the load demand better; the operation cost can be reduced by 21% and 20% under snowy and cloudy conditions, and the new energy consumption rate is 90.85% and 87.39%, respectively.