Using Artificial Neural Network to Execute Demand Response Considering Indoor Thermal Comfort and Forecast Load-Shedding

This paper used artificial neural network to forecast the cooling load in the building in 24 hours. The unloading experiment kept the indoor thermal comfort at the ideal range of PMV=0~0.5 and PPD=5~10. Finally, dry bulb temperature, relative humidity, wet-bulb temperature and forecast cooling load were used for modeling by neural network. We can use this model to forecast how much load can be unloaded in summer peak hours accurately. This method controls the demand response for central air conditioning system, not only maintaining comfortable indoor environment, but also attaining the goals for reducing the electric expenses.