Load prediction and optimal control for green buildings with ice storage systems

Load forecasting models have been presented for a central energy plant in Tianjin based on time sequence technologies. By optimizing parameters and correcting error of models, the short-term forecasting load for energy plant has been obtained. On the basis of the forecasting load, the optimal objective function and state parameters models of ice storage tanks have been developed to determine the minimum operation costs of ice storage system. With constraints for storage discharging and charging rates and rated capacity of chillers, dynamic programming of multi-objective optimization was used to obtain optimal control strategies of 100%, 80%, 60%, 40% peak load.