A day-ahead planning for multi-energy system in building community

A novel integrated energy system is proposed to provide stable electricity and cold energy for a multi-building community, which comprises wind turbines, solar photovoltaic panels and a liquid natural gas station as energy sources, and liquid air energy storage and packed bed for energy storage. A predictive algorithm based on long short-term memory is adopted to forecast the electricity and cooling load, wind speed and available cold energy for day-ahead scheduling. Furthermore, a comparative study of energy allocation schemes including electricity and cooling are carried out. The thermal inertia of room temperature is considered in the cooling allocation scheme. The pivotal results are presented as follows. The forecast errors are no more than 5%, proving the reliable accuracy of the following analysis. The robust schemes based on forecast method and uncertainty analysis have lower energy mismatch rates compared to the conservative schemes and radical schemes, which are 4.6% for electricity scheduling and 5.0% for cooling scheduling, respectively.