基于离线优化和在线决策的光伏智能楼宇能量管理算法

作为综合能源的重要发展形态,光伏智能楼宇的用电需求及光功率具有极大的不确定性,现有的能量管理方法很难完全适用,因此提出了基于离线优化和在线决策的光伏智能楼宇能量管理算法。首先,结合光伏智能楼宇的历史运行数据,建立了以运营收益最大化为目标的离线优化模型,通过离线优化为在线学习提供知识库;其次,为了实现分时电价条件下光伏智能楼宇的实时调度,建立了在线学习与认知规则相结合的在线决策算法,实时决策电动汽车充电功率以及可平移负荷的工作状况;最后,以某商业楼宇为例进行了仿真测试,结果表明所提算法在未来光功率、充电需求及可平移负荷未知的情况下具有良好的运行效果。

An Energy Management Algorithm of PV-Assisted Smart Building Based on Offline Optimization and Online Decision

As an important development form of comprehensive energy, the charging demand and PV power of PV-assisted smart buildings have great uncertainty. The existing methods of energy management of PV-assisted smart buildings are not completely applicable. An energy management algorithm of PV-assisted smart building based on offline optimization and online decision is proposed in this paper. Firstly,combined with the historical operation data of photovoltaic intelligent buildings, an offline optimization model aiming at maximizing operating income is established, which provides a knowledge base for online learning through offline optimization. Then, in order to dispatch the smart building energy in real-time, under the condition of time-of-use price, the online algorithm is established combined with online learning and rule-based. The charging power of EVs and working conditions of shiftable loads can be decided using the online algorithm. Finally, taking a business building as an example, the proposed algorithm is tested. The simulation results show that the method can be operated without future information of PV power, charging demand and shiftable loads.