基于表达谱的时间粒度自适应调度模式

随着可再生能源渗透率不断提高，常规火电机组小时级日前调度可能因可再生能源的波动而不再适用。为此引入粒计算理论，定义并分析了风电、负荷时间序列的信息粒度表达谱，同时提出了一种基于表达谱的时间粒度自适应调度模式。该模式提供了调度时段内包含多种时间粒度的火电机组启停和出力安排，以更好地适应净负荷变化。首先，建立了电力系统信息粒子，并在分析表达谱中粒子变化规律的基础上，采用Ward法实现功率-时间粒子的横纵方向双层粒化，形成典型电力场景和自适应时间粒度，以此设置火电机组自适应运行模式。其次，利用双重粒子群方法对5，15，30，60 min时间粒度下的调度模式进行机组组合和经济调度模型的双层求解。最后，以改进IEEE 39节点互联区域测试系统为例,验证了所提调度模式的合理性和有效性。

Time-granularity Adaptive Dispatch Mode Based on Expression Spectrum

With the continuously increasing penetration rate of renewable energy, the hourly day-ahead dispatch of conventional thermal power units may no longer be suitable due to the fluctuation of renewable energy sources. Therefore, the granular computing theory is introduced to define and analyze the expression spectrum of information granularities of wind power and load time series. Meanwhile, a time-granularity adaptive dispatch mode is proposed based on the expression spectrum, which provides start up-shut down and output power arrangement for thermal power units with multi-time granularities during the dispatch period, so as to better adapt to the net load changes. Firstly, the information granularities of power system are generated. Based on the analysis of granularity change rules in expression spectrum, the Ward method is used to achieve the horizontal and vertical two-layer granulation of power-time particles and form typical power scenes and adaptive time granularity. Thus an adaptive operation mode for thermal power units is set up. Secondly, the dual particle swarm method is applied to perform the two-layer solving of unit commitment and economic dispatch models at the time granularity of 5, 15, 30, and 60 minutes. Finally, taking the modified IEEE 39-bus regional interconnection test system as example, the rationality and effectiveness of the proposed dispatch mode are verified.