针对火力发电机组面对自动发电控制(AGC)指令的响应速度慢，爬坡率低的问题，提出了基于负荷指令智能优化分解的火储协同控制方法。首先，针对火电机组与电池储能面对二次调频响应速度不同，应用改进型自适应噪声完备集合经验模态分解AGC指令，根据分解指令的样本熵进行Kmeans自适应聚类，由此建立功率分配器，将AGC指令分为高频与低频两种指令。其次，建立火电机组与储能电池控制模型，高频指令由储能电池承担负荷。低频指令主要由火力发电机组承担负荷，储能电池承担部分快速跟踪负荷。然后，依据电池和火电机组状态作为约束条件，应用模型预测控制器自适应滚动优化控制储能电池和火电机组的出力。最后，对某1000MW机组的AGC实际指令进行仿真实验，仿真结果表明，所提策略较于传统火电机组与电池储能的控制，能有效提高AGC指令的跟踪速度，提升火电机组的爬坡率，减少跟踪指令时的震荡次数。

A thermal and energy storage coordinated control method based on intelligent optimization decomposition of load commands is proposed to address the slow response speed and low ramping rate of thermal power units facing automatic generation control (AGC) commands. Firstly, an improved adaptive noise complete ensemble empirical mode decomposition is applied to AGC commands. Considering the different responses of thermal units and battery energy storage to secondary frequency regulation. The commands are decomposed, and K-means adaptive clustering is conducted based on the sample entropy of the decomposed commands to establish a power allocator. This allocator divides AGC commands into high-frequency and low-frequency commands. Secondly, control models for thermal units and energy storage batteries are established. The high-frequency commands are undertaken by the energy storage battery, while the low-frequency commands are mainly undertaken by the thermal unit, with the energy storage battery also handling some rapid load tracking. Then, a model predictive controller is used to adaptively optimize the output of the energy storage battery and thermal unit in real time, with the states of the battery and thermal unit serving as constraints. Finally, simulation experiments are conducted on the AGC commands of a 1000MW unit. Results show that the proposed strategy effectively improves the tracking speed of AGC commands, enhances the ramping rate of thermal units, and reduces oscillations during command tracking, compared to conventional control methods.

国家杰出青年科学基金