| 引用本文: | 杨冬锋,周苏荃,魏剑啸,张实.基于MPC的超短期优化调度策略研究[J].电力系统保护与控制,2015,43(11):21-26.[点击复制] |
| YANG Dongfeng,ZHOU Suquan,WEI Jianxiao,ZHANG Shi.Ultra-short term optimal dispatch method based on MPC[J].Power System Protection and Control,2015,43(11):21-26[点击复制] |
|
| 摘要: |
| 为实现多时间尺度协调调度模式，基于模型预测控制(MPC)理论对超短期优化调度策略进行了研究。以系统发电机组运行成本和排放量费用最小为目标函数，通过求解前瞻时间内的多时段优化问题，为实时调度提供初始策略，以实际调度结果和新的预测信息作为反馈信息进行滚动优化调度。该方法将日前调度时间尺度和实时调度时间尺度的信息联系起来，可有效应对预测信息波动对系统调度的影响。算例仿真结果表明：考虑多时段整体优化，可以更好地在机组之间合理的优化分配负荷，提高整个火电厂运行的经济性。同时验证了MPC方法的鲁棒性和收敛性。 |
| 关键词: 有功调度 实时调度 多时间尺度 模型预测控制 |
| DOI：10.7667/j.issn.1674-3415.2015.11.004 |
| 投稿时间：2014-10-20修订日期：2014-12-01 |
| 基金项目: |
|
| Ultra-short term optimal dispatch method based on MPC |
| YANG Dongfeng,ZHOU Suquan,WEI Jianxiao,ZHANG Shi |
| (School of Electrical Engineering and Automation, Harbin Institute of Technology, Harbin 150001, China;Xuji
Transformer Co., Ltd., Xuchang 461000, China;Beijing Jingqiao Thermal Power Co., Ltd., Beijing 100067, China) |
| Abstract: |
| To realize the multiple time-scale coordination dispatching mode, the extended short-term optimal dispatching strategy based on the model predictive control method (MPC) is studied. The object functions of the model are the minimizations of the units’ cost and emission expense. The initial strategy of the real-time dispatching is decided by solution of the multi-period optimal problem in the periods ahead, and then the real dispatching condition and the new forecasting information being taken as the feedback information are used to execute circling optimal dispatch. The day-ahead dispatch time-scale and the real-time time scale are cooperated to deal with the effect of forecasting information fluctuation on system dispatch. The case study results show that considering the multiple period optimizations as a whole could allocate load among the units reasonably and improve the economy of the power plant operation. At the same time, the robustness and convergence of MPC method is testified. |
| Key words: active power dispatch real time dispatch multi-time scale model prediction control |
