模型预测控制在1030 MW超超临界发电机组的应用

Applications of Model Predictive Control in 1030 MW Ultra-superelectrical Power Plant

介绍模型预测控制(MPC)在某1030 MW超超临界燃煤发电机组中的应用。MPC应用于该机组的协调控制(CCS)和主、再热蒸汽温度控制。所用MPC技术是模型辨识和预测控制一体化控制系统。模型辨识使用自主开发的多变量闭环系统辨识技术;预测控制使用自主开发的双层优化控制算法,且动态控制部分使用新研发的三项MPC算法。MPC系统投运后已运行一年多,投运率接近100%。与原有基于PID的控制系统相比,MPC系统显著改进负荷跟踪性能,负荷跟踪速率从0.8%/min提高到1.5%/min;同时负荷和汽压的标准差分别降低35%和38%;主汽温和再热汽温标准差降低40%以上,主汽温和再热汽温均值分别提高2℃和4℃。

The model predictive control(MPC)in a 1030 MW ultra-supercritical coal power plant was described,in which,the MPC answers for the coordinated control(CCS)of the power set and main steam and reheat steam temperature control.Of which,the MPC technology boasts of model identification and predictive control and the model identification employs independently-developed multivariate closed-loop system identification technique.The predictive control adopts self-developed bi-level optimization control algorithm,and the dynamic control part uses newly-developed three MPC algorithms.The MPC system has been running for more than half a year after it’s put into operation and the operation rate is close to 100%.Compared with original PID control system,the MPC system can significantly improve the load tracking performance,and the load tracking velocity increased from 0.8%/min to 1.5%/min.Meanwhile,the standard deviation of the load and steam pressure can be reduced by 35%and 38%,respectively and that of main steam and reheat steam temperature can be reduced by more than 40%,and the mean value of main steam and reheat steam temperature increased by 2℃and 4℃,respectively.