基于MSVR的MCFC/GT发电系统最优运行轨迹预测研究

PREDICTION RESEARCH ON OPTIMIZED OPERATION TRAJECTORY OF MCFC/GT HYBRID POWER GENERATION SYSTEM BASED ON MSVR

利用离线优化运行数据,基于多输出支持向量机回归(MSVR)方法建立指导混合系统高效运行的预测器(函数逼近器),实现根据系统负荷情况实时给出系统高效运行的最佳设定值和前馈控制输入值。实验表明基于MSVR的预测器能预测不同负荷下系统最优运行轨迹,实时指导系统运行。

With the available off-line optimization data, the multi-output support vector machine regression （MSVR） predictor was developed for online estimation optimal setpoints and feedforward control inputs for the hybrid power generation system according to the reahime load. The simulation tests show that the developed MSVR predictor is capable of predicting the setpoints and the feedforward controller inputs, and can timely guide the optimizing operation.