基于设备状态的机组带负荷能力预测

Prediction of unit load capacity based on equipment status

满足电网实时调度要求是火电厂负荷预测得到推广应用的关键。综合考虑近期炉、机、电及其主要辅机性能其中主要相关参数包括主、再热蒸汽压力、高压缸胀差、辅机电流等,对燃煤机组的负荷进行预测。引入改进粒子群算法(PSO)用于优化SVM模型中参数,建立改进粒子群支持向量机模型(PSOSVM)。将改进PSOSVM用于某300 MW燃煤机组并与SVM进行比较。结果表明,改进PSOSVM模型和SVM模型均能够辨别出各参数与机组负荷之间的复杂关系,实现对机组负荷的预测;但改进PSOSOVM可以有效地降低SVM模型的建模误差和预测误差,而且改进PSOSVM模型比SVM模型具有更高的预测精度和更好的泛化能力。

Meeting the real-time scheduling requirements of power grids is the key to the promotion and application of thermal power plant load forecasting.Considering the performance of recent furnaces,machines,electricity and its main auxiliary machines and the main relevant parameters include,mainly reheat steam pressure,high pressure cylinder expansion difference,auxiliary machine current etc,to predict the load of coal-fired units.An improved particle swarm optimization(PSO)algorithm was introduced to optimize the parameters in the SVM model,The improved PSOSVM is used for a 300MW coal-fired unit and compared to the SVM.The improved PSOSVM will be used for a 300MW coalfired unit.The results show that the improved PSOSVM model and SVM model can distinguish the complex relationship between parameters and unit load,and realize the prediction of unit load;However,the improved PSOSOVM can effectively reduce the modeling error and prediction error of SVM model.Moreover,the improved PSOSVM model has higher prediction accuracy and better generalization ability than the SVM model.