论文部分内容阅读
为了控制燃煤锅炉的NOx排放量并提高锅炉效率,对某超超临界1 000 MW机组锅炉的热态运行数据进行分析,基于支持向量回归机(SVM),建立了NOx排放和锅炉热效率的FOASVM模型,采用果蝇优化算法(FOA)对模型中的惩罚因子C、核函数参数g和不敏感损失系数ε这3个参数寻优,并与遗传算法(GA)优化参数的预测模型进行比较。结果表明,FOASVM模型的预测精度更高,泛化能力更强,其中误差最大的NOx排放模型测试集数据的平均相对误差仅3.59%,能够精准地预测锅炉热效率和NOx排放,适合于在线建模预测,为大容量锅炉的进一步优化运行提供了良好的模型基础。
In order to control the NOx emissions of coal-fired boilers and improve the boiler efficiency, the thermal operation data of an ultra-supercritical 1 000 MW boiler were analyzed. Based on the support vector regression machine (SVM), the FOASVM for NOx emission and boiler thermal efficiency Model was used to optimize the three parameters of penalty factor C, kernel function parameter g and insensitive loss coefficient ε in the model by using the fruit fly optimization algorithm (FOA), and compared with the genetic algorithm (GA) optimization parameter prediction model. The results show that the prediction accuracy of FOASVM model is higher and the generalization ability is stronger. The average relative error of FOXVM model test set data is only 3.59%, which can accurately predict the boiler thermal efficiency and NOx emission and is suitable for online modeling Predictions provide a good model base for the further optimization of large capacity boilers.