【摘要】 针对传统的建模方法无法建立精准的锅炉燃烧数学模型及其氮氧化物(NOx)排放预测精度不高等问题，提出一种差分量子灰狼(DEQGWO)算法来优化无迹卡尔曼滤波(UKF)的预测模型。利用不同工况下1000MW超超临界机组锅炉历史数据验证DEQGWO-UKF模型预测结果并与基本UKF模型、PSO-UKF模型进行预测精度对比。结果证明：DEQGWO算法比其它仿生算法在优化UKF预测NOx模型中有更好的预测精度和泛化能力，对火电厂SCR系统NOx预测提供有效解决方法。更多还原

【Abstract】 Aiming at problems such as the inability of traditional modeling methods to establish an accurate mathematical model of boiler combustion and the low accuracy of nitrogen oxide(NOx) emission prediction, a differential quantum gray wolf(DEQGWO) algorithm is proposed to optimize the unscented Kalman filter(UKF) prediction model. The historical data of 1000MW ultra-supercritical unit boilers under different operating conditions were used to verify the prediction results of the DEQGWO-UKF model and compare the prediction accuracy with the basic UKF model and PSO-UKF model. The results prove that the DEQGWO algorithm has better prediction accuracy and generalization ability in optimizing the UKF NOx prediction model than other bionic algorithms, and provides an effective solution for the NOx prediction of the SCR system of thermal power plants.更多还原