论文部分内容阅读
为了保证热力系统稳定运行,提高锅炉安全寿命,控制污染物,该文利用多模型思想,对煤种低位发热值进行初步辨识和精确辨识。初步辨识中,采用改进的K均值聚类算法,快速辨识出煤种类型;精确辨识中,利用初步辨识的结果优化发热量辨识模型,减少模型搜索范围,采用自动调节隐节点和参数的径向基函数(RBF)神经网络算法。仿真结果表明,该辨识方法的辨识误差在1.5%以内,具有良好的辨识精度,在速度上也优于单独的RBF辨识算法,可以应用于热力系统煤种发热量在线辨识。
In order to ensure the stable operation of the thermodynamic system, improve the safe life of the boiler and control the pollutants, this paper uses the multi-model idea to preliminarily recognize and precisely identify the low calorific value of coal. In the initial identification, an improved K-means clustering algorithm is used to quickly identify coal types. In the precise identification, the identification model of calorific value is optimized with the result of preliminary identification to reduce the search range of the model. By adopting the method of automatically adjusting the hidden nodes and the radial Basis Function (RBF) Neural Network Algorithm. The simulation results show that the identification error of the identification method is within 1.5%, which has good identification accuracy and is superior to the RBF identification algorithm alone in speed, which can be applied to on-line identification of coal heat in thermal system.