基于PCA-BP神经网络的锅炉煤质的软测量

Research on Soft-sensing of Boiler Coal Quality Based on Pca-bp Neural Network

采用主元分析法(PCA)与BP神经网络相结合的方法,为电站锅炉入炉煤质中的挥发分和低位热值建立了软测量模型。应用主元分析法对与入炉煤质相关的运行参数进行降维处理,再将处理过后的综合变量作为BP神经网络的输入变量,方便和简化了过程数据的处理,亦使得煤质预测的精度得到了有效提高。

To combine the Principal Component Analysis and the BP Neural Network, establish a soft-sensing model for volatile and lower calorific value of boiler coal in power plant. In this paper, the PCA is applied to reducing the dimension of operational parameter interrelated with boiler coal quality, and then the processed comprehensive factors are required as the input variables of the BP Neural Network, treatment of the process data is facilitated and simplified, and the predicting precision of coal quality can be also raised effectively.