382 / 2018-12-18 14:20:56

Bed Temperature Prediction Model of Circulating Fluidized Bed Boiler Based on BP Neural Network

BP neural networks; Bed temperature; Sensitivity analysis

全体主题 > 锅炉

基于BP神经网络的循环流化床锅炉床温预测模型
朱琎琦，肖显斌
(华北电力大学 生物质发电成套设备国家工程实验室，北京市 102206)
摘要：床温是直接影响循环流化床锅炉安全经济高效运行的重要参数。本文基于BP神经网络，得出以 CFB 锅炉各煤质参数、给煤量、烟气含氧量等参数为输入，床温为输出的9输入1输出系统模型。利用我国某150MW电厂实际运行数据对模型进行检验，结果表明，模型的拟合效果良好，预测值的离散程度较小，R2=0.9882。利用Garson算法进行敏感性分析，结果表明，煤的收到基低位发热量是最敏感的因素，对床温影响最大。
关键词：BP神经网络；床温；敏感性分析

Bed Temperature Prediction Model of Circulating Fluidized Bed Boiler Based on BP Neural Network
Zhu Jinqi，Xiao Xianbin
(National Engineering Laboratory of Biomass Power Generation Equipment，North China Electric Power University，Beijing 102206)
Abstract:　Bed temperature is an important parameter that directly affects the safe, economical and efficient operation of circulating fluidized bed boilers. Based on BP neural network, a 9-input (CFB boiler coal quality parameters, coal supply, flue gas oxygen content, etc.) and 1-output (bed temperature) system model was obtained. The model was tested by the actual operation data of a 150MW power plant in China. The prediction results showed that the model had a good fitting effect and the dispersion of the predicted value was small, with R2 equal to 0.9882. Sensitivity analysis conducted by the Garson Algorithm showed that the low calorific value of coal was the most sensitive factor and had the greatest impact on bed temperature.
Key words: BP neural networks; Bed temperature; Sensitivity analysis