摘要
利用热分析(TG-DTG)数据建立了煤粉着火稳定性指数CI,它是煤粉着火温度与燃烧强度的综合反映。采用遗传算法(GA)对BP神经网络结构进行了优化,获得了影响煤粉着火稳定性指数CI的主要煤质指标(Mad、Aad、Qnet,ad、Oad、焦渣特征CRC)和最优BP神经网络的隐层数、神经元数、激活函数,建立了煤粉着火稳定性指数的优化BP神经网络预测模型。对20个校验样本进行了预测,得到了较高的预测精度。
An ignition stability index for pulverized coal （CI） has been contrived with IG-DTG analysis data. It comprehensively reflects the pulverized coal＇s ignition temperature and its combustion intensity. After optimizing, the BP network＇s structure with the help of genetic algorithm, the main coal indices （Mad, Aad, Qnet,ad, Oad, coke slag characteristics） and the optimized BP neural network＇ s concealed layers, neural cell number, as well as its activation function are obtained and the optimized BP neural network prediction model for predicting ignition indices of pulverized coal formed. Checks with 20 samples showed relatively high prediction precision.
出处
《动力工程》
CSCD
北大核心
2006年第1期81-83,115,共4页
Power Engineering
关键词
动力机械工程
预测模型
遗传算法
BP神经网络
煤粉
着火稳定性
power and mechanical engineering
prediction model
genetic algorithm
BP neural network
pulverized coal
ignition stability
