摘要
利用3层BP神经网络对气流床粉煤气化炉进行模拟研究。以Gibbs自由能最小化方法建立粉煤气化炉数学模型的模拟结果作为BP神经网络训练数据,训练后的BP神经网络模型对模拟数据的预测准确度较好。以Shell粉煤气化炉和国内首套粉煤加压气化中试装置上的实际生产数据作为BP神经网络的训练数据,训练后的BP神经网络模型能预测实际生产数据。
3-Layer BP neural network was used to simulate entrained-flow pulverized coal gasifier. The simulated results of gasifier by the way of Gibbs free energy minimization were used as training samples of BP neural networks, and the trained BP neural network model could accurately predict these simulated data. Meanwhile, actual operating data of Shell gasifier and the first domestic pilot plant of pulverized gasification were used as training samples of BP neural networks, and the trained BP neural network model could predict actual operating data.
出处
《华东理工大学学报（自然科学版）》
CAS
CSCD
北大核心
2009年第5期688-692,共5页
Journal of East China University of Science and Technology
基金
国家重点基础研究发展计划(973计划)(2004CB217700)
长江学者和创新团队发展计划资助(IRT0620)
中石化科技项目(407077)
关键词
粉煤气化
BP神经网络
Gibbs自由能最小化
pulverized coal gasification
BP neural networks
Gibbs free energy minimization
