摘要
蒸气管网是具有典型大时滞特点的非线性网络系统结构,提高管网运行预测能力,对管网的安全高效运行有很好的指导意义。贝叶斯神经网络具有良好的泛化能力和准确计算能力,在网络目标函数中引入表示网络结构复杂性的惩罚项,以便能够在训练优化过程中降低网络结构的复杂性,达到避免网络过拟合的目的。实例验证表明,模型计算结果和泛化能力均有良好表现,优于传统BP算法计算性能,可提高企业蒸气管网运行管理水平,对流程工业节能减排建设有一定的帮助。
Steam pipe network is typical non-linear network structure.It is instructive to increase predictive capacity to steam pipe network highly effective.Bayesian neural networks is well generalization and better calculate capabili-ty.A penalty term which could be interpreted as an indication of the complexity of the network was introduced into the obj ective function to present the occurrence of “overfitting”.Compared with the conventional BP neural net-work,it has the advantages of faster convergence rate,higher stability and ability for generalization.The result had certain guided signification to accelerate the construction of hybrid process energy-saving and emission-reduction.
出处
《中国冶金》
CAS
2014年第6期53-57,共5页
China Metallurgy
基金
国家高技术研究发展计划(863计划)专项基金资助项目(2012AA050215)
关键词
蒸气管网
贝叶斯神经网络
数据归一化
预测建模
steam pipe network
Bayesian neural networks
data normalization
predictive modeling
