基于改进型S算子BP神经网络的钢材大气腐蚀影响因子评估模型

中国腐蚀与防护学报, 2010, 30(3): 227-230.

基于改进型S算子BP神经网络的钢材大气腐蚀影响因子评估模型

1.1. Department of Computer Science, Navy University of Engineering, Wuhan 4300332. Institute of New Weaponry Technology and Application, Navy University of Engineering, Wuhan 430033>

基金项目: 两栖装甲装备海洋环境适应性研究项目（海装办LD0701C）资助

关键词：双曲正切-S算子 , BP neural network , Bayesian-regularization , atmospheric corrosion

A CAUSE-FACTORS EVALUATION FOR ATMOSPHERIC CORROSION OF STEELS BASED ON IMPROVED SIGMOID FUNCTION BP NEURAL NETWORK

LUAN Ruipeng，BEN Kerong，XIAO Yuxing，TIAN Liye ,

Keywords： hyperbolic tangent sigmoid function , BP神经网络 , 贝叶斯正则化 , 大气腐蚀

通过对双曲正切-S算子的改进，提出了一种用于钢的大气腐蚀影响因子评估的BP神经网络模型，采用零均值标准化使输入数据符合模型要求，引入贝叶斯正则化算法解决了小样本泛化问题。仿真试验表明，该模型能在无任何先验知识的情况下较好的反映诸因子对大气腐蚀的影响。

On the basis of improved hyperbolic tangent sigmoid transfer function, a cause-factors evaluation BP neural network model for estimating the atmospheric corrosion of steels was proposed. Using the zero mean stand method to preprocess the input data, Bayesian regularization arithmetic was introduced to solve the generalization problem on sparse data. The simulated results showed that the model provided good evaluation for atmospheric corrosion cause-factors without any prior knowledge.

参考文献

[1]

上一张下一张

计量

下载量()
访问量()

作者相关

文章评分

您的评分：
1

0%
2

0%
3

0%
4

0%
5

0%

材料期刊网