全部 标题 作者
关键词 摘要

OALib Journal期刊
ISSN: 2333-9721
费用:99美元

查看量下载量

相关文章

更多...
化工学报  2012 

基于对称Alpha稳定分布概率神经网络的铝电解槽况诊断

DOI: 10.3969/j.issn.0438-1157.2012.10.027, PP. 3196-3201

Keywords: 对称Alpha稳定分布,概率神经网络,故障诊断,铝电解槽,概率密度函数

Full-Text   Cite this paper   Add to My Lib

Abstract:

在铝电解槽非稳态情况下,槽参数易发生局部突变,呈现非高斯概率分布,且各种槽参数相关性较强,无法满足概率神经网络中训练样本必须服从独立同分布的假设条件,影响槽况诊断的精确度。提出一种基于对称Alpha稳定分布概率神经网络的铝电解槽况诊断方法,利用其对非高斯分布数据的良好近似拟合能力,改进模式层的径向基函数,提高概率神经网络对槽参数局部突变的适应性。通过取自某厂170kA大型预焙槽的样本进行检验表明,该方法能够对5种槽况做出正确的诊断,具有较强的分类精度和收敛速度。

References

[1]  Li Hesong(李贺松),Yin Xiaobao(殷小宝),Huang Yongbo(黄涌波),Ding Liwei(丁立伟),Jiang Changwei(姜昌伟).Diagnosis system of different status of aluminum reduction cells based on anode current fluctuation[J].CIESC Journal(化工学报),2011,62(6):1770-1777
[2]  Yin Ensheng(殷恩生).The 160kA Center Prebake Aluminum Cutting Production Technology and Management(160 kA中心下料预焙铝电解槽生产工艺及管理)[M].Changsha:Central South University Press,2004
[3]  Jiang Shaolong(姜绍龙).Research on intelligent health diagnosis system for aluminum electrolysis cells.Beijing:Tsinghua University,2005
[4]  Specht D F.Probabilistic neural networks[J].Neural Networks,1990,3(1):109-118
[5]  Cai Qulin(蔡曲林),Liu Puyin(刘普寅).A new supervised learning algorithm for probabilistic neural network[J].Fuzzy Systems and Mathematics(China)(模糊系统与数学),2006,20(6):83-87
[6]  Li Chunfang(李春芳),Liu Lianzhong(刘连忠),Lu Zhen(陆震).Probabilistic neural network based on data field[J].Acta Electronica Sinica(China)(电子学报),2011,8:1739-1745
[7]  Samorodnitsky G,Taqqu M.Stable Non-Gaussian Random Processes:Stochastic Models with Infinite Variance[M].London:Chapman and Hall,1994
[8]  Mehdi K,Mehdi B,Gholam A R A.Improvement of auto-regressive integrated moving average models using fuzzy logic and artificial neural networks(ANNs)[J].Neurocomputing,2009,72(4/5/6):956-967

Full-Text

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133