CIESC Journal ›› 2017, Vol. 68 ›› Issue (3): 1032-1040.DOI: 10.11949/j.issn.0438-1157.20161613

Previous Articles     Next Articles

Soft-sensor method for effluent ammonia nitrogen based on interval type-2 fuzzy neural networks

HAN Honggui1,2, CHEN Zhiyuan1,2, QIAO Junfei1,2, ZHANG Huiqing1,2   

  1. 1 Faculty of Information Technology, Beijing University of Technology, Beijing 100124, China;
    2 Beijing Key Laboratory of Computational Intelligence and Intelligent System, Beijing 100124, China
  • Received:2016-11-15 Revised:2016-11-25 Online:2017-03-05 Published:2017-03-05
  • Contact: 10.11949/j.issn.0438-1157.20161613
  • Supported by:

    supported by the National Natural Science Foundation of China (61622301,61533002) and the Beijing Municipal Education Commission Science and Technology Development Program (KZ201410005002,km201410005001).

基于区间二型模糊神经网络的出水氨氮软测量

韩红桂1,2, 陈治远1,2, 乔俊飞1,2, 张会清1,2   

  1. 1 北京工业大学信息学部, 北京 100124;
    2 计算智能与智能系统北京市重点实验室, 北京 100124
  • 通讯作者: 韩红桂,1307441474@emails.bjut.edu.cn
  • 基金资助:

    国家自然科学基金项目(61622301,61533002);北京市教育委员会科研计划项目(KZ201410005002,km201410005001)。

Abstract:

A soft-sensor method for online detection of effluent ammonia nitrogen (NH4-N) in waste water treatment process was proposed on the basis of interval type-2 fuzzy neural networks (IT2FNN). First, actual operation data related to pre-treatment process variables was collected and process variables having strong correlation to effluent NH4-N were selected by principal component analysis (PCA) technique. Second, a self-sensor model between principal component variables and effluent NH4-N was established via IT2FNN and model parameters were adjusted by gradient algorithm. Finally, the proposed soft-sensor method was used in a real waste water treatment process (WWTP). The experimental results show that the new method can predict effluent NH4-N online with better accuracy than traditional methods.

Key words: effluent ammonia nitrogen, soft-sensor, interval type-2 fuzzy neural network, dynamic modeling, waste water treatment process

摘要:

针对污水处理过程出水氨氮(ammonia nitrogen,NH4-N)难以实时检测的问题,提出了一种基于区间二型模糊神经网络(interval type-2 fuzzy neural networks,IT2FNN)的软测量方法,建立了出水NH4-N的软测量模型,实现了出水NH4-N的实时检测。首先,采集和预处理相关过程变量的实际运行数据,通过主元分析法筛选出与出水NH4-N相关性较强的过程变量。其次,利用IT2FNN建立所选变量与出水NH4-N的软测量模型,通过梯度下降算法对模型相关参数进行修正。最后,将基于IT2FNN的出水NH4-N软测量模型应用于实际污水处理过程。实验结果表明,提出的出水NH4-N软测量方法不仅能够实现污水处理过程出水NH4-N的实时检测,而且具有较高的检测精度。

关键词: 出水氨氮, 软测量, 区间二型模糊神经网络, 动态建模, 污水处理过程

CLC Number: