基于DLNS-DSVDD的故障检测方法
作者:
作者单位:

作者简介:

通讯作者:

中图分类号:

基金项目:

国家自然科学基金(61673279); 辽宁省自然科学基金(2019-MS-262); 辽宁省教育厅基金一般项目(LJ2019013)


Fault detection method based on DLNS-DSVDD
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    为了解决支持向量数据描述方法对多模态动态过程故障检测率低的问题,提出了基于双局部近邻标准化的动态支持向量数据描述的故障检测方法。首先,应用DLNS方法标准化样本生成标准数据集;其次,利用动态方法将标准数据集构造成动态矩阵;最后,使用SVDD方法建立基于空间距离的统计量进行故障检测,应用DLNS-DSVDD方法对一个数值例子和田纳西-伊斯曼过程(TE过程)进行仿真测试,并与传统方法对比分析。结果表明,DLNS-DSVDD方法在具有更低的误报率的前提下拥有更高的故障检测率,平均误报率为2.2%,相比于SVDD方法降低了2.1%,平均故障检测率为88.2%,相比于SVDD方法提高了38.5%,证明了所提出的方法可以有效地监控具有多模态动态特征的化工生产过程。DLNS-DSVDD方法可以有效地剔除数据的多模态结构并且准确地捕获样本中的动态信息,为提高SVDD方法在多模态动态过程中的故障检测性能提供了参考。

    Abstract:

    Aiming at the problem of low fault detection rate of multi-modal dynamic process by support vector data description (SVDD) method,a fault detection method based on dynamic SVDD based on double local neighborhood standardization (DLNS-DSVDD) was proposed.Firstly,the DLNS method was used to standardize samples to generate a standard data set.Secondly,the dynamic method was used to construct the standard data set into a dynamic matrix.Finally,the SVDD method was used to establish statistics based on spatial distance for fault detection,and the DLNS-DSVDD method was used to simulate a numerical example and the Tennessee-Eastman (TE) process,which was compared with the traditional method.The results show that the DLNS-DSVDD method has a higher fault detection rate on the premise of lower false alarm rate.The average false alarm rate is 2.2%,which is 2.1% lower than that of the SVDD method,and the average fault detection rate is 88.2%,which is 38.5% higher than the SVDD method.It is proved that the method proposed in this paper can effectively monitor the chemical production process with multi-modal dynamic characteristics.The DLNS-DSVDD method can effectively eliminate the multi-modal structure of the data and accurately capture the dynamic information in the sample,which provides a reference for improving the fault detection performance of the SVDD method in the multi-modal dynamic process.

    参考文献
    相似文献
    引证文献
引用本文

谢彦红,张浩然,张 成,李 元.基于DLNS-DSVDD的故障检测方法[J].河北工业科技,2022,39(4):300-309

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2021-12-01
  • 最后修改日期:2022-05-21
  • 录用日期:
  • 在线发布日期: 2022-08-01
  • 出版日期: