CIESC Journal ›› 2018, Vol. 69 ›› Issue (10): 4292-4301.DOI: 10.11949/j.issn.0438-1157.20180492

Previous Articles     Next Articles

Fault monitoring of high-sulfur gas sweetening process by combined indicator of CW-HCA

GU Xiaohua1,2, ZHANG Kun1, WANG Tian1, HOU Song1, SONG Hongfei1, LI Taifu1, QIU Kui1   

  1. 1. School of Electrical & Information Engineering, Chongqing University of Science and Technology, Chongqing 401331, China;
    2. Artificial Intelligence of Key Laboratory of Sichuan Province, Sichuan University of Science & Engineering, Zigong 643000, Sichuan, China
  • Received:2018-05-09 Revised:2018-06-19 Online:2018-10-05 Published:2018-10-05
  • Supported by:

    supported by the National Science and Technology Major Project of the Ministry of Science and Technology of China (2016ZX05017004), the Chongqing National Science Foundation (cstc2015jcyjBX0089), the Research Foundation of Chongqing University of Science and Technology (CK2016Z16) and the Research Foundation of Chongqing University of Science and Technology (YKJCX1620413).

基于CW-HCA联合指标的高含硫天然气净化过程故障监测

辜小花1,2, 张堃1, 王甜1, 候松1, 宋鸿飞1, 李太福1, 邱奎1   

  1. 1. 重庆科技学院电气与信息工程学院, 重庆 401331;
    2. 四川理工学院四川省人工智能实验室, 四川 自贡 643000
  • 通讯作者: 辜小花
  • 基金资助:

    国家科技重大专项(2016ZX05017004);重庆市基础科学与前沿技术研究项目(cstc2015jcyjBX0089);重庆科技学院校内基金项目(CK2016Z16);重庆科技学院研究生科技创新基金项目(YKJCX1620413)。

Abstract:

The high-sulfur gas (HSG) sweetening process is large-scale and complicated. The acid component may cause great security risk. Applying real-time monitoring of the production process has important significance to ensure the system normal work and safety. Regarding to such chemical processes, the higher-order cumulants analysis (HCA) used the higher-order cumulants and their polyspectras of samples to construct the statistical index, which greatly improves the detection rate. Nevertheless, the different importance of independent components is not concerned in the construction of HS statistical indicators, which may lead to a certain degree of deviation of the monitoring results. Meanwhile, using multiple-indicators may cause the conflicting results among indicators so that the accuracy cannot be guaranteed. Therefore, a fault monitoring method based on contribution weighted higher-order cumulants analysis (CW-HCA) is proposed. The approach weights the sample's third order cumulant according to the contribution of independent components respectively, afterwards, the new weighted index and the residual spatial index are combined to realize the real-time monitoring. The results in TE and HSG sweetening process indicate that, compared with ICA and HCA, the proposed algorithm shows significant efficiency and superiority.

Key words: dynamic modeling, principal component analysis, natural gas, contribution weight, combined indices, fault monitoring

摘要:

高含硫天然气(HSG)净化过程复杂,导致安全开发风险极高。因此保障净化系统可靠运行、实现过程安全生产具有重要意义。对于类似化工过程,在独立成分分析(ICA)方法基础上,高阶累积量分析(HCA)用样本三阶累积量代替均值方差构造统计指标大大提高了检测率。然而,HCA构造独立分量空间指标时未考虑不同独立分量间重要性差异,这可能致使监测结果经样本高阶累积后出现一定程度的偏差。同时,采用多指标监测策略可能出现指标间监测结果相互冲突问题。为此,提出一种基于贡献度加权高阶累积量分析(CW-HCA)联合指标的故障监测方法。该方法根据独立分量的贡献度对样本的三阶累积量进行加权;再将加权后的指标与残差空间指标联合获得联合指标,实现监测。TE以及HSG净化过程的实验结果表明,所提算法相比ICA算法、HCA算法具有有效性和优越性。

关键词: 动态建模, 主元分析, 天然气, 贡献度加权, 联合指标, 故障监测

CLC Number: