On Generative Topographic Mapping and Graph Theory combined approach for unsupervised non-linear data visualization and fault identification
On Generative Topographic Mapping and Graph Theory combined approach for unsupervised non-linear data visualization and fault identification
复制标题
DOI：
10.1016/j.compchemeng.2016.12.009
复制
发表时间：
2017-03
期刊：
影响因子：
--
通讯作者：
Matheus S. Escobar;H. Kaneko;K. Funatsu
中科院分区：
文献类型：
--
作者：
Matheus S. Escobar;H. Kaneko;K. Funatsu
化工厂的过程监控依赖于两个步骤：识别异常(故障检测)和表征异常(故障识别)。提出了一种结合生成地形图(GTM)和图论(GT)的方法。GTM突出了系统特征，减少了可变维度，并提供了一种计算样本之间相似性的策略。然后，GT使用网络对它们进行聚类，区分正常和异常条目。然而，由于有偏见的正常和异常标签，所提出的方法是无监督的，这意味着标签是不存在的。三个案例研究被考虑：一个模拟数据集，田纳西伊士曼过程和一个工业数据集。采用主成分分析(PCA)、动态主成分分析和核主成分分析指数(QANT2)以及GTM和GT独立监测方法进行比较，考虑了监督和非监督方法。对于工业场景，使用软测量来评估歧视性能。所提出的方法虽然没有监督，但与监督策略类似地区分正常状态，从而证明其发展是合理的。
Process monitoring of chemical plants relies on two steps: discriminating anomalies (fault detection) and characterizing them (fault identification). This work proposes a combined Generative Topographic Mapping (GTM) and Graph Theory (GT) approach. GTM highlights system features, reducing variable dimensionality and providing a strategy for calculating similarity between samples. GT then clusters them using networks, discriminating normal and anomalous entries. Because of biased normal and anomalous labeling, however, the methodology proposed is unsupervised, meaning that labels are inexistent. Three case studies were considered: a simulation data set, Tennessee Eastman process and an industrial data set. Principal Component Analysis (PCA), dynamic PCA and kernel PCA indexes (QandT2) alongside GTM and GT independent monitoring methodologies were used for comparison, considering supervised and unsupervised approaches. For the industrial scenario, soft sensors were used for assessing discrimination performance. The proposed method, while unsupervised, discriminated normal states similarly to supervised strategies, justifying its development.