Extracting knowledge from historical databases for process monitoring using feature extraction and data clustering
Document Type
Article
Publication Date
1-1-2016
Abstract
For most chemical plants, a major obstacle inhibiting the application of cutting edge fault detection and diagnosis is that many of the best methods require data organized into groups before training is possible. Data clustering and non-linear dimensionality reduction are underutilized tools for this task and this study evaluates how they can work in tandem to extract knowledge from chemical process data sets. Two non-linear dimensionality reduction techniques and principal component analysis as well as two clustering techniques are studied on industrial case studies and a simulation
Publication Source (Journal or Book title)
Computer Aided Chemical Engineering
First Page
859
Last Page
864
Recommended Citation
Thomas, M., & Romagnoli, J. (2016). Extracting knowledge from historical databases for process monitoring using feature extraction and data clustering. Computer Aided Chemical Engineering, 38, 859-864. https://doi.org/10.1016/B978-0-444-63428-3.50148-X