Batch-Wise Unfolding PCA for Fault Detection and Identification of a continuously operated UF system

The aim of this work is to design a fault-detection and identification system for an industrial-scale Ultrafiltration process and to propose the adopted methodology to other membrane applications. The model was created using a multivariate batch analysis tool, namely, Unfolded Principal Component an...

Full description

Saved in:
Bibliographic Details
Main Authors: Shams, Mohamed Bin, Aldeeb, Eman, Elshereef, Rand, Rezk, Sara
Format: Conference Proceeding
Language:English
Subjects:
Aspen ProMVTM
Chemical engineering
Data models
Fault detection
Fault detection and identification (FDI)
Fault diagnosis
Monitoring
Optimization
Principal component analysis
Principal Component Analysis (PCA)
Ultrafiltration
Unfolded PCA
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The aim of this work is to design a fault-detection and identification system for an industrial-scale Ultrafiltration process and to propose the adopted methodology to other membrane applications. The model was created using a multivariate batch analysis tool, namely, Unfolded Principal Component analysis (UPCA), which reduces the high dimensionality of the data to facilitate the analysis. Although ultrafiltration system is a continuous process, modelling it as a batch process proved effective in overcoming some limitations of the observation wise unfolding e.g. the high false alarms. In addition, to ensure the practicability of the proposed fault detection scheme, the recently launched AspenTech's Asset Performance Management (APM) suite, namely, Aspen ProMV™ was used for modelling, analysis and virtual online monitoring.
ISSN:2573-5276
DOI:10.1109/ICMSAO.2019.8880409