Real-Time Chemical Process Monitoring with UMAP

Document Type

Article

Publication Date

1-1-2021

Abstract

In the modern chemical world, engineers have access to millions of data points at their fingertips. Using this data properly can help us to recognize vast improvements in the way plants are run, especially in process monitoring. In this paper three aspects of process monitoring are considered: visualization, fault identification, and fault diagnosis. For each aspect, the use of modern machine learning techniques for addressing these issues is discussed, and improvements over outdated methods are illustrated. This proposed approach is tested using the Tennessee Eastman Process (TEP) for several types of faults.

Publication Source (Journal or Book title)

Computer Aided Chemical Engineering

First Page

2077

Last Page

2082

This document is currently not available here.

Share

COinS