Partial Least Squares based process monitoring for sensor faults in a process reactor

Kini, Ramakrishna and Madakyaru, Muddu (2016) Partial Least Squares based process monitoring for sensor faults in a process reactor. In: International conference on inter disciplinary mathematics and computational techniques, 22/12/2016, Manipal University, Jaipur.

[img] PDF
541.pdf - Published Version
Restricted to Registered users only

Download (621kB) | Request a copy

Abstract

Monitoring modern chemical processes and diagnosing for any abnormalities in those set ups are crucial in process control domain since they are needed for safety and betterment of production quality in the process. Various statistical based malfunction detection methods have been used over the years, namely that are classified into univariate and multivariate techniques. Univariate techniques have been used popularly but are limited for monitoring only a single variable at a time whereas multivariate techniques have been used with multiple correlated variables. Partial Least Squares (PLS), a multi-variate technique, has been successfully used in the domain of process monitoring for handling correlated variables. In the work, PLS model is developed to capture the behavior of the plant and T2 statistics is used to detect fault indices. The effectiveness of the developed FD algorithm is demonstrated and presented for various sensor faults on benchmark continuous stirred tank reactor (CSTR).

Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: Fault detection; Partial Least Squares; T2 statistics; CSTR.
Subjects: Engineering > MIT Manipal > Chemical
Engineering > MIT Manipal > Instrumentation and Control
Depositing User: MIT Library
Date Deposited: 11 Jan 2017 09:36
Last Modified: 11 Jan 2017 09:36
URI: http://eprints.manipal.edu/id/eprint/147993

Actions (login required)

View Item View Item