Identification of faulty sensors using principal component analysis

Ricardo Dunia, S. Joe Qin, Thomas F. Edgar, Thomas J. McAvoy

AIChE Journal · 1996 · 524 citations · 16 references

Concepts

TL;DR

Although PCA has recently attracted attention for sensor fault detection, few schemes account for abnormal plant conditions, and the principal component model captures measurement correlations and reconstructs variables while transient behavior of sensor faults in various residuals has been analyzed. The article proposes using PCA for sensor fault identification via reconstruction and introduces a sensor validity index (SVI) to determine each sensor’s status. The authors propose a sensor validity index (SVI) and examine its online implementation for various sensor faults, noting that filtering the index is a key tuning parameter. An example using boiler process data demonstrates the attractive features of the SVI.

Abstract

Abstract Even though there has been a recent interest in the use of principal component analysis (PCA) for sensor fault detection and identification, few identification schemes for faulty sensors have considered the possibility of an abnormal operating condition of the plant. This article presents the use of PCA for sensor fault identification via reconstruction. The principal component model captures measurement correlations and reconstructs each variable by using iterative substitution and optimization. The transient behavior of a number of sensor faults in various types of residuals is analyzed. A sensor validity index (SVI) is proposed to determine the status of each sensor. On‐line implementation of the SVI is examined for different types of sensor faults. The way the index is filtered represents an important tuning parameter for sensor fault identification. An example using boiler process data demonstrates attractive features of the SVI.

References

16