Abstract Process measurements do not satisfy conservation laws such as mass and energy balances because they are subject to random and sometimes gross errors. Sources of gross errors are process disturbances, leaks, departures from steady state, and malfunctioning or miscalibrated instruments. A challenging task in dealing with contaminated data is to correctly detect and identify them as such before reconciliation. Recently developed gross error detection and identification methods based on Principal Component Analysis have received attention from process industries and have been implemented in industrial software packages. This paper reviews a few cases that were solved using the Principal Component Tests with this software.
[1]
Ajit C. Tamhane,et al.
Detection of gross errors in process data
,
1982
.
[2]
C. M. Crowe,et al.
Recursive identification of gross errors in linear data reconciliation
,
1988
.
[3]
Hongwei Tong,et al.
Detection of gross erros in data reconciliation by principal component analysis
,
1995
.
[4]
Jose A. Romagnoli,et al.
Use of orthogonal transformations in data classification-reconciliation
,
1996
.
[5]
C. M. Crowe,et al.
Reconciliation of process flow rates by matrix projection. Part II: The nonlinear case
,
1983
.