Partial Least Square Modeling for the Control of Refining Processes on Mid-Distillates by Near Infrared Spectroscopy

S. Aji 1 N. Schildknecht-Szydlowski 1 A. Faraj 1
1 IFPEN - IFP Energies nouvelles
Abstract : Partial Least Squares regression (PLS) was used to elaborate the prediction models of the different chemical families i.e. wt% paraffins, naphthenes, and wt% and mol/100 g monoaromatics, diaromatics+ and total aromatics from Near InfraRed spectra (NIR) of mid-distillates covering a large range of chemical compositions. Cluster analysis was used to reveal the chemical similarities between samples and to organize in clusters the calibration data base. The relationships between NIR spectra and modeled properties were well adapted for most of the prediction models in twice the interval of confidence at 95% of the reference methods after clustering of the data base into three clusters. Cluster analysis was necessary to improve the prediction quality of the PLS models.
Document type :
Journal articles
Complete list of metadatas

Cited literature [14 references]  Display  Hide  Download
https://hal-ifp.archives-ouvertes.fr/hal-02017303
Contributor : Edp Sciences <>
Submitted on : Wednesday, February 13, 2019 - 10:01:16 AM
Last modification on : Thursday, February 14, 2019 - 1:16:54 AM
Long-term archiving on : Tuesday, May 14, 2019 - 3:03:37 PM

File

Aji_vol59n3.pdf
Publication funded by an institution

Identifiers

Collections

IFP | OGST

Citation

S. Aji, N. Schildknecht-Szydlowski, A. Faraj. Partial Least Square Modeling for the Control of Refining Processes on Mid-Distillates by Near Infrared Spectroscopy. Oil & Gas Science and Technology - Revue d'IFP Energies nouvelles, Institut Français du Pétrole, 2004, 59 (3), pp.303-321. ⟨10.2516/ogst:2004022⟩. ⟨hal-02017303⟩

Export

BibTeX TEI DC DCterms EndNote

Share

Metrics

Record views

12

Files downloads

8

Contact

support.ccsd.cnrs.fr
hal.support@ccsd.cnrs.fr
Logo CCSD