We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Online Determination on the Properties of Naphtha as the Ethylene Feedstock Using Near-Infrared Spectroscopy.
- Authors
Fan, Chen; Liu, Tianbo; Hu, Guihua; Yang, Minglei; Long, Jian
- Abstract
Providing real-time information on the properties of naphtha as the ethylene feedstock within the minimal time is significant for improvement of the process simulation, control, and real-time optimization. To develop models predicting naphtha properties for different pre-processing methods, an online full transmittance near-infrared (NIR) spectrum measurement system has been used along with the principal component regression and partial least squares (PLS) methods. The results show that the Savitzky-Golay smoothing combined with the first-derivative pre-processing provides the best denoising effect compared to other methods. The predicted relative errors of the NIR models developed by PLS, especially for the cutting temperature points of the test set, basically make 1‒5% indicating it can be used to create good NIR prediction models for the on-line determination of naphtha properties.
- Subjects
PARTIAL least squares regression; NAPHTHA; IMAGE denoising; ETHYLENE; NEAR infrared spectroscopy; METAL cutting
- Publication
Petroleum Chemistry, 2023, Vol 63, Issue 9, p1069
- ISSN
0965-5441
- Publication type
Article
- DOI
10.1134/S0965544123060208