Open AccessSpectra selection methods: A novel optimization way for treating dynamic spectra and in-line near infrared modeling
Author(s) -
Haiyan Wang,
Ronghua Liu,
Lei Nie,
Dongbo Xu,
Wenping Yin,
Lian Li,
Hengchang Zang
Publication year - 2020
Publication title -
journal of innovative optical health sciences/journal of innovation in optical health science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.421
H-Index - 24
eISSN - 1793-5458
pISSN - 1793-7205
DOI - 10.1142/s1793545820500157
Subject(s) - near infrared spectroscopy , partial least squares regression , process analytical technology , computer science , spectral line , mean squared error , selection (genetic algorithm) , biological system , process (computing) , fluidized bed , calibration , process engineering , mathematics , artificial intelligence , statistics , chemistry , machine learning , optics , engineering , work in process , physics , operations management , astronomy , biology , operating system , organic chemistry
Near infrared (NIR) spectroscopy is now widely used in fluidized bed granulation. However, there are still some demerits that should be overcome in practice. Valid spectra selection during modeling process is now a hard nut to crack. In this study, a novel NIR sensor and a cosine distance method were introduced to solve this problem in order to make the fluidized process into “visualization”. A NIR sensor was fixed on the side of the expansion chamber to acquire the NIR spectra. Then valid spectra were selected based on a cosine distance method to reduce the influence of dynamic disturbances. Finally, spectral pretreatment and wavelength selection methods were investigated to establish partial least squares (PLS) models to monitor the moisture content. The results showed that the root mean square error of prediction (RMSEP) was 0.124% for moisture content model, which was much lower than that without valid spectra selection treatment. All results demonstrated that with the help of valid spectra selection treatment, NIR sensor could be used for real-time determination of critical quality attributes (CQAs) more accurately. It makes the manufacturing easier to understand than the process parameter control.