Predicting % of crystallinity in FCC catalysts by FT‐MIR and PLS

Abstract

a This paper describes an analytical procedure for prediction of percent of crystallinity of fluidized catalytic cracking catalysts (FCC) using Fourier transform mid infrared spectroscopy (FT-MIR) and partial least-squares (PLS) multivariate calibration technique. In order to make a robust regression model, multiplicative scatter correction (MSC) and smoothed second derivative pre-processing methods were tested. Root mean squared error of prediction (RMSEP) of an independent test set was used to measure the performance of the models. The comparison shows that reasonable values of RMSEP and RMSECV were obtained for PLS-MSC model (RMSEP ¼ 0.8% and RMSECV ¼ 1.3%). The accuracy of the results obtained by the PLS-MSC regression model is in accordance with the uncertainty of the XRPD reference method. The developed method can be implemented in a refinery laboratory environment with ease.

Regularized sliced inverse regression for determination of the percentage of crystallinity in FCC catalysts

“…All these approaches are compared among themselves and with the results obtained using PLS, reported previously [4].…”

“…The spectra were recorded in transmission mode. Powder samples were prepared in KBr pressed disks, according to an internal laboratory procedure [4]. The spectra were collected over the range of 4000-400 cm À1 at a resolution of 4 cm À1 , as an average of 32 scans.…”

“…All these aspects make the experimental procedure extremely laborious and often result in inaccurate assessments of the phase proportions. Owing to these problems, in a previous work [4], the authors presented an analytical procedure for the prediction of the percentage of crystallinity in FCC catalysts using FT-MIR spectroscopy and Partial least square (PLS) regression as the multivariate calibration technique. Although results obtained in Reference [4] were satisfactory, in this paper the authors further study other multivariate calibration methods in order to obtain a better prediction performance.…”

“…Owing to these problems, in a previous work [4], the authors presented an analytical procedure for the prediction of the percentage of crystallinity in FCC catalysts using FT-MIR spectroscopy and Partial least square (PLS) regression as the multivariate calibration technique. Although results obtained in Reference [4] were satisfactory, in this paper the authors further study other multivariate calibration methods in order to obtain a better prediction performance. A new calibration model based on regularized slice inverse regression (RSIR) [5][6][7][8] is proposed.…”

“…Multiplicative scatter correction (MSC) was used as a preprocessing method. Further details on the analytical procedure can be found in Reference [4].…”