“…These score plots clearly addressed the concern of representativeness of calibration and validation sets which has been voiced in previous calibrations. 24,40 Figure 2.Principal component analysis score plots of the diesel samples: (a) multiplicative-scattering corrected spectra, (b) Savitzky-Golay derivatised spectra, and (c) phsioco-chemical properties; where ● donates the calibration samples and ○ donates the validation samples.…”
Section: Resultsmentioning
confidence: 99%
“…As shown in Table 5, the models obtained from this study are fit for the purpose of online CI assessment, with excellent degree of precision and unbiased when compared to those reported models including the ASTM reference method. 21,24,38,42,44,45 In particular, the PLS regression model developed using teh Savitzky-Golay derivative that explained 98% of the CI variation with 4 LVs showed the best predictive performance. When considering CI measurement at the limit specified in the local requirement of diesel fuel, i.e., 47, the relative prediction error of the model was 0.77%.…”
Section: Modelling Of Cetane Indexmentioning
confidence: 98%
“…These score plots clearly addressed the concern of representativeness of calibration and validation sets which has been voiced in previous calibrations. 24,40 Figure Modelling the boiling point at 95% recovery Diesel fuel comprises hydrocarbon derivatives, including paraffin, naphthene, olefins, and aromatics. In general, paraffin has a lower boiling point than naphthenic and aromatic compounds of the same carbon number; together, they express the boiling range distribution.…”
Section: Multivariate Calibrationsmentioning
confidence: 99%
“…To boost the performance of NIR-based measurements, numerous novel algorithms that focused on individual calibration component have been proposed. [20][21][22][23][24][25][26] Nevertheless, the fitness of a supervised learning model relies on the entire workflow, from signal pre-processing to model validation; and this key element has not been highlighted in many literature reports.…”
It is essential to control the quality of diesel products so that they comply with relevant fuel specifications, however, the quality assessments rely upon conventional wet chemical analyses that are costly and time consuming. Rapid, simultaneous quality measurement enabling immediate online optimisation for process control and blending offers tremendous cost savings by minimising product quality give-away, shipment demurrage, tank inventory, and laboratory analysis. In this study, the use of near infrared spectroscopy and chemometrics demonstrates a straightforward workflow for simultaneous determination of the petroleum diesel’s boiling point at 95% recovery (T95), flash point (FP), cloud point (CP), and cetane index (CI) calibration development. It involved appropriate spectral region selection, calibration/validation set partition, data pre-processing, regression modelling and validation. Based on the calibration and validation results, the supervised learning models that are obtained from a combination region of 4000–4800 cm−1 on a randomly selected calibration set managed to deliver promising predictive performance in terms of coefficient of determination for prediction (r2P/T95 ≥ 0.94, r2P/FP ≥ 0.89, r2P/CP ≥ 0.89, r2P/CI ≥ 0.993), root mean square error of prediction (RMSEP (T95) ≤ 5.2°C, RMSEP (FP) ≤ 2.0°C, RMSEP (CP) ≤ 2.4°C, RMSEP (CI) ≤ 0.3), and ratio of performance deviation (RPD (T95) ≥ 3.7, RPD (FP) ≥ 3.0, RPD (CP) ≥ 2.9, RPD (CI) ≥ 11). Regardless of principal component regression or partial least square regression on either the multiplicative scattering corrected spectra or Savitzky Golay second derivative spectra, the developed models met respective ASTM reproducibility requirements, and were considered adequate for immediate quality assessment of diesel.
“…These score plots clearly addressed the concern of representativeness of calibration and validation sets which has been voiced in previous calibrations. 24,40 Figure 2.Principal component analysis score plots of the diesel samples: (a) multiplicative-scattering corrected spectra, (b) Savitzky-Golay derivatised spectra, and (c) phsioco-chemical properties; where ● donates the calibration samples and ○ donates the validation samples.…”
Section: Resultsmentioning
confidence: 99%
“…As shown in Table 5, the models obtained from this study are fit for the purpose of online CI assessment, with excellent degree of precision and unbiased when compared to those reported models including the ASTM reference method. 21,24,38,42,44,45 In particular, the PLS regression model developed using teh Savitzky-Golay derivative that explained 98% of the CI variation with 4 LVs showed the best predictive performance. When considering CI measurement at the limit specified in the local requirement of diesel fuel, i.e., 47, the relative prediction error of the model was 0.77%.…”
Section: Modelling Of Cetane Indexmentioning
confidence: 98%
“…These score plots clearly addressed the concern of representativeness of calibration and validation sets which has been voiced in previous calibrations. 24,40 Figure Modelling the boiling point at 95% recovery Diesel fuel comprises hydrocarbon derivatives, including paraffin, naphthene, olefins, and aromatics. In general, paraffin has a lower boiling point than naphthenic and aromatic compounds of the same carbon number; together, they express the boiling range distribution.…”
Section: Multivariate Calibrationsmentioning
confidence: 99%
“…To boost the performance of NIR-based measurements, numerous novel algorithms that focused on individual calibration component have been proposed. [20][21][22][23][24][25][26] Nevertheless, the fitness of a supervised learning model relies on the entire workflow, from signal pre-processing to model validation; and this key element has not been highlighted in many literature reports.…”
It is essential to control the quality of diesel products so that they comply with relevant fuel specifications, however, the quality assessments rely upon conventional wet chemical analyses that are costly and time consuming. Rapid, simultaneous quality measurement enabling immediate online optimisation for process control and blending offers tremendous cost savings by minimising product quality give-away, shipment demurrage, tank inventory, and laboratory analysis. In this study, the use of near infrared spectroscopy and chemometrics demonstrates a straightforward workflow for simultaneous determination of the petroleum diesel’s boiling point at 95% recovery (T95), flash point (FP), cloud point (CP), and cetane index (CI) calibration development. It involved appropriate spectral region selection, calibration/validation set partition, data pre-processing, regression modelling and validation. Based on the calibration and validation results, the supervised learning models that are obtained from a combination region of 4000–4800 cm−1 on a randomly selected calibration set managed to deliver promising predictive performance in terms of coefficient of determination for prediction (r2P/T95 ≥ 0.94, r2P/FP ≥ 0.89, r2P/CP ≥ 0.89, r2P/CI ≥ 0.993), root mean square error of prediction (RMSEP (T95) ≤ 5.2°C, RMSEP (FP) ≤ 2.0°C, RMSEP (CP) ≤ 2.4°C, RMSEP (CI) ≤ 0.3), and ratio of performance deviation (RPD (T95) ≥ 3.7, RPD (FP) ≥ 3.0, RPD (CP) ≥ 2.9, RPD (CI) ≥ 11). Regardless of principal component regression or partial least square regression on either the multiplicative scattering corrected spectra or Savitzky Golay second derivative spectra, the developed models met respective ASTM reproducibility requirements, and were considered adequate for immediate quality assessment of diesel.
“…They are fast, non-destructive, cost-effective and environmentally friendly methods. NIR spectroscopy has been successfully applied as an analytical technique in different areas, such as the food [ 17 , 18 ], agriculture [ 19 , 20 ], fuel [ 21 ] and paper industries [ 22 ], among others, due to several advantages, including fast response, being reagent-free and noninvasiveness. MIR has been applied to identify cellulose-based fibers [ 23 ], natural rubber samples [ 24 ], changes in engine-oil properties [ 25 ] and to determine the thermal stability of explosives [ 26 ], among many other applications.…”
There is an urgent need to reuse and recycle textile fibers, since today, low recycling rates are achieved. Accurate classification methods for post-consumer textile waste are needed in the short term for a higher circularity in the textile and fashion industries. This paper compares different spectroscopic data from textile samples in order to correctly classify the textile samples. The accurate classification of textile waste results in higher recycling rates and a better quality of the recycled materials. The data fusion of near- and mid-infrared spectra is compared with single-spectrum information. The classification results show that data fusion is a better option, providing more accurate classification results, especially for difficult classification problems where the classes are wide and close to one another. The experimental results presented in this paper prove that the data fusion of near- and mid-infrared spectra is a good option for accurate textile-waste classification, since this approach allows the classification results to be significantly improved.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
