Search

The aim of this study was to evaluate and compare the performance of multivariate classification algorithms, specifically Partial Least Squares Discriminant Analysis (PLS-DA) and machine learning algorithms, in the classification of Monthong durian pulp based on its dry matter content (DMC) and soluble solid content (SSC), using the inline acquisition of near-infrared (NIR) spectra. A total of 415 durian pulp samples were collected and analyzed. Raw spectra were preprocessed using five different combinations of spectral preprocessing techniques: Moving Average with Standard Normal Variate (MA+SNV), Savitzky–Golay Smoothing with Standard Normal Variate (SG+SNV), Mean Normalization (SG+MN), Baseline Correction (SG+BC), and Multiplicative Scatter Correction (SG+MSC). The results revealed that the SG+SNV preprocessing technique produced the best performance with both the PLS-DA and machine learning algorithms. The optimized wide neural network algorithm of machine learning achieved the highest overall classification accuracy of 85.3%, outperforming the PLS-DA model, with overall classification accuracy of 81.4%. Additionally, evaluation metrics such as recall, precision, specificity, F1-score, AUC ROC, and kappa were calculated and compared between the two models. The findings of this study demonstrate the potential of machine learning algorithms to provide similar or better performance compared to PLS-DA in classifying Monthong durian pulp based on DMC and SSC using NIR spectroscopy, and they can be applied in the quality control and management of durian pulp production and storage.

Geographic origin of durian (Durio zibethinus Murray) can influence consumers purchasing decisions. A guaranteed geographical origin is extremely useful to maintain the credibility and popularity of this fruit. The objective of this work is to confirm geographical identification (GI) of durian Monthong from Prachuap Kiri Khan Province from south of Thailand which achieved the GI lable from the department of intellectual property by near infrared spectroscopy (NIRs). Total of 60 durian Monthong samples were acquired in the wavenumber range of 12,500-4,000 cm

This research aimed to propose an online system based on multispectral images for the real-time estimation of the moisture content (MC) of sugarcane bagasse. The system consisted of a conveyor belt, four halogen bulbs, and a multispectral camera. The MC models were developed using machine learning algorithms, i.e., multiple linear regression (MLR), principal component regression (PCR), artificial neural network (ANN), PCA-ANN, Gaussian process regression (GPR), PCA-GPR, random forest regression (RFR), and PCA-GPR. The models were developed using 150 samples (calibration set) meanwhile the remaining 50 samples were applied as a validation set. The comparison of all developed models showed that the PCA-RFR model achieved better detection with a higher accuracy of MC prediction. The PCA-RFR model showed the best results which were a coefficient of determination of prediction (r2) of 0.72, root mean square error of prediction (RMSEP) of 11.82 wt%, and a ratio of the standard error of prediction to standard deviation (RPD) of 1.85. The results show that this technique was very useful for MC rapid screening of the sugarcane bagasse.

The aim of this research was to investigate the effect of torrefaction temperature on 4 energy properties as high heating value (HHV), enhancement factor, solid yield, and energy yield of spent coffee ground (SCG). Four different torrefaction temperatures (200, 250, 300, and 350°C) were selected. Torrefaction process was conducted at the heating rate of 10°C/min. HHV and enhancement factor were the highest when SCG was torrefied at the highest temperature of 350°C. However, at this temperature, solid and energy yields were the lowest. Torrefaction temperature highly affected these four energy properties with R of higher than 0.9. Regression models representing the relationship between torrefaction temperature and HHV and energy yield were HHV = 0.0519T+15.917, R2 = 0.9483 and energy yield = -0.2743T+155.1, R2 = 0.9976. These models are helpful for prediction of the energy properties of SCG undergoing torrefaction process in the studied temperature range.

The prediction of dry matter content (DMC) and soluble solids content (SSC) in durian pulp were performed using a small laboratory scale in-line visible and short wave near infrared (Vis-SWNIR) spectroscopic system. The fiber optic diode array spectrometer with a charged coupled device (CCD) detector in a wavelength range of 450−1000 nm was used for spectral data acquisition. The spectra of the sample were acquired on the moving conveyor belt in two different orientations, including scanning in the upright position of pulps collected in 2018 and the stable position by scanning on the side of the pulps collected in 2019. Partial least squares regression (PLSR) was used to establish the relationship between the spectra and observed DMC and SSC values using the different wavelength ranges, including 450−1000, 700−1000, and 800−1000 nm for the comparison. The results showed that the durian pulp should be scanned in the upright position at the center of the pulp. Moving average smoothing preprocessing combined with the standard normal variate (SNV) for DMC and multiple scatter correction (MSC) for SSC gave the best result. The suitable wavelength range for model development to predict the DMC and SSC was 700−1000 nm and 800−1000 nm, respectively. After comparing the results, the optimum model showed the coefficient of determination of calibration (RC2), and prediction (RP2), root mean square error of prediction (RMSEP), bias, and the ratio of performance to interquartile distance (RPIQ) of 0.88, 0.83, 4.32 %, 1.25 %, and 3.52 for DMC and 0.70, 0.70, 4.0 %, 0.4 %, and 2.2 for SSC prediction.

This paper reports the development of a rapid and low-cost method based on near-infrared spectroscopy as an alternative for thermogravimetric determination of the pyrolysis characteristics, including T onset , T sh , T peak , T offset and DTG peak , of milled bamboo. T onset is the extrapolated onset temperature that is calculated from the partial peak resulting from the decomposition of the hemicellulose component, T sh is the temperature corresponding to the overall maximum of the hemicellulose decomposition rate, DTG peak is the overall maximum of the cellulose decomposition rate, T peak is the temperature corresponding to the overall maximum of the cellulose decomposition rate and T offset is the extrapolated offset temperature of the DTG peak curves determined using thermogravimetric analysis (TGA). The models may be used to control the pyrolysis processes of bamboo to achieve the most economical and environmental conditions. 80 samples of bamboo with various circumferences of culms in the ranges of approximately 16–18, 18–20, 20–22, 22–24, 24–26, 26–28, 28–30, 30–32, 32–34, 34–36, 36–38 and 38–40 cm were randomly collected for optimization of the models. The models were optimized by partial least squares regression (PLSR) with 80% of samples for the calibration set and 20% for the validation set. For T onset , T sh , T peak , T offset and DTG peak, the models showed coefficients of determination (R 2 ) of 0.566, 0.845, 0.917, 0.973, and 0.671; root mean square errors of prediction (RMSEP) of 9.7 °C, 4.36 °C, 3.77 °C, 2.66 °C, and 0.428 wt loss %/min; ratios of prediction to deviation (RPD) of 1.52, 2.58, 3.48, 3.55, and 1.75; and biases of −0.344 °C, −0.765 °C, 0.349 °C, −5.41 °C, and 0.045 wt loss %/min, respectively. In addition, the results showed that pyrolysis characteristics did not depend on the circumference. The vibrational bands of water and CH 3 , O H stretch, first overtones of Ar–OH, CH 2 and HC CH in the cellulose and lignin structures, O H hydrogen bonds of polyvinyl alcohol and C H stretch corresponding to the first overtone of CH 2 had the highest influence on the values of T onset , T sh , T peak , and T offset , respectively. The vibrational band of the C O C asymmetrical stretches of cellulose and hemicellulose, and the combination of O H stretch and HOH bend of polysaccharides influenced the DTG peak value. These results are beneficial for studying the thermal behaviour of milled bamboo as a potential resource for producing biofuels, especially in the pyrolysis process.

The MC (moisture content) and HHV (higher heating value) of Leucaena leucocephala pellets using NIR (near infrared) spectroscopy was investigated in this study. The MC of the pellets was adjusted by subjecting the samples to different relative humidity environments. The samples were scanned in diffuse reflection mode at wavenumbers of 12,500–4000 cm−1. Partial least squares regression models correlating the MC and HHV with the NIR spectra were developed and validated by full cross validation. The model for MC and HHV provided coefficients of determination (R2) of 0.995 and 0.964, a root mean square error of cross validation (RMSECV) of 0.187%wb and 79.2 J g−1, bias of −0.0008%wb and 1.29 J g−1 and a RPD (ratio of prediction to deviation) of 13.9 and 5.30, respectively. The models had excellent accuracy. This rapid quality evaluation method may be used for trading of biomass pellets. An equation related MC and HHV was also developed.

The higher heating value (HHV) plays a significant role in determining of the energy potential of biomass. The method to receive this value using a bomb calorimeter takes a long time approximately of 15 minutes. The implementation of near infrared spectroscopy (NIR) technique by generating a prediction model can be replaced the former method in order to reduce time into 2 minutes. Therefore, the precision of both techniques have to be concerned in order to get the reliable values. The repeatability and reproducibility processes were carried out to evaluate the precision of both methods. It was found that the percentage difference of standard deviation (SD) from the mean of repeatability for the NIR was less than 5 % and for reproducibility was slightly over 5 % (but not over than 10%). The SD of repeatability and reproducibility of the bomb calorimeter were 54.6 J/g (0.3 % difference of average HHV value) and 81.3 J/g (0.5 % difference of average HHV value) respectively. The average HHV from the bomb calorimeter was 18,166.77±236.29 J/g. Both techniques showed very low percentage difference of the SD from the mean that revealed very high reliability and precision of measurements.