Application of infrared spectroscopy as Process Analytics Technology (PAT) approach in biodiesel production process utilizing Multivariate Curve Resolution Alternative Least Square (MCR-ALS)

Different modeling approaches for inline biochemical monitoring over the VLP-making upstream stages using Raman spectroscopy

This work aimed to set inline Raman spectroscopy models to monitor biochemically (viable cell density, cell viability, glucose, lactate, glutamine, glutamate, and ammonium) all upstream stages of a virus-like particle-making process. Linear (Partial least squares, PLS; Principal components regression, PCR) and nonlinear (Artificial neural networks, ANN; supported vector machine, SVM) modeling approaches were assessed. The nonlinear models, ANN and SVM, were the more suitable models with the lowest absolute errors. The mean absolute error of the best models within the assessed parameter ranges for viable cell density (0.01-8.83 × 106 cells/mL), cell viability (1.3-100.0 %), glucose (5.22-10.93 g/L), lactate (18.6-152.7 mg/L), glutamine (158-1761 mg/L), glutamate (807.6-2159.7 mg/L), and ammonium (62.8-117.8 mg/L) were 1.55 ± 1.37 × 106 cells/mL (ANN), 5.01 ± 4.93 % (ANN), 0.27 ± 0.22 g/L (SVM), 4.7 ± 2.6 mg/L (SVM), 51 ± 49 mg/L (ANN), 57 ± 39 mg/L (SVM) and 2.0 ± 1.8 mg/L (ANN), respectively. The errors achieved, and best-fitted models were like those for the same bioprocess using offline data and others, which utilized inline spectra for mammalian cell lines as a host.

Raman spectra soft modeling of the biodiesel oxidation through evolving factor analysis & multivariate curve resolution

The oxidative stability of biodiesel is defined by its relative resistance to the action of oxygen at room temperature. Its determination is an essential reference to the quality of biofuel and a significant parameter to be determined. This parameter concerns the quality of the biodiesel to be supplied to the consumer, and its determination is fundamental to maintaining the engine's proper functioning. Raman spectroscopy allows the rapid obtaining of structural information regarding biodiesel quality and, when aided by multivariate analysis methods, allows a quantitative determination of specific properties. This work uses Raman spectroscopy, Multivariate Curve Resolution with Alternative Least Squares (MCR-ALS) method, and Evolving Factor Analysis (EFA) to study biodiesel's oxidation kinetics. Also, the vibrational modes C = C, CH2, and CH3 were identified as the main structural groups involved in this process, corroborating previous studies. The MCR-ALS & EFA combination allowed modeling of the degradation kinetics following an A → B → C mechanism, where A corresponds to the biodiesel (starting material), B is related to the hydroperoxide mixture, and C is the final product. The results also suggested that this process follows a first-order reaction, with kinetic constant values of k1 = 0.0056 min-1 and k2 = 0.0031 min-1.

Calibration with virtual standards for monitoring biodiesel production using a miniature NIR spectrometer

This work describes the use of virtual standards as calibration samples in an innovative multivariate calibration approach for the on-line monitoring of alkyl-esters content during biodiesel production process using a miniature near infrared (NIR) spectrometer. For comparison purposes, a partial least squares (PLS) model was built using synthetic blends prepared in laboratory with different concentrations of oil, glycerol, biodiesel, and ethanol and resulted in a satisfactory predictive ability (root mean square error of prediction, RMSEP, of 1.51% w/w). When compared to conventional methods, calibration with synthetic blends has the advantage of simplifying the experimental procedure and reducing the need for reference analysis. Nevertheless, it still requires the preparation of a considerable number of blends in laboratory. To overcome this limitation, this study proposed an innovative approach where a PLS model was constructed based on virtual standards: representative calibration spectra were created by mathematically mixing spectra from pure components and performing an adjustment using the Piecewise Direct Standardization (PDS) method. This significantly reduced the need for calibration synthetic blends and led to similar results (RMSEP of 1.75% w/w), compared to the previous approach. This work also demonstrates the use of the constructed models to predict the concentration profiles of alkyl-esters during the batch transesterification process.

Machine Learning Applications in Biofuels’ Life Cycle: Soil, Feedstock, Production, Consumption, and Emissions

Machine Learning (ML) is one of the major driving forces behind the fourth industrial revolution. This study reviews the ML applications in the life cycle stages of biofuels, i.e., soil, feedstock, production, consumption, and emissions. ML applications in the soil stage were mostly used for satellite images of land to estimate the yield of biofuels or a suitability analysis of agricultural land. The existing literature have reported on the assessment of rheological properties of the feedstocks and their effect on the quality of biofuels. The ML applications in the production stage include estimation and optimization of quality, quantity, and process conditions. The fuel consumption and emissions stage include analysis of engine performance and estimation of emissions temperature and composition. This study identifies the following trends: the most dominant ML method, the stage of life cycle getting the most usage of ML, the type of data used for the development of the ML-based models, and the frequently used input and output variables for each stage. The findings of this article would be beneficial for academia and industry-related professionals involved in model development in different stages of biofuel’s life cycle.

A Review on Machine Learning Application in Biodiesel Production Studies

The consumption of fossil fuels has exponentially increased in recent decades, despite significant air pollution, environmental deterioration challenges, health problems, and limited resources. Biofuel can be used instead of fossil fuel due to environmental benefits and availability to produce various energy sorts like electricity, power, and heating or to sustain transportation fuels. Biodiesel production is an intricate process that requires identifying unknown nonlinear relationships between the system input and output data; therefore, accurate and swift modeling instruments like machine learning (ML) or artificial intelligence (AI) are necessary to design, handle, control, optimize, and monitor the system. Among the biodiesel production modeling methods, machine learning provides better predictions with the highest accuracy, inspired by the brain’s autolearning and self-improving capability to solve the study’s complicated questions; therefore, it is beneficial for modeling (trans) esterification processes, physicochemical properties, and monitoring biodiesel systems in real-time. Machine learning applications in the production phase include quality optimization and estimation, process conditions, and quantity. Emissions composition and temperature estimation and motor performance analysis investigate in the consumption phase. Fatty methyl acid ester stands as the output parameter, and the input parameters include oil and catalyst type, methanol-to-oil ratio, catalyst concentration, reaction time, domain, and frequency. This paper will present a review and discuss various ML technology advantages, disadvantages, and applications in biodiesel production, mainly focused on recently published articles from 2010 to 2021, to make decisions and optimize, model, control, monitor, and forecast biodiesel production.

Portable and benchtop Raman spectrometers coupled to cluster analysis to identify quinine sulfate polymorphs in solid dosage forms and antimalarial drug quantification in solution by AuNPs-SERS with MCR-ALS

This paper proposes for the first time: (a) a qualitative analytical method based on portable and benchtop backscattering Raman spectrometers coupled to hierarchical cluster analysis (HCA) and multivariate curve resolution - alternating least-squares (MCR-ALS) to identify two polymorphs of antimalarial quinine sulfate in commercial pharmaceutical tablets in their intact forms and (b) a quantitative analytical method based on gold nanoparticles (AuNPs) as active substrates for surface-enhanced Raman scattering (SERS) in combination with MCR-ALS to quantify quinine sulfate in commercial pharmaceutical tablets in solution. The pure concentration and spectral profiles recovered by MCR-ALS proved that both formulations present different polymorphs. These results were also confirmed by two clusters observed in the HCA model, according to their similarities within and among the samples that provided useful information about the homogeneity of different pharmaceutical manufacturing processes. AuNPs-SERS coupled to MCR-ALS was able to quantify quinine sulfate in the calibration range from 150.00 to 200.00 ng mL-1 even with the strong overlapping spectral profile of the background SERS signal, proving that it is a powerful ultrahigh sensitivity analytical method. This reduced linearity was validated throughout a large calibration range from 25.00 to 175.00 μg mL-1 used in a reference analytical method based on high performance liquid chromatography with a diode array detector (HPLC-DAD) coupled to MCR-ALS for analytical validation purposes, even in the presence of a coeluted compound. The analytical methods developed herein are fast, because second-order chromatographic data and first-order SERS spectroscopic data were obtained in less than 6 and 2 min, respectively. Concentrations of quinine sulfate were estimated with low root mean square error of prediction (RMSEP) values and a low relative error of prediction (REP%) in the range 1.8-4.5%.