Machine learning technology in biodiesel research: A review

Biodiesel has the potential to significantly contribute to making transportation fuels more sustainable. Due to the complexity and nonlinearity of processes for biodiesel production and use, fast and accurate modeling tools are required for their design, optimization, monitoring, and control. Data-driven machine learning (ML) techniques have demonstrated superior predictive capability compared to conventional methods for modeling such highly complex processes. Among the available ML techniques, the artificial neural network (ANN) technology is the most widely used approach in biodiesel research. The ANN approach is a computational learning method that mimics the human brain's neurological processing ability to map input-output relationships of ill-defined systems. Given its high generalization capacity, ANN has gained popularity in dealing with complex nonlinear real-world engineering and scientific problems. This paper is devoted to thoroughly reviewing and critically discussing various ML technology applications, with a particular focus on ANN, to solve function approximation, optimization, monitoring, and control problems in biodiesel research. Moreover, the advantages and disadvantages of using ML technology in biodiesel research are highlighted to direct future R&D efforts in this domain. ML technology has generally been used in biodiesel research for modeling (trans)esterification processes, physico-chemical characteristics of biodiesel, and biodiesel-fueled internal combustion engines. The primary purpose of introducing ML technology to the biodiesel industry has been to monitor and control biodiesel systems in real-time; however, these issues have rarely been explored in the literature. Therefore, future studies appear to be directed towards the use of ML techniques for real-time process monitoring and control of biodiesel systems to enhance production efficiency, economic viability, and environmental sustainability.     

An immunomagnetic-bead-based enzyme-linked immunosorbent assay for sensitive quantification of fumonisin B1

A sensitive enzyme-linked immunosorbent assay (ELISA) based on immunomagnetic beads (IMB-ELISA) was established using a magnetic-bead signal-enrichment system. The immunomagnetic beads were coated with polyclonal antibody directed against keyhole limpet hemocyanin (KLH), which were then coupled with a KLH–fumonisin B1 (FB1) conjugate. Anti-FB1 monoclonal antibody and sample extract were mixed and added to the immunomagnetic-bead solution. After the addition of horseradish peroxidase (HRP)-labeled goat anti-mouse antibody and the substrate solution, stop solution was added and the optical density of the reaction mixture was determined. To improve the performance of this method, the dilution of the immunomagnetic beads, the concentrations of the monoclonal antibody and HRP-labeled goat anti-mouse antibody, and the incubation time for the competition reaction were optimized. Based on the optimum conditions, the regression equation for this IMB-ELISA in quantifying FB1 was y = −0.3538x + 0.703 (R² = 0.9988). The detection limit and IC₅₀ were 0.24 ng/mL and 3.17 ng/mL, respectively. The working range was 0.54–26.3 ng/mL. The recovery rates were 80.4–114.7%, when the spiked concentrations ranged from 19.5 to 156.3 μg/kg. This IMB-ELISA is accurate and more sensitive and less time-consuming than the conventional ELISA.     

Ultrasonic vibration-assisted pelleting for cellulosic biofuel manufacturing: Investigation on power consumption

Cellulosic ethanol produced from cellulosic biomass is an alternative to petroleum-based transportation fuels. Raw cellulosic biomass has low density, causing high costs in their storage, transportation, and handling. Ultrasonic vibration-assisted (UV-A) pelleting can increase the density of cellulosic biomass. Effects of UV-A pelleting variables on pellet quality (density, durability, stability, and strength) and sugar yield have been reported. However, power consumption in UV-A pelleting has not been fully investigated. This paper presents an experimental investigation on power consumption in UV-A pelleting of wheat straw. Effects of input variables (biomass moisture content, biomass particle size, pelleting pressure, and ultrasonic power) on power consumption are investigated. Results show that power consumption in UV-A pelleting increases as moisture content and particle size decrease, and as pelleting pressure and ultrasonic power increase.     

Movement and distribution of Sitophilus zeamais adults and relationship between their density and trapping frequency in wheat bulks under different grain temperatures and moisture contents

Movement and distribution of Sitophilus zeamais adults and relationship between their trapping frequency and insect density were determined in one tonne wheat with 11.5 ± 0.2, 12.5 ± 0.1 and 13.5 ± 0.2% moisture contents (wet basis) inside cylindrical bins at 20, 25 and 30 °C. The introduced adult densities were 0.5, 1.0, and 2.0 adults/kg wheat (A/kg). Inside each bin, 15 probe traps were vertically installed in three layers. In each layer, five traps were installed at the center and half radii of the bin individually. The trapped adults were counted daily up to 7 d, and then 15 kg wheat was sampled at each trap location. The insect number in each sample was counted.The recovered adults in the top layer were less than that in the middle layer, while no significant difference between at middle and bottom layers. In each layer, adults preferred the center location at 0.5 A/kg, while this trend was unidentified at other densities. There was an autocorrelation of insect densities at adjacent locations at ≥ 1.0 A/kg, while no autocorrelation in the vertical direction. The adults had an aggregation distribution at any temperature, moisture content and adult density. The normal and Poisson model could not be used to describe the distribution of the count frequency of the adults in the samples, while the negative binomial model could describe this distribution in 18 out of 27 sampling sets at ≤ 1.0 A/kg. The trapping frequency was strongly influenced by the trapping period and increased with the increase of insect densities. The relationship between adult densities and trapping frequencies at the sampling locations had a significant difference at different grain temperatures, while no difference at different moisture contents, grain depths, and locations.     

Comparison of the volatile composition and sensory characteristics of Spanish PDO cheeses manufactured from ewes’ raw milk and animal rennet

Idiazabal, Manchego, Roncal and Zamorano varieties are the Spanish cheeses manufactured under Protected Denomination of Origin (PDO) from ewes’ raw milk and coagulated with animal rennet. Two batches of each cheese variety were ripened for six months and the cheeses were compared for sensory characteristics and composition of volatile compounds. Seventy-six volatile compounds were identified by GC-MS analysis of headspaces of the cheeses. The volatile profile of the four PDO cheeses differed significantly. Acids were the most abundant volatile compounds in Idiazabal, Roncal, and Zamorano cheeses, whereas alcohols were the main volatile compounds in Manchego cheese due to the large percentage of 2-butanol. Aldehydes, ketones and esters were minor compounds in all the cheese varieties, whereas terpenes and unsaturated hydrocarbons were only found in Manchego cheese. The sensory profiles, of the four PDO cheeses also differed significantly. The highest scores for sharp, brine and rennet odours, and rancid and rennet flavours were assessed in Idiazabal and Zamorano cheeses, whereas milky and buttery odour scores were higher in Manchego, Roncal and Zamorano cheeses than in Idiazabal cheese. Principal component analysis was applied to sensory attributes and volatile composition of the cheeses. Differences in flavour and odour attributes were correlated with differences in the volatile compounds. Two principal components correlating sensory attributes and volatile compounds were defined as “strength factor” and “sweetness factor”. The “strength factor” distinguished among the four cheese varieties, except between Roncal and Manchego cheeses, and the “sweetness factor” distinguished the Zamorano cheese from the Idiazabal, Manchego and Roncal cheeses.     

Site-dependent doping effects on quantum transport in zigzag graphene nanoribbons

We investigate the site-dependent effects of a substitutional nitrogen or boron atom on quantum transport in zigzag graphene nanoribbons from first principles and tight-binding model calculations. The former show three characteristics in transmission spectra: drops around the Fermi level, dips at the bottom of the second conduction (nitrogen) or valence (boron) band, and sharp peaks at the Fermi level. Comparing with the latter, the origins of the transmission features are revealed. The drops are attributed to the impurity onsite potential or its Coulomb interaction, depending on its location. The dips come from the interaction between the impurity and its neighbor atoms. The peaks are associated with the resonant long-rang components of the Coulomb interaction.     

Impacts of urea-formaldehyde resin residue on recycling and reconstitution of wood-based panels

Recycling and reuse of wood-based panels has been attracting great attention since it is an environmentally friendly means of managing wood waste. The recycled wood-based panels usually still contain adhesive residue. In the current research, the impact of urea-formaldehyde (UF) resin, one of the most widely used adhesives, on recycling and reconstitution of wood-based products has been evaluated by adding cured UF resin particles into a new adhesive system. The morphology, chemical structure, curing behavior, and bonding quality of an adhesive system containing 0–6% previously cured resin particles of varied sizes (< 600 μm) were investigated and compared to those properties of a regular adhesive without any impurity. Three-ply plywood glued with a UF adhesive containing 0–10 wt% previously cured resin of varied particle sizes (< 600 μm) were prepared and characterized. The results indicated that the resin residue could significantly affect the curing behavior and bonding quality of the new adhesive system. Higher content and greater size of the particles caused more severe deterioration in the bonding strength.     

Influence of hydrothermal carbonization on formation of curved graphite structures obtained from a lignocellulosic precursor

In this letter we report the preparation of curved graphite structures from crude biomass by using hydrothermal carbonization followed by pyrolysis. High resolution transmission electron microscopy (HRTEM) showed clearly the determinant role of the hydrothermal carbonization as an intermediate process, since the highly curved graphite structures were formed only when the crude biomass is pre-treated hydrothermally. X-ray diffraction (XRD) and Raman Spectroscopy techniques confirmed the presence of (0 0 2) graphite planes and sp² carbon network, thus evidencing a higher organization degree for samples hydrothermally treated prior to the pyrolysis step.