Cleaning-In-Place of heat exchangers in Ultra-High Temperature dairy production remains a resource-intensive process, typically operated without explicit information on the underlying cleaning mechanisms. In laboratory experiments from the literature, grayscale image sequences of soil removal were used to extract statistical features describing local and temporal variations in pixel intensity. Building on this framework, the present work investigates the feasibility of identifying cleaning mechanisms directly from industrial dairy process data using machine learning models retrained on image-based features. A fully connected neural network was trained to classify three dominant cleaning mechanisms – cohesive separation, adhesive detachment, and viscous shifting – achieving a classification accuracy of 97.9% on independent experimental test data. The same feature definitions were subsequently transferred to industrial cleaning-run data of a shell-and-tube heat exchanger. When applied to more than 800 real cleaning runs of mainly pudding, the models consistently identified adhesive detachment as the dominant mechanism across the two main cleaning stages. Additionally, laboratory cleaning experiments conducted with the corresponding pudding products under near-industrial conditions supported these findings by reproducing similar phase-dependent removal dynamics. The study demonstrates the potential of mechanism-resolving machine learning for interpretable, data-driven monitoring of cleaning efficiency in industrial heat exchangers.
This study optimised the formulation of a freeze-dried functional powder produced from goat milk, goat whey, and green banana biomass (Musa paradisiaca L.), two agro-industrial by-products combined with fresh goat milk as the base matrix. A full factorial experimental design varying goat whey (17-51% w/w) and green banana biomass (10-30% w/w) concentrations was used to evaluate 18 physicochemical response variables. Multi-response desirability optimisation identified 34% w/w goat whey with 20% w/w green banana biomass as the best-compromise formulation, combining high starch content, adequate solubility, and acceptable hygroscopicity. Freezing kinetics modelling across four cryogenic temperatures identified −140 °C as the optimal pre-freezing condition, yielding the finest pore microstructure. The optimised powder showed a green banana biomass-dependent increase in phenolic content and antioxidant activity, together with a consistent mineral profile, supporting its potential use as a functional dairy ingredient in reconstituted beverages or fortified foods.
Stable front-end feeding is a key unit operation in machine-vision-based walnut-kernel grading lines because kernel accumulation, overlap, and jamming can reduce image-acquisition reliability and cause mechanical damage during food processing. This study developed a walnut-kernel dispersion feeder composed of an inclined guiding chamber and a reciprocating multi-layer dispersion pillar group. Representative DEM models of 1/2 kernels, 1/4 kernels, and broken kernels were constructed using three-dimensional scanning and the multi-sphere clump method. DEM analysis showed that kernel dispersion resulted from repeated collision-induced deceleration, inter-layer re-acceleration, posture deflection, and lateral migration under gravity-driven conveying, reciprocating pillar disturbance, inter-kernel contact, and geometric constraints. The guiding-chamber inclination angle, number of pillar layers, and horizontal pillar spacing were optimized using single-factor simulations and response surface methodology. The optimized parameters were 44.1°, eight pillar layers, and 40.5mm, respectively. Verification simulations showed relative errors of 3.27% for dispersion rate and 1.31% for feeding time. Bench-scale tests identified 0.3m/s conveyor-belt speed and 2.0Hz pillar reciprocating frequency as suitable operating parameters. Under the matched structural and operating conditions, the DEM and experimental feeding times differed by 1.00%, whereas the DEM dispersion rate was 7.09% lower than the experimental value. Under this condition, the feeder achieved a dispersion rate of 89.33%, a jamming rate of 3.67%, a feeding time of 19.93s, and a capacity of approximately 5 kernels/s. The newly generated damage rate was 2.44% ± 0.88%, and the successful visual recognition rate reached 92.78% ± 1.39%. These findings indicate that the DEM-assisted design workflow, followed by experimental evaluation, can support the development of stable and efficient front-end feeding systems for irregular and brittle food particulates while preserving product quality.
The concentrated disposal of agro-industrial waste represents a critical environmental challenge, driving the search for circular economy strategies that valorize underutilized biomass. Dragon fruit (Hylocereus spp.) peel, a byproduct rich in betalains, remains unexplored as a natural dye in various applications. This study aimed to develop a sustainable strategy for recovering betalains from lyophilized peels and evaluate its direct application in gelatin formulations. The peel powder presented a bimodal particle size distribution (peaks between 0.2–0.6 mm) and a porous microstructure, as evidenced by scanning electron microscopy, characteristics that favored mass transfer. Ultrasound-assisted removal was conducted in aqueous medium (pH 2, 4, and 6), demonstrating that mild conditions maximize recovery in the initial 5 min. Among the conditions evaluated, pH 6 confers greater stability to the pigment. Thermal release kinetics revealed activation energy (Ea) between 36 and 107 kJ·mol⁻¹ , confirming the strong thermal dependence and enhanced stability at pH 6. Incorporation of the extract into concentrated gelatins resulted in an intense and pure red color (a* ≈ 30.23; h° ≈ 357.96, CIELAB system). Texture profile analysis did not indicate significant changes in hardness (1.68 ± 2.62) and elasticity (97.18 ± 7.28) (p > 0.05), although a reduction in cohesiveness occurred (0.98 → 0.82). In sensory evaluation, the obtained formulation showed a significantly superior visual appearance (7.17) compared to the without dye without compromising the flavor profile. In summary, the results validate dragon fruit peel as a promising source of betalains for the food industry.
Hydrothermolysis pretreatment expedited the depolymerization, denaturation, fragmentation, degradation and bond fission/cleavage of biomass sample’s matrix complexes. Filterability anchored on diverse molecular weight differences of constituent compounds. 10 kilodalton (kDa) permeates showed high yield of fructose (83.49000 ± 0.00816%) alongside traces of lignin and other sugar compounds. More purified fructose (56.07000 ± 0.01732%), glucose (13.99000 ± 0.00816%) and sucrose (62.25000 ± 0.03317%) were obtained as 5kDa permeates which were free from traces of polymeric compounds. A combined flux vs. time comparison for the fluxes gave insights on how fouling progression varied across different operating conditions which is significant in guiding cleaning protocols, pretreatment strategies and process optimizations when implemented in a biorefinery process. Notably, preflux declined sharply, flux remained stable and postflux held average flux values with slight variations. 5kDa preflux run started high (~48.5L/m²·h) but dropped quickly to ~35.1L/m²·h within 20minutes, indicating rapid fouling that can be attributed to pore blocking or concentration polarization. Flux run maintained a low flux (~4.9L/m²·h) across nearly two hours, which was suggestive that the membrane was experiencing low fouling or the feed-in (biomass sample) had higher foulant load. Here, biomass hydrolysate feed-in was responsible for the low fouling which later stabilized and was overcome. This underscores the process can effectively be integrated into a biorefinery process for fractionation/separation of lignocelluloses biomass constituent compounds after pretreatment. Postflux run was average (~9.2L/m²·h) with slight improvement over time indicating average fouling. However, the system reached a quasi-steady flow state where flux stabilized. These comparisons demonstrated how fouling can manifest differently depending on feed conditions, pretreatment or membrane history when integrated into green-biorefinery engineering bioprocess design. For energy input during 10kDa ultrafiltration, about 197.16J (≈ 0.2kJ) was expended in processing feedstock sample of 328.6g (water + solubilized 3.1g sample) that operated for a duration of about 100minutes. During 5kDa process, about 176.4J (≈ 0.18kJ) was expended in processing feedstock of 294.24g (water + solubilized 1.24g sample) for a duration of about 119minutes. These are indicators of good energy efficiencies because the ultrafiltration bioprocesses expended less than 1kJ during each fractionation process. Data Metrics of Techno-economic Assessment (TEA) affirmed the feasibility/viability of the conceptualized green-biorefinery model because profits were to be made after sales of about 45 units of products within a short period of 5.4months with a net profit of about $55,436/year (100 units x $554.36). Life cycle analysis (LCA) and product valorization metrics underscored that valorizations of byproducts of retentate and permeate streams are significant in circular biorefinery design because they advance ecological sustainability.
To address chilled meat spoilage induced by rapid depletion of active ingredients in conventional packaging, starch-gelatin composite films incorporated with octenyl succinic anhydride-modified β-cyclodextrin/eugenol (OSA-β-CD/EUG) inclusion complexes were fabricated as sustained-release active packaging. The formation of stable host-guest complexes via hydrogen bonds and hydrophobic interactions was jointly confirmed by FT-IR, XRD, ¹H NMR and molecular docking, while favorable binding energy was quantified by docking simulations to elucidate intermolecular affinity. An encapsulation efficiency of 63.40 ± 1.01% was achieved by the optimal 1:1 molar complex, and 80.27 ± 0.15% DPPH radical scavenging activity was retained. First-order release kinetics were observed at 4 °C, and only 35.70 ± 1.69% of eugenol was cumulatively released within 12 days. When films were loaded with 1.0% OSA-β-CD/EUG, compact networks were formed, and enhanced flexibility (86.60 ± 1.13% elongation at break) and reduced oxygen permeability were exhibited, along with prominent antioxidant and antibacterial capacity. During chilled duck breast preservation, lipid and protein oxidation were retarded, TVB-N and TBARS accumulation was suppressed, and total viable counts were drastically reduced by the films over storage. Benefiting from the slow eugenol delivery, the composite film successfully extended the shelf life of duck meat by over 2 days.
Apricot (Prunus armeniaca) seeds, a valuable byproduct of apricot processing, are rich in carbohydrates, proteins, vitamins, and bioactive compounds. It is known for its remedial properties due to the existence of carotenoids, phytosterols, tocopherols, terpenoids, flavonoids, and phenolic compounds. Despite their nutritional value, apricot seeds are generally discarded as waste, leading to environmental concerns and the loss of beneficial nutrients. Therefore, the extraction of high-value compounds from apricot seeds through conventional (mechanical pressing, solvent extraction, and two-phase extraction) and green techniques (ultrasound-assisted extraction, microwave-assisted extraction, enzyme-assisted extraction, and supercritical fluid extraction), offers an efficient approach for valorization. The extracted compounds exhibited antioxidant, antimicrobial, antiallergic, antidiabetic, and anticarcinogenic properties, etc., supporting their application in food, nutraceutical, cosmetics, and pharmaceutical formulations. This review primarily focuses on the nutritional and phytochemical composition of apricot seed, extraction techniques for recovering high-value compounds and their application in different food sectors (bakery, dairy products, snacks, and beverages) and development of biofuel. In addition, the potential utilization of apricot seed-derived compounds in biofuel production and the development of bio-based polyurethane composites has been discussed. The efficient utilization of apricot seeds not only fulfil the nutritional demand but also helps to promote circular bioeconomy principles.
The valorization of agricultural biomass as a source of high-value bioactive compounds represents an important strategy for advancing in sustainable food and bioproduct processing. In this study, spinach (Spinacia oleracea) leaf biomass was investigated as a renewable feedstock for the recovery of coenzyme Q10 (CoQ10) and its subsequent application for the green synthesis of silver nanoparticles (AgNPs). A CoQ10-enriched extract was obtained through solvent extraction that was characterized by HPLC, confirming the presence of the coenzyme. The extract was applied as a natural reducing and stabilizing agent for AgNPs formation, providing an integrated biomass-to-bioproduct approach.AgNPs formation was observed by a visible color change and it was identified by the surface plasmon resonance with a characteristic peak at 410-420nm. Dynamic light scattering analysis revealed an average hydrodynamic diameter of approximately 38nm, while the zeta potential (−44mV) indicated high colloidal stability. FTIR spectra demonstrated the involvement of spinach-derived phytochemicals in both silver ions reduction and AgNPs stabilization. The antioxidant potential of the recovered extract and synthesized AgNPs was assessed using DPPH radical scavenging and FRAP assays, revealing significant antioxidant activity and suggesting the preservation of bioactive functionality throughout processing. The proposed approach demonstrates the potential of spinach biomass as a sustainable source of value-added feedstock and functional nanobioproducts. Considering the green synthesis of AgNPs using CoQ10-enriched extract, this work contributes to biomass valorization aligned with circular bioeconomy principles and development of innovative food and nutraceutical bioproducts.
Drying is a critical step for the industrial application of lactic acid bacteria, yet how drying-induced microstructural differences in starter powders subsequently regulate fermentation behavior and gel network formation remains poorly understood, particularly in plant-based systems. In this study, Lactobacillus helveticus MB2–1 powders prepared by spray drying (SD), freeze drying (FD), and spray-freeze-drying (SFD) were used to ferment coconut-based yogurt alternatives (CBYAs), and the effects of drying strategy on microbial viability, fermentation kinetics, protein interaction pathways, water distribution, and gel structure development were systematically investigated. The results showed that SD and FD led to structurally compromised powders characterized by heterogeneous aggregation and unbalanced intermolecular interactions, whereas SFD produced uniformly dispersed, viable powders that supported controlled acidification and homogeneous gel network development. Additionally, fermentation induced protein unfolding, redistribution of sulfhydryl groups, and progressive reorganization of intermolecular interactions across all systems. Notably, the SFD system uniquely maintained a balanced interplay among hydrophobic interactions (41.81%) and hydrogen bonding (30.93%), avoiding overreliance on any single dominant interaction. This synergistic interaction pattern facilitated gradual gel network assembly, promoted efficient conversion of free water into immobilized water, and resulted in a cohesive yet adaptive weak gel structure with superior rheological performance. Overall, this study demonstrates that SFD holds strong potential for the production of high-quality plant-based fermented foods by synchronously regulating microbial activity, protein network assembly, and final product microstructure. These findings provide a scientific basis for the controlled industrial development of plant-based fermented systems.
The global food packaging industry is shifting toward sustainable, bio-based materials that combine active protection with intelligent monitoring capabilities. This study aimed to develop and characterize starch-based films incorporating 20, 30, and 40% of red cabbage anthocyanins (RCA) as dual-functional active and intelligent packaging systems, coupled with an artificial neural network (ANN) model for predicting chicken breast freshness based on film colorimetric responses. Increasing RCA content significantly influenced their physicochemical properties, leading to higher thickness and transparency, alongside reduced density, hydrophobicity, and water vapor permeability. The FTIR spectra confirmed intermolecular interactions between RCA and the starch matrix through hydrogen bonding. RCA integration markedly enhanced antioxidant activity with a 14.1-fold increase in DPPH scavenging over the control, although no antimicrobial activity was observed against E. coli or S. aureus as all films produced inhibition zones of 0.0 mm. Principal component analysis identified S-RCA40 as the optimal formulation, balancing structural integrity and functional sensitivity. The S-RCA40 film exhibited distinct color responses over a pH range of 2–12 and high sensitivity to ammonia and dimethylamine vapors. When applied to chicken breast stored at 25 °C and 4 °C for 72 h, the indicator transitioned from violet–purple to dark blue, greenish, and brown/orange-red at room temperature, correlating with increases in pH and total volatile basic nitrogen (TVBN) beyond spoilage thresholds, while remaining stable under refrigeration. The ANN model (6–6–4–2–2), trained on RGB and CIELab inputs, achieved excellent predictive performance (R = 0.99998) for simultaneous estimation of pH and TVBN. This integrated system demonstrates strong potential for non-destructive, real-time monitoring of poultry freshness.
Wheat bran aqueous extracts (WBAE) promote the biomimetic formation of spherical brushite particles; however, the biochemical mechanism linking wheat bran processing to calcium phosphate mineralization remains unknown. This study investigated whether endogenous phytases activated during bran drying regulate phosphate availability and thereby control brushite formation. Wheat bran was dried at 24, 40, or 55 °C before aqueous extraction, while autoclaved bran served as an enzyme-inactivated control. Phytase activity, free phosphorus, and phytic acid contents were quantified in wheat bran from each pretreatment and the corresponding WBAE. After adding Ca2+ to each WBAE, the resulting particles were characterized by FTIR, XRD, SEM–EDS, and particle size analysis. Drying temperature markedly influenced phytase activity and phosphate availability. WBAE prepared from bran dried at 40 °C exhibited the highest phytase activity (207.3 μmol P/min/ g dry extract) and free phosphorus concentration (1.88g/100g), whereas autoclaving reduced enzymatic activity by 94% and completely prevented mineral precipitation. Phytic acid hydrolysis increased inorganic phosphate availability, providing the chemical precursor required for brushite precipitation under the experimental conditions. Highly monodisperse spherical brushite particles (mean diameter = 80.4 ± 13 μm; PDI = 0.03) were precipitated with WBAE from bran dried at 40 °C, displaying a characteristic hierarchical microstructure composed of radially organized lamellae, whereas drying at 24 or 55 °C yielded smaller and more polydisperse particles. These findings establish endogenous phytases as the mechanistic link between wheat bran processing and brushite biomimetic mineralization, identifying bran drying as a critical process parameter.
Beer, as one of the most consumed beverages in the world, is produced in large quantities. Its production requires huge amounts of energy and raw materials. Nowadays, one of the challenges for the brewing sector is sustain-ability in terms of energy consumption, transport, packaging, and byproducts valorization. This research work focuses on utilising brewing byproducts in the production cycle to create a new beer. Therefore, spent grains were used again in the mashing process, hopping was performed using spent hops from another beer batch, and fermentation was conducted using yeasts obtained from previous beer batches. The obtained beverage had comparable parameters to classical beer, and consumer tests showed high acceptability of the final product.
Astaxanthin is a high-value xanthophyll carotenoid with strong antioxidant activity and wide applications in nutraceuticals, cosmetics, and aquaculture. Microbial production using the red yeast Phaffia rhodozyma (syn. Xanthophyllomyces dendrorhous) represents a sustainable alternative to chemical synthesis, but production costs are a key obstacle to its large-scale deployment. In this study, an integrated biorefinery strategy is proposed to valorise artichoke (Cynara scolymus L.) by-products as a low-cost substrate for P. rhodozyma fermentation. These by-products, rich in lignocellulosic polysaccharides and inulin, were subjected to a combined ultrasound and enzymatic pre-treatment under mild conditions, increasing the sugar release by 146% compared to the untreated control. The resulting hydrolysate supported yeast growth and astaxanthin production in a 5 L bioreactor, reaching 310.3 mg L-1, with a final biomass of 15.8 g L-1 and almost complete depletion of sugars after 168 h. In a second valorisation step, the recovered fibrous fraction was used to prepare citric acid-crosslinked chitosan biocomposites with potential application as biodegradable food contact materials, which showed good thermal stability (T5% 170.87 degrees C; Tmax 257.47 degrees C) and a more ductile mechanical response (Young's modulus 189.33 MPa; elongation at break 17.83%) compared to the control consisting of chitosan, citric acid, and glycerol. Overall, this single-feedstock platform co-produces a high-value carotenoid and biodegradable composites, supporting circular bioeconomy options for artichoke supply chains.
Porang (Amorphophallus muelleri) flour is a potential source of glucomannan, but its functional quality is often limited by impurities, including calcium oxalate, starch, proteins, pigments, and non-glucomannan fibers. This study aimed to investigate both impurity removal and glucomannan enrichment through an equilibrium-based multistage ethanol purification process and to evaluate their effects on the physicochemical and morphological characteristics of purified porang flour. Purification was conducted using three stages of cross-current ethanol washing at graded ethanol concentrations (50-60-70%, 60-70-80%, and 70-80-90%) with an ethanol-to-flour ratio of 2:1 (v/w). Each washing stage was followed by centrifugation and drying before the subsequent purification stage. The multistage washing process was designed to selectively remove ethanol-soluble impurities while retaining glucomannan in the solid phase. An equilibrium-based approach using the distribution coefficient (KD) was applied to evaluate glucomannan retention and purification selectivity between solid and ethanol phases during purification. Higher KD values indicated more selective impurity removal with greater glucomannan retention in the solid phase. Both sulfur-treated and non-sulfur-treated samples were evaluated to determine sulfur pretreatment effects on purification performance. The highest quality was obtained in non-sulfur-treated samples purified using the 70-80-90% ethanol sequence, resulting in a glucomannan content of 90.02% db, transparency of 10.83%, apparent viscosity of 29,067 mPa.s, whiteness of 58.60, solubility of 58.02%, density 0.63 g/mL, and low calcium oxalate 0.08%, and the highest KD value of 3.60. SEM analysis confirmed a cleaner and more open glucomannan structure with reduced impurity coverage.
Thermal processing of ice cream mixes is carried out in conventional heat exchangers to assure safety during manufacturing. Significant ice cream related food safety issues are observed in the last two decades while use of green energy and reduction of carbon footprint of processes with a sustainable approach have become a strategic target for the UN 2030 Sustainable Development Goals (SDGs). Therefore, this study aimed to determine the effects of microwave heating (MW) on thermal processing of ice cream mixes. Listeria monocytogenes was chosen as the target microorganism, and MW processing on temperature evolution, microbial inactivation and color change were determined using a batch 1 kW 2450 MHz custom-designed system. A computational model was developed to predict the temperature change of the ice cream mix by coupling heat transfer with electromagnetic field distribution and fluid flow. Dielectric and rheological properties of the mix were measured as a function of temperature, and thermophysical properties were determined with respect to its composition. Model validation studies were completed with the experimentally obtained temperature data. MW heating of 500 mL sample for 600 s reduced the initial Listeria monocytogenes load from 6.51 +/- 0.03-3.81 +/- 0.01 log CFU/g, while the count decrease was below the detection limit after 750 s. Although statistically significant changes were observed in L*, a*, and b* values after the process, visually color deterioration was not detected. This study demonstrated the effect of MW heating for thermal processing of ice cream mixes, and knowing design parameters of an industrial process would be an important aspect.
