Pub Date : 2026-02-07DOI: 10.1016/j.cherd.2026.02.012
Luc Dewulf , Jordan M. MacInnes , Michael K. Hausmann , Annabel Bozon , Gerhard Niederreiter , Stefan Palzer , Agba D. Salman
Fat migration from compacted particulate foods into fibrous paper-based wrappers causes undesired fat stains on packaging and is a major concern for food manufacturers that are increasingly moving towards more sustainable paper-based packaging. While fibre-based materials are prone to absorb fats by capillary sorption, mechanisms of fat migration from the food matrix are dependent on the underlying food microstructure and full understanding is still lacking. Here, we developed a first-principle capillary flow model predicting liquid fat flow from model seasoning compacts (95 w/w% salt, 5 w/w% palm kernel fat) into contacting blotting paper. Compacts with systematic variations in salt particle size from 5 ≤ d50 ≤ 500 µm were produced in ternary design of experiments assessing the pore microstructure effect on capillarity and permeability. Measurements from x-ray microtomography and fat wicking kinetics were used to evaluate microstructural information for model parameters. Model validation was then performed in a physical set up characterising the fat migration behaviour on the compact side via Raman chemical imaging and on the paper side via optical stain imaging. Experiment and model were in better agreement (R2 up to 0.96) for compacts from coarse particles than for compacts with small porosity features. Yet, the model directed development towards using smaller particle sizes achieving almost 0 % fat migration into paper packaging for optimal samples.
{"title":"Fat migration from a particulate food system into fibrous material via capillary flow – first-principle modelling and experimental validation","authors":"Luc Dewulf , Jordan M. MacInnes , Michael K. Hausmann , Annabel Bozon , Gerhard Niederreiter , Stefan Palzer , Agba D. Salman","doi":"10.1016/j.cherd.2026.02.012","DOIUrl":"10.1016/j.cherd.2026.02.012","url":null,"abstract":"<div><div>Fat migration from compacted particulate foods into fibrous paper-based wrappers causes undesired fat stains on packaging and is a major concern for food manufacturers that are increasingly moving towards more sustainable paper-based packaging. While fibre-based materials are prone to absorb fats by capillary sorption, mechanisms of fat migration from the food matrix are dependent on the underlying food microstructure and full understanding is still lacking. Here, we developed a first-principle capillary flow model predicting liquid fat flow from model seasoning compacts (95 w/w% salt, 5 w/w% palm kernel fat) into contacting blotting paper. Compacts with systematic variations in salt particle size from 5 ≤ d<sub>50</sub> ≤ 500 µm were produced in ternary design of experiments assessing the pore microstructure effect on capillarity and permeability. Measurements from x-ray microtomography and fat wicking kinetics were used to evaluate microstructural information for model parameters. Model validation was then performed in a physical set up characterising the fat migration behaviour on the compact side via Raman chemical imaging and on the paper side via optical stain imaging. Experiment and model were in better agreement (R<sup>2</sup> up to 0.96) for compacts from coarse particles than for compacts with small porosity features. Yet, the model directed development towards using smaller particle sizes achieving almost 0 % fat migration into paper packaging for optimal samples.</div></div>","PeriodicalId":10019,"journal":{"name":"Chemical Engineering Research & Design","volume":"227 ","pages":"Pages 493-505"},"PeriodicalIF":3.9,"publicationDate":"2026-02-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146171264","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-07DOI: 10.1016/j.cherd.2026.02.010
Simon Ranthe Filtenborg , Peter Galsøe , Julie Senius Mølgaard , Dan Asbjørn Linnemann Axelsen , Carsten Skovmose Kallesøe , Maryam Tavakolmoghadam , Morten Lykkegaard Christensen , Astrid Ræbild Kjul , Mads Koustrup Jørgensen
Membrane filtration is a widely applied technology for water and wastewater treatment and for separation and purification in e.g. food and pharmaceutical industry. However, the applicability is severely limited by fouling. Several methods have been proposed to monitor membrane fouling, yet none have proven effective for full-scale implementation. The 3ω sensing is introduced as a novel approach for monitoring membrane fouling and shows promising potential to scale for in situ fouling monitoring. Promising results have been obtained for measuring filter-cake build up and compression (fouling) in dead-end filtration. In the current study, 3ω sensing is investigated for monitoring fouling in crossflow filtration to simultaneously measure how heat convection from the surface of the membrane depends on crossflow and formation of organic and inorganic fouling. A 3ω sensor was integrated onto the surface of a microfiltration membrane, and crossflow filtrations of kaolin and E. coli suspensions were conducted. It was observed that application of crossflow leads to a reduction of 3ω signal as it enhances heat transfer from the sensor. Measurements of 3ω signals at stagnant conditions (no crossflow) showed lower signals for membranes with inorganic fouling (thermally conducting) compared to a clean membrane, while measurements of a membrane fouled with E. coli shows a signal similar to that of a clean membrane due to the similarity in thermal conductivity between the feed and the fouling layer. Hence, the E. coli fouling layer could not be sensed in stagnant conditions. However, measurements in crossflow mode showed increasing 3ω signals by the formation of both kaolin and E. coli fouling layers. This happens because the fouling layer acts as a protective barrier against heat convection from the 3ω sensor, initially increasing the 3ω signal, regardless of the thermal conductivity. This phenomenon is coined shielding and has the notable consequence of increasing resolution of 3ω sensing for a foulant with thermal properties similar to those of water. This makes 3ω sensing an effective technique for detecting membrane fouling, with the potential to characterize both the type and thickness of the fouling layer with high resolution in crossflow filtration. These findings pave the way for advanced fouling diagnostics, predictive maintenance, and optimized cleaning strategies, offering substantial benefits for full-scale membrane operations in water and wastewater treatment, food, and pharmaceutical industries.
{"title":"In situ detection of fouling in crossflow filtration using 3ω sensing","authors":"Simon Ranthe Filtenborg , Peter Galsøe , Julie Senius Mølgaard , Dan Asbjørn Linnemann Axelsen , Carsten Skovmose Kallesøe , Maryam Tavakolmoghadam , Morten Lykkegaard Christensen , Astrid Ræbild Kjul , Mads Koustrup Jørgensen","doi":"10.1016/j.cherd.2026.02.010","DOIUrl":"10.1016/j.cherd.2026.02.010","url":null,"abstract":"<div><div>Membrane filtration is a widely applied technology for water and wastewater treatment and for separation and purification in e.g. food and pharmaceutical industry. However, the applicability is severely limited by fouling. Several methods have been proposed to monitor membrane fouling, yet none have proven effective for full-scale implementation. The 3ω sensing is introduced as a novel approach for monitoring membrane fouling and shows promising potential to scale for <em>in situ</em> fouling monitoring. Promising results have been obtained for measuring filter-cake build up and compression (fouling) in dead-end filtration. In the current study, 3ω sensing is investigated for monitoring fouling in crossflow filtration to simultaneously measure how heat convection from the surface of the membrane depends on crossflow and formation of organic and inorganic fouling. A 3ω sensor was integrated onto the surface of a microfiltration membrane, and crossflow filtrations of kaolin and <em>E. coli</em> suspensions were conducted. It was observed that application of crossflow leads to a reduction of 3ω signal as it enhances heat transfer from the sensor. Measurements of 3ω signals at stagnant conditions (no crossflow) showed lower signals for membranes with inorganic fouling (thermally conducting) compared to a clean membrane, while measurements of a membrane fouled with <em>E. coli</em> shows a signal similar to that of a clean membrane due to the similarity in thermal conductivity between the feed and the fouling layer. Hence, the <em>E. coli</em> fouling layer could not be sensed in stagnant conditions. However, measurements in crossflow mode showed increasing 3ω signals by the formation of both kaolin and <em>E. coli</em> fouling layers. This happens because the fouling layer acts as a protective barrier against heat convection from the 3ω sensor, initially increasing the 3ω signal, regardless of the thermal conductivity. This phenomenon is coined shielding and has the notable consequence of increasing resolution of 3ω sensing for a foulant with thermal properties similar to those of water. This makes 3ω sensing an effective technique for detecting membrane fouling, with the potential to characterize both the type and thickness of the fouling layer with high resolution in crossflow filtration. These findings pave the way for advanced fouling diagnostics, predictive maintenance, and optimized cleaning strategies, offering substantial benefits for full-scale membrane operations in water and wastewater treatment, food, and pharmaceutical industries.</div></div>","PeriodicalId":10019,"journal":{"name":"Chemical Engineering Research & Design","volume":"227 ","pages":"Pages 454-464"},"PeriodicalIF":3.9,"publicationDate":"2026-02-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146171201","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1016/j.cherd.2026.01.066
Achu Govind K.R.
Precise thermal regulation of rotary kilns is essential for maintaining product quality and energy efficiency in ZnO production. However, the strong nonlinear dynamics, distributed thermal behavior, and persistent disturbances in kiln operation pose significant challenges to conventional control strategies. These difficulties are compounded by model uncertainty, actuator limitations, and long-term drift in process characteristics. To address these issues, this work proposes a hybrid Active Disturbance Rejection Control (ADRC) with Long Short-Term Memory (LSTM) control framework. This integrates the robustness of ADRC with a data-driven residual correction learned by an LSTM network. The extended state observer (ESO) provides real-time disturbance estimation, while the LSTM compensates for unmodeled dynamics and estimator bias. The controller is designed to remain computationally lightweight and compatible with industrial implementation. A comprehensive evaluation is conducted, including an ablation study, parameter-uncertainty tests, gain-sensitivity analysis, actuator-fault simulations, Monte Carlo robustness assessment, and Lyapunov-based stability verification. The results show that the proposed ADRC-LSTM controller significantly improves transient and steady-state performance compared with nominal ADRC and a nonlinear MPC baseline. The hybrid controller achieves faster disturbance recovery, reduced undershoot, and smoother actuator usage. Robustness is maintained under -20% parameter variations and gain perturbations, and the system exhibits stable behavior under noise, actuator degradation, and fault conditions. Monte Carlo analysis confirms consistent closed-loop performance, while Lyapunov analysis verifies satisfaction of stability conditions. Overall, the proposed architecture provides a reliable, efficient, and fault-tolerant solution for advanced rotary-kiln temperature control.
{"title":"Learning-assisted active disturbance rejection control for robust temperature regulation of industrial rotary kilns","authors":"Achu Govind K.R.","doi":"10.1016/j.cherd.2026.01.066","DOIUrl":"10.1016/j.cherd.2026.01.066","url":null,"abstract":"<div><div>Precise thermal regulation of rotary kilns is essential for maintaining product quality and energy efficiency in ZnO production. However, the strong nonlinear dynamics, distributed thermal behavior, and persistent disturbances in kiln operation pose significant challenges to conventional control strategies. These difficulties are compounded by model uncertainty, actuator limitations, and long-term drift in process characteristics. To address these issues, this work proposes a hybrid Active Disturbance Rejection Control (ADRC) with Long Short-Term Memory (LSTM) control framework. This integrates the robustness of ADRC with a data-driven residual correction learned by an LSTM network. The extended state observer (ESO) provides real-time disturbance estimation, while the LSTM compensates for unmodeled dynamics and estimator bias. The controller is designed to remain computationally lightweight and compatible with industrial implementation. A comprehensive evaluation is conducted, including an ablation study, parameter-uncertainty tests, gain-sensitivity analysis, actuator-fault simulations, Monte Carlo robustness assessment, and Lyapunov-based stability verification. The results show that the proposed ADRC-LSTM controller significantly improves transient and steady-state performance compared with nominal ADRC and a nonlinear MPC baseline. The hybrid controller achieves faster disturbance recovery, reduced undershoot, and smoother actuator usage. Robustness is maintained under <span><math><mrow><mo>±</mo><mn>10</mn></mrow></math></span>-20% parameter variations and <span><math><mrow><mo>±</mo><mn>20</mn><mtext>%</mtext></mrow></math></span> gain perturbations, and the system exhibits stable behavior under noise, actuator degradation, and fault conditions. Monte Carlo analysis confirms consistent closed-loop performance, while Lyapunov analysis verifies satisfaction of stability conditions. Overall, the proposed architecture provides a reliable, efficient, and fault-tolerant solution for advanced rotary-kiln temperature control.</div></div>","PeriodicalId":10019,"journal":{"name":"Chemical Engineering Research & Design","volume":"227 ","pages":"Pages 400-415"},"PeriodicalIF":3.9,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146171262","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1016/j.cherd.2026.02.008
Zifeng Jin
As the core separation unit in petrochemical processes, the real-time optimization of distillation columns is crucial for energy efficiency and product quality. Traditional mechanistic models are computationally expensive for dynamic conditions, while data-driven approaches often ignore the mass/heat transfer topology between trays, leading to significant errors. Although graph neural networks (GNNs) can model system structure, they face challenges like unstable training, inadequate heterogeneous graph modeling, and high latency. To address these issues, this study proposes a real-time decision-making system using a Dynamic Spatiotemporal Graph Convolutional Network (DST-GCN). It employs heterogeneous node encoding to distinguish tray states and component properties, a dynamic adjacency matrix to capture spatiotemporal evolution of energy transfer, and uses orthogonal convolution with regularization to stabilize training. Integrated with feedforward-feedback control and edge computing, the system achieves millisecond response and lightweight deployment. Industrial tests in ethylene and methanol separation demonstrate improved purity control, reduced steam and power consumption, high fault detection accuracy, strong robustness, and cross-condition generalization. The results confirm the feasibility of GNNs in real-time optimization of complex industrial systems. Future work will focus on transfer learning, equipment degradation integration, and federated learning to advance from single-column to plant-wide intelligence.
{"title":"Optimization model of distillation tower operation based on graph neural network: Real time decision system for multi component separation","authors":"Zifeng Jin","doi":"10.1016/j.cherd.2026.02.008","DOIUrl":"10.1016/j.cherd.2026.02.008","url":null,"abstract":"<div><div>As the core separation unit in petrochemical processes, the real-time optimization of distillation columns is crucial for energy efficiency and product quality. Traditional mechanistic models are computationally expensive for dynamic conditions, while data-driven approaches often ignore the mass/heat transfer topology between trays, leading to significant errors. Although graph neural networks (GNNs) can model system structure, they face challenges like unstable training, inadequate heterogeneous graph modeling, and high latency. To address these issues, this study proposes a real-time decision-making system using a Dynamic Spatiotemporal Graph Convolutional Network (DST-GCN). It employs heterogeneous node encoding to distinguish tray states and component properties, a dynamic adjacency matrix to capture spatiotemporal evolution of energy transfer, and uses orthogonal convolution with regularization to stabilize training. Integrated with feedforward-feedback control and edge computing, the system achieves millisecond response and lightweight deployment. Industrial tests in ethylene and methanol separation demonstrate improved purity control, reduced steam and power consumption, high fault detection accuracy, strong robustness, and cross-condition generalization. The results confirm the feasibility of GNNs in real-time optimization of complex industrial systems. Future work will focus on transfer learning, equipment degradation integration, and federated learning to advance from single-column to plant-wide intelligence.</div></div>","PeriodicalId":10019,"journal":{"name":"Chemical Engineering Research & Design","volume":"227 ","pages":"Pages 465-479"},"PeriodicalIF":3.9,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146171330","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1016/j.cherd.2026.02.007
Jiawei zhou, Shihao Lv, Xi Wei, Yichen Liu, Shun Guo, Lijian Li
To investigate the influence of material properties on the silo discharge characteristics, a Plexiglas gravity silo experiment system was constructed. Wall pressure measurements and material flow patterns were conducted for analysis. Based on these, the discharge behavior and discharge mass rates of various bulk materials were systematically investigated under both conical and sidewall orifice conditions. Six typical bulk materials were tested, including soybean, carbon black masterbatch, coarse sand, fine sand, industrial salt and white corundum. Quantitative relationships of the discharge coefficient and both the internal and wall friction angles of materials were established based on the Beverloo equation. It indicates that the discharge coefficient is strongly influenced by the internal friction angle of material. Discharge experiments through sidewall openings of internal pipe bundles in a blending silo were carried out using soybeans, carbon black masterbatch, and industrial salt as the test materials. A comparative analysis of mass flow rate characteristics at different material layer depths was performed. A predictive equation for sidewall orifice mass flow rate was developed by dimensional analysis, accounting for material properties and material pressure. The prediction error of the proposed model was within 10 %.
{"title":"Analysis of discharge mass flow rate for gravity blending silo","authors":"Jiawei zhou, Shihao Lv, Xi Wei, Yichen Liu, Shun Guo, Lijian Li","doi":"10.1016/j.cherd.2026.02.007","DOIUrl":"10.1016/j.cherd.2026.02.007","url":null,"abstract":"<div><div>To investigate the influence of material properties on the silo discharge characteristics, a Plexiglas gravity silo experiment system was constructed. Wall pressure measurements and material flow patterns were conducted for analysis. Based on these, the discharge behavior and discharge mass rates of various bulk materials were systematically investigated under both conical and sidewall orifice conditions. Six typical bulk materials were tested, including soybean, carbon black masterbatch, coarse sand, fine sand, industrial salt and white corundum. Quantitative relationships of the discharge coefficient and both the internal and wall friction angles of materials were established based on the Beverloo equation. It indicates that the discharge coefficient is strongly influenced by the internal friction angle of material. Discharge experiments through sidewall openings of internal pipe bundles in a blending silo were carried out using soybeans, carbon black masterbatch, and industrial salt as the test materials. A comparative analysis of mass flow rate characteristics at different material layer depths was performed. A predictive equation for sidewall orifice mass flow rate was developed by dimensional analysis, accounting for material properties and material pressure. The prediction error of the proposed model was within 10 %.</div></div>","PeriodicalId":10019,"journal":{"name":"Chemical Engineering Research & Design","volume":"227 ","pages":"Pages 435-445"},"PeriodicalIF":3.9,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146171269","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-02DOI: 10.1016/j.cherd.2026.02.003
Xuefang Gao , Dewu Wang , Yan Liu , Ruojin Wang , Baisong Hu , Lei Wang , Hongrui Wei , Shaofeng Zhang , Meng Tang
Fluid mixing plays a significant role in the mass transfer process. Static mixers are core insert elements in the mixing industry. The relationship between the vortex evolution induced by the perforated structures and the resultant chaotic mixing performance in a rotational–perforated static mixer (RPSM) remains to be quantitatively elucidated. To elucidate the impact of perforated structures on chaotic mixing, we employed Planar Laser Induced Fluorescence (PLIF) to visualize the vortex evolution in RPSMs under different installations. The transient and spatial development of vortices was traced using optical flow and edge detection algorithms, while the mixing performance was assessed by the coefficient of variation (CoV) and the largest Lyapunov exponent (LLE). It was found that the perforated structures promoted rotational strengthening in the central area and periodic near-wall vortex changes; however, a larger aperture ratio adversely affected mixing. The outer ring area exhibited superior mixing performance to the center. The backward installation conferred a significant advantage, reducing the CoV by 11–56 % compared to the forward installation. Subsequent factors analysis enabled the calculation of CoV, and the close agreement (relative error < 15 %) between calculated and experimental values validated the findings.
{"title":"Experimental study of vortex evolution process and chaotic mixing analysis of rotational–perforated static mixers with different perforated structures","authors":"Xuefang Gao , Dewu Wang , Yan Liu , Ruojin Wang , Baisong Hu , Lei Wang , Hongrui Wei , Shaofeng Zhang , Meng Tang","doi":"10.1016/j.cherd.2026.02.003","DOIUrl":"10.1016/j.cherd.2026.02.003","url":null,"abstract":"<div><div>Fluid mixing plays a significant role in the mass transfer process. Static mixers are core insert elements in the mixing industry. The relationship between the vortex evolution induced by the perforated structures and the resultant chaotic mixing performance in a rotational–perforated static mixer (RPSM) remains to be quantitatively elucidated. To elucidate the impact of perforated structures on chaotic mixing, we employed Planar Laser Induced Fluorescence (PLIF) to visualize the vortex evolution in RPSMs under different installations. The transient and spatial development of vortices was traced using optical flow and edge detection algorithms, while the mixing performance was assessed by the coefficient of variation (<em>CoV</em>) and the largest Lyapunov exponent (<em>LLE</em>). It was found that the perforated structures promoted rotational strengthening in the central area and periodic near-wall vortex changes; however, a larger aperture ratio adversely affected mixing. The outer ring area exhibited superior mixing performance to the center. The backward installation conferred a significant advantage, reducing the <em>CoV</em> by 11–56 % compared to the forward installation. Subsequent factors analysis enabled the calculation of <em>CoV</em>, and the close agreement (relative error < 15 %) between calculated and experimental values validated the findings.</div></div>","PeriodicalId":10019,"journal":{"name":"Chemical Engineering Research & Design","volume":"227 ","pages":"Pages 416-434"},"PeriodicalIF":3.9,"publicationDate":"2026-02-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146171271","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The kinetics of n-hexane oxidative cracking to olefins using lattice oxygen on (V/Si)-ZSM-5 were examined. The catalyst shows a total acidity of 0.44 mmol g⁻¹ , comprising 47 % weak sites that promote olefin selectivity and 53 % strong sites that enhance n-hexane conversion. TPR confirms that (V/Si)-ZSM-5 is readily reducible. In a CREC Riser simulator, oxidative cracking resulted in ∼72.2 % olefin selectivity at 74.8 % n-hexane conversion. A kinetic model was formulated, including (1) catalytic cracking and (2) oxidative dehydrogenation. The cracking pathway treats adsorption, C–H/C–C bond cleavage, and desorption as elementary steps under a pseudo–steady-state assumption, while the ODH reaction follows a Langmuir–Hinshelwood mechanism. The model reproduces the experiments with strong statistical agreement, and the estimated rate constantly aligns with the observed product selectivity.
{"title":"Phenomenological-based kinetics of oxidative cracking of n-hexane to light olefins over tandem (V/Si)-ZSM-5 catalysts","authors":"Ariel Hazril Gursida , Sagir Adamu , Shaikh Abdur Razzak , Mohammad Mozahar Hossain","doi":"10.1016/j.cherd.2026.02.002","DOIUrl":"10.1016/j.cherd.2026.02.002","url":null,"abstract":"<div><div>The kinetics of <em>n</em>-hexane oxidative cracking to olefins using lattice oxygen on (V/Si)-ZSM-5 were examined. The catalyst shows a total acidity of 0.44 mmol g⁻¹ , comprising 47 % weak sites that promote olefin selectivity and 53 % strong sites that enhance <em>n</em>-hexane conversion. TPR confirms that (V/Si)-ZSM-5 is readily reducible. In a CREC Riser simulator, oxidative cracking resulted in ∼72.2 % olefin selectivity at 74.8 % <em>n</em>-hexane conversion. A kinetic model was formulated, including (1) catalytic cracking and (2) oxidative dehydrogenation. The cracking pathway treats adsorption, C–H/C–C bond cleavage, and desorption as elementary steps under a pseudo–steady-state assumption, while the ODH reaction follows a Langmuir–Hinshelwood mechanism. The model reproduces the experiments with strong statistical agreement, and the estimated rate constantly aligns with the observed product selectivity.</div></div>","PeriodicalId":10019,"journal":{"name":"Chemical Engineering Research & Design","volume":"227 ","pages":"Pages 480-491"},"PeriodicalIF":3.9,"publicationDate":"2026-02-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146171202","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-01DOI: 10.1016/j.cherd.2026.01.059
Haotong Pang , Chenqiang Qin , Shuzhen Hu , Xuefeng Yan , Fakai Zhang , Rui Sun , Youbin Zhao , Chaolan Tang , Jiajia Ren
This study presents a cooling structure for a low-temperature micro-grinding machine and conducts a numerical simulation of the cooling system using CFD and DEM to determine the optimal operating parameters. The study investigated the effects of variations in cooling fluid temperature, flow rate, and inner wall thickness of the flow channels on temperature uniformity in the grinding chamber, particle temperature distribution, cooling efficiency, and the load-bearing capacity of the inner walls. The results show that injecting cooling fluid at −20°C to −30°C results in better temperature uniformity in the grinding chamber at 0°C to 20°C compared to −30°C to −10°C, while the cooling effect is inversely proportional. Additionally, the consistency of particle temperature distribution and cooling efficiency across six tests was assessed, model error ＜ 7.27 %. A coolant flow rate of 0.6–2.1 m/s was used to simulate the cooling effect on particles after 30 s of operation at 0°C, and the optimal economic solution was identified, model error ＜ 3.94 %. The optimal inner wall thickness of the cooling channel was determined to be 13 mm. The prototype test results indicate a particle size distribution with D50 = 31 ± 3 µm and D90 = 75 ± 8 µm. These findings will support the investigation of low-temperature cooling effects on micro-grinding and provide a theoretical foundation for its practical applications.
{"title":"Simulation analysis of low-temperature micro-grinding system based on CFD-DEM","authors":"Haotong Pang , Chenqiang Qin , Shuzhen Hu , Xuefeng Yan , Fakai Zhang , Rui Sun , Youbin Zhao , Chaolan Tang , Jiajia Ren","doi":"10.1016/j.cherd.2026.01.059","DOIUrl":"10.1016/j.cherd.2026.01.059","url":null,"abstract":"<div><div>This study presents a cooling structure for a low-temperature micro-grinding machine and conducts a numerical simulation of the cooling system using CFD and DEM to determine the optimal operating parameters. The study investigated the effects of variations in cooling fluid temperature, flow rate, and inner wall thickness of the flow channels on temperature uniformity in the grinding chamber, particle temperature distribution, cooling efficiency, and the load-bearing capacity of the inner walls. The results show that injecting cooling fluid at −20°C to −30°C results in better temperature uniformity in the grinding chamber at 0°C to 20°C compared to −30°C to −10°C, while the cooling effect is inversely proportional. Additionally, the consistency of particle temperature distribution and cooling efficiency across six tests was assessed, model error ＜ 7.27 %. A coolant flow rate of 0.6–2.1 m/s was used to simulate the cooling effect on particles after 30 s of operation at 0°C, and the optimal economic solution was identified, model error ＜ 3.94 %. The optimal inner wall thickness of the cooling channel was determined to be 13 mm. The prototype test results indicate a particle size distribution with D<sub>50</sub> = 31 ± 3 µm and D<sub>90</sub> = 75 ± 8 µm. These findings will support the investigation of low-temperature cooling effects on micro-grinding and provide a theoretical foundation for its practical applications.</div></div>","PeriodicalId":10019,"journal":{"name":"Chemical Engineering Research & Design","volume":"227 ","pages":"Pages 374-387"},"PeriodicalIF":3.9,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146171263","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This study reports the preparation of high-quality, large-particle ammonium sulfate crystals via vacuum evaporation crystallization, employing ammonium sulfamate and manganese sulfate as composite additives. The process was conducted under both large seed crystal and decelerated crystallization systems, significantly increasing the yield of large-particle crystals. Building on prior systematic investigations of ammonium sulfate crystallization without additives, the authors utilized orthogonal and single-factor experiments to optimize parameters for two novel preparation methods. In the large seed crystal system, the mass fractions of crystals larger than 2.0 mm and 1.4 mm reach 31.36 % and 67.30 %, respectively. In the decelerated system, despite a 50.00 % reduction in additive amount, a 79.17 % decrease in seed crystal size, a 46.15 % reduction in seed loading, and a 20.00 % extension of crystallization time, the mass fractions of crystals exceeding 2.0 mm and 1.4 mm decrease only marginally by 3.66 % and 6.27 %, respectively. The resulting crystals exhibit a crystallinity of 89.38 %, an aspect ratio of 1.02, and demonstrate superior mechanical strength, flowability, and sustained-release properties compared to those produced in the large seed crystal system. By analyzing crystallization kinetic curves across different systems, the microscopic mechanisms of the additives were elucidated, and a high-accuracy crystallization kinetic equation was derived.
{"title":"A novel method for preparing high-quality large-particle ammonium sulfate crystals","authors":"Peng Zhang, Xueru Wang, Lei Xu, Meiqi Zhang, Yuting Weng, Baozeng Ren","doi":"10.1016/j.cherd.2026.01.065","DOIUrl":"10.1016/j.cherd.2026.01.065","url":null,"abstract":"<div><div>This study reports the preparation of high-quality, large-particle ammonium sulfate crystals via vacuum evaporation crystallization, employing ammonium sulfamate and manganese sulfate as composite additives. The process was conducted under both large seed crystal and decelerated crystallization systems, significantly increasing the yield of large-particle crystals. Building on prior systematic investigations of ammonium sulfate crystallization without additives, the authors utilized orthogonal and single-factor experiments to optimize parameters for two novel preparation methods. In the large seed crystal system, the mass fractions of crystals larger than 2.0 mm and 1.4 mm reach 31.36 % and 67.30 %, respectively. In the decelerated system, despite a 50.00 % reduction in additive amount, a 79.17 % decrease in seed crystal size, a 46.15 % reduction in seed loading, and a 20.00 % extension of crystallization time, the mass fractions of crystals exceeding 2.0 mm and 1.4 mm decrease only marginally by 3.66 % and 6.27 %, respectively. The resulting crystals exhibit a crystallinity of 89.38 %, an aspect ratio of 1.02, and demonstrate superior mechanical strength, flowability, and sustained-release properties compared to those produced in the large seed crystal system. By analyzing crystallization kinetic curves across different systems, the microscopic mechanisms of the additives were elucidated, and a high-accuracy crystallization kinetic equation was derived.</div></div>","PeriodicalId":10019,"journal":{"name":"Chemical Engineering Research & Design","volume":"227 ","pages":"Pages 309-327"},"PeriodicalIF":3.9,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146171257","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The absence of reliable online sensors for substrate () and ethanol () concentrations remains a critical limitation in industrial ethanol fermentation. An energy-balance-based soft sensor was developed to estimate and using temperature and water flow data in combination with thermodynamic and stoichiometric parameters. The model was calibrated under non-stripping fed-batch conditions at 34 °C, yielding a heat yield coefficient () of 0.0946 g kJ−1 (corresponding to a heat generation per gram of biomass of = 10.57 kJ g−1), an ethanol yield coefficient () of 0.463 ± 0.024 g ethanol (g TRS)−1, and a biomass yield coefficient () of 0.0451 ± 0.0039 g biomass (g TRS)−1. Validation in fermentations under distinct operational conditions, including extractive CO2 stripping and very high gravity (VHG) feeding, confirmed high predictive accuracy. Coefficients of determination (R²) exceeded 0.98 and mean absolute percentage errors (MAPE) remained below 8 %. These results demonstrate that the proposed framework provides a reliable basis for real-time estimation of substrate and ethanol concentrations under anaerobic conditions.
{"title":"Heat balance soft sensor for ethanol and substrate monitoring in very high gravity fed-batch fermentations with CO2 stripping","authors":"V.T. Mazziero , I.I.K. Veloso , B.G. Campos , M.V. Santos , A.J.G. Cruz , A.C. Badino , M.O. Cerri","doi":"10.1016/j.cherd.2026.01.064","DOIUrl":"10.1016/j.cherd.2026.01.064","url":null,"abstract":"<div><div>The absence of reliable online sensors for substrate (<span><math><msub><mrow><mi>C</mi></mrow><mrow><mi>S</mi></mrow></msub></math></span>) and ethanol (<span><math><msub><mrow><mi>C</mi></mrow><mrow><mi>E</mi></mrow></msub></math></span>) concentrations remains a critical limitation in industrial ethanol fermentation. An energy-balance-based soft sensor was developed to estimate <span><math><msub><mrow><mi>C</mi></mrow><mrow><mi>S</mi></mrow></msub></math></span> and <span><math><msub><mrow><mi>C</mi></mrow><mrow><mi>E</mi></mrow></msub></math></span> using temperature and water flow data in combination with thermodynamic and stoichiometric parameters. The model was calibrated under non-stripping fed-batch conditions at 34 °C, yielding a heat yield coefficient (<span><math><msub><mrow><mi>Y</mi></mrow><mrow><mi>H</mi></mrow></msub></math></span>) of 0.0946 g kJ<sup>−1</sup> (corresponding to a heat generation per gram of biomass <span><math><msub><mrow><mi>Y</mi></mrow><mrow><mrow><mi>Q</mi></mrow><mo>/</mo><mrow><mi>X</mi></mrow></mrow></msub></math></span> of = 10.57 kJ g<sup>−1</sup>), an ethanol yield coefficient (<span><math><msub><mrow><mi>Y</mi></mrow><mrow><mrow><mi>E</mi></mrow><mo>/</mo><mrow><mi>S</mi></mrow></mrow></msub></math></span>) of 0.463 ± 0.024 g ethanol (g TRS)<sup>−1</sup>, and a biomass yield coefficient (<span><math><msub><mrow><mi>Y</mi></mrow><mrow><mrow><mi>X</mi></mrow><mo>/</mo><mrow><mi>S</mi></mrow></mrow></msub></math></span>) of 0.0451 ± 0.0039 g biomass (g TRS)<sup>−1</sup>. Validation in fermentations under distinct operational conditions, including extractive CO<sub>2</sub> stripping and very high gravity (VHG) feeding, confirmed high predictive accuracy. Coefficients of determination (R²) exceeded 0.98 and mean absolute percentage errors (MAPE) remained below 8 %. These results demonstrate that the proposed framework provides a reliable basis for real-time estimation of substrate and ethanol concentrations under anaerobic conditions.</div></div>","PeriodicalId":10019,"journal":{"name":"Chemical Engineering Research & Design","volume":"227 ","pages":"Pages 297-308"},"PeriodicalIF":3.9,"publicationDate":"2026-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146171267","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号