Pub Date : 2024-12-26DOI: 10.1016/j.dche.2024.100210
Priscila Marques da Paz , Caroline Satye Martins Nakama , Galo Antonio Carrillo Le Roux
Increasing the productivity of a biotechnological process becomes feasible through the development of Process Systems Engineering tools, which integrate experimental data with mathematical modeling. This work aims to develop a software architecture for modeling bioprocesses that is accessible to a multidisciplinary group. To achieve this aim, the software must be thoroughly designed based on an ontology that describes bioprocesses that can be apprehended by researchers from different fields. The ontological representation is carried out using Unified Modeling Language diagrams, whose use is demonstrated by a parameter estimation case study. It is concluded that good software development practices can be provided through the proposed architecture, since it guides simulations and parameter estimations of biotechnological processes in a structured way.
{"title":"Development of a software architecture for bioprocess modeling","authors":"Priscila Marques da Paz , Caroline Satye Martins Nakama , Galo Antonio Carrillo Le Roux","doi":"10.1016/j.dche.2024.100210","DOIUrl":"10.1016/j.dche.2024.100210","url":null,"abstract":"<div><div>Increasing the productivity of a biotechnological process becomes feasible through the development of Process Systems Engineering tools, which integrate experimental data with mathematical modeling. This work aims to develop a software architecture for modeling bioprocesses that is accessible to a multidisciplinary group. To achieve this aim, the software must be thoroughly designed based on an ontology that describes bioprocesses that can be apprehended by researchers from different fields. The ontological representation is carried out using Unified Modeling Language diagrams, whose use is demonstrated by a parameter estimation case study. It is concluded that good software development practices can be provided through the proposed architecture, since it guides simulations and parameter estimations of biotechnological processes in a structured way.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100210"},"PeriodicalIF":3.0,"publicationDate":"2024-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143159038","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-23DOI: 10.1016/j.dche.2024.100212
Peter Jul-Rasmussen , Mads Stevnsborg , Xiaodong Liang , Jakob Kjøbsted Huusom
Digital twins are frequently discussed in a bio-manufacturing context, but actual realisations of digital twins are rare. To use digital twin instances, significant investments in digital infrastructure and high-fidelity mathematical models are required. This work presents a real-time implementation of an ensemble hybrid model with incremental learning for predicting dissolved oxygen concentration in a pilot-scale bubble column. A bootstrap-aggregated hybrid modelling framework is applied for constructing an ensemble of 1000 hybrid models using different partitions of the training/validation data, providing a measure of the parameter distributions and prediction uncertainty. Each model in the ensemble hybrid model has the same model structure relying on first-principles material balances and an Artificial Neural Network for prediction of the liquid phase volumetric mass transfer coefficient. Incremental learning is applied, efficiently enabling the model to adapt to new data acquired during runtime. The software implementation follows recent ISO issues using a modular structure allowing for flexible allocation of server resources and an intuitive User-Interface is developed for controlling the application. From a real-time prediction study, the models using incremental learning are found to have superior performance both at normal operating conditions, when interpolating, and when extrapolating compared to using only the pre-trained model.
{"title":"Implementation of real-time incremental learning for ensemble hybrid model prediction in pilot scale bubble column aeration","authors":"Peter Jul-Rasmussen , Mads Stevnsborg , Xiaodong Liang , Jakob Kjøbsted Huusom","doi":"10.1016/j.dche.2024.100212","DOIUrl":"10.1016/j.dche.2024.100212","url":null,"abstract":"<div><div>Digital twins are frequently discussed in a bio-manufacturing context, but actual realisations of digital twins are rare. To use digital twin instances, significant investments in digital infrastructure and high-fidelity mathematical models are required. This work presents a real-time implementation of an ensemble hybrid model with incremental learning for predicting dissolved oxygen concentration in a pilot-scale bubble column. A bootstrap-aggregated hybrid modelling framework is applied for constructing an ensemble of 1000 hybrid models using different partitions of the training/validation data, providing a measure of the parameter distributions and prediction uncertainty. Each model in the ensemble hybrid model has the same model structure relying on first-principles material balances and an Artificial Neural Network for prediction of the liquid phase volumetric mass transfer coefficient. Incremental learning is applied, efficiently enabling the model to adapt to new data acquired during runtime. The software implementation follows recent ISO issues using a modular structure allowing for flexible allocation of server resources and an intuitive User-Interface is developed for controlling the application. From a real-time prediction study, the models using incremental learning are found to have superior performance both at normal operating conditions, when interpolating, and when extrapolating compared to using only the pre-trained model.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100212"},"PeriodicalIF":3.0,"publicationDate":"2024-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143160286","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-21DOI: 10.1016/j.dche.2024.100215
Foteini Michalopoulou , Maria M. Papathanasiou
Chromatographic separation process models are described by nonlinear partial differential and algebraic equations, often leading to high computational cost that limits their applicability in real-time applications. To address this, in this work we propose a hybrid modelling approach that integrates artificial neural networks with process knowledge to describe the system nonlinear dynamics. Specifically, the separation isotherm is maintained in its mechanistic form, while the need for spatial discretisation is eliminated, reducing computational effort by 97 % in the open-loop simulation. The resulting hybrid model relies solely on experimentally measurable variables and performs well both in interpolation and extrapolation tests. It is further utilised within a process optimisation framework, for the maximisation of process yield and product purity. The results demonstrate that the hybrid model accurately captures the intricate dynamics of chromatographic separations while providing a computationally efficient alternative, making it an effective tool for development in industrial applications.
{"title":"An approach to hybrid modelling in chromatographic separation processes","authors":"Foteini Michalopoulou , Maria M. Papathanasiou","doi":"10.1016/j.dche.2024.100215","DOIUrl":"10.1016/j.dche.2024.100215","url":null,"abstract":"<div><div>Chromatographic separation process models are described by nonlinear partial differential and algebraic equations, often leading to high computational cost that limits their applicability in real-time applications. To address this, in this work we propose a hybrid modelling approach that integrates artificial neural networks with process knowledge to describe the system nonlinear dynamics. Specifically, the separation isotherm is maintained in its mechanistic form, while the need for spatial discretisation is eliminated, reducing computational effort by 97 % in the open-loop simulation. The resulting hybrid model relies solely on experimentally measurable variables and performs well both in interpolation and extrapolation tests. It is further utilised within a process optimisation framework, for the maximisation of process yield and product purity. The results demonstrate that the hybrid model accurately captures the intricate dynamics of chromatographic separations while providing a computationally efficient alternative, making it an effective tool for development in industrial applications.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100215"},"PeriodicalIF":3.0,"publicationDate":"2024-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143159672","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this study, hydrated lime (Ca(OH)₂) sorbent was prepared from industrial brine sludge waste using simple laboratory procedures and utilized in a laboratory-scale spray dry scrubber for desulfurization tests. The effects of key process parameters in spray drying (sorbent particle size, inlet gas phase temperature, and Ca:S ratio) on desulfurization efficiency were investigated using central composite design (CCD). Three machine learning (ML) models, multilayer perceptron (MLP), support vector regressor (SVR), and light gradient boosting machine (LightGBM), were assessed for their output estimation accuracy and compared to the CCD prediction model. The computational framework utilized experimental variables structured by CCD software as input metadata. Model performance was evaluated through generalization and accuracy measurements, including the coefficient of determination (R²), root mean square error (RMSE), mean square error (MSE), and mean squared logarithmic error (MSLE). Analysis of variance revealed that the Ca:S ratio had the most significant influence on SO₂ absorption. A quadratic model correlating the process variables with desulfurization efficiency was developed, yielding an R-squared value of 93.47%. Characterization of the final desulfurization products, particularly using XRD, showed the emergence of new phases such as hannebachite (CaSO3.0·5H2O), while FTIR analysis identified unreacted portlandite and calcite. Among the ML models, the MLP demonstrated superior performance over SVR and LightGBM, highlighting its efficacy in extracting and decoding information from the input data. The response surface methodology (RSM) model also proved to be a reliable forecasting tool, indicating its potential as a practical alternative to complex algorithmic computations in scenarios with limited raw data.
{"title":"Machine learning and response surface methodology forecasting comparison for improved spray dry scrubber performance with brine sludge-derived sorbent","authors":"B.J. Chepkonga , L. Koech , R.S. Makomere , H.L. Rutto","doi":"10.1016/j.dche.2024.100214","DOIUrl":"10.1016/j.dche.2024.100214","url":null,"abstract":"<div><div>In this study, hydrated lime (Ca(OH)₂) sorbent was prepared from industrial brine sludge waste using simple laboratory procedures and utilized in a laboratory-scale spray dry scrubber for desulfurization tests. The effects of key process parameters in spray drying (sorbent particle size, inlet gas phase temperature, and Ca:S ratio) on desulfurization efficiency were investigated using central composite design (CCD). Three machine learning (ML) models, multilayer perceptron (MLP), support vector regressor (SVR), and light gradient boosting machine (LightGBM), were assessed for their output estimation accuracy and compared to the CCD prediction model. The computational framework utilized experimental variables structured by CCD software as input metadata. Model performance was evaluated through generalization and accuracy measurements, including the coefficient of determination (R²), root mean square error (RMSE), mean square error (MSE), and mean squared logarithmic error (MSLE). Analysis of variance revealed that the Ca:S ratio had the most significant influence on SO₂ absorption. A quadratic model correlating the process variables with desulfurization efficiency was developed, yielding an R-squared value of 93.47%. Characterization of the final desulfurization products, particularly using XRD, showed the emergence of new phases such as hannebachite (CaSO<sub>3</sub>.0·5H<sub>2</sub>O), while FTIR analysis identified unreacted portlandite and calcite. Among the ML models, the MLP demonstrated superior performance over SVR and LightGBM, highlighting its efficacy in extracting and decoding information from the input data. The response surface methodology (RSM) model also proved to be a reliable forecasting tool, indicating its potential as a practical alternative to complex algorithmic computations in scenarios with limited raw data.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100214"},"PeriodicalIF":3.0,"publicationDate":"2024-12-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143160284","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-19DOI: 10.1016/j.dche.2024.100209
Jesse Y. Rumbo-Morales , Gerardo Ortiz-Torres , Felipe D.J. Sorcia-Vázquez , Carlos Alberto Torres-Cantero , Jair Gómez Radilla , Mario Martínez García , Julio César Rodríguez-Cerda , Antonio Márquez Rosales , Moises Ramos-Martinez , Juan Carlos Mixteco-Sánchez , Mayra G. Mena-Enriquez , Mario A. Juarez
Biofuels produced from renewable raw materials, in this case bioethanol, provide a sustainable and renewable energy source for the future, as bioethanol positively impacts the economy, the environment, and society. Bioethanol is an alternative and immediate solution to mitigate the main greenhouse gases generated by transportation and industries that use fossil fuels. However, to produce bioethanol it is necessary to use advanced dehydration processes or technologies. Currently, azeotropic distillation, extractive distillation, and the Pressure Swing Adsorption (PSA) process using selective zeolites on water molecules are used. This PSA process has shown high selectivity, high yield, and high energy efficiency for producing anhydrous ethanol compared to other technologies. This work aims to implement automatic control laws (geometric and PID) to maintain stable the desired purity (99.5%), have higher bioethanol recovery and generate higher productivity using less energy. Both controllers performed adequately on the PSA bioethanol-producing plant, however, the geometric control presented greater robustness against disturbances, achieving to maintain stable bioethanol purity above 99% by wt, generating a recovery of 73.62%, with productivity of 59.07 and using an energy efficiency of 59.21%. Using this control law, it was possible to use the entire length of the columns to adsorb a greater amount of water molecules and achieve higher production.
{"title":"Energy efficiency and productivity of a Pressure Swing Adsorption plant to purify bioethanol: Disturbance attenuation through geometric control","authors":"Jesse Y. Rumbo-Morales , Gerardo Ortiz-Torres , Felipe D.J. Sorcia-Vázquez , Carlos Alberto Torres-Cantero , Jair Gómez Radilla , Mario Martínez García , Julio César Rodríguez-Cerda , Antonio Márquez Rosales , Moises Ramos-Martinez , Juan Carlos Mixteco-Sánchez , Mayra G. Mena-Enriquez , Mario A. Juarez","doi":"10.1016/j.dche.2024.100209","DOIUrl":"10.1016/j.dche.2024.100209","url":null,"abstract":"<div><div>Biofuels produced from renewable raw materials, in this case bioethanol, provide a sustainable and renewable energy source for the future, as bioethanol positively impacts the economy, the environment, and society. Bioethanol is an alternative and immediate solution to mitigate the main greenhouse gases generated by transportation and industries that use fossil fuels. However, to produce bioethanol it is necessary to use advanced dehydration processes or technologies. Currently, azeotropic distillation, extractive distillation, and the Pressure Swing Adsorption (PSA) process using selective zeolites on water molecules are used. This PSA process has shown high selectivity, high yield, and high energy efficiency for producing anhydrous ethanol compared to other technologies. This work aims to implement automatic control laws (geometric and PID) to maintain stable the desired purity (99.5%), have higher bioethanol recovery and generate higher productivity using less energy. Both controllers performed adequately on the PSA bioethanol-producing plant, however, the geometric control presented greater robustness against disturbances, achieving to maintain stable bioethanol purity above 99% by wt, generating a recovery of 73.62%, with productivity of 59.07 <span><math><mrow><mi>k</mi><mi>m</mi><mi>o</mi><mi>l</mi></mrow></math></span> and using an energy efficiency of 59.21%. Using this control law, it was possible to use the entire length of the columns to adsorb a greater amount of water molecules and achieve higher production.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100209"},"PeriodicalIF":3.0,"publicationDate":"2024-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143160285","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-18DOI: 10.1016/j.dche.2024.100208
Fernando Arrais R.D. Lima , Marcellus G.F. de Moraes , Amaro G. Barreto Jr , Argimiro R. Secchi , Martha A. Grover , Maurício B. de Souza Jr
Crystallization is a separation method relevant to the production of medicines, food and many other products. An efficient crystallization process must obtain a product with the desired size, length, and purity. Therefore, models and control schemes are applied to achieve this goal. Artificial intelligence techniques, such as machine learning (ML), are applied for modeling and controlling these processes. The current review aims to present the use of ML for modeling and advanced control of crystallization processes. Considering modeling crystallization processes, this paper presents the advances and different uses of ML, such as neural networks, symbolic regression, and transformer algorithms. This review also presents the development of hybrid models combining ML with physical laws for crystallization processes. For the advanced control of crystallization processes, this review presents the development of advanced control strategies based on ML approaches, such as applying neural networks in a nonlinear model predictive controller and based on reinforcement learning. This work can be a relevant reference for the progress of the application of ML in the process systems engineering (PSE) to crystallization processes. It is also expected to encourage industry and academy to use these approaches for different crystallization processes.
{"title":"Applications of machine learning for modeling and advanced control of crystallization processes: Developments and perspectives","authors":"Fernando Arrais R.D. Lima , Marcellus G.F. de Moraes , Amaro G. Barreto Jr , Argimiro R. Secchi , Martha A. Grover , Maurício B. de Souza Jr","doi":"10.1016/j.dche.2024.100208","DOIUrl":"10.1016/j.dche.2024.100208","url":null,"abstract":"<div><div>Crystallization is a separation method relevant to the production of medicines, food and many other products. An efficient crystallization process must obtain a product with the desired size, length, and purity. Therefore, models and control schemes are applied to achieve this goal. Artificial intelligence techniques, such as machine learning (ML), are applied for modeling and controlling these processes. The current review aims to present the use of ML for modeling and advanced control of crystallization processes. Considering modeling crystallization processes, this paper presents the advances and different uses of ML, such as neural networks, symbolic regression, and transformer algorithms. This review also presents the development of hybrid models combining ML with physical laws for crystallization processes. For the advanced control of crystallization processes, this review presents the development of advanced control strategies based on ML approaches, such as applying neural networks in a nonlinear model predictive controller and based on reinforcement learning. This work can be a relevant reference for the progress of the application of ML in the process systems engineering (PSE) to crystallization processes. It is also expected to encourage industry and academy to use these approaches for different crystallization processes.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100208"},"PeriodicalIF":3.0,"publicationDate":"2024-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143159059","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-16DOI: 10.1016/j.dche.2024.100213
Shahina Riaz , Nabeel Ahmad , Wasif Farooq , Imtiaz Ali , Mohd Sajid , Muhammad Naseem Akhtar
Kinetic study is crucial in digital chemical engineering as a foundation for understanding and optimizing chemical processes. By analyzing reaction rates and mechanisms, kinetic models provide essential data for designing reactors, scaling processes, and predicting performance under various conditions. This study is part of a broader research series focused on efficiently converting waste plastics into hydrocarbons through catalytic pyrolysis. As the first study from the series, it investigates the thermal degradation of pristine high-density polyethylene (HDPE), aiming to understand its reaction kinetics under catalytic and non-catalytic conditions. The research employs iso-conventional methods to estimate the activation of energy HDPE and leverages a machine learning algorithm, specifically the BRT model, to effectively predict activation energy and optimize pyrolysis parameters. The activation energy (323 kJ/mol) of non-catalytic pyrolysis of HDPE was reduced to 164 kJ/mol during catalytic pyrolysis. Thermodynamic parameters such as change in activation enthalpy (), activation Gibbs free energy () and, activation entropy () were also significantly reduced during catalytic reaction. The statistical and machine-learning approaches were used in the kinetic analyses. Boosted regression trees (BRT) were used to predict the during conversion at different heating rates for non-catalytic and catalytic processes. The liquid and gas fractions obtained from HDPE at different temperatures were characterized. The increase in yield of hydrocarbons at elevated temperatures indicated the reuse potential of plastic waste. The comprehensive analysis of HDPE exhibited 86 % of carbon and 14 % of hydrogen contributing to high heating value (HHV) of 44.41 MJ/kg.
{"title":"Catalytic pyrolysis of HDPE for enhanced hydrocarbon yield: A boosted regression tree assisted kinetics study for effective recycling of waste plastic","authors":"Shahina Riaz , Nabeel Ahmad , Wasif Farooq , Imtiaz Ali , Mohd Sajid , Muhammad Naseem Akhtar","doi":"10.1016/j.dche.2024.100213","DOIUrl":"10.1016/j.dche.2024.100213","url":null,"abstract":"<div><div>Kinetic study is crucial in digital chemical engineering as a foundation for understanding and optimizing chemical processes. By analyzing reaction rates and mechanisms, kinetic models provide essential data for designing reactors, scaling processes, and predicting performance under various conditions. This study is part of a broader research series focused on efficiently converting waste plastics into hydrocarbons through catalytic pyrolysis. As the first study from the series, it investigates the thermal degradation of pristine high-density polyethylene (HDPE), aiming to understand its reaction kinetics under catalytic and non-catalytic conditions. The research employs iso-conventional methods to estimate the activation of energy HDPE and leverages a machine learning algorithm, specifically the BRT model, to effectively predict activation energy and optimize pyrolysis parameters. The activation energy (323 kJ/mol) of non-catalytic pyrolysis of HDPE was reduced to 164 kJ/mol during catalytic pyrolysis. Thermodynamic parameters such as change in activation enthalpy (<span><math><mrow><mstyle><mi>Δ</mi></mstyle><msup><mrow><mi>H</mi></mrow><mo>‡</mo></msup></mrow></math></span>), activation Gibbs free energy (<span><math><mrow><mstyle><mi>Δ</mi></mstyle><msup><mrow><mi>G</mi></mrow><mo>‡</mo></msup></mrow></math></span>) and, activation entropy (<span><math><mrow><mstyle><mi>Δ</mi></mstyle><msup><mrow><mi>S</mi></mrow><mo>‡</mo></msup></mrow></math></span>) were also significantly reduced during catalytic reaction. The statistical and machine-learning approaches were used in the kinetic analyses. Boosted regression trees (BRT) were used to predict the<span><math><mrow><mspace></mspace><msub><mi>E</mi><mi>a</mi></msub></mrow></math></span> during conversion at different heating rates for non-catalytic and catalytic processes. The liquid and gas fractions obtained from HDPE at different temperatures were characterized. The increase in yield of hydrocarbons at elevated temperatures indicated the reuse potential of plastic waste. The comprehensive analysis of HDPE exhibited 86 % of carbon and 14 % of hydrogen contributing to high heating value (HHV) of 44.41 MJ/kg.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100213"},"PeriodicalIF":3.0,"publicationDate":"2024-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143159063","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-13DOI: 10.1016/j.dche.2024.100207
Eslam G. Al-Sakkari , Ahmed Ragab , Mostafa Amer , Olumoye Ajao , Marzouk Benali , Daria C. Boffito , Hanane Dagdougui , Mouloud Amazouz
Several processes and strategies have been developed to promote the utilization of lignin and to facilitate its market adoption across a broad spectrum of applications within the expanding lignin bioeconomy. However, the inherent variability in lignin properties, resulting from diverse feedstock sources and varied recovery and downstream processing methods, remains a significant challenge. This highlights the critical need to investigate lignin's miscibility and reactivity with polymers and solvents, as most lignin valorization pathways involve mixing, blending, or solubilization. Accurate estimation of Hansen solubility parameters (HSP) is crucial for solvent selection in several fields such as polymer science, coatings, adhesives, lignin-based biorefineries and solvent-based carbon capture. Traditional methods for predicting HSP are time-consuming and involve complex experiments, especially in applications dealing with carbon dioxide and lignin solubility. This paper introduces a novel ensemble modeling methodology based on machine learning (ML) techniques for accurate HSP prediction using Simplified Molecular Input Line Entry System (SMILES) codes as entries. The methodology integrates different ML approaches, including deep and shallow learning, to enhance prediction accuracy. Decision fusion of individual ML models is achieved through a hybrid approach combining non-learnable and learnable methods, resulting in reduced errors and enhanced accuracy. The results highlight the effectiveness of the ensemble-based methodology, which achieved 99% accuracy in predicting dispersion solubility parameters, outperforming other individual ML techniques. The proposed generic methodology, from data preprocessing to decision fusion through diverse ML algorithms, can be applied to various chemical analytics beyond HSP prediction.
{"title":"Ensemble machine learning to accelerate industrial decarbonization: Prediction of Hansen solubility parameters for streamlined chemical solvent selection","authors":"Eslam G. Al-Sakkari , Ahmed Ragab , Mostafa Amer , Olumoye Ajao , Marzouk Benali , Daria C. Boffito , Hanane Dagdougui , Mouloud Amazouz","doi":"10.1016/j.dche.2024.100207","DOIUrl":"10.1016/j.dche.2024.100207","url":null,"abstract":"<div><div>Several processes and strategies have been developed to promote the utilization of lignin and to facilitate its market adoption across a broad spectrum of applications within the expanding lignin bioeconomy. However, the inherent variability in lignin properties, resulting from diverse feedstock sources and varied recovery and downstream processing methods, remains a significant challenge. This highlights the critical need to investigate lignin's miscibility and reactivity with polymers and solvents, as most lignin valorization pathways involve mixing, blending, or solubilization. Accurate estimation of Hansen solubility parameters (HSP) is crucial for solvent selection in several fields such as polymer science, coatings, adhesives, lignin-based biorefineries and solvent-based carbon capture. Traditional methods for predicting HSP are time-consuming and involve complex experiments, especially in applications dealing with carbon dioxide and lignin solubility. This paper introduces a novel ensemble modeling methodology based on machine learning (ML) techniques for accurate HSP prediction using Simplified Molecular Input Line Entry System (SMILES) codes as entries. The methodology integrates different ML approaches, including deep and shallow learning, to enhance prediction accuracy. Decision fusion of individual ML models is achieved through a hybrid approach combining non-learnable and learnable methods, resulting in reduced errors and enhanced accuracy. The results highlight the effectiveness of the ensemble-based methodology, which achieved 99% accuracy in predicting dispersion solubility parameters, outperforming other individual ML techniques. The proposed generic methodology, from data preprocessing to decision fusion through diverse ML algorithms, can be applied to various chemical analytics beyond HSP prediction.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100207"},"PeriodicalIF":3.0,"publicationDate":"2024-12-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143159065","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-13DOI: 10.1016/j.dche.2024.100203
Guilherme V. Espinosa, Amanda L.T. Brandão
Oxidative dehydrogenation of propane using CO2 (ODPC) is among the most investigated on-purpose processes to meet the increased propylene demand, due to the necessity to reduce CO2 emissions. In this context, the present work simulated an ODPC reactor integrated with chemical looping combustion (CLC) of biogas, which provides the necessary heat, and CO2 capture technology in Aspen Plus. The simulation was evaluated based on economic and sustainability criteria. In addition, a kinetic model was proposed and validated for a sufficient range of operation. It was possible to achieve net present value (NPV) of -14.86 106 US$, over a 15-year operational period, based on current carbon pricing policies. However, the potential profitability of the process was demonstrated by investigating the effects of more favorable carbon credit policies, with an increase from 50 to 120 US$ tCO2eq-1 resulting in a NPV of 164.15 106 US$ and 4 years payback period.
{"title":"Economic and sustainability evaluation of green CO2-assisted propane dehydrogenation design","authors":"Guilherme V. Espinosa, Amanda L.T. Brandão","doi":"10.1016/j.dche.2024.100203","DOIUrl":"10.1016/j.dche.2024.100203","url":null,"abstract":"<div><div>Oxidative dehydrogenation of propane using CO<sub>2</sub> (ODPC) is among the most investigated on-purpose processes to meet the increased propylene demand, due to the necessity to reduce CO<sub>2</sub> emissions. In this context, the present work simulated an ODPC reactor integrated with chemical looping combustion (CLC) of biogas, which provides the necessary heat, and CO<sub>2</sub> capture technology in Aspen Plus. The simulation was evaluated based on economic and sustainability criteria. In addition, a kinetic model was proposed and validated for a sufficient range of operation. It was possible to achieve net present value (NPV) of -14.86 10<sup>6</sup> US$, over a 15-year operational period, based on current carbon pricing policies. However, the potential profitability of the process was demonstrated by investigating the effects of more favorable carbon credit policies, with an increase from 50 to 120 US$ tCO<sub>2</sub>eq<sup>-1</sup> resulting in a NPV of 164.15 10<sup>6</sup> US$ and 4 years payback period.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100203"},"PeriodicalIF":3.0,"publicationDate":"2024-12-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143158979","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-10DOI: 10.1016/j.dche.2024.100206
Henrik Wang , Feiyang Ou , Julius Suherman , Gerassimos Orkoulas , Panagiotis D. Christofides
Control methods for Atomic Layer Etching (ALE) processes are constantly evolving due to the increasing level of precision needed to manufacture next-gen semiconductor devices. This work presents a novel, real-time Endpoint-based (EP) control approach for an Al2O3 ALE process in a discrete feed reactor. The proposed method dynamically adjusts the process time of both ALE half-cycles to ensure an optimal process outcome. The EP controller uses a machine learning-based transformer to take in variable-length, time-series pressure profiles to identify when the ALE process is complete. However, this model requires a large amount of process data to ensure that it will perform well even when under a variety of kinetic and pressure disturbances that mimic common issues in a real-world manufacturing environment. Thus, this work uses a multiscale modeling method that integrates a macroscopic Computational Fluid Dynamics (CFD) and a mesoscopic kinetic Monte Carlo (kMC) simulation to generate process data and test the proposed controllers. After testing the performance of the EP controller on individual runs, various combinations of ex-situ Run-to-Run (R2R) and EP controllers are examined in order to determine the strongest control strategy in a manufacturing environment. The final results show that the EP controller is highly accurate when trained on conditions that are representative of its implementation environment. Compared to traditional EWMA controllers, it has significantly fewer misprocesses, which enhances the overall control performance and efficiency of the ALE process.
{"title":"Integration of on-line machine learning-based endpoint control and run-to-run control for an atomic layer etching process","authors":"Henrik Wang , Feiyang Ou , Julius Suherman , Gerassimos Orkoulas , Panagiotis D. Christofides","doi":"10.1016/j.dche.2024.100206","DOIUrl":"10.1016/j.dche.2024.100206","url":null,"abstract":"<div><div>Control methods for Atomic Layer Etching (ALE) processes are constantly evolving due to the increasing level of precision needed to manufacture next-gen semiconductor devices. This work presents a novel, real-time Endpoint-based (EP) control approach for an Al<sub>2</sub>O<sub>3</sub> ALE process in a discrete feed reactor. The proposed method dynamically adjusts the process time of both ALE half-cycles to ensure an optimal process outcome. The EP controller uses a machine learning-based transformer to take in variable-length, time-series pressure profiles to identify when the ALE process is complete. However, this model requires a large amount of process data to ensure that it will perform well even when under a variety of kinetic and pressure disturbances that mimic common issues in a real-world manufacturing environment. Thus, this work uses a multiscale modeling method that integrates a macroscopic Computational Fluid Dynamics (CFD) and a mesoscopic kinetic Monte Carlo (kMC) simulation to generate process data and test the proposed controllers. After testing the performance of the EP controller on individual runs, various combinations of ex-situ Run-to-Run (R2R) and EP controllers are examined in order to determine the strongest control strategy in a manufacturing environment. The final results show that the EP controller is highly accurate when trained on conditions that are representative of its implementation environment. Compared to traditional EWMA controllers, it has significantly fewer misprocesses, which enhances the overall control performance and efficiency of the ALE process.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100206"},"PeriodicalIF":3.0,"publicationDate":"2024-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143160282","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","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