Lichao Ge, Can Zhao, Yang Wang, Zhifu Qi, Ruikun Wang, Qianqian Yin, Yuli Zhang, Chang Xu
Copyrolysis of lignin and cellulose can effectively improve pore structure and optimize product distribution. Therefore, the distribution, characteristics, components, and formation mechanism of the copyrolysis products of cellulose and sodium lignosulfonate were studied. The pyrolysis of sodium lignosulfonate was effectively inhibited by cellulose, especially when the amount of doped cellulose was 40 wt.%, and tubes presumed to be carbon nanotubes were prepared under these conditions. For bio‐oil, the contents of phenol, 2‐methoxy‐, and 4‐aminopyridine increased with decreasing amounts of doped cellulose. However, cellulose substantially reduced the content of 2‐furanmethanol. H2, CO2, CO, and CH4 were the main components of the biogas; among them, H2 was the most abundant component in the biogas. Considering the characteristics of the three‐phase product, a higher C content in the volatiles (especially bio‐oil) can promote the formation of carbon nanotubes. Finally, the formation mechanism and interactions of the main components in the volatiles of cellulose and sodium lignosulfonate were proposed.
{"title":"Effects of cellulose addition on sodium lignosulfonate pyrolysis: Product distribution and formation pathway","authors":"Lichao Ge, Can Zhao, Yang Wang, Zhifu Qi, Ruikun Wang, Qianqian Yin, Yuli Zhang, Chang Xu","doi":"10.1002/cjce.25450","DOIUrl":"https://doi.org/10.1002/cjce.25450","url":null,"abstract":"Copyrolysis of lignin and cellulose can effectively improve pore structure and optimize product distribution. Therefore, the distribution, characteristics, components, and formation mechanism of the copyrolysis products of cellulose and sodium lignosulfonate were studied. The pyrolysis of sodium lignosulfonate was effectively inhibited by cellulose, especially when the amount of doped cellulose was 40 wt.%, and tubes presumed to be carbon nanotubes were prepared under these conditions. For bio‐oil, the contents of phenol, 2‐methoxy‐, and 4‐aminopyridine increased with decreasing amounts of doped cellulose. However, cellulose substantially reduced the content of 2‐furanmethanol. H<jats:sub>2</jats:sub>, CO<jats:sub>2</jats:sub>, CO, and CH<jats:sub>4</jats:sub> were the main components of the biogas; among them, H<jats:sub>2</jats:sub> was the most abundant component in the biogas. Considering the characteristics of the three‐phase product, a higher C content in the volatiles (especially bio‐oil) can promote the formation of carbon nanotubes. Finally, the formation mechanism and interactions of the main components in the volatiles of cellulose and sodium lignosulfonate were proposed.","PeriodicalId":501204,"journal":{"name":"The Canadian Journal of Chemical Engineering","volume":"32 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141931507","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mohammad Amin Moradkhani, Ali Reza Miroliaei, Nasim Ghasemi, Seyyed Hossein Hosseini, Mikel Tellabide, Martin Olazar
The present study concerns the development of new models to estimate the minimum spouting velocity (Ums) in various configurations of fountain‐confined conical spouted beds (FC‐CSBs) with fine particles. Existing literature correlations were found to be inaccurate for FC‐CSBs. Therefore, smart modelling techniques were employed to design more accurate predictive tools. The radial basis function (RBF) approach provided the best predictions for systems without draft tubes as well as those with open‐sided draft tubes. Additionally, the Gaussian process regression (GPR) approach yielded the best predictions for systems with nonporous draft tubes. The mean absolute percentage error (MAPE) values for the testing phase were 5.80%, 5.67%, and 5.59%, respectively. These models consider how bed shape and particle properties affect Ums. The sensitivity analysis was conducted to determine the factors with more importance in controlling Ums. Finally, simpler correlations were derived for Ums prediction in different FC‐CSB configurations, with accuracy around 12% error.
{"title":"Minimum spouting velocity of fine particles in fountain confined conical spouted beds using machine learning and least square fitting approaches","authors":"Mohammad Amin Moradkhani, Ali Reza Miroliaei, Nasim Ghasemi, Seyyed Hossein Hosseini, Mikel Tellabide, Martin Olazar","doi":"10.1002/cjce.25429","DOIUrl":"https://doi.org/10.1002/cjce.25429","url":null,"abstract":"The present study concerns the development of new models to estimate the minimum spouting velocity (<jats:italic>U</jats:italic><jats:sub>ms</jats:sub>) in various configurations of fountain‐confined conical spouted beds (FC‐CSBs) with fine particles. Existing literature correlations were found to be inaccurate for FC‐CSBs. Therefore, smart modelling techniques were employed to design more accurate predictive tools. The radial basis function (RBF) approach provided the best predictions for systems without draft tubes as well as those with open‐sided draft tubes. Additionally, the Gaussian process regression (GPR) approach yielded the best predictions for systems with nonporous draft tubes. The mean absolute percentage error (MAPE) values for the testing phase were 5.80%, 5.67%, and 5.59%, respectively. These models consider how bed shape and particle properties affect <jats:italic>U</jats:italic><jats:sub>ms</jats:sub>. The sensitivity analysis was conducted to determine the factors with more importance in controlling <jats:italic>U</jats:italic><jats:sub>ms</jats:sub>. Finally, simpler correlations were derived for <jats:italic>U</jats:italic><jats:sub>ms</jats:sub> prediction in different FC‐CSB configurations, with accuracy around 12% error.","PeriodicalId":501204,"journal":{"name":"The Canadian Journal of Chemical Engineering","volume":"94 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141931455","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This manuscript presents a proof of concept for the estimation of parameters in a bioprocess while providing reliable confidence intervals. Specifically, Bayesian inference is used to estimate the uncertainty in the prediction of a parameter due to the presence of measurement noise in the process. The resultant joint probability distribution is utilized to infer the confidence interval of the resultant estimates. This method is numerically applied using a technique known as nested sampling. This algorithm iteratively samples parameters from a pre‐determined range of values to compare model predictions and obtain a probability density function. One challenge typically associated with this algorithm is in the determination of the prediction error, especially when a high‐fidelity dynamic model is being utilized. For the motivating example in the present manuscript, where a high‐fidelity simulated bioprocess is being considered, the use of the high‐fidelity model provided by Sartorius AG as part of the estimation algorithm poses computational challenges. To overcome this challenge, a universal approximator such as a parameterized neural network is used. This neural network is designed to simulate the results of the first principles model (while also capturing the dependence of the model parameters on the output), and once trained can provide near instantaneous results making the use of nested sampling computationally tractable for the application. Simulation results demonstrate the feasibility and capability of the proposed approach.
本手稿提出了一个概念验证，用于估算生物过程中的参数，同时提供可靠的置信区间。具体来说，贝叶斯推理用于估算由于过程中存在测量噪声而导致的参数预测不确定性。利用由此产生的联合概率分布来推断结果估计值的置信区间。这种方法使用一种称为嵌套采样的技术进行数值计算。该算法从预先确定的数值范围内反复采样参数，以比较模型预测值并获得概率密度函数。这种算法通常面临的一个挑战是如何确定预测误差，尤其是在使用高保真动态模型时。在本手稿的激励性示例中，考虑了高保真模拟生物过程，使用 Sartorius AG 提供的高保真模型作为估算算法的一部分带来了计算上的挑战。为了克服这一挑战，我们使用了参数化神经网络等通用近似器。这种神经网络旨在模拟第一原理模型的结果（同时还能捕捉模型参数对输出的依赖性），一旦经过训练，就能提供近乎瞬时的结果，从而使嵌套采样的应用在计算上变得简单易行。仿真结果证明了所提方法的可行性和能力。
{"title":"Leveraging neural networks to estimate parameters with confidence intervals","authors":"Nigel Mathias, Lauren Weir, Brandon Corbett, Prashant Mhaskar","doi":"10.1002/cjce.25438","DOIUrl":"https://doi.org/10.1002/cjce.25438","url":null,"abstract":"This manuscript presents a proof of concept for the estimation of parameters in a bioprocess while providing reliable confidence intervals. Specifically, Bayesian inference is used to estimate the uncertainty in the prediction of a parameter due to the presence of measurement noise in the process. The resultant joint probability distribution is utilized to infer the confidence interval of the resultant estimates. This method is numerically applied using a technique known as nested sampling. This algorithm iteratively samples parameters from a pre‐determined range of values to compare model predictions and obtain a probability density function. One challenge typically associated with this algorithm is in the determination of the prediction error, especially when a high‐fidelity dynamic model is being utilized. For the motivating example in the present manuscript, where a high‐fidelity simulated bioprocess is being considered, the use of the high‐fidelity model provided by Sartorius AG as part of the estimation algorithm poses computational challenges. To overcome this challenge, a universal approximator such as a parameterized neural network is used. This neural network is designed to simulate the results of the first principles model (while also capturing the dependence of the model parameters on the output), and once trained can provide near instantaneous results making the use of nested sampling computationally tractable for the application. Simulation results demonstrate the feasibility and capability of the proposed approach.","PeriodicalId":501204,"journal":{"name":"The Canadian Journal of Chemical Engineering","volume":"2 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141931454","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pranita A. Karekar, Vishwanath H. Dalvi, Chandrakanth R. Gadipelly, Ashwin W. Patwardhan
This work reports hydrodynamic and mass transfer studies on a novel microreactor that can passively break up liquid–liquid slugs using judiciously placed internals. The reactors were fabricated in stainless steel (SS‐316 L, hereafter SS) and PMMA (hereafter acrylic). The performance of both is comparable to the current state‐of‐the‐art in microreactor technologies. A separated flow model is proposed to estimate the pressure drop for two‐phase flows, with a mean absolute error (MAE) of 15.44% in SS and 19.83% in acrylic, respectively. Pulse tracer experiments were performed for residence time distribution (RTD) studies. They are fitted to a model for the prediction of RTD for single and two‐phase flows. The results obtained from mass transfer experiments show that the volumetric mass transfer coefficient () in the case of SS reactor is, on average, 2.4 times higher than acrylic. A correlation is developed for estimating the based on total velocity and phase fraction, providing better fits than the models based on energy dissipation. All studies show that wall characteristics significantly impact the hydrodynamics and mass transfer phenomena since the pressure drop and the are greater in (the rougher) SS than in acrylic.
{"title":"Hydrodynamics and mass transfer studies on plate‐type microchannel reactor for liquid–liquid systems","authors":"Pranita A. Karekar, Vishwanath H. Dalvi, Chandrakanth R. Gadipelly, Ashwin W. Patwardhan","doi":"10.1002/cjce.25441","DOIUrl":"https://doi.org/10.1002/cjce.25441","url":null,"abstract":"This work reports hydrodynamic and mass transfer studies on a novel microreactor that can passively break up liquid–liquid slugs using judiciously placed internals. The reactors were fabricated in stainless steel (SS‐316 L, hereafter SS) and PMMA (hereafter acrylic). The performance of both is comparable to the current state‐of‐the‐art in microreactor technologies. A separated flow model is proposed to estimate the pressure drop for two‐phase flows, with a mean absolute error (MAE) of 15.44% in SS and 19.83% in acrylic, respectively. Pulse tracer experiments were performed for residence time distribution (RTD) studies. They are fitted to a model for the prediction of RTD for single and two‐phase flows. The results obtained from mass transfer experiments show that the volumetric mass transfer coefficient () in the case of SS reactor is, on average, 2.4 times higher than acrylic. A correlation is developed for estimating the based on total velocity and phase fraction, providing better fits than the models based on energy dissipation. All studies show that wall characteristics significantly impact the hydrodynamics and mass transfer phenomena since the pressure drop and the are greater in (the rougher) SS than in acrylic.","PeriodicalId":501204,"journal":{"name":"The Canadian Journal of Chemical Engineering","volume":"128 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141931511","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Non‐linear model predictive control (NMPC) is increasingly seen as a promising tool to tackle the problem of handling process nonlinearity and achieve optimal operation. One roadblock to NMPC implementation, however, is the lack of a good model, whether a first‐principles‐based or a non‐linear data‐driven‐based model such as artificial neural networks (ANN). This manuscript proposes a data‐driven modelling approach that integrates an autoencoder‐like network and dynamic mode decomposition (DMD) methods to result in a non‐linear modelling technique where the non‐linearity in the model stems from the modelling of the measured variables. The proposed method results in a semi‐linear state‐space model where the mapping between the model state and outputs are non‐linear (via the autoencoder‐like network) while the model dynamics are linear. In the subsequent model predictive controller (MPC) implementation, the autoencoder translates setpoints and outputs to the states of a state space model. The proposed approach is illustrated using a continuously stirred tank reactor simulation example. For comparison, a linear MPC and non‐linear MPC based on a traditional neural network (NN) model, a classic Koopman operator‐based MPC, and (to benchmark) a perfect model‐based NMPC are implemented and tested on several setpoint tracking tasks. The proposed MPC design outperforms the other data driven MPCs, and has similar performance as the first‐principles‐based NMPC while requiring less computational time and without requiring first principles knowledge.
{"title":"Integrating autoencoder with Koopman operator to design a linear data‐driven model predictive controller","authors":"Xiaonian Wang, Sheel Ayachi, Brandon Corbett, Prashant Mhaskar","doi":"10.1002/cjce.25445","DOIUrl":"https://doi.org/10.1002/cjce.25445","url":null,"abstract":"Non‐linear model predictive control (NMPC) is increasingly seen as a promising tool to tackle the problem of handling process nonlinearity and achieve optimal operation. One roadblock to NMPC implementation, however, is the lack of a good model, whether a first‐principles‐based or a non‐linear data‐driven‐based model such as artificial neural networks (ANN). This manuscript proposes a data‐driven modelling approach that integrates an autoencoder‐like network and dynamic mode decomposition (DMD) methods to result in a non‐linear modelling technique where the non‐linearity in the model stems from the modelling of the measured variables. The proposed method results in a semi‐linear state‐space model where the mapping between the model state and outputs are non‐linear (via the autoencoder‐like network) while the model dynamics are linear. In the subsequent model predictive controller (MPC) implementation, the autoencoder translates setpoints and outputs to the states of a state space model. The proposed approach is illustrated using a continuously stirred tank reactor simulation example. For comparison, a linear MPC and non‐linear MPC based on a traditional neural network (NN) model, a classic Koopman operator‐based MPC, and (to benchmark) a perfect model‐based NMPC are implemented and tested on several setpoint tracking tasks. The proposed MPC design outperforms the other data driven MPCs, and has similar performance as the first‐principles‐based NMPC while requiring less computational time and without requiring first principles knowledge.","PeriodicalId":501204,"journal":{"name":"The Canadian Journal of Chemical Engineering","volume":"128 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141931508","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Fisseha A. Bezza, Hendrik G. Brink, Evans M. N. Chirwa
In the face of the continuous development of novel adsorbents, developing robust adsorbents with high efficiency, strong phosphate selectivity, high regenerability, and cost effectiveness is a scientific challenge. In the present study, an activated carbon‐supported MgFe2O4‐layered double hydroxide (AC@ MgFe2O4‐LDH) derived Mg–Fe layered double oxide (AC@ MgFe2O4‐LDO) nanocomposite was synthesized at various temperatures and its potential application for phosphate adsorption was investigated. The nanocomposite exhibited a hierarchical mesoporous structure with a Brunauer–Emmett–Teller (BET) specific surface area of 193 m2/g and a narrow per‐size distribution of ~2 nm. AC@MgFe2O4‐LDO exhibited a high point of zero charge (pHpzc) value of 9.8 and robust phosphate adsorption potential over a wide pH range of 4–9 due to its high pH buffering capacity. The effects of adsorbent dose, layered double hydroxides (LDH) calcination temperature, initial phosphate concentration, contact time, and temperature on the phosphate adsorption capacity of the adsorbent were investigated. In the present study, up to 99.0% removal of phosphate was achieved at a 4 g/L adsorbent dosage in 4 h at pH 7 and 30°C. An adsorption kinetics study revealed that the adsorption of phosphate by AC@MgFe2O4‐LDO reached equilibrium within 240 min, with the kinetic experimental data fitting well with pseudo‐first‐order kinetics (r2 >0.99). The Langmuir adsorption isotherm model fit the experimental data well, with a maximum adsorption capacity of 25.81 mg/g. The adsorbent displayed strong phosphate selectivity in the presence of competing anions, and the study demonstrated that AC@MgFe2O4‐LDO has promising potential for efficient phosphate adsorption over a wide pH range.
{"title":"Selective and efficient removal of phosphate from aqueous solution using activated carbon‐supported Mg–Fe layered double oxide nanocomposites","authors":"Fisseha A. Bezza, Hendrik G. Brink, Evans M. N. Chirwa","doi":"10.1002/cjce.25440","DOIUrl":"https://doi.org/10.1002/cjce.25440","url":null,"abstract":"In the face of the continuous development of novel adsorbents, developing robust adsorbents with high efficiency, strong phosphate selectivity, high regenerability, and cost effectiveness is a scientific challenge. In the present study, an activated carbon‐supported MgFe<jats:sub>2</jats:sub>O<jats:sub>4</jats:sub>‐layered double hydroxide (AC@ MgFe<jats:sub>2</jats:sub>O<jats:sub>4</jats:sub>‐LDH) derived Mg–Fe layered double oxide (AC@ MgFe<jats:sub>2</jats:sub>O<jats:sub>4</jats:sub>‐LDO) nanocomposite was synthesized at various temperatures and its potential application for phosphate adsorption was investigated. The nanocomposite exhibited a hierarchical mesoporous structure with a Brunauer–Emmett–Teller (BET) specific surface area of 193 m<jats:sup>2</jats:sup>/g and a narrow per‐size distribution of ~2 nm. AC@MgFe<jats:sub>2</jats:sub>O<jats:sub>4</jats:sub>‐LDO exhibited a high point of zero charge (pH<jats:sub>pzc</jats:sub>) value of 9.8 and robust phosphate adsorption potential over a wide pH range of 4–9 due to its high pH buffering capacity. The effects of adsorbent dose, layered double hydroxides (LDH) calcination temperature, initial phosphate concentration, contact time, and temperature on the phosphate adsorption capacity of the adsorbent were investigated. In the present study, up to 99.0% removal of phosphate was achieved at a 4 g/L adsorbent dosage in 4 h at pH 7 and 30°C. An adsorption kinetics study revealed that the adsorption of phosphate by AC@MgFe<jats:sub>2</jats:sub>O<jats:sub>4</jats:sub>‐LDO reached equilibrium within 240 min, with the kinetic experimental data fitting well with pseudo‐first‐order kinetics (<jats:italic>r</jats:italic><jats:sup>2</jats:sup> >0.99). The Langmuir adsorption isotherm model fit the experimental data well, with a maximum adsorption capacity of 25.81 mg/g. The adsorbent displayed strong phosphate selectivity in the presence of competing anions, and the study demonstrated that AC@MgFe<jats:sub>2</jats:sub>O<jats:sub>4</jats:sub>‐LDO has promising potential for efficient phosphate adsorption over a wide pH range.","PeriodicalId":501204,"journal":{"name":"The Canadian Journal of Chemical Engineering","volume":"25 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141931510","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jianwen Wang, Fei Chu, Jianyu Zhao, Wenchao Bao, Fuli Wang
The effectiveness of control decisions provided by the safe operation control model for the dense medium coal preparation process may decline due to its inability to adapt to changing working conditions. To address this issue, this paper investigates a safe operation control model update strategy based on Bayesian network and incremental learning. This strategy can update the model structure and parameters according to different conditions, ensuring the effectiveness of the updated model. Considering that the old model has effective information to explain the new working conditions, the Bayesian network structure update learning method based on incremental learning is proposed. This method retains the components of the old model that can describe the joint probability distribution of the sampled data under the new working conditions while updating the remaining structure. This approach improves the efficiency of model updating. The simulation results show that the updated model obtained by the proposed method can effectively deal with new abnormal conditions.
{"title":"Updating strategy of safe operation control model for dense medium coal preparation process based on Bayesian network and incremental learning","authors":"Jianwen Wang, Fei Chu, Jianyu Zhao, Wenchao Bao, Fuli Wang","doi":"10.1002/cjce.25418","DOIUrl":"https://doi.org/10.1002/cjce.25418","url":null,"abstract":"The effectiveness of control decisions provided by the safe operation control model for the dense medium coal preparation process may decline due to its inability to adapt to changing working conditions. To address this issue, this paper investigates a safe operation control model update strategy based on Bayesian network and incremental learning. This strategy can update the model structure and parameters according to different conditions, ensuring the effectiveness of the updated model. Considering that the old model has effective information to explain the new working conditions, the Bayesian network structure update learning method based on incremental learning is proposed. This method retains the components of the old model that can describe the joint probability distribution of the sampled data under the new working conditions while updating the remaining structure. This approach improves the efficiency of model updating. The simulation results show that the updated model obtained by the proposed method can effectively deal with new abnormal conditions.","PeriodicalId":501204,"journal":{"name":"The Canadian Journal of Chemical Engineering","volume":"32 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141931458","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper, we aim to correlate various process and product quality attributes of a mammalian cell culture process with process parameters. To achieve this, we employed physics‐informed neural networks that solve the governing ordinary differential equations comprising independent variables (inputs‐ time, flow rates, and volume) and dependent variables (outputs‐ viable cell density, dead cell density, glucose concentration, lactate concentration, and monoclonal antibody concentration). The proposed model surpasses the prediction and accuracy capabilities of other commonly used modelling approaches, such as the multilayer perceptron model. It has higher R‐squared (R2), lower root mean square error, and lower mean absolute error than the multilayer perceptron model for all output variables (viable cell density, viability, glucose concentration, lactate concentration, and monoclonal antibody concentration). Furthermore, we incorporate a Bayesian optimization study to maximize viable cell density and monoclonal antibody concentration. Single objective optimization and weighted sum multiobjective optimization were carried out for viable cell density and monoclonal antibody concentration in separate (single objective optimization) and combined (multiobjective optimization) forms. An increment of 13.01% and 18.57% for viable cell density and monoclonal antibody concentration, respectively, were projected under single objective optimization, and 46.32% and 67.86%, respectively, for multiobjective optimization as compared to the base case. This study highlights the potential of the physics‐informed neural networks‐based modelling and optimization of upstream processing of mammalian cell‐based monoclonal antibodies in biopharmaceutical operations.
本文旨在将哺乳动物细胞培养过程的各种过程和产品质量属性与过程参数联系起来。为此，我们采用了物理信息神经网络来求解由自变量（输入--时间、流速和体积）和因变量（输出--存活细胞密度、死亡细胞密度、葡萄糖浓度、乳酸浓度和单克隆抗体浓度）组成的常微分方程。所提出的模型在预测能力和准确性方面超过了其他常用的建模方法，如多层感知器模型。就所有输出变量（存活细胞密度、存活率、葡萄糖浓度、乳酸浓度和单克隆抗体浓度）而言，它比多层感知器模型具有更高的 R 平方（R2）、更低的均方根误差和更低的平均绝对误差。此外，我们还进行了贝叶斯优化研究，以最大限度地提高存活细胞密度和单克隆抗体浓度。我们以单独（单目标优化）和组合（多目标优化）的形式对存活细胞密度和单克隆抗体浓度进行了单目标优化和加权和多目标优化。与基本情况相比，单目标优化预测的存活细胞密度和单克隆抗体浓度分别增加了 13.01% 和 18.57%，多目标优化预测的存活细胞密度和单克隆抗体浓度分别增加了 46.32% 和 67.86%。这项研究凸显了基于物理信息神经网络的建模和优化哺乳动物细胞单克隆抗体上游处理在生物制药操作中的潜力。
{"title":"Physics‐informed neural networks guided modelling and multiobjective optimization of a mAb production process","authors":"Md Nasre Alam, Anurag Anurag, Neelesh Gangwar, Manojkumar Ramteke, Hariprasad Kodamana, Anurag S. Rathore","doi":"10.1002/cjce.25446","DOIUrl":"https://doi.org/10.1002/cjce.25446","url":null,"abstract":"In this paper, we aim to correlate various process and product quality attributes of a mammalian cell culture process with process parameters. To achieve this, we employed physics‐informed neural networks that solve the governing ordinary differential equations comprising independent variables (inputs‐ time, flow rates, and volume) and dependent variables (outputs‐ viable cell density, dead cell density, glucose concentration, lactate concentration, and monoclonal antibody concentration). The proposed model surpasses the prediction and accuracy capabilities of other commonly used modelling approaches, such as the multilayer perceptron model. It has higher <jats:italic>R</jats:italic>‐squared (<jats:italic>R</jats:italic><jats:sup>2</jats:sup>), lower root mean square error, and lower mean absolute error than the multilayer perceptron model for all output variables (viable cell density, viability, glucose concentration, lactate concentration, and monoclonal antibody concentration). Furthermore, we incorporate a Bayesian optimization study to maximize viable cell density and monoclonal antibody concentration. Single objective optimization and weighted sum multiobjective optimization were carried out for viable cell density and monoclonal antibody concentration in separate (single objective optimization) and combined (multiobjective optimization) forms. An increment of 13.01% and 18.57% for viable cell density and monoclonal antibody concentration, respectively, were projected under single objective optimization, and 46.32% and 67.86%, respectively, for multiobjective optimization as compared to the base case. This study highlights the potential of the physics‐informed neural networks‐based modelling and optimization of upstream processing of mammalian cell‐based monoclonal antibodies in biopharmaceutical operations.","PeriodicalId":501204,"journal":{"name":"The Canadian Journal of Chemical Engineering","volume":"8 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141884319","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ning Yang, Zundong Xiao, Hanyang Liu, Junan Jiang, Fei Liu, Xiaoxia Yang, Rijie Wang
Micro/milli‐scale annular reactor with straight and helical forms has excellent heat and mass transfer performance due to the short molecular diffusion distance and dual‐wall surface transport. The annular gap spacing is scalable by adjusting the inner and outer tube diameter. The influence of diffusion and convection effects on axial dispersion as expanding the flow scale requires further elucidation with the help of residence time distribution (RTD) curves and Péclet (Pe) numbers. The correlation of RTD characteristics with annulus ratio γ = Dh/D (ratio of annulus characteristic size to outer diameter) is investigated using computational fluid dynamics. Results show that with enlarging the straight annular gap from micro‐scale to milli‐scale, RTD characteristics exhibit opposing patterns. This can be attributed to the transition from diffusion‐dominated to convection‐dominated on momentum transfer, and the transition interval is 0.167 < γ < 0.250. Correlation equations of Pe number with Reynolds (Re) number and γ are established under diffusion‐dominated and convection‐dominated states. The symmetrically distributed secondary flow in the helical annular gap effectively elevates the Pe (Pemax > 100). Correlation equations of Pe with Re and γ are established in helical annular gaps with 0.083 < γ < 0.208 and 0.167 < γ < 0.500. The above results contribute to understanding the annular flow RTD characteristics for better applications of tube‐in‐tube reactors.
采用直管和螺旋管形式的微米/毫微米级环形反应器由于分子扩散距离短和双壁表面传输，具有出色的传热和传质性能。环形间隙间距可通过调整内外管直径进行扩展。随着流动尺度的扩大，扩散和对流效应对轴向分散的影响需要借助停留时间分布（RTD）曲线和佩克莱特（Pe）数来进一步阐明。利用计算流体动力学研究了 RTD 特性与环形比 γ = Dh/D（环形特性尺寸与外径之比）的相关性。结果表明，随着直环间隙从微米级扩大到毫米级，热电阻特性呈现出相反的模式。这可归因于动量传递从扩散主导型过渡到对流主导型，过渡区间为 0.167 < γ < 0.250。在扩散主导和对流主导状态下，建立了 Pe 值与雷诺（Re）值和 γ 的相关方程。螺旋环形间隙中对称分布的二次流有效地提高了 Pe 值（Pemax > 100）。在 0.083 < γ < 0.208 和 0.167 < γ < 0.500 的螺旋环形间隙中，建立了 Pe 与 Re 和 γ 的相关方程。上述结果有助于了解环流热电阻特性，从而更好地应用管中管反应器。
{"title":"Effect of annulus ratio on the residence time distribution and Péclet number in micro/milli‐scale reactors","authors":"Ning Yang, Zundong Xiao, Hanyang Liu, Junan Jiang, Fei Liu, Xiaoxia Yang, Rijie Wang","doi":"10.1002/cjce.25428","DOIUrl":"https://doi.org/10.1002/cjce.25428","url":null,"abstract":"Micro/milli‐scale annular reactor with straight and helical forms has excellent heat and mass transfer performance due to the short molecular diffusion distance and dual‐wall surface transport. The annular gap spacing is scalable by adjusting the inner and outer tube diameter. The influence of diffusion and convection effects on axial dispersion as expanding the flow scale requires further elucidation with the help of residence time distribution (RTD) curves and Péclet (Pe) numbers. The correlation of RTD characteristics with annulus ratio <jats:italic>γ = D</jats:italic><jats:sub>h</jats:sub>/<jats:italic>D</jats:italic> (ratio of annulus characteristic size to outer diameter) is investigated using computational fluid dynamics. Results show that with enlarging the straight annular gap from micro‐scale to milli‐scale, RTD characteristics exhibit opposing patterns. This can be attributed to the transition from diffusion‐dominated to convection‐dominated on momentum transfer, and the transition interval is 0.167 < <jats:italic>γ</jats:italic> < 0.250. Correlation equations of Pe number with Reynolds (Re) number and <jats:italic>γ</jats:italic> are established under diffusion‐dominated and convection‐dominated states. The symmetrically distributed secondary flow in the helical annular gap effectively elevates the Pe (Pe<jats:sub>max</jats:sub> > 100). Correlation equations of Pe with Re and <jats:italic>γ</jats:italic> are established in helical annular gaps with 0.083 < <jats:italic>γ</jats:italic> < 0.208 and 0.167 < <jats:italic>γ</jats:italic> < 0.500. The above results contribute to understanding the annular flow RTD characteristics for better applications of tube‐in‐tube reactors.","PeriodicalId":501204,"journal":{"name":"The Canadian Journal of Chemical Engineering","volume":"100 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141887311","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
There is currently a huge imbalance between the demand and supply of freshwater resources. The shortage of fresh water can be mitigated by seawater desalination. Reverse osmosis (RO) is currently the most popular desalination technology around the world. Despite its various advantages, fouling has been one of its major limitations of RO. Membrane fouling can be divided into four categories: colloidal fouling, inorganic fouling, organic fouling, and biofouling. Precipitation of inorganic salts of small solubility, among which CaCO3, CaSO4, BaSO4, and SiO2 are the most common ones, are the cause of inorganic fouling, which is commonly referred to as scaling. Pretreatment technologies for prevention or mitigation of scaling in the RO process can be classified as conventional pretreatment technologies, which include water softening and scale inhibitors, and membrane‐based pretreatment technologies which include nanofiltration, forward osmosis, and membrane surface modification.
{"title":"Scaling in reverse osmosis seawater desalination: Mechanism and prevention—A literature review","authors":"Jiaxuan Shen, Xiaodong Wang, Xiaoyi Zhu, Bojin Tang, Cong Liu, Wan Li, Xueqiang Gao","doi":"10.1002/cjce.25427","DOIUrl":"https://doi.org/10.1002/cjce.25427","url":null,"abstract":"There is currently a huge imbalance between the demand and supply of freshwater resources. The shortage of fresh water can be mitigated by seawater desalination. Reverse osmosis (RO) is currently the most popular desalination technology around the world. Despite its various advantages, fouling has been one of its major limitations of RO. Membrane fouling can be divided into four categories: colloidal fouling, inorganic fouling, organic fouling, and biofouling. Precipitation of inorganic salts of small solubility, among which CaCO<jats:sub>3</jats:sub>, CaSO<jats:sub>4</jats:sub>, BaSO<jats:sub>4</jats:sub>, and SiO<jats:sub>2</jats:sub> are the most common ones, are the cause of inorganic fouling, which is commonly referred to as scaling. Pretreatment technologies for prevention or mitigation of scaling in the RO process can be classified as conventional pretreatment technologies, which include water softening and scale inhibitors, and membrane‐based pretreatment technologies which include nanofiltration, forward osmosis, and membrane surface modification.","PeriodicalId":501204,"journal":{"name":"The Canadian Journal of Chemical Engineering","volume":"10 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141884318","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：