Pub Date : 2022-08-09DOI: 10.1080/10618562.2022.2146677
Hai V. Nguyen, T. Bui-Thanh
Real-time accurate solutions of large-scale complex dynamical systems are in critical need for control, optimisation, uncertainty quantification, and decision-making in practical engineering and science applications, especially digital twin applications. This paper contributes in this direction a model-constrained tangent slope learning (mcTangent) approach. At the heart of mcTangent is the synergy of several desirable strategies: (i) a tangent slope learning to take advantage of the neural network speed and the time-accurate nature of the method of lines; (ii) a model-constrained approach to encode the neural network tangent slope with the underlying governing equations; (iii) sequential learning strategies to promote long-time stability and accuracy; and (iv) data randomisation approach to implicitly enforce the smoothness of the neural network tangent slope and its likeliness to the truth tangent slope up second order derivatives in order to further enhance the stability and accuracy of mcTangent solutions. Rigorous results are provided to analyse and justify the proposed approach. Several numerical results for transport equation, viscous Burgers equation, and Navier–Stokes equation are presented to study and demonstrate the robustness and long-time accuracy of the proposed mcTangent learning approach.
{"title":"A Model-Constrained Tangent Slope Learning Approach for Dynamical Systems","authors":"Hai V. Nguyen, T. Bui-Thanh","doi":"10.1080/10618562.2022.2146677","DOIUrl":"https://doi.org/10.1080/10618562.2022.2146677","url":null,"abstract":"Real-time accurate solutions of large-scale complex dynamical systems are in critical need for control, optimisation, uncertainty quantification, and decision-making in practical engineering and science applications, especially digital twin applications. This paper contributes in this direction a model-constrained tangent slope learning (mcTangent) approach. At the heart of mcTangent is the synergy of several desirable strategies: (i) a tangent slope learning to take advantage of the neural network speed and the time-accurate nature of the method of lines; (ii) a model-constrained approach to encode the neural network tangent slope with the underlying governing equations; (iii) sequential learning strategies to promote long-time stability and accuracy; and (iv) data randomisation approach to implicitly enforce the smoothness of the neural network tangent slope and its likeliness to the truth tangent slope up second order derivatives in order to further enhance the stability and accuracy of mcTangent solutions. Rigorous results are provided to analyse and justify the proposed approach. Several numerical results for transport equation, viscous Burgers equation, and Navier–Stokes equation are presented to study and demonstrate the robustness and long-time accuracy of the proposed mcTangent learning approach.","PeriodicalId":56288,"journal":{"name":"International Journal of Computational Fluid Dynamics","volume":"4 1","pages":"655 - 685"},"PeriodicalIF":1.3,"publicationDate":"2022-08-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87149520","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2022-08-09DOI: 10.1080/10618562.2022.2154758
K. Fuchi, Eric M. Wolf, D. Makhija, Christopher R. Schrock, P. Beran
A machine learning method to predict steady external fluid flows using elliptic input features is introduced. Using data from as few as one high-fidelity simulation, the proposed method produces models generalisable under changes to boundary geometry by using solutions to elliptic boundary value problems over the flow domain as the model input, instead of Cartesian coordinates of the domain. Training data is generated through pointwise evaluation of flow features at points selected through a quad-tree adaptive sampling method to concentrate training points in areas with large field gradients. Models are trained within a training window around the body, while predictions are smoothly extended to freestream conditions using a Partition-of-Unity extension. Predictive capabilities of the machine learning model are demonstrated in steady-state flow of incompressible fluid around a cylinder and a Joukowski airfoil. The predicted flow field is used to warm-start CFD simulations to achieve acceleration in solver convergence.
{"title":"Multi-Fidelity Machine Learning Applied to Steady Fluid Flows","authors":"K. Fuchi, Eric M. Wolf, D. Makhija, Christopher R. Schrock, P. Beran","doi":"10.1080/10618562.2022.2154758","DOIUrl":"https://doi.org/10.1080/10618562.2022.2154758","url":null,"abstract":"A machine learning method to predict steady external fluid flows using elliptic input features is introduced. Using data from as few as one high-fidelity simulation, the proposed method produces models generalisable under changes to boundary geometry by using solutions to elliptic boundary value problems over the flow domain as the model input, instead of Cartesian coordinates of the domain. Training data is generated through pointwise evaluation of flow features at points selected through a quad-tree adaptive sampling method to concentrate training points in areas with large field gradients. Models are trained within a training window around the body, while predictions are smoothly extended to freestream conditions using a Partition-of-Unity extension. Predictive capabilities of the machine learning model are demonstrated in steady-state flow of incompressible fluid around a cylinder and a Joukowski airfoil. The predicted flow field is used to warm-start CFD simulations to achieve acceleration in solver convergence.","PeriodicalId":56288,"journal":{"name":"International Journal of Computational Fluid Dynamics","volume":"1 1","pages":"618 - 640"},"PeriodicalIF":1.3,"publicationDate":"2022-08-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81636575","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2022-08-09DOI: 10.1080/10618562.2022.2154918
Saeed Akbari, Suraj Pawar, O. San
Recent developments in diagnostic and computing technologies offer to leverage numerous forms of nonintrusive modelling approaches from data where machine learning can be used to build computationally cheap and accurate surrogate models. To this end, we present a nonlinear proper orthogonal decomposition (POD) framework, denoted as NLPOD, to forge a nonintrusive reduced-order model for the Boussinesq equations. In our NLPOD approach, we first employ the POD procedure to obtain a set of global modes to build a linear-fit latent space and utilise an autoencoder network to compress the projection of this latent space through a nonlinear unsupervised mapping of POD coefficients. Then, long short-term memory (LSTM) neural network architecture is utilised to discover temporal patterns in this low-rank manifold. While performing a detailed sensitivity analysis for hyperparameters of the LSTM model, the trade-off between accuracy and efficiency is systematically analysed for solving a canonical Rayleigh–Bénard convection system.
{"title":"Numerical Assessment of a Nonintrusive Surrogate Model Based on Recurrent Neural Networks and Proper Orthogonal Decomposition: Rayleigh–Bénard Convection","authors":"Saeed Akbari, Suraj Pawar, O. San","doi":"10.1080/10618562.2022.2154918","DOIUrl":"https://doi.org/10.1080/10618562.2022.2154918","url":null,"abstract":"Recent developments in diagnostic and computing technologies offer to leverage numerous forms of nonintrusive modelling approaches from data where machine learning can be used to build computationally cheap and accurate surrogate models. To this end, we present a nonlinear proper orthogonal decomposition (POD) framework, denoted as NLPOD, to forge a nonintrusive reduced-order model for the Boussinesq equations. In our NLPOD approach, we first employ the POD procedure to obtain a set of global modes to build a linear-fit latent space and utilise an autoencoder network to compress the projection of this latent space through a nonlinear unsupervised mapping of POD coefficients. Then, long short-term memory (LSTM) neural network architecture is utilised to discover temporal patterns in this low-rank manifold. While performing a detailed sensitivity analysis for hyperparameters of the LSTM model, the trade-off between accuracy and efficiency is systematically analysed for solving a canonical Rayleigh–Bénard convection system.","PeriodicalId":56288,"journal":{"name":"International Journal of Computational Fluid Dynamics","volume":"42 1","pages":"599 - 617"},"PeriodicalIF":1.3,"publicationDate":"2022-08-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74112692","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2022-08-09DOI: 10.1080/10618562.2022.2152014
P. H. Dabaghian, Shady E. Ahmed, O. San
In this paper, we formulate three nonintrusive methods and systematically explore their performance in terms of the ability to reconstruct the quantities of interest and their predictive capabilities. The methods include deterministic dynamic mode decomposition, randomised dynamic mode decomposition and nonlinear proper orthogonal decomposition (NLPOD). We apply these methods to a convection dominated fluid flow problem governed by the Boussinesq equations. We analyse the reconstruction results primarily at two different times for considering different noise levels synthetically added into the data snapshots. Overall, our results indicate that, with a proper selection of the number of retained modes and neural network architectures, all three approaches make predictions that are in a good agreement with the full order model solution. However, we find that the NLPOD approach seems more robust for higher noise levels compared to both dynamic mode decomposition approaches.
{"title":"Nonintrusive Reduced Order Modelling of Convective Boussinesq Flows","authors":"P. H. Dabaghian, Shady E. Ahmed, O. San","doi":"10.1080/10618562.2022.2152014","DOIUrl":"https://doi.org/10.1080/10618562.2022.2152014","url":null,"abstract":"In this paper, we formulate three nonintrusive methods and systematically explore their performance in terms of the ability to reconstruct the quantities of interest and their predictive capabilities. The methods include deterministic dynamic mode decomposition, randomised dynamic mode decomposition and nonlinear proper orthogonal decomposition (NLPOD). We apply these methods to a convection dominated fluid flow problem governed by the Boussinesq equations. We analyse the reconstruction results primarily at two different times for considering different noise levels synthetically added into the data snapshots. Overall, our results indicate that, with a proper selection of the number of retained modes and neural network architectures, all three approaches make predictions that are in a good agreement with the full order model solution. However, we find that the NLPOD approach seems more robust for higher noise levels compared to both dynamic mode decomposition approaches.","PeriodicalId":56288,"journal":{"name":"International Journal of Computational Fluid Dynamics","volume":"35 1","pages":"578 - 598"},"PeriodicalIF":1.3,"publicationDate":"2022-08-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80461486","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2022-08-09DOI: 10.1080/10618562.2023.2175788
P. Orkwis, Mahdi Pourbagian
Fluid dynamics research has continuously benefitted from advances in applied mathematics, computer science, and computer engineering; the development of CFD being a prime example of this multidisciplinary contribution. Improved numerical analysis techniques led to stable and efficient algorithms, while vectorization and parallelization fueled the expansion of CFD from research codes to validated design software. As computing power has grown, so too has the generation of data and the complexity of flows simulated. Gaining insight into fundamental physics is now growing, but with that comes enormous amounts of data that represent opportunities for insight and exploitation. Machine learning (ML) is the next iteration in the aforementioned multidisciplinary collaboration. This field encompasses many techniques and approaches, including the classification of flow fields, prediction of nonlinear processes like aerodynamic stall, modeling of fine-scale structures, and unsupervised exploration of complex, voluminous data for hidden physics. This field should be rightly thought of as a new toolbox that will allow us to further understand and utilize fluid dynamics through increasingly complex simulations. In short, it is ready for application to distinctly nonlinear, unsteady, multiscale problems in fluid dynamics. In this special issue, readers will find applications of machine learning to modeling complex phenomena, classifying flows, and increasing the fidelity of existing methods. The readers will also find that much of the new jargon used in the ML community is remarkably similar to ideas used for decades in the CFD community. The guest editors have chosen papers that are representative of leading-edge CFD applications. In each case, ML is a means to obtain more than could normally be accomplished with traditional methods. We hope that this special issue will encourage readers to begin new studies in this growing field. Finally, we would like to thank Prof. Habashi, Editor-in-Chief of the IJCFD for providing us with supportive guidelines and comments throughout the process of this special issue. We would also like to express our gratitude to the editorial staff at Taylor & Francis and the reviewers for their valuable effort. Respectfully,
{"title":"Machine Learning in CFD","authors":"P. Orkwis, Mahdi Pourbagian","doi":"10.1080/10618562.2023.2175788","DOIUrl":"https://doi.org/10.1080/10618562.2023.2175788","url":null,"abstract":"Fluid dynamics research has continuously benefitted from advances in applied mathematics, computer science, and computer engineering; the development of CFD being a prime example of this multidisciplinary contribution. Improved numerical analysis techniques led to stable and efficient algorithms, while vectorization and parallelization fueled the expansion of CFD from research codes to validated design software. As computing power has grown, so too has the generation of data and the complexity of flows simulated. Gaining insight into fundamental physics is now growing, but with that comes enormous amounts of data that represent opportunities for insight and exploitation. Machine learning (ML) is the next iteration in the aforementioned multidisciplinary collaboration. This field encompasses many techniques and approaches, including the classification of flow fields, prediction of nonlinear processes like aerodynamic stall, modeling of fine-scale structures, and unsupervised exploration of complex, voluminous data for hidden physics. This field should be rightly thought of as a new toolbox that will allow us to further understand and utilize fluid dynamics through increasingly complex simulations. In short, it is ready for application to distinctly nonlinear, unsteady, multiscale problems in fluid dynamics. In this special issue, readers will find applications of machine learning to modeling complex phenomena, classifying flows, and increasing the fidelity of existing methods. The readers will also find that much of the new jargon used in the ML community is remarkably similar to ideas used for decades in the CFD community. The guest editors have chosen papers that are representative of leading-edge CFD applications. In each case, ML is a means to obtain more than could normally be accomplished with traditional methods. We hope that this special issue will encourage readers to begin new studies in this growing field. Finally, we would like to thank Prof. Habashi, Editor-in-Chief of the IJCFD for providing us with supportive guidelines and comments throughout the process of this special issue. We would also like to express our gratitude to the editorial staff at Taylor & Francis and the reviewers for their valuable effort. Respectfully,","PeriodicalId":56288,"journal":{"name":"International Journal of Computational Fluid Dynamics","volume":"23 1","pages":"519 - 519"},"PeriodicalIF":1.3,"publicationDate":"2022-08-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80225181","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2022-07-07DOI: 10.1080/10618562.2023.2171021
Ettore Saetta, R. Tognaccini, G. Iaccarino
A convolutional autoencoder is trained using a database of airfoil aerodynamic simulations and assessed in terms of overall accuracy and interpretability. The goal is to predict the stall and to investigate the ability of the autoencoder to distinguish between the linear and non-linear response of the airfoil pressure distribution to changes in the angle of attack. After a sensitivity analysis of the learning infrastructure, we investigate the latent space identified by the autoencoder targeting extreme compression rates, i.e. very low-dimensional reconstructions. We also propose a strategy to use the decoder to generate new synthetic airfoil geometries and aerodynamic solutions by interpolation and extrapolation in the latent representation learned by the autoencoder.
{"title":"Machine Learning to Predict Aerodynamic Stall","authors":"Ettore Saetta, R. Tognaccini, G. Iaccarino","doi":"10.1080/10618562.2023.2171021","DOIUrl":"https://doi.org/10.1080/10618562.2023.2171021","url":null,"abstract":"A convolutional autoencoder is trained using a database of airfoil aerodynamic simulations and assessed in terms of overall accuracy and interpretability. The goal is to predict the stall and to investigate the ability of the autoencoder to distinguish between the linear and non-linear response of the airfoil pressure distribution to changes in the angle of attack. After a sensitivity analysis of the learning infrastructure, we investigate the latent space identified by the autoencoder targeting extreme compression rates, i.e. very low-dimensional reconstructions. We also propose a strategy to use the decoder to generate new synthetic airfoil geometries and aerodynamic solutions by interpolation and extrapolation in the latent representation learned by the autoencoder.","PeriodicalId":56288,"journal":{"name":"International Journal of Computational Fluid Dynamics","volume":"9 1","pages":"641 - 654"},"PeriodicalIF":1.3,"publicationDate":"2022-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83770255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2022-07-03DOI: 10.1080/10618562.2022.2153835
Behzad Saberali, Kai Zhang, N. Golsanami
In the water flooding process, determining the location of the injected water front is as one of the most critical variables, which is the basis of many subsequent predictions. Despite the importance and use of this parameter in a vast range of flooding-related assessments, there are no alternative methods to traditional analytical modeling or time-consuming numerical 3D simulation for its determination. This study introduces a data-driven proxy modeling approach based on two powerful deep learning algorithms for real-time determination of the injected water front location on the grid scale. The developed proxy models have realized the possibility of modeling the location of the flow front by minimally using the data extracted from the numerical simulators and only relying on commonly available field data. The proposed proxy models successfully simulated the breakthrough time in production wells and water arrival time in certain reservoir grids in new blind scenarios.
{"title":"Data-Driven Proxy Modeling of Water Front Propagation in Porous Media","authors":"Behzad Saberali, Kai Zhang, N. Golsanami","doi":"10.1080/10618562.2022.2153835","DOIUrl":"https://doi.org/10.1080/10618562.2022.2153835","url":null,"abstract":"In the water flooding process, determining the location of the injected water front is as one of the most critical variables, which is the basis of many subsequent predictions. Despite the importance and use of this parameter in a vast range of flooding-related assessments, there are no alternative methods to traditional analytical modeling or time-consuming numerical 3D simulation for its determination. This study introduces a data-driven proxy modeling approach based on two powerful deep learning algorithms for real-time determination of the injected water front location on the grid scale. The developed proxy models have realized the possibility of modeling the location of the flow front by minimally using the data extracted from the numerical simulators and only relying on commonly available field data. The proposed proxy models successfully simulated the breakthrough time in production wells and water arrival time in certain reservoir grids in new blind scenarios.","PeriodicalId":56288,"journal":{"name":"International Journal of Computational Fluid Dynamics","volume":"262 1","pages":"465 - 487"},"PeriodicalIF":1.3,"publicationDate":"2022-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79674640","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2022-07-03DOI: 10.1080/10618562.2022.2152013
A. Fracassi, R. De Donno, A. Ghidoni, G. Noventa
This work presents a Delayed version of the hybrid RANS-LES eXtra Large Eddy Simulation (SST DX-LES) model for the simulation of turbulent flows. In particular, in the proposed model the Shear Stress Transport (SST) k-ω turbulence model replaces the TNT k-ω model, and a shielding function is introduced to avoid the Modelled Stress Depletion (MSD) and the related Grid-Induced Separation (GIS), typical of DES-like hybrid models. Moreover, a different definition of the filter is used, which blends the formula based on the maximum element spacing and element volume. The proposed hybrid model is implemented in the open-source CFD software OpenFOAM, calibrated, validated and assessed on several benchmark cases. The results are compared with both experimental data and reference numerical results. Simulations are performed also with the original X-LES model to spotlight the accuracy improvement.
{"title":"Implementation and Validation of the SST Delayed eXtra-LES Model for Complex Turbulent Flows","authors":"A. Fracassi, R. De Donno, A. Ghidoni, G. Noventa","doi":"10.1080/10618562.2022.2152013","DOIUrl":"https://doi.org/10.1080/10618562.2022.2152013","url":null,"abstract":"This work presents a Delayed version of the hybrid RANS-LES eXtra Large Eddy Simulation (SST DX-LES) model for the simulation of turbulent flows. In particular, in the proposed model the Shear Stress Transport (SST) k-ω turbulence model replaces the TNT k-ω model, and a shielding function is introduced to avoid the Modelled Stress Depletion (MSD) and the related Grid-Induced Separation (GIS), typical of DES-like hybrid models. Moreover, a different definition of the filter is used, which blends the formula based on the maximum element spacing and element volume. The proposed hybrid model is implemented in the open-source CFD software OpenFOAM, calibrated, validated and assessed on several benchmark cases. The results are compared with both experimental data and reference numerical results. Simulations are performed also with the original X-LES model to spotlight the accuracy improvement.","PeriodicalId":56288,"journal":{"name":"International Journal of Computational Fluid Dynamics","volume":"189 1","pages":"441 - 464"},"PeriodicalIF":1.3,"publicationDate":"2022-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86923309","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2022-07-03DOI: 10.1080/10618562.2022.2153834
S. Channouf, M. Jami
The present work focuses on the development of the 3D MRT-LBM computational method to simulate a droplet formed by the condensation of a gas at a saturation temperature. The droplet falls by the effect of its volume (body force) and collides with a solid surface of parallelepiped shape for different values of 0.25 0.95. This study presents the evolution of the drop from its formation on the upper cold surface until it falls on the lower wall. For this purpose, we present the behaviour of the dropwise on a horizontal surface as well as the parameters which characterise its evolution over time: its radius R, its height H, and its maximum distance during its contact with the obstacle. Moreover, the distribution of its local heat flux in 2D during its first contact with the upper face of the solid obstacle is presented for each value of κ to describe their heat exchange.
{"title":"Study of Falling Condensate Droplets on Parallelepiped Solid Surface Using Hybrid 3D MRT-LBM","authors":"S. Channouf, M. Jami","doi":"10.1080/10618562.2022.2153834","DOIUrl":"https://doi.org/10.1080/10618562.2022.2153834","url":null,"abstract":"The present work focuses on the development of the 3D MRT-LBM computational method to simulate a droplet formed by the condensation of a gas at a saturation temperature. The droplet falls by the effect of its volume (body force) and collides with a solid surface of parallelepiped shape for different values of 0.25 0.95. This study presents the evolution of the drop from its formation on the upper cold surface until it falls on the lower wall. For this purpose, we present the behaviour of the dropwise on a horizontal surface as well as the parameters which characterise its evolution over time: its radius R, its height H, and its maximum distance during its contact with the obstacle. Moreover, the distribution of its local heat flux in 2D during its first contact with the upper face of the solid obstacle is presented for each value of κ to describe their heat exchange.","PeriodicalId":56288,"journal":{"name":"International Journal of Computational Fluid Dynamics","volume":"60 1","pages":"488 - 505"},"PeriodicalIF":1.3,"publicationDate":"2022-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84699230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2022-07-03DOI: 10.1080/10618562.2022.2159026
Zhiyong Li, Lingying Zhao, M. Ye
The simulation of engineering research is difficult, especially the engineering problem of the large differences between the size of the equipment and materials processed. At present, two methods are used to solve this problem, i.e. the equal scale reduction model and the study of only a part of it, which makes it inconsistent with the actual situation. To find a better way to improve this problem, the multi-scale is introduced. In this study, the heat transfer of the particles in a drying drum with engineering size is studied by multi-scale and fluid-solid coupling methods. The general situation of the drying drum is introduced, and the fluid-solid coupling mechanism based on multi-scale is established. A method of establishing a particle micro model is proposed. The feasibility of this method is proved by simulation and experiment, and the accuracy of the proposed model is improved by 15.62% compared with the traditional model.
{"title":"Heat Transfer of Aggregate in a Drying Drum Based on the Multi-Scale Model and Fluid-Solid Coupling","authors":"Zhiyong Li, Lingying Zhao, M. Ye","doi":"10.1080/10618562.2022.2159026","DOIUrl":"https://doi.org/10.1080/10618562.2022.2159026","url":null,"abstract":"The simulation of engineering research is difficult, especially the engineering problem of the large differences between the size of the equipment and materials processed. At present, two methods are used to solve this problem, i.e. the equal scale reduction model and the study of only a part of it, which makes it inconsistent with the actual situation. To find a better way to improve this problem, the multi-scale is introduced. In this study, the heat transfer of the particles in a drying drum with engineering size is studied by multi-scale and fluid-solid coupling methods. The general situation of the drying drum is introduced, and the fluid-solid coupling mechanism based on multi-scale is established. A method of establishing a particle micro model is proposed. The feasibility of this method is proved by simulation and experiment, and the accuracy of the proposed model is improved by 15.62% compared with the traditional model.","PeriodicalId":56288,"journal":{"name":"International Journal of Computational Fluid Dynamics","volume":"119 1","pages":"506 - 517"},"PeriodicalIF":1.3,"publicationDate":"2022-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81677077","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"工程技术","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：