Hybrid Computational Fluid Dynamics Research Articles

Enhancing CFD solver with Machine Learning techniques

This study addresses the computational challenges in fluid flow simulations arising from demanding computational grids, required to capture the temporal and length scales involved. Our approach focuses on the pressure solver, as this is a resource-intensive component in Computational Fluid Dynamics (CFD) solvers. We achieve this by integrating a Machine Learning (ML) surrogate model with an incompressible fluid flow solver. We created two variants of an ML-enhanced CFD solver which were able to reduce the number of iterations required by the CFD pressure solver during unsteady flow simulations. Consequently, the simulations yielded comparable drag coefficients and Strouhal numbers, accompanied by an eightfold decrease in execution time. The performance enhancements are attributed to reduced computational effort per temporal iteration and early-stage forcing on the simulation dynamical behavior when using the ML-based surrogate models. This research introduces an approach to enhance the computational efficiency of fluid flow analyses by incorporating surrogate models to aid the pressure solver in CFD simulations. We propose a Hybrid CFD solver, ie. a physics-informed solver enhanced with data-driven surrogate models.

Experimental investigation and numerical simulation of intense plume impingement effects from a 10 N bipropellant thruster

The intense interaction between thruster plumes and spacecraft surfaces presents substantial challenges for the advancement of space technologies. The impingement effects from a 10 N bipropellant thruster plume on a vertical instrumented plate were experimentally and numerically investigated in this study. Impingement pressure and heat flux were measured using micro differential pressure transducers and Gardon gauges, respectively, while the spatial distribution of the plume was captured using the high-speed camera. The phenomena and the distribution characteristics of plume impingement have been summarized, thereby providing a comprehensive validation dataset for subsequent predictions. Numerical simulations are conducted utilizing the coupled hybrid Computational Fluid Dynamics and Direct Simulation Monte Carlo (CFD-DSMC) algorithm. The dependence of the simulation outcomes on grid resolution was critically assessed, and the accuracy of the results was validated by comparison with experimental data. An integrated methodology, combining experimental and numerical approaches, was developed to calculate the maximum heat flux under realistic grid configurations. This method has proven to be effective and supports the accurate and efficient evaluation of intense plume impingement effects for engineering applications.

Graph and convolutional neural network coupling with a high-performance large-eddy simulation solver

Computational Fluid Dynamics (CFD) traditionally relies on long-standing numerical simulation strategies for the Navier–Stokes equations. Recently, interest in data-driven hybrid CFD solvers has spiked, leveraging pre-computed datasets to enhance various weak links inside existing solvers, such as closure models, under-resolved physics, or even to guide numerical resolution strategies. Running these hybrid solvers, notably in High Performance Computing (HPC) environments, presents specific challenges. In particular, context-aware deep learning (e.g. Convolutional (CNN) and Graph (GNN) Neural Networks) is promising for this task, but requires passing data representations between the physics solver and the neural network. In relevant industrial configurations, CFD meshes can be Cartesian but highly irregular, or unstructured, both of which do not match the pixel/voxel structure needed to run CNNs. In addition, discrepancies in programming language and libraries are common between CFD and machine learning applications. This work explores the many challenges of running a parallel hybrid solver in an HPC context, through the coupling of the AVBP CFD solver with neural networks in turbulent combustion and wall friction modeling applications. The knowledge gained is showcased in this article, as well as assembled in an actionable open-source library.

Numerical Investigation of the Flow Structure and Acoustical Noise in a Suppressor

The noise created when a firearm is fired has a lot of adverse effects on humans and the environment, so analyzing and attenuating this noise is essential. This study aimed to examine propellant flow and the sound generated by this flow using a hybrid computational fluid dynamics and computational aeroacoustics method. The compatibility of this numerical study was also validated by comparing it with the experimental result. The impact of critical parameters, such as the shape and number of baffles of the suppressor, was studied. Finally, the overpressure and acoustics results of different suppressors were compared with each other and with the unsuppressed condition. According to the result, the suppressor with curved baffles shows a better performance. When the exit pressure at the tip of suppressor compared to the initial inlet pressure to the suppressor the percentage drop in overpressure was 76.01%, 78.79% and 81.3% for the suppressor with one, three and five curved baffles, respectively. For condition without suppressor, 169.49[Formula: see text]dB of SPL was recorded. When using a suppressor without a baffle, this value was reduced to 162.13[Formula: see text]dB. For the suppressor with one, three and five curved baffles, the SPL value was 160.234[Formula: see text]dB, 159.43[Formula: see text]dB and 158.11[Formula: see text]dB, respectively. Generally, this study shows that the suppressor’s efficiency is directly proportional to its shape and number of baffles.

Chemical reactor network modeling in the context of solid fuel combustion under oxy-fuel atmospheres

A multi-phase chemical reactor network (CRN) is developed and applied to investigate solid fuel combustion under oxy-fuel atmospheres in a laboratory-scale combustion chamber. The development of a novel solid-gas plug flow reactor (sPFR) allows the full coupling of solid and gas phase processes to increase the fidelity in CRN modeling. Together with the recently developed solid-gas perfectly stirred reactor (sPSR), the new sPFR is applied within a multi-phase reactor network to model the complex features of the investigated application. To design an initial reactor network, a hybrid experimental and computational fluid dynamics (CFD)-driven approach is being pursued, taking into account both global and local information on flow and temperature fields. Initially, an in-depth investigation of the physical processes is conducted, covering solid fuel conversion processes such as devolatilization and char conversion, along with the thermochemical characteristics of the flue gas. In particular, the prediction of carbon monoxide (CO) emissions is analyzed in detail by means of sensitivity analyses. Based on the findings of the sensitivity study, the network complexity of the initial reactor network is increased to enable an accurate prediction of the CO formation in comparison to the experiments. Finally, the formation of nitric oxides (NOx), sulfur oxides (SOx), and aromatic pollutants is discussed, highlighting the importance of detailed chemistry in solid fuel combustion at technically relevant scales.

Hole formation mechanisms in double-sided laser drilling of Ti6Al4V-C/SiC stacked materials

This is the first investigation on the potential of doubled-sided laser drilling for hole making in Ti6Al4V-C/SiC, a hard-to-machine stacked structure that holds critical significance in high-temperature structural applications. While previous literature reported laser drilling-induced microstructural change in Ti6Al4V and C/SiC single sheets respectively, a comprehensive understanding of the melt pool formation process and its influence on hole morphology/quality is still lacking. The complexity increases when Ti6Al4V is coupled with C/SiC in a stacked structure, highlighting a significant yet understudied knowledge gap in understanding the intricate material interactions at the interface. Here we present a multifaceted study combining novel computational fluid dynamics (CFD) modeling approach with ablation mechanism investigation through advanced microstructure analysis. A hybrid CFD modeling strategy integrating volume of fluid method and level set method has been developed to successfully simulate the evolution of hole shape, recast layer formation, and melt pool flow behaviour in Ti6Al4V-C/SiC. This model provides comprehensive insights into molten pool dynamics under various laser drilling conditions, fostering a fundamental understanding of the mechanism that drives recast layer formation and influencing surface quality. To validate our modeling outcomes, extensive experimental investigations were undertaken across a range of laser drilling conditions. Whenever possible, samples were analysed using advanced materials characterization techniques, including scanning electron microscopy, energy dispersive spectroscopy, and electron backscatter diffraction, to elucidate the microstructural evolution within the recast layer and the formation of the heat-affected zone. Whenever possible, the results were also compared with single side drilled Ti6Al4V-C/SiC stacks, Ti6Al4V or C/SiC single sheets (i.e., the benchmark references). Our experimental findings align well with the modeling results, systematically illustrating that doubled-sided laser drilling of the stacked structure outperforms the conventional single-sided laser drilling in terms of hole morphology, surface roughness, and reduced recast layer formation. Despite these promising results, the proposed technology is currently limited to laboratory scales and lacks capacity in manufacturing certain parts with complex geometry. Future outlook has therefore been proposed with potential solution to address these challenges and enable real world applications in future.

Leaping fishbot by combustion and propulsion hybrid actuation

AbstractAutonomous underwater vehicles have been extensively developed in the past decades and deployed for various ocean research activities. While the leaping‐out motions of underwater creatures have motivated the development of underwater robots with the motion ability to jump out of water, the dynamics of such jumping robots have not been fully described. In this study, we developed a leaping fishbot, a flying fish‐inspired underwater soft robot that can leap out of the water and further sail on the surface using the combustion and propulsion hybrid actuation. Specifically, the fishbot uses the transient driving method, in which a premixed oxygen–propane gas generates an instant pushing impulse to launch the robot out of the water at a transiently high speed while the attached jetting propellers provide a sustained thrust as it swims on the water surface at a stable velocity. A series of flume experiments and solid–fluid hybrid computational fluid dynamics numerical simulations were conducted to study the kinematic performance of the robot. Experimental results established a quantitative relationship between the kinematic performance, gas amount, and launch angle. The numerical results accurately predicted the water–air multiphase jumping motions, the numerical result of kinematic was significantly agreeing with the experimental results, the error in the horizontal direction is less than 15%, and that in the vertical direction is less than 10%. We further analyzed the flow field generated by the leaping fishbot to gain a better understanding of the velocity and vortex distribution around it. Our research findings provide valuable insights into bionic application of underwater robots in the future.

Computational simulation of mass transfer in membranes using hybrid machine learning models and computational fluid dynamics

In this work, a modeling approach has been developed for describing component separation by selective membrane systems. The developed modeling strategy in this research is based on computational fluid dynamics (CFD) integrated with machine learning (ML) approach to reduce the computational costs and evaluate the possibility of ML model for integration to CFD models. The considered case study is a membrane-based molecular separation for removal of species from water. CFD was performed to solve mass transfer equations, and the data in form of concentration profiles have been used for ML computations. The data set in this research has about two thousand data vectors, each of which contains two inputs r and z and one output C which is the species concentration in the feed channel of membrane. Three models, Multi-Layer Perceptron (MLP), Support Vector Machines with RBF kernel (RBF-SVM), and Decision Tree (DT) were chosen for modeling since they were anticipated to perform well on the provided data by adjusting their hyper-parameters. After optimizing the models, their results were carefully checked with different criteria. All three models showed acceptable results with R2 score criteria higher than 0.9. MAPE decreased to 3.57 × 10−3 and improved to 3.6 with the RMSE criterion. We can introduce the best model in this research as the MLP model for description of the mass transfer in the membrane.

A Simplified GPU Implementation of the Hybrid Lattice Boltzmann Model for Three-Dimensional High Rayleigh Number Flows

This paper provides an analysis of the numerical performance of a hybrid computational fluid dynamics (CFD) solver for 3D natural convection. We propose to use the lattice Boltzmann equations with the two-relaxation time approximation for the fluid flow, whereas thermodynamics is described by the macroscopic energy equation with the finite difference solution. An in-house parallel graphics processing unit (GPU) code is written in MATLAB. The execution time of every single step of the algorithm is studied. It is found that the explicit finite difference scheme is not as stable as the implicit one for high Rayleigh numbers. The most time-consuming steps are energy and collide, while stream, boundary conditions, and macroscopic parameters recovery are executed in no time, despite the grid size under consideration. GPU code is more than 30 times faster than a typical low-end central processing unit-based code. The proposed hybrid model can be used for real-time simulation of physical systems under laminar flow behavior and on mid-range segment GPUs.

A hybrid CFD – Deep Learning methodology to improve the accuracy of cut-off diameter prediction in coarse-grid simulations for cyclone separators

In many industries, cyclone separators are frequently employed to remove solid particles from the fluid flow. Cut-off diameter is recognized as a significant parameter to evaluate the performance of cyclone separators in addition to pressure drop. Computational Fluid Dynamics (CFD), a powerful computer-based method, can precisely estimate the cut-off diameter of cyclone separators. There is no arguing, however, that the CFD technique is computationally expensive and practically difficult. This research has suggested a more precise, computationally proficient hybrid CFD–DL method to improve the accuracy of cut-off diameter prediction in coarse-grid simulations for cyclone separators. It has been demonstrated that the proposed method not only requires less computational cost than typical CFD, but also delivers more accuracy results (with mean error less than 5.1% compared to experimental data). In other words, it takes advantage of the promise of a novel approach to decrease computational time while enhancing accuracy for CFD simulations.

Assessment of The Acoustic Scattering Matrix of a Heat Exchanger Using ssCFD-LNSE Simulation

The risk of thermoacoustic instability is present in any combustion appliance. The instability results from a closed loop feedback between unsteady combustion, heat-transfer and acoustic modes of the system. To predict the system acoustics all constituting elements of the appliance need to be modelled. A heat-exchanger is the element where the gas faces complex fluid dynamics and heat transfer processes. Therefore, modelling of (thermo)-acoustic properties of a heat exchanger is challenging. In this paper, a computational approach is proposed to characterize the acoustic properties of a generic heat exchanger in both laminar and turbulent flow regimes. A hybrid Computational Fluid Dynamics - Computational Aero-Acoustics (CFD-CAA) method is used based on full linearized Navier-Stokes equation, called ssCFD-LNSE. The fundamental idea in this approach is to efficiently model acoustic wave propagation with inclusion of mean flow and temperature fields. ssCFD-LNSE is performed by splitting the quantities of the total field into a mean part (obtained from CFD) and a (acoustic) perturbation part modelled within the LNSE solver. The goal of this research is to assess the two-port acoustic scattering matrix of an array of tubes, as a generic model of heat-exchanger, with a hot cross flow.

Numerical investigation of wall effects on combustion noise from a lean-premixed hydrogen/air low-swirl flame

A hybrid computational fluid dynamics (CFD)/computational aero-acoustics (CAA) approach, in which large-eddy simulation (LES) and APE-RF (solution of the acoustic perturbation equations for reacting flows) are employed for the CFD and CAA, respectively, calling it the hybrid LES/APE-RF approach, is used to analyze the influence of a wall on the combustion noise from a lean-premixed gaseous hydrogen/air low-swirl turbulent jet flame. The wall boundary conditions pertaining to the APE-RF system are formulated to account for acoustic reflection from the wall. The results show that the sound pressure level (SPL) spectrum obtained from the LES/APE-RF is in good agreement with that measured in the experiment. In the LES/APE-RF, the SPL spectrum of combustion noise with the wall plate explicitly changes compared to that without the wall plate. Specifically, the presence of the wall plate tends to ease the peaks that appeared in the case without the wall plate and create a nearly constant SPL within a specific frequency band. The analysis of the heat release rate fluctuation reveals that these phenomena are caused by the absence of a single periodic oscillation of heat release rate. The presence of the wall plate creates an asymmetric flow around the flame and distorts the flame structure, thereby altering the flame fluctuation phenomena.

A hybrid CFD – Deep learning methodology for improving the accuracy of pressure drop prediction in cyclone separators

Cyclone separators are widely used in many industries to separate solid particles from the fluid flow. As same as separation efficiency, pressure drop is considered as a primary parameter to evaluate performance of cyclone separators. Computational Fluid Dynamics (CFD) is a powerful method to predict the pressure drop of cyclone separators. However, it is indisputable that CFD technique is mathematically complicated and computationally expensive. This study has proposed a more accurate, computationally efficient hybrid method to predict cyclone pressure drop by merging CFD and Deep Learning algorithms. Its generalization has been validated by k-fold cross validation. The proposed hybrid method has also been proved to be more superior than the DNN regression method. The results verify that the proposed method not only predicts cyclone pressure drop accurately (with a maximum error less than 5% in comparison with experimental data), but also requires less computational time than traditional CFD. In other words, it leverages the potential of novel approaches to decrease CFD computational cost while increasing the accuracy level.

Dual-Solver Hybrid Computational Approach to Integrated Propulsion Aerodynamics

The demand for high-efficiency air vehicles and the development of distributed electric propulsion systems for aircraft are exposing the need for accurate aerodynamic prediction of multiple integrated propulsion systems. Conventional high-fidelity computational methods are too expensive, and low-cost low-fidelity methods make too many aerodynamic assumptions to be applied to these analyses. Dual-solver hybrid computational fluid dynamics (CFD) methods bridge the gap between these two categories of solvers, but many lack capabilities that are needed for complex vehicles. Recent improvements to the hybrid solver OVERFLOW-CHARM provide efficient and accurate analysis of integrated propulsion systems with hybrid CFD methods, as demonstrated on the wing–propeller system of the AIAA Workshop for Integrated Propeller Prediction. Predictions of propeller thrust are within 1% of conventional CFD data, and wing pressure coefficient distributions agree well with both experimental data and conventional CFD methods. The dual-solver results were obtained at half the computational cost of the full CFD simulation.

Red blood cells tracking and cell-free layer formation in a microchannel with hyperbolic contraction: A CFD model validation

Background and ObjectiveIn recent years, progress in microfabrication technologies has attracted the attention of researchers across disciplines. Microfluidic devices have the potential to be developed into powerful tools that can elucidate the biophysical behavior of blood flow in microvessels. Such devices can also be used to separate the suspended physiological fluid from whole in vitro blood, which includes cells. Therefore, it is essential to acquire a detailed description of the complex interaction between erythrocytes (red blood cells; RBCs) and plasma. RBCs tend to undergo axial migration caused by occurrence of the Fåhræus-Lindqvist effect. These dynamics result in a cell-free layer (CFL), or a low volume fraction of cells, near the vessel wall. The aim of the paper is to develop a numerical model capable of reproducing the behavior of multiphase flow in a microchannel obtained under laboratory conditions and to compare two multiphase modelling techniques Euler-Euler and Euler-Lagrange. MethodsIn this work, we employed a numerical Computational Fluid Dynamics (CFD) model of the blood flow within microchannels with two hyperbolic contraction shapes. The simulation was used to reproduce the blood flow behavior in a microchannel under laboratory conditions, where the CFL formation is visible downstream of the hyperbolic contraction. The multiphase numerical model was developed using Euler-Euler and hybrid Euler-Lagrange approaches. The hybrid CFD simulation of the RBC transport model was performed using a Discrete Phase Model. Blood was assumed to be a nonhomogeneous mixture of two components: dextran, whose properties are consistent with plasma, and RBCs, at a hematocrit of 5% (percent by volume of RBCs). ResultsThe results show a 5 μm thick CFL in a microchannel with a broader contraction and a 35 μm thick CFL in a microchannel with a narrower contraction. The RBC volume fraction in the CFL is less than 2%, compared to 7–8% in the core flow. The results are consistent for both multiphase simulation techniques used. The simulation results were then validated against the experimentally-measured CFL in each of the studied microchannel geometries. ConclusionsReasonable agreement between experiments and simulations was achieved. A validated model such as the one tested in this study can expedite the microchannel design process by minimizing the need to prefabricate prototypes and test them under laboratory conditions.

Modeling the interplay between water capillary transport and species diffusion in gas diffusion layers of proton exchange fuel cells using a hybrid computational fluid dynamics formulation

Improved modeling of the membrane electrode assembly (MEA) and operation is essential to optimize proton exchange fuel cells (PEFCs). In this work, a hybrid model, which includes a pore network formulation to describe water capillary transport and a continuum formulation to describe gas diffusion, is presented. The model is validated with previous data of carbon-paper gas diffusion layers (GDL), including capillary pressure curve, relative effective diffusivity, g ( s ) , and saturation profile. The model adequately captures the increase of capillary pressure with compression, the nearly cubic dependency of g ( s ) on average saturation, s avg , and the shape of the saturation profile in conditions dominated by capillary fingering (e.g., running PEFC at low temperature). Subsequently, an analysis is presented in terms of the area fraction of water at the inlet and the outlet of the GDL, A w i n and A w o u t , respectively. The results show that gas diffusion is severely hindered when A w i n is exceedingly high (¿80%), a situation that can arise due to the bottleneck created by flooded interfacial gaps. Furthermore, it is found that s avg increases with A w o u t , reducing the GDL effective diffusivity. Overall, the work shows the importance of an appropriate design of MEA porous media and interfaces in PEFCs. • A hybrid model for gas diffusion layers in proton exchange fuel cells is presented. • Pore network & continuum models are used for water transport & gas diffusion, resp. • The model is extensively compared with experimental data, showing good agreement. • The limiting current density dramatically reduces when interfacial flooding occurs. • More isolated evacuation sites toward the channel decrease effective diffusivity.

Numerical and experimental study on spectrum and temporal coherence analyses of flow noise caused by sinusoidal vertical motion of sonobuoy

Flow noise caused by sinusoidal vertical motion of the sonobuoy reduces the signal-to-noise ratio (SNR) of the recorded signal. Therefore, understanding the physical properties of the flow noise is important for improving the working performance of the sonobuoy. The present work is dedicated to investigate the characteristics of the flow noise of sonobuoy generated from vertical motion based on the combination of numerical calculations and experimental results. The flow noise of the hydrophone under time varying stream velocity is predicted using a hybrid computational fluid dynamics (CFD)-Ffowcs Williams-Hawkings (FW–H) equation technique. The flow noise calculated with the hybrid computational method is in good agreement with that analyzed by the experimental measurement and empirical formula. The numerical calculation results of the flow noise under time-varying stream velocity condition reveal that the spectrum and temporal coherence of the flow noise exhibit the periodic changes same as that of the stream velocity. These analyses and conclusions are verified by experimental data. The study in this paper will contribute to signal processing of the sonobuoy and thus improving SNR of the measure signal.

Effect of fracture roughness on the hydrodynamics of proppant transport in hydraulic fractures

The effect of fracture roughness is investigated on proppant transport in hydraulic fractures using Joint Roughness Coefficient and a three-dimensional multiphase modelling approach. The equations governing the proppant transport physics in the fracturing fluid is solved using the hybrid computational fluid dynamics model. The reported proppant transport models in the literature are limited to the assumption of a smooth fracture domain with no fluid leak-off or fluid flow from fracture to rock matrix interface. In this paper, a proppant transport model is proposed that accounts for the proppant distribution in rough fracture geometry with fluid leak-off effect to surrounding porous rock. The hydrodynamic and mechanical behaviour of proppant transport was found directly related to the fracture roughness and flow regime especially under the influence of low viscosity fracturing fluid typically used in shale gas reservoirs. For the proppant transport in smooth fractures, the fracture walls employ mechanical retardation effects and reduce the proppant horizontal velocity resulting in more significant proppant deposition. On the contrary, for the proppant transport in rough fractures, the inter-proppant and proppant wall interactions become dominant that adds turbulence to the flow. It results in mechanical interaction flow effects becoming dominant and consequently higher proppants suspended in the slurry and greater horizontal transport velocity. Furthermore, the mechanical interaction flow effects were found to be principally dependant on the proppant transport regime and become significant at higher proppant Reynolds number.

Fouling in a Steam Cracker Convection Section Part 2: Coupled Tube Bank Simulation using an Improved Hybrid CFD-1D Model

Performing an adequate fouling study for the heat exchangers in the convection section of a steam cracker requires reliable data on circumferential tube wall temperature profiles. A hybrid Computational Fluid Dynamics (CFD)-1D convection section model, developed to perform coupled flue gas/process gas side simulations of convection sections, is improved by the implementation of flue gas radiation modeling and extended to include typical tube banks. A complete naphtha cracker convection section is simulated with the improved hybrid CFD-1D model. All tubes show distinct maximum heat fluxes on the tube walls due to the high flue gas velocity. Based on the calculated circumferential heat flux profiles, the maximum heat flux value is calculated to be 1.8 times the average tube heat flux value. As computational costs associated with a hybrid CFD-1D simulation are high, a convective heat flux profile reconstruction scheme is developed. Using the scheme, circumferential heat flux profiles are reconstructed, based on the heat fluxes calculated when performing a fully 1D coupled convection section simulation. The heat flux reconstruction profile scheme enables fast retrieval of circumferential heat flux profiles and, thus, tube wall temperature profiles. Optimization and/or design of a steam cracker convection section becomes less computationally demanding.

Optimizing the 3D Distributed Climate inside Greenhouses Using Multi-Objective Optimization Algorithms and Computer Fluid Dynamics

As one of the major production facilities in agriculture, a greenhouse has many spatial distributed factors influencing crop growth and energy consumption, such as temperature field, air flow pattern, CO 2 concentration distribution, etc. By introducing a hybrid computational fluid dynamics–evolutionary algorithm (CFD-EA) method, this paper constructs a micro-climate model of greenhouse with main environmental parameters optimized. Considering environmental factors’ spatial influences together with energy usage simultaneously, the optimal solutions of control variables for crop growth are calculated. A commercial greenhouse located in east China is chosen for the method validation. Field experiments using temperature/velocity sensor matrix are carried out for CFD accuracy investigation. On this basis, the proposed optimization method is employed to search for the optimal control variables and parameters corresponding to the environmental Pareto frontier. By the proposed multi-objective scheme, we believe the method can provide set point basis for the design and regulation of large/medium-sized greenhouse production with high spatial resolution.