Steady-state Computational Fluid Dynamics Research Articles

Analysis of the shear-driven flow in a scale model of a phase separator: Validation of a coupled CFD approach using experimental data from a physical model

The effects of shear stress at the interface between a flowing gaseous phase and a liquid-filled cavity, a scenario commonly found in flue gas/slag phase separators are investigated in this study. This effect is frequently disregarded in computational fluid dynamics (CFD) simulations within metallurgy, primarily due to its limited influence on the motion of liquids (e.g., slag, liquid metal) when compared to other factors such as stirring by natural convection or bubbling. This study was undertaken to investigate shear-driven flow in a 1:4 scale model of a phase separator in which a flue gas stream coming from a furnace is mobilizing the flow of a slag melt. To cover a wide range of operating conditions, silicone oil with varying viscosities (20, 50, and 70 cSt) was considered for mimicking viscous slag. Additionally, different volumetric airflow rates (18, 27, and 36 m3/h) were considered to simulate the flue gas flow. Particle image velocimetry (PIV) measurements were conducted on the model. The experimental results were utilized to validate computationally efficient, coupled steady-state simulations of the system. Notably, the steady-state simulations, requiring approximately one hour of CPU time, demonstrated a computation time decrease of two magnitudes compared to conventional Volume of Fluid (VOF) modeling. A stable, laminar flow pattern of the liquid phase with surface velocities of up to approximately 30 mm/s was observed for the case with the lowest oil viscosity and the highest airflow rate, while a steady, turbulent flow was observed in the gas phase. Conversely, the lowest surface velocities, measuring around 5 mm/s, were recorded and calculated for the scenario with the highest oil viscosity and the lowest airflow rate. It was noted that velocities increased proportionally with the airflow rate but exhibited a disproportionate relationship with oil viscosity. The prediction of velocity magnitudes by the steady-state CFD model was highly accurate, with only minor discrepancies observed between the measured and calculated flow patterns. In summary, valuable insights into the often-overlooked shear stress effects at the phase boundary in gaseous-liquid flow over a cavity are provided by our study.

Development of a steady-state computational fluid dynamics solver for transonic rotorcraft flows

The capability for solving compressible fluid flows in the rotating frame of reference is added to an existed open-source CFD solver, namely, HiSA solver. The new implementation is explained and validated using the experimental data of the Sikorsky S-76 rotor. A comparison is presented between the moving mesh results obtained from the original HiSA code and the single rotating frame results achieved through the new implementation. The comparison includes an analysis of torque and thrust values, as well as computational costs. The results imply that, for evaluating the performance of an isolated rotor or for shape optimization purposes at the transonic regime, the single rotating frame method, like the one introduced in the current work, can provide accurate results within an acceptable computational budget. Furthermore, the results show that, at least 25 revolutions are required for the transient analysis to reach an acceptable steady-state converged solution like the one obtained by the single rotating frame method.

Influence of posture during gliding flight in the flying lizard Draco volans

The agamid lizards of the genus Draco are undoubtedly the most renown reptilian gliders, using their rib-supported patagial wings as lifting surfaces while airborne. Recent investigations into these reptiles highlighted the role of body posture during gliding, however, the aerodynamics of postural changes in Draco remain unclear. Here, we examine the aerodynamics and gliding performances of Draco volans using a numerical approach focusing on three postural changes: wing expansion, body camber, and limb positioning. To this aim, we conducted 70 three-dimensional steady-state computational fluid dynamics simulations of gliding flight and 240 two-dimensional glide trajectory calculations. Our results demonstrate that while airborne, D. volans generates a separated turbulent boundary layer over its wings characterized by a large recirculation cell that is kept attached to the wing surface by interaction with wing–tip vortices, increasing lift generation. This lift generating mechanism may be controlled by changing wing expansion and shape to modulate the generation of aerodynamic force. Furthermore, our trajectory simulations highlight the influence of body camber and orientation on glide range. This sheds light on how D. volans controls its gliding performance, and conforms to the observation that these animals plan their glide paths prior to take off. Lastly, D. volans is mostly neutral in pitch and highly maneuverable, similar to other vertebrate gliders. The numerical study presented here thus provides a better understanding of the lift generating mechanism and the influence of postural changes in flight in this emblematic animal and will facilitate the study of gliding flight in analogous gliding reptiles for which direct observations are unavailable.

Acoustic attenuation prediction and analysis of perforated hybrid mufflers with non-uniform flow based on frequency domain linearized Navier-Stokes equations

The three-dimensional frequency-domain linearized Navier-Stokes equations (LNSEs) with consideration of eddy viscosity are developed to evaluate the acoustic attenuation performance of perforated hybrid mufflers in the presence of non-uniform flow. The computations are performed in two steps: time averaged flow variables are acquired by using steady-state computational fluid dynamics (CFD) method and then mapped into the acoustic mesh, and the acoustic perturbation variables are obtained by solving frequency-domain LNSEs, where the sound-absorbing material is treated as an equivalent fluid with complex sound speed and density. The predictions of transmission losses of the two-pass perforated hybrid mufflers in the presence of non-uniform flow are in good consistencies with the measurements, which verifies the correctness of LNSEs. The effect of Mach numbers on acoustic attenuation performance of the mufflers with different filling densities and perforated components is investigated in detail. The transmission loss of the mufflers with various filling densities are increased by complex airflow in the lower frequency range. For the mufflers where not all components have been perforated, the flow lowers low-frequency resonance peak. The influence of complex airflow on acoustic attenuation performance of all configurations is weakened at higher frequencies.

Acoustic Attenuation Prediction of Perforated Reactive and Dissipative Mufflers With Flow by Using Frequency-Domain Linearized Navier-Stokes Equations

To comprehensively consider the impact of complex airflow on sound propagation, a numerical approach by solving the three-dimensional (3-D) frequency-domain linearized Navier-Stokes equations (LNSEs) is proposed to accurately predict the transmission losses of perforated tube reactive and dissipative mufflers in the presence of complex airflow. The numerical computations are performed in two steps: (1) solve the time averaged flow by steady-state computational fluid dynamics (CFD) simulation and then map the results into acoustic mesh; (2) treat sound-absorbing material as an equivalent fluid and obtain acoustic perturbation variables by solving the frequency-domain LNSEs. The transmission losses of perforated tube reactive and dissipative mufflers are measured in the presence of flow and compared with the numerical predictions. The excellent consistencies between predictions and measurements validate the present approach. The influence of Mach number on acoustic attenuation performance of perforated tube reactive and dissipative mufflers is then investigated. With Mach number increases, the transmission loss of perforated tube reactive muffler increases somewhat, while the transmission loss of dissipative mufflers decreases obviously in the lower frequency range and change is limited at higher frequencies.

The importance of unsteady phenomena of ammonia/methane combustion in an experimental swirl burner: Comparison of steady-state and transient simulation results

Ammonia is a potential advanced green fuel. Combustion simulations play a crucial role in facilitating the usage of ammonia in existing combustion systems as a drop-in fuel. Choosing between the unsteady and steady simulation methods should be done carefully depending on the investigated phenomena. In this study, steady-state and transient Computational Fluid Dynamics simulations were performed using detailed chemistry of the Okafor mechanism for a wide range of ammonia/methane blend combustion under lean conditions. The results revealed that the steady-state simulations could provide close values to the unsteady mean results at the outlet in magnitude, while outlet temperature and intermediate species concentration slightly differ with ammonia concentration-dependent extents. The unsteady simulations provided valuable insights into the temporal evolution of the combustion and flow field characteristics, and the Root Mean Square and mean properties were evaluated comprehensively, indicating that the oscillations highly depend on the fuel composition. The model was validated using measurement data from the literature, including Particle Image Velocimetry, OH* chemiluminescence, and pollutant emission. The velocity and OH* distribution matched acceptably, while the CO, NO, and NO2 emissions were underpredicted, which implies necessary reaction model improvements.

Validation of Rotating Detonation Combustor Computational Fluid Dynamics Simulations for Predicting Unsteady Supersonic–Subsonic Flow Field at the Exit

Abstract Rotating detonation combustors (RDCs) have gained increased interest for integration with power-generating gas turbines due to the potential to increase thermal efficiency. The unsteady flow field exiting the RDC is fundamentally different compared to traditional swirl-stabilized combustors. Successful integration of RDC with gas turbines will depend on the ability to properly condition the unsteady flow to achieve performance levels comparable to swirl-stabilized combustors. RDC simulations require significant computational resources due to the small spatial and temporal time scales required to resolve the detonation phenomenon. Furthermore, traditional steady-state computational fluid dynamics (CFD) analyses are not possible for RDC simulations. The present study develops and validates a computationally efficient approach for predicting unsteady flow fields exiting the combustor using 2D, transient reacting CFD with periodic boundary conditions in the combustor and a downstream plenum. Validation is performed by comparing the CFD results to various experimental measurements: (i) wave speed obtained from high-speed ion probe and dynamic pressure data, (ii) average wall static pressure measurements, and (iii) time-resolved particle image velocimetry (PIV) at 100 kHz at the RDC exit. Results indicate good agreement between CFD and experiments with respect to velocity field exiting the RDC, detonation wave speed, and static pressure distribution.

The data assimilation of volumetric velocity and temperature data to extract boundary conditions for high-fidelity simulation of the indoor air

The contemporary challenge of the air conditioning sector is to continuously improve the provided comfort conditions while reducing the energy consumption of the devices. This necessity can only be satisfied by improving research and design methods. Therefore, on the one hand, researchers worldwide develop and utilize experimental methods for more accurate measurements. On the other hand, researchers develop numerical methods to investigate design iterations effectively and quickly.The main idea of this study is to create a novel data assimilation method to utilize all the advantages of experimental and numerical methods to develop a high-fidelity investigation method of an indoor air model.In the present study, the three-dimensional velocity distribution at the split air conditioning indoor unit's outlet section was measured by using the Stereo Particle Image Velocity Measurement method. The temperature distribution at the outlet section of the split air conditioning indoor unit was also determined using the Meshed Infrared Thermography method. The data from these experimental measurements were assimilated to generate a highly realistic boundary condition of the steady-state and time-dependent Computational Fluid Dynamics (CFD) models. This is achieved for the first time in the literature.The results of these analyses compared with conventional CFD and the aforementioned experimental methods. The study's results show that the novel data assimilation approach (3.13% average error) is significantly more accurate than the traditional average boundary condition approach (13.19% average error).

A Computationally Efficient Method for the Thermo-Structural Analysis of a Turbocharger Turbine Wheel Under Engine Transient Operation

Abstract The accurate evaluation of thermally-induced stresses is of particular importance for life-cycle analysis of modern turbochargers. Conjugated heat transfer (CHT), steady-state calculations represent the state-of-the-art, whose temperature distribution is usually imported into a structural solver as a thermal load for carrying out thermo-structural analyses. In the case of transient analyses, however, issues arise. Indeed, due to the different timescales of convective and conductive heat-transfer mechanisms, unsteady CHT simulations are burdened by high computational costs. Moreover, in the specific case of turbocharged engines, the timescales of engine load transients and turbocharger response may differ by a factor of up to 105. The present study proposes an approach for carrying out transient thermo-structural analyses of the solid domain only, characterized by a physical time in the order of 101 s, without the need for unsteady fluid flow calculations, which would imply the use of a solver time-step in the order of 10−5 s. In the proposed methodology, the effects of the fluid flow in the heat-transfer process are modeled by imposing adequate boundary conditions to wet surfaces of the solid domain with minimum computational effort. Indeed, by performing two steady-state CHT calculations and one steady-state computational fluid dynamics (CFD) calculation, time-dependent boundary conditions can be derived in terms of convective heat transfer coefficient and near-wall fluid temperature. The case study used for model development is a high-performance turbocharger turbine for automotive applications, responding to engine load transients of tenths of seconds. Initially, steady-state CFD and CHT models of the turbine stage are validated against experimental data. Then, transient analyses on the turbine wheel are carried out, and their results are imported as thermal loads into an unsteady structural solver, allowing for accurate transient thermo-structural analyses at sustainable computational cost.

Assessment of Computational Fluid Dynamics Reynolds-Averaged Navier–Stokes Models for Bluff Bodies Aerodynamics

<div>Since the steady-state computational fluid dynamics (CFD) Reynolds-averaged Navier–Stokes (RANS) turbulence models offer low-cost and sensible accuracy, they are frequently utilized for bluff bodies’ external aerodynamics investigations (e.g., upwind, crosswind, and shape optimization). However, no firm certainty is made regarding the best model in terms of accuracy and cost. Based on cost and accuracy aspects, four RANS turbulence models were studied, which are Spalart–Allmaras, realizable k-ε, RNG k-ε, and SST k-ω. Ahmed body with a 25° slant angle benchmark case was introduced for this investigation. Two grids were generated to satisfy the near-wall treatment of each turbulence model. All grid settings were proposed and discussed in detail. Fluid-structure analysis was performed on five different planes. Regarding flow field prediction, realizable k-ε and renormalization group (RNG) k-ε models demonstrated a remarkable consistency with experimental data, while Menter’s shear stress transport (SST) k-ω showed a poor agreement. The obtained computational values of drag and lift coefficients were compared with experimental results. All investigated RANS turbulence models had reported results in excellent agreement with experimental drag coefficient values. The SST k-ω model has underestimated lift coefficient value with an error of about −45% with experimental value. Only realizable k-ε and RNG k-ε presented an error <10% for predicting drag and lift coefficients.</div>

Design and analysis of cooling jacket developed for vacuum power tubes by multiphase cooling

The vacuum tubes are used extensively in electronic industries. The MW level of vacuum tube are used in RF amplifiers and microwave sources in different sectors like defence, nuclear energy, satellites and medical diagnostics. The first step in the indigenous development of MW level of vacuum tube is the design of its cooling jacket. The cooling jacket of vacuum tubes use hypervapotron (HV) fins which are known for providing efficient cooling in compact space. The key objectives of this paper are to finalize the jacket’s construction material, evaluate the cooling jacket’s heat transfer performance at the MW level of RF (radiofrequency) power, and forecast safe operating conditions for given conditions of high heat flux and water flow rates. In this design, the vacuum tubes’ external dimensions are taken into consideration when designing the cooling jacket. ETP copper (Electrolytic Tough Pitch copper) and copper-chromium-zirconium (CuCrZr) are likely materials for the jacket’s construction. For the aforementioned design of the cooling jacket, FEA (Finite Element Analysis) simulations using a steady-state thermal module and computational fluid dynamics (CFD) simulations using an ANSYS CFX module are described for flow rates of 15, 30, and 60 lpm (liters per minute) and the heat flux 1, 2, 4, and 6 MW m−2. The results of CFD and FEA simulations are found to be in close agreement. A small deviation in results is seen after the start of nucleate boiling. 30 and 60 lpm are desirable flow rates for high heat flux values greater than 2 MW m−2. CuCrZr is the ideal material for a cooling jacket’s construction as it has safe working temperature limit of 350 °C at high heat flux values. Also, 15 lpm flow rate should be avoided while using this cooling jacket at heat flux values greater than or equal to 6 MW m−2.

In-depth investigation of natural convection thermal characteristics of BALI experiment through Eulerian computational fluid dynamics code and comparison with Lagrangian code

In-vessel retention through external reactor vessel cooling (IVR-ERVC) is a severe accident management (SAM) strategy that has been adopted and used in many nuclear reactors such as AP1000, APR1400, and light water reactor etc. Some reactor accidents have raised concerns about nuclear reactors among residents, leading to a decrease in residents' acceptability and many studies on SAM are being conducted. Experiments on IVR-ERVC are almost impossible due to its specificity, so fluid characteristics are analyzed through BALI experiments with similar condition. In this study, computational fluid dynamics (CFD) via Reynolds-averaged Navier-Stokes (RANS) and large eddy simulation (LES) for BALI experiments were performed. Steady-state CFD analysis was performed on three turbulence models, and SST k-ω model was in good agreement with the experimental measurement temperature within the maximum error range of 1.9%. LES CFD analysis was performed based on the RANS analysis results and it was confirmed that the temperature and wall heat flux for depth was consistent within an error range of 1.0% with BALI experiment. The LES CFD analysis results were compared with those of the Lagrangian-based solver. LES matched the temperature distribution better than SOPHIA, but SOPHIA calculated the position of boundary between stratified layer and convective layer more accurately. On the other hand, Lagrangian-based solver predicted several small eddy behaviors of the convective layer and LES predicted large vortex behavior. The vibration characteristics near the cooling part of the BALI experimental device were confirmed through Fast Fourier Transform (FFT) investigation. It was found that the power spectral density for pressure at least 10 times higher near the side cooling than near the top cooling.

Hostile Hemodynamics in Distal Stent Graft-Induced New Entry Prior to Aortic Rupture: A Comparison of Transient versus Steady-State CFD Simulations.

Computational fluid dynamics (CFD) simulations model blood flow in aortic pathologies. The aim of our study was to understand the local hemodynamic environment at the site of rupture in distal stent graft-induced new entry (dSINE) after frozen elephant trunk with a clinically time efficient steady-flow simulation versus transient simulations. Steady-state simulations were performed for dSINE, prior and after its development and prior to aortic rupture. To account for potential turbulences due geometric changes at the dSINE location, Reynolds-averaged Navier-Stokes equations with the realizable k-ε model for turbulences were applied. Transient simulations were performed for comparison. Hemodynamic parameters were assessed at various locations of the aorta. Post-dSINE, jet-like flow due to luminal narrowing was observed which increased prior to rupture and resulted in focal neighbored regions of high and low wall shear stress (WSS). Prior to rupture, aortic diameter at the rupture site increased lowering WSS at the entire aortic circumference. Concurrently, WSS and turbulence increased locally above the entry tear at the inner aortic curvature. Turbulent kinetic energy and WSS elevation in the downstream aorta demonstrated enhanced stress on the native aorta. Results of steady-state simulations were in good qualitative agreement with transient simulations. Steady-flow CFD simulations feasible at clinical time scales prior to aortic rupture reveal a hostile hemodynamic environment at the dSINE rupture site in agreement with lengthy transient simulations. Consequently, our developed approach may be of value in treatment planning where a fast assessment of the local hemodynamic environment is essential.

Performance of single-layer paper-based co-laminar flow microbial fuel cells

This paper presents a novel design and operation of microbial fuel cells (ΜFCs), which contain monolayer paper-based substrate/electrodes and microchannels with co-laminar flow. The electrodes with multi-wall carbon nanotubes are fabricated by the screen-printing method and the microchannels are patterned using photo-lithography. A double-inlet and diverging channel design is incorporated in the fuel cell configuration and demonstrated significantly improved performance. The fluid flows of electrolytes through the porous paper media are simulated using steady-state and transient computational fluid dynamics The best performance is achieved under the following conditions: an electroactive microbial (S. oneidensis) concentration of OD6001.5, 50 mM electron donor (lactate), and direct immobilized of the inoculum on the anode surface. The developed MFCs achieves a peak power density of 19.4 ± 0.23 μW cm−2 and maximum current density of 190.4 ± 1.39 μA cm−2, surpassing the performance of all previously reported paper-based single MFCs that utilize paper-based electrodes. In addition, hybrid-type MFCs that containing enzymatic air-breathing cathodes are investigated to enhance their performance. The novel paper based self-pumping MFC has the potential to make lab-on-a-chip type portable medical diagnosis devices with integrated power sources practical and feasible.

Steady-state sound propagation through hot exhaust jets in cooler cross-flow: A computational study

An analysis has been carried out to investigate the sound radiation through a heated jet in cooler cross-flow, which is representative of many industrial exhaust systems, using a hybrid steady-state computational fluid dynamics and computational acoustic model. The mean flow and temperature fields are modelled using steady-state computational fluid dynamics, with the turbulence modelled using Reynolds Averaged Navier-Stokes equations. The corresponding mean flow and temperature fields are used in the computational sound propagation model using linearised acoustic wave equation with mean flow based on a scalar flow potential. The results obtained from the computational simulations show that the flow significantly changes the sound propagation path and that the sound levels downstream of the duct outlet are higher than expected from using an acoustic monopole radiation pattern. The dominant mechanism affecting the propagation of sound is the refraction arising from the plume's temperature and velocity gradients. The sound propagation is highly dependent on the proximity from the duct outlet, normalised wavenumber, temperature and the jet to cross-flow mean velocity ratio. This computational study builds upon previous experimental work to analyse the fluid-acoustic interaction for heated jets in cooler cross-flow to understand the complex radiation pattern that leads to higher-than-expected sound levels downstream of the duct outlet.

Experimental and computational investigation of the vortical structures generated from a blended-wing-body UAV model

The current study presents a combination of experimental and computational investigations of a scaled-down Blended Wing Body (BWB) Unmanned Aerial Vehicle (UAV) model. The BWB model features highly swept wings. The experiments are conducted in a subsonic windtunnel in the Laboratory of Fluid Mechanics and Turbomachinery, in the Aristotle University of Thessaloniki. More specifically, surface oil flow visualization and Laser-Doppler Anemometry (LDA) are employed to investigate the vortical structures emanating from the leading edge and the wingtips of the model and provide the corresponding velocity and vorticity profiles. Using the windtunnel measurement data as a reference point, both steady-state and transient computational fluid dynamics (CFD) analyses are also conducted. The flow around the BWB scaled-down model is modeled by adopting four turbulence models, three of them based on the concept of the eddy-viscosity and one, on the Reynolds-stress transport equation modeling. The modeling results are compared to the experimental measurements, emphasizing on the streamwise and spanwise velocity profiles, axial vorticity profiles and separation patterns. The results of this study provide insight on the vortical phenomena dominating the flowfield over and around BWB configurations and conclusions concerning the tradeoffs between accuracy and computational efficiency are drawn.

A hybrid experimental-computational study: Prediction of flow fields and full-field pressure distributions on measured shapes of three-tab asphalt roofing shingles subjected to hurricane velocity winds

A novel hybrid experimental-computational study is performed to predict the flow fields and pressure distributions on the measured three-dimensional shapes of flexible, three-tab asphalt roofing shingles undergoing increasing uplift when exposed to hurricane velocity winds for two hours. To quantify the evolution of shingle shapes, StereoDIC analysis is used to measure the transient, full-field deformed shapes of full-sized, three-tab asphalt shingles that did not show separation or failure when subjected to hurricane velocity winds for two hours. Based on physical observations during wind loading, the authors performed steady state computational fluid dynamics (CFD) simulations to predict the full-field pressure distributions on as-measured, uplifted three-dimensional shingle shapes at selected time instances during wind loading.Simulation predictions clearly show flow recirculation regions on both the front and top of the shingles that remain attached throughout wind loading and control the full-field uplift pressure distribution. For low velocity flow with maximum uplift ≤ 8.4 mm, CFD-predicted pressures are in good agreement with prior measurements. For both low and high-speed flows, the model predictions indicate that high pressures are formed at the leading-edge, upstream of the sealant layer, with maximum pressure occurring near the tab cutouts along the leading-edge of the shingle, providing a physical basis for the observed higher uplift and increased potential for shingle failure in these regions. The combined experimental-computational studies provide a contemporary way to eliminate the difficulties associated with attachment of pressure sensors to flexible materials that can alter shingle response, providing the basis for future design improvements by delineating the physical processes controlling pressure loading and shingle uplift in hurricane velocity winds.

End-to-end wind turbine wake modelling with deep graph representation learning

Wind turbine wake modelling is of crucial importance to accurate resource assessment, to layout optimisation, and to the operational control of wind farms. This work proposes a surrogate model for the representation of wind turbine wakes based on a state-of-the-art graph representation learning method termed a graph neural network. The proposed end-to-end deep learning model operates directly on unstructured meshes and has been validated against high-fidelity data, demonstrating its ability to rapidly make accurate 3D flow field predictions for various inlet conditions and turbine yaw angles. The specific graph neural network model employed here is shown to generalise well to unseen data and is less sensitive to over-smoothing compared to common graph neural networks. A case study based upon a real world wind farm further demonstrates the capability of the proposed approach to predict farm scale power generation. Moreover, the proposed graph neural network framework is flexible and highly generic and as formulated here can be applied to any steady state computational fluid dynamics simulations on unstructured meshes.

Analysis of Surrogate Models for Vapour Transport and Distribution in a Hollow Fibre Membrane Humidifier

To achieve high efficiency and low degradation of a polymer electrolyte fuel cell (PEMFC), it is necessary to maintain an appropriate level of humidification in the fuel cell membrane. Thus, membrane humidifiers are typically used in PEMFC systems. Parameter studies are important to evaluate membrane humidifiers under various operating conditions to reduce the amount of physical tests. However, simulative studies are computationally expensive when using detailed models. To reduce the computational cost, surrogate models are set up. In our study, a 3D computational fluid dynamics (CFD) model of a hollow fibre membrane humidifier is presented and validated using measurement data. Based on the results of the validated CFD model, a surrogate model of the humidifier is constructed using proper orthogonal decomposition (POD) in combination with different interpolation methods. To evaluate the surrogate models, their results are compared against reference solutions from the CFD model. Our results show that a Halton design combined with a thin-plate-spline interpolation results in the most accurate surrogate humidifier model. Its normalised mean absolute error for 18 test points when predicting the water mass fraction in the membrane humidifier is 0.58%. Furthermore, it is demonstrated that the solutions of the POD model can be used to initialise CFD calculations and thus accelerate the calculation of steady state CFD solutions.

Modelling blood flow in patients with heart valve disease using deep learning: A computationally efficient method to expand diagnostic capabilities in clinical routine.

The computational modelling of blood flow is known to provide vital hemodynamic parameters for diagnosis and treatment-support for patients with valvular heart disease. However, most diagnosis/treatment-support solutions based on flow modelling proposed utilize time- and resource-intensive computational fluid dynamics (CFD) and are therefore difficult to implement into clinical practice. In contrast, deep learning (DL) algorithms provide results quickly with little need for computational power. Thus, modelling blood flow with DL instead of CFD may substantially enhances the usability of flow modelling-based diagnosis/treatment support in clinical routine. In this study, we propose a DL-based approach to compute pressure and wall-shear-stress (WSS) in the aorta and aortic valve of patients with aortic stenosis (AS). A total of 103 individual surface models of the aorta and aortic valve were constructed from computed tomography data of AS patients. Based on these surface models, a total of 267 patient-specific, steady-state CFD simulations of aortic flow under various flow rates were performed. Using this simulation data, an artificial neural network (ANN) was trained to compute spatially resolved pressure and WSS using a centerline-based representation. An unseen test subset of 23 cases was used to compare both methods. ANN and CFD-based computations agreed well with a median relative difference between both methods of 6.0% for pressure and 4.9% for wall-shear-stress. Demonstrating the ability of DL to compute clinically relevant hemodynamic parameters for AS patients, this work presents a possible solution to facilitate the introduction of modelling-based treatment support into clinical practice.