Large Eddy Simulation Research Articles

Evaluation of flow and flame characteristics of turbulent bluff-body CH4-H2 flame using LES-FPV approach

ABSTRACT This study investigated the flow and flame structure of a bluff-body stabilised CH4-H2 flame using the large eddy simulation (LES) and the flamelet progress variable (FPV) model. The performance of three subgrid-scale (SGS) LES models, i.e. the standard Smagorinsky model (SSM), the wall-adapting local eddy viscosity (WALE) model, and the dynamic Smagorinsky model (DSM) in predicting turbulent and reactive flow characteristics was assessed. Parameters, namely y+ , Pope index (M), vorticity field, and resolved turbulent kinetic energy (kt ), were analysed to determine the predictive capability of the three models. The results demonstrated that DSM and WALE successfully resolved 80% of kt for chosen grid configurations, whereas SSM exhibited discrepancies. The overall trends captured by all three models matched well with the experimental trend with DSM and WALE outperforming SSM in predicting reacting flow fields in the recirculation zone, while SSM produced better predictions in the regions away from the recirculating zone.

Large Eddy Simulation and Acoustic Analysis of Combustion Instability in an Annular Combustor with Low-Swirl Flames

Large Eddy Simulation and Acoustic Analysis of Combustion Instability in an Annular Combustor with Low-Swirl Flames

Validation Assessment of Turbulent Reacting Flow Model Using the Area-Validation Metric on Medium-Scale Methanol Pool Fire Results

Abstract Accident analysis and ensuring power plant safety are pivotal in the nuclear energy sector. Significant strides have been achieved over the past few decades regarding fire protection and safety, primarily centered on design and regulatory compliance. Yet, after the Fukushima accident a decade ago, the imperative to enhance measures against fire, internal flooding, and power loss has intensified. Hence, a comprehensive, multilayered protection strategy against severe accidents is needed. Consequently, gaining a deeper insight into pool fires and their behavior through extensive validated data can greatly aid in improving these measures using advanced validation techniques. A model validation study was performed at Sandia National Laboratories (SNL) in which a 30-cm diameter methanol pool fire was modeled using the SIERRA/Fuego turbulent reacting flow code. This validation study used a standard validation experiment to compare model results against, and conclusions have been published. The fire was modeled with a large eddy simulation (LES) turbulence model with subgrid turbulent kinetic energy closure. Combustion was modeled using a strained laminar flamelet library approach. Radiative heat transfer was accounted for with a model utilizing the gray-gas approximation. In this study, additional validation analysis is performed using the area validation metric (AVM). These activities are done on multiple datasets involving different variables and temporal/spatial ranges and intervals. The results provide insight into the use of the area validation metric on such temporally varying datasets and the importance of physics-aware use of the metric for proper analysis.

LES of the upper plenum of the liquid metal fast reactor under the forced flow conditions

A Large Eddy Simulation (LES) of the upper plenum of the E-SCAPE (European SCAled Pool Experiment) facility under the normal operations has been conducted to gain deeper insights into the thermal hydraulics phenomena in liquid metal fast reactors (LMFRs). The results unravel the overall flow features in the upper plenum, the thermal instability in the above core structure region and the impact of the jets from the barrel holes on the large flow circulation and thermal mixing.The above core structure region is represented using a homogeneous porous model. It is shown that during the normal operations, most of the hot stream of the lead bismuth eutectic (LBE) from the core center rises to the top and disperses from the upper barrel holes. Conversely, the cold LBE spread out from the lower barrel holes. Mixing between the two streams only affect the fluids leaving from the middle heights demonstrating limited mixing. These jets help to create a large circulation in the upper plenum region which prevents any thermal stratification. At the interface between the hotter and colder streams, there are unsteady large-scale structures resulting in a mixing layer in the thermal field. This strong mixing is not conventional turbulence and potentially cannot be captured by Reynolds Averaged Navier Stokes average (RANS) modelling.The flow in the upper plenum is dominated by the influences of the different types of jets, which include some free jets issued horizontally or angled upwards and some jets impinging onto the structures in the plenum. The jets overall behave similar to ‘standard’ jets, though significant interactions between the top and bottom jets and the background circulations have strong influences at later stages of the jets, typically after four jet diameters leading to the jets not reaching self-similarity. In addition, the turbulence that is observed in the upper plenum is largely generated by the jet flows and hence the distribution of turbulence is very non-uniform. Away from the jets, turbulence is minimal with the mixing largely driven by the large-scale circulation.

Large eddy simulations of sheet-to-cloud cavitation transitions with special emphasis on the simultaneous existence of the re-entrant jet and shock waves

Sheet-to-cloud cavitation transitions are very complex owing to the simultaneous appearance of the re-entrant jet and shock waves. The objective of this paper is to investigate the physics of compressible cavitating flows with emphasis on the simultaneous existence of the re-entrant jet and shock waves. A compressible cavitating solver, which considers the compressibility effects of both the liquid and the vapour, was used to account for the different shedding characteristics induced by the re-entrant jet mechanism (RJM), the shock wave mechanism (SWM), and both the re-entrant jet and the shock waves (RJM-SMW). The solver used Large Eddy Simulations (LES) and the Schnerr-Sauer cavitation model. The numerical results are first compared with available experimental data [1] which showed satisfactory agreement for various aspects of the flow, including the pressure distribution, cavity shapes, condensation front speed, and shedding frequency. The results first show three different cavity shedding processes induced by the re-entrant jet and the shock waves. Frequency analyses show that the shock waves produce a wider frequency distribution which indicates that shock waves induce more shedding than the re-entrant jet. The quasi-periodic characteristics of the transitions between the various cavity shedding characteristics are mainly induced by the variations of the pressure, residual cavity volume and medium compressibility. Further analyses show that the different pressure, residual cavity volume and medium compressibility amplitudes determine the cavity shedding characteristics in the subsequent shedding cycle. When the conditions favour both the re-entrant jets and the shock wave effects, both the re-entrant jet and the shock waves impact the cavity shedding.

Acoustic tones generated by impinging jets: Differences between laminar and highly-disturbed nozzle-exit boundary layers

The differences between the acoustic tones generated by impinging jets with laminar and highly-disturbed nozzle-exit boundary layers are investigated. For that, jets at Mach numbers between 0.6 and 1.3 impinging on a flat plate at a distance of 8 nozzle radii from the nozzle exit are computed using large-eddy simulations. The amplitudes of the tones generated by the jets through feedback loops establishing between the nozzle and the plate are found to be significantly affected by the exit turbulent disturbances. In the present study, overall, they are lower for the initially laminar jets than for the initially disturbed ones. The level decrease varies from a few dB up to 15 dB, depending on the tones, which can change the frequencies of the dominant tones and the numbers and azimuthal structures of their associated feedback modes. For Mach numbers 0.75 and 0.8, for instance, the dominant tone frequencies are approximately two times lower for the initially laminar jets than for the other ones, yielding a better agreement with experiments of the literature in the former case. For a Mach number of 1.1, as a second example, the dominant tone is associated with the axisymmetric third feedback mode in the laminar case but with the helical fifth feedback mode in the disturbed case. The differences in the tone amplitude are finally discussed by estimating the power gains of the shear-layer instability waves between the nozzle and the plate using linear stability analysis for the axisymmetric mode. In most cases, at the frequency of a specific tone, the higher the gain, the stronger the acoustic tone.

Numerical prediction of mean flow and acoustic field of a supersonic impinging jet

The three-dimensional turbulent mean flow and acoustic field of a supersonic jet impinging on a solid plate is studied computationally using the general purpose CFD code Ansys Fluent. A pressure-based coupled solver formulation with the second order weighted central-upwind spatial discretization is applied to compute transient solutions. Cold and hot jet thermal conditions are considered. Mean flow characteristics are investigated by a steady-state modeling approach. Acoustic radiation of impingement tones is simulated using a transient time-domain formulation. The effects of turbulence in steady-state are modeled by the SST k-ω turbulence model. The Wall-Modeled Large-Eddy Simulation (WMLES) model is applied to compute transient solutions. The near-wall mesh on the impingement plate is fine enough to resolve the viscosity-affected near-wall region all the way to the laminar sublayer. Nozzle-to-plate distance is parameterized in the model for automatic re-generation of the mesh and results. Steady-state predictions of hover lift loss and mean jet velocity distributions are compared with experimental data, and favorable agreement is reported. The transient solution reproduces the mechanism of impingement tone generation by the interaction of large scale vortical structures with the impingement plate. The acoustic near-field is directly resolved by Computational Aeroacoustics (CAA) to accurately propagate impingement tone waves to near-field microphone locations. Calculated impingement tone frequencies and sound pressure levels agree with experimental values.

LES and RANS Spray Combustion Analysis of OME3-5 and n-Dodecane

Clean-burning oxygenated and synthetic fuels derived from renewable power, so-called e-fuels, are a promising pathway to decarbonize compression–ignition engines. Polyoxymethylene dimethyl ethers (PODEs or OMEs) are one candidate of such fuels with good prospects. Their lack of carbon-to-carbon bonds and high concentration of chemically bound oxygen effectively negate the emergence of polycyclic aromatic hydrocarbons (PAHs) and even their precursors like acetylene (C2H2), enabling soot-free combustion without the soot-NOx trade-off common for diesel engines. The differences in the spray combustion process for OMEs and diesel-like reference fuels like n-dodecane and their potential implications on engine applications include discrepancies in the observed ignition delay, the stabilized flame lift-off location, and significant deviations in high-temperature flame morphology. For CFD simulations, the accurate modeling and prediction of these differences between OMEs and n-dodecane proved challenging. This study investigates the spray combustion process of an OME3 − 5 mixture and n-dodecane with advanced optical diagnostics, Reynolds-Averaged Navier–Stokes (RANS), and Large-Eddy Simulations (LESs) within a constant-volume vessel. Cool-flame and high-temperature combustion were measured simultaneously via high-speed (50 kHz) imaging with formaldehyde (CH2O) planar laser-induced fluorescence (PLIF) representing the former and line-of-sight OH* chemiluminescence the latter. Both RANS and LES simulations accurately describe the cool-flame development process with the formation of CH2O. However, CH2O consumption and the onset of high-temperature reactions, signaled by the rise of OH* levels, show significant deviations between RANS, LES, and experiments as well as between n-dodecane and OME. A focus is set on the quality of the simulated results compared to the experimentally observed spatial distribution of OH*, especially in OME fuel-rich regions. The influence of the turbulence modeling is investigated for the two distinct ambient temperatures of 900 K and 1200 K within the Engine Combustion Network Spray A setup. The capabilities and limitations of the RANS simulations are demonstrated with the initial cool-flame propagation and periodic oscillations of CH2O formation/consumption during the quasi-steady combustion period captured by the LES.

Efficient prediction of tidal turbine fatigue loading using turbulent onset flow from Large Eddy Simulations

AbstractTo maximise the availability of power extraction from a tidal stream site, tidal turbines need to be able to operate reliably when located within arrays. This requires a thorough understanding of the operating conditions, which include turbulence, velocity shear due to bed proximity and roughness, ocean waves and due to upstream turbine wakes, over the range of flow speeds that contribute to the loading experienced by the devices. High-fidelity models such as Large Eddy Simulation (LES) can be used to represent these complex flow conditions and turbine device models can be embedded to predict loading. However, to inform micro-siting of multiple turbines with an array, the computational cost of performing multiple simulations of this type is impractical. Unsteady onset conditions can be generated from the LES to be used in an offline coupling fashion as input to lower-fidelity load prediction models to enable computationally efficient array design. In this study, an in-house Blade Element Momentum (BEM) method is assessed for prediction of the unsteady loads on the turbines of a floating tidal device with unsteady inflow developed with the in-house LES solver DOFAS. Load predictions are compared to those obtained using the same unsteady inflow to the commercial tool Tidal Bladed and from an Actuator Line Model (ALM) embedded in the LES solver. Estimates of fatigue loads differ by up to 3% for mean thrust and 11% for blade root bending moment for a turbine subject to a turbulent channel flow. When subjected to more complex flows typical of a turbine wake, the predictions of rotor thrust fatigue differ by up to 10%, with loads reduced by the inclusion of a pitch controller.

Research on wake characteristics and HCWF blade number matching of a new coupled propeller

For the purpose of wake energy recovery and hub vortex control in a new Coupled Propeller (CP), three hub cap with fins (HCWF) schemes with different blade numbers are designed for a specific multi-blade CP. A numerical investigation of the CP-HCWF system is conducted using the Large Eddy Simulation (LES) method. Firstly, the analysis focuses on the distribution of axial and tangential flow velocity characteristics of the CP. The wake spectral characterization analysis results reveal that the 10th-order component is equivalent to the 5th-order components within 0.7 R of the HCWF. However, beyond 0.7 R of the HCWF, the 10th-order component becomes more significant. Then, the influence of the HCWF blade number on the CP performance is further analyzed. Hydrodynamic efficiency, post-hub pressure recovery, wake energy recovery, and hub vortex elimination improve with increased HCWF blade numbers. The 10-blade HCWF demonstrates the best performance. Ultimately, the reliability of the numerical findings is verified through the model test. The efficiency variation with the fin number exhibits a consistent trend in the test and calculation results. It is suggested that the HCWF blade number is selected by considering the sum of the CP blades as a reference.

An Open–Closed Flow Separation in a Compressor Radial Diffuser

Abstract Large eddy simulations (LES) results are used to track flow separation zones in a splittered vaned radial diffuser near the compressor best efficiency at high rotation speed. The topology analysis reveals some closed separations at the corner of the blade and hub or shroud surface. In addition, an open–closed flow separation at the main blade pressure side is observed for the first time. The surface separation has the shape of a sheet at the blade mid-span, and the associated separation line is delimited only at its downstream extremity by a trailing edge node. This flow separation process is robust to a small shift in operating condition, and the separation surface remains essentially unchanged. A large flowrate variation converts the open–closed separation into a more classic closed separation. The evolution in time of the separation surface is also observed, and the authors conclude to the unsteady existence of the open–closed separation.

Sweeping JET actuator for the control of flow separation on a hump airfoil

Flow control can effectively improve lift and delay flow separation. Sweeping jet (SWJ) actuators have attracted increasing interest for as active flow control actuators. In this paper, the characteristics of the SWJ actuator were studied through large-eddy simulation and experiments. It was discovered that the oscillation frequency of the SWJ actuator is directly proportional to the mass flow rate, and its working principle was determined. Subsequently, SWJ actuators with [Formula: see text], 1.9%, and 6.6% were applied to flow control on the hump airfoil. It was observed that with an increase in momentum coefficient, the SWJ actuator was able to push the shear layer toward the hump airfoil surface, leading to a reduction in the size of the separation vortex. At [Formula: see text], it was achieved that the reattachment point of the separation vortex shifted upstream from [Formula: see text] to [Formula: see text], and the separation vortex size essentially disappeared. During the investigation of the flow control mechanism of the SWJ actuator, it was found that the SWJ actuator periodically generates a separation vortex V1, which has entrainment effects, resulting in the formation of an external jet on the airfoil surface, effectively inhibiting flow separation.

The multi-scale coupled model: a new framework capturing wind farm–atmosphere interaction and global blockage effects

Abstract. The growth in the number and size of wind energy projects in the last decade has revealed structural limitations in the current approach adopted by the wind industry to assess potential wind farm sites. These limitations are the result of neglecting the mutual interaction of large wind farms and the thermally stratified atmospheric boundary layer. While currently available analytical models are sufficiently accurate to conduct site assessments for isolated rotors or small wind turbine clusters, the wind farm's interaction with the atmosphere cannot be neglected for large-size arrays. Specifically, the wind farm displaces the boundary layer vertically, triggering atmospheric gravity waves that induce large-scale horizontal pressure gradients. These perturbations in pressure alter the velocity field at the turbine locations, ultimately affecting global wind farm power production. The implication of such dynamics can also produce an extended blockage region upstream of the first turbines and a favorable pressure gradient inside the wind farm. In this paper, we present the multi-scale coupled (MSC) model, a novel approach that allows the simultaneous prediction of micro-scale effects occurring at the wind turbine scale, such as individual wake interactions and rotor induction, and meso-scale phenomena occurring at the wind farm scale and larger, such as atmospheric gravity waves. This is achieved by evaluating wake models on a spatially heterogeneous background velocity field obtained from a reduced-order meso-scale model. Verification of the MSC model is performed against two large-eddy simulations (LESs) with similar average inflow velocity profiles and a different capping inversion strength, so that two distinct interfacial gravity wave regimes are produced, i.e. subcritical and supercritical. Interfacial waves can produce high blockage in the first case, as they are allowed to propagate upstream. On the other hand, in the supercritical regime their propagation speed is less than their advection velocity, and upstream blockage is only operated by internal waves. The MSC model not only proves to successfully capture both local induction and global blockage effects in the two analyzed regimes, but also captures the interaction between the wind farm and gravity waves, underestimating wind farm power by about only 2 % compared with the LES results. Conversely, wake models alone cannot distinguish between differences in thermal stratification, even if combined with a local induction model. Specifically, they are affected by a first-row overprediction bias that leads to an overestimation of the wind farm power by 13 % to 20 % for the modeled regimes.

Embedded large eddy simulation of boundary layer transition behind a micro-ramp

Boundary layer transition induced by surface roughness elements plays an important role in aerodynamic and aero-thermodynamic design of subsonic/supersonic/hypersonic vehicles. However, the effect of three-dimensional isolated roughness element in the process of promoting/suppressing boundary layer transition is far from being fully understood, particularly in the incompressible laminar flow. In the present study, the laminar-to-turbulent transition induced by a three-dimensional isolated micro-ramp element immersed in an incompressible laminar boundary layer is investigated numerically. The embedded large eddy simulation (ELES) combining the intermittency transition model and the wall-modeled large eddy simulation (WMLES) S − Ω model is employed for the first time. Numerical results on the time-averaged/instantaneous flow field and the statistical flow fluctuations are analysed and validated thoroughly by the existing experimental measurements. It is found that the interaction of the secondary and the tertiary streamwise vortices causes a high level of wall shear and an inflectional velocity distribution in the near-wall region. An additional eddy system that develops nonlinearly from the unstable inflectional flow may trigger the boundary layer transition. The present study validates that the ELES combined with the WMLES S − Ω model is an efficient simulation tool for industrial wall-bounded flows allowing effective compromise between flexibility, cost, and accuracy.

U-Plume: automated algorithm for plume detection and source quantification by satellite point-source imagers

Abstract. Current methods for detecting atmospheric plumes and inferring point-source rates from high-resolution satellite imagery are labor-intensive and not scalable with regard to the growing satellite dataset available for methane point sources. Here, we present a two-step algorithm called U-Plume for automated detection and quantification of point sources from satellite imagery. The first step delivers plume detection and delineation (masking) with a U-Net machine learning architecture for image segmentation. The second step quantifies the point-source rate from the masked plume using wind speed information and either a convolutional neural network (CNN) or a physics-based integrated mass enhancement (IME) method. The algorithm can process 62 images (each measuring 128 pixels × 128 pixels) per second on a single 2.6 GHz Intel Core i7-9750H CPU. We train the algorithm using large-eddy simulations of methane plumes superimposed on noisy and variable methane background scenes from the GHGSat-C1 satellite instrument. We introduce the concept of point-source observability, Ops=Q/(UWΔB), as a single dimensionless number to predict plume detectability and source rate quantification error from an instrument as a function of source rate Q, wind speed U, instrument pixel size W, and instrument-dependent background noise ΔB. We show that Ops can powerfully diagnose the ability of an imaging instrument to observe point sources of a certain magnitude under given conditions. U-Plume successfully detects and masks plumes from sources as small as 100 kg h−1 in GHGSat-C1 images over surfaces with low background noise and successfully handles larger point sources over surfaces with substantial background noise. We find that the IME method for source quantification is unbiased over the full range of source rates, while the CNN method is biased towards the mean of its training range. The total error in source rate quantification is dominated by wind speed at low wind speeds and by the masking algorithm at high wind speeds. A wind speed of 2–4 m s−1 is optimal for detection and quantification of point sources from satellite data.

Large-eddy simulation investigation on the effects of inlet swirl/Reynolds number and fuel heating value on a turbulent kerosene spray flame

To progress toward the high efficiency and low emission aviation engine technology, it is mandatory to understand the complex phenomena including kerosene spray atomization, evaporation, turbulent mixing and combustion in the engine combustor. Nowadays with the rapid growth of computational capacity over the world, large-eddy simulation (LES) performs a more and more important role in understanding the complex multi-phase combustion process of kerosene fuels in aviation engine configuration. In the present work, LES is employed to investigate a swirling kerosene spray jet flame in a model combustor, which has been studied experimentally. The turbulent combustion is modeled by the flamelet generated manifold (FGM) approach with a 3-component surrogate kerosene skeletal mechanism. The averaged gas velocity and temperature predicted by LES generally agrees well with the experimental measurements. Local extinction is observed at the inner shear layer of the swirling spray flame, due to droplet-turbulence-flame interactions. A parametric study with 11 cases is then performed to investigate the effects of inlet swirl/Reynolds number and fuel heating value on the kerosene flame. As the swirl number increases, the enhanced turbulent mixing between kerosene fuel and air facilitates the combustion process, but combustion instabilities may occur if the swirl number is too high. A higher Reynolds number can promote the turbulent combustion via a stronger recirculation zone, but it also has strong dilution effects on the combustion field. Variations in the fuel heating value directly impact the temperature field, but the flow field, major species concentrations and droplet phase statistics are only slightly affected. These LES results give clues to achieve optimal combustion performance in realistic aviation engine configurations via better arrangements of fuel and air injections.

Large-eddy simulation for indoor flow in a multizone environment

Large-eddy simulation for indoor flow in a multizone environment

Direct-Noise of an Ultrahigh-Bypass-Ratio Turbofan: Periodic-Sector vs Full-Annulas Large-Eddy Simulations

The tonal and broadband noise of an ultrahigh-bypass-ratio fan stage, which has been developed at École Centrale de Lyon, is studied using large-eddy simulations (LES). Wall-modeled LES of a periodic sector and a 360-deg full fan stage have been performed at approach conditions, which is a relevant operating point for aircraft noise certifications. The fan noise is directly obtained from the fully compressible LES, using a well-refined unstructured mesh, and compared with state-of-the-art analytical models. The impact of the periodic boundary conditions, which are often used for high-fidelity simulations of turbofan engines, is assessed. The results from the 360-deg and periodic-sector LES are compared from aerodynamic and acoustic perspectives, including an analysis of the mean and turbulent flow quantities and sound-pressure spectra. Aerodynamic parameters show similar results for both configurations. However, the fan blade loading is slightly reduced in the 360-deg- LES near the blade tip. Acoustically, lower sound power levels at the intake and exhaust sections of the fan stage are obtained in the 360-deg LES, when compared to the periodic sector LES, particularly at low and middle frequencies. This can be associated with lower coherence levels in the fan wakes and smaller spanwise correlation lengths at the trailing edge of the blades. The modal content of the acoustic field has also been analyzed in detail and shows that the periodic sector LES cannot correctly simulate the modal content of the fan noise.

Characteristics of very-large-scale motions in natural rivers based on large eddy simulation

Exploring very-large-scale motions (VLSMs) in open channel flows (OCFs) is crucial for comprehensively understanding material transport and energy exchange. While previous studies focused on OCFs in ideal flumes with simple boundaries, this paper presents large eddy simulation (LES) results on the existence and scale characteristics of VLSMs in complex river morphologies with the Minjiang River in southwestern China as a case study. This research demonstrates that the OpenFOAM-based LES model is capable of accurately reconstructing the time-averaged flow field and providing instantaneous velocity data that capture VLSMs with sufficient resolution. Spectrum analysis of the streamwise fluctuating velocity shows that VLSMs are present in the straight segment of the river, with streamwise wavelengths approximately (16–22) times the water depth but are absent in the bend and confluence segments due to the suppression by secondary currents. This article validates the effectiveness of LES in examining the characteristics of VLSMs in natural rivers, thereby laying a good foundation for further studies on the impacts of such structures on sediment transport and pollutant dispersion.

Corner Vortex Structures Within Recirculation Zones of Confined Bluff-Body Flows

The vortical structures of recirculation zones in turbulent nonreacting and premixed reacting flows around confined equilateral-triangle bluff bodies are investigated using large-eddy simulations. A three-dimensional flow structure is observed within the recirculation zones of reacting and nonreacting cases; the structure includes both the “spanwise” vortices that extend the length of the recirculation zone and “corner vortex structures” that are situated adjacent to the bluff-body trailing edge and spanwise walls. The corner vortex structures enhance the mixing and residence times of fluid inside the recirculation zone. Fluid is circulated between the spanwise vortices and corner vortex structures. Corner vortex structures always appear downstream (relative to the local flow) of the boundary-layer separation location on the spanwise walls within the recirculation zone, and they are absent when boundary-layer formation on the spanwise walls is prevented. Laminar boundary-layer theory predicts the boundary-layer separation locations within ∼12% of the calculated values from the large-eddy simulations in all but one case. These results suggest that 1) boundary-layer separation on the spanwise walls is a necessary condition for the formation of corner vortex structures, and 2) boundary-layer separation on the spanwise walls is caused by an adverse pressure gradient in the recirculation zone.