Reynolds-averaged Navier-Stokes Model Research Articles

Comparison of RANS and LES for a cross-flow turbine in confined and unconfined flow

This work examines the dynamic stall process and resulting wake features of cross-flow turbines under confined configurations using two computational modeling approaches, Reynolds-averaged Navier-Stokes (RANS) and large-eddy simulation (LES). Cross-flow turbines harvest energy from wind or water currents via rotation about an axis perpendicular to the flow and are a complementary technology to the more common axial-flow turbine. During their 360° rotation cross-flow turbine blades experience a cyclical variation in the angle of attack and velocity relative to the oncoming flow, leading to flow separation and reattachment, otherwise known as dynamic stall. The dynamic stall process causes an instantaneous loss in torque generation and unsteady force fluctuations which pose a challenge to accurate predictions of both the performance and the resulting unsteady flow field. This research compares RANS simulations to higher fidelity LES of a straight-bladed two-blade cross-flow turbine at a moderate Reynolds number (Rec = 45,000) in a confined configuration. The RANS model is shown to be very sensitive to confinement at the simulated tip speed ratio as it over-predicts power generation due to suppression of flow separation, while the flow field from LES matches well with the experimental validation. Results are compared with an unconfined configuration for which the RANS model successfully predicts a power curve; however, it displays significant differences in the evolution of flow structures such as premature shedding of the dynamic stall vortex and a lack of vortex diffusion during convection in the wake.

Validation and assessment of different RANS turbulence models for simulating turbulent flow through an orifice plate

In the present study, numerical simulations using different Reynolds-Averaged Navier–Stokes (RANS) turbulence models are carried out to investigate the turbulent flow through the orifice plate at Reynolds number (Re) of 23000. The orifice thickness to pipe diameter ratio (t) and the orifice diameter to pipe diameter ratio (β) are fixed and equal to 0.1 and 0.5, respectively. The objective is to evaluate the behaviour of various RANS models with respect to the relevant flow parameters such as the pressure drop, velocity distributions and turbulence intensity profiles in the pipe by comparing the results with available published experimental data. The following turbulence models are studied: the k – ε, the k – ε Low Re, the k – ε RNG, the k – ε Realizable, the k – ω SST, the γ – SST, the EARSM and the k – ε Cubic models. It is found that based on the validation study of the flow through the orifice plate, the following models are in good agreement with experimental measurements: the k – ω SST, the γ – SST and the EARSM. They show a better performance than the k – ε model family in predicting the flow features which are important for the orifice flowmeter design.

Investigation of vortical and near-acoustic fields in three-stream jets.

The connections between the vortical and near-acoustic fields of three-stream, high-speed jets are investigated for the ultimate purpose of developing linear surface-based models for the noise source. Those models would be informed by low-cost, Reynolds-averaged Navier-Stokes (RANS) computations of the flow field. The study uses two triple-stream jets, one is coaxial and the other has eccentric tertiary flow that yields noise suppression in preferred directions. Large eddy simulations (LES) validate the RANS-based models for the convective velocity Uc of the noise-generating turbulent eddies. In addition, the LES results help define a "radiator surface" on which the jet noise source model would be prescribed. The radiator surface is located near the boundary between the rotational and irrotational fields and defined as the surface on which the Uc distribution, obtained from the space-time correlations of the pressure, matches that inferred from the RANS model. The edge of the mean vorticity field is nearly coincident with the radiator surface, which suggests a RANS-based criterion for locating this surface. The two-dimensional space-time correlations show how the asymmetry of the tertiary stream and the resulting thicker low-speed flow weakens the generation of acoustic disturbances from the vortical field.

Data-driven turbulence modelling using symbolic regression

The study is focused on the performance of machine-learning methods applied to improve the velocity field predictions in canonical turbulent flows by the Reynolds-averaged Navier–Stokes (RANS) equation models. A key issue here is to approximate the unknown term of the Reynolds stress (RS) tensor needed to close the RANS equations. A turbulent channel flow with the curved backward-facing step on the bottom has the high-fidelity LES data set. It is chosen as the test case to examine possibilities of GEP (gene expression programming) of formulating the enhanced RANS approximations. Such a symbolic regression technique allows us to get the new explicit expressions for the RS anisotropy tensor. Results obtained by the new model produced using GEP are compared with those from the LES data (serving as the target benchmark solution during the machine-learning algorithm training) and from the conventional RANS model with the linear gradient Boussinesq hypothesis for the Reynolds stress tensor.

On integrated experimentation and simulation of a counter-turning swirling mixer, part I: Experimentation

Success in developing advanced combustor concepts relies significantly on achieving rapid and uniform mixing of fuel with combustion air through a fuel/air preparation device called mixer. To guide the design and development of gas turbine combustor expeditiously, detailed understanding of the flow characteristics and performance of a mixer of interest is essential. Leveraging recent advances in computational fluid dynamics (CFD) and diagnostic techniques, a series of studies have been initiated to develop and calibrate best practices on integrated experimentation and simulation for two generic families of mixers, namely for lean-direct injection and rich-dome combustors. Here, a set of three papers have been prepared to summarize our research activities on a mixer relevant to rich-dome combustors. This Part I paper presents the method by which a high-quality Time-Resolved Particle Image Velocimetry (TR-PIV) two-dimensional velocity dataset is established for a counter-turning multi-swirlers (CTS) mixer. The non-reacting flow field characteristics at 3% pressure drop across this CTS mixer are quantified by collecting more than 10,000 image pairs captured and processed via a high-accuracy sub-pixel adaptive correlation algorithm. Mean and root-mean-square velocity profiles are presented along with quantification of critical flow structure metrics. Proper Orthogonal Deconvolution structure analysis is also presented for the dominant modes. A full description of the uncertainty analysis of the TR-PIV velocity data is further provided, as is a detailed convergence study. The dataset is determined to be of sufficient quality to be used for quantitatively anchoring CFD simulations as summarized in our two companion papers Part II (Reynolds-Averaged Navier-Stokes models) and Part III (Large Eddy Simulation).

Comparative study between turbulence models in unsteady cavitating flow with special emphasis on shock wave propagation

The generation and propagation of shock waves are important sources of cavitation instability and material damage. This paper describes an evaluation of the predictive ability of different turbulence models in a compressible cavitation flow for shock wave propagation. The Schnerr–Sauer cavitation model is used to model cavitation and the volume of fluid method is used to capture the water/vapor interface. The pressure pulsations and cavity evolution given by numerical simulations and experimental results are compared to evaluate the correctness of the numerical method and the prediction accuracy of different turbulence models. The propagation mechanism of the shock wave, the evolution of vortex structures, and the temporal and spatial distribution characteristics of the cavitating flow are analyzed. The results show that the results predicted by the large-eddy simulation (LES) model are in good agreement with the experimental results. The shock wave is caused by a pressure wave that is generated by the collapse of the previously detached cloud cavity. This pressure wave hits the trailing edge of the cavity, thus inhibiting the development of the cavity and delaying the period of the cavity. In the process of shock wave propagation, the peak pressure and the peak lift (drag) coefficient appear at the same time. The propagation of the shock wave strongly disturbs the motion of the vortices, causing the large-scale vortex structure on the trailing edge of the hydrofoil to be depressed and the stable vortex on the hydrofoil to lift. The vortex stretching term and vortex dilatation term dominate the transmission process of the eddy current, and the vortex dilatation term is the most important in the process of shock wave propagation. The pressure fluctuations in the evolution of the cavity lead to changes in the statistical structure of the velocity field. Moreover, the anisotropy of velocity fluctuations in the cloud-cavity region is higher than that in the region of the sheet cavity. The Reynolds-averaged Navier–Stokes (RANS) model underestimates the cavity volume and the pulsation characteristics of the flow field, and cannot capture the negative X-direction velocity and vortex structure changes in the shock wave propagation process. Moreover, the RANS model and detached-eddy simulation (DES) model cannot accurately predict the characteristics of the flow field near the wall, and overestimate the dominant frequency of the cavity oscillations. The LES model is better at capturing the flow field characteristics in unsteady dynamics, and reproduces the shock wave propagation process well.

CFD Evaluation of Pressure Change Along Coolant Passages in Sodium-Cooled Fast Reactor with Nek5000

To support the design efforts of advanced sodium-cooled fast reactors (SFRs), a series of computational fluid dynamics (CFD) simulations are performed to investigate the pressure change along various flow passages in the proposed SFR system. The simulations are carried out with the state-of-the-art spectral element flow solver, Nek5000. Two specific case studies are presented in this paper: the flow exiting the axial neutron reflector channels and the flow entering the fuel pin bundle. Due to the high Reynolds numbers expected, a Reynolds-averaged Navier-Stokes (RANS) approach is necessary to model the turbulence. A newly developed regularized RANS model is adopted in the related CFD calculations. The first case study explores the effect of Reynolds number on the pressure change when flow exits the reflector channels. The pressure change in this case has two major contributors: the change due to wall friction and the Bernoulli effect. It is noted that the nondimensional pressure loss follows a log-linear trend up to Re = 105, and then the trend is flattened. In the second case study, the advanced NekNek coupling capability is tested where an integral domain can be divided into multiple subdomains with coupling interfaces, which would greatly ease the meshing process of complex engineering geometries and potentially save computational resources. The preliminary results obtained so far confirm the consistency between the NekNek results and those produced by regular Nek5000 simulation. The presented work demonstrates the readiness and flexibility of the related CFD techniques, which is part of the broader effort to leverage cutting-edge CFD to inform the advanced nuclear reactor designs.

Mesh sensitivity of RANS simulations on film cooling flow

Accurate prediction of film cooling flow is a necessity in the design of high-pressure turbine components given the temperature of hot gas exceeds the melting point of the component material; temperatures that are necessary to achieve high thermal efficiency. Nowadays, Computational Fluid Dynamics (CFD) methods are becoming increasingly popular to predict film cooling flows. When compared to large eddy simulation (LES) and direct numerical simulation (DNS), the Reynolds Averaged Navier-Stokes (RANS) method is most frequently employed due to its much lower computational cost, though it struggles to accurately solve the highly three-dimensional and anisotropic cooling flows. However, few publications have focused on the mesh sensitivity problem of RANS simulations. And yet mesh generation is fundamental to CFD simulations and directly affects the accuracy of the calculated solution. Two typical film cooling hole geometries, namely cylindrical and fan-shaped, are utilised to investigate the mesh sensitivity of RANS simulations with low, medium, and high blowing ratios. Seven mesh sizes ranging from super coarse to super fine (4000 times larger) are employed for analysis. In general, the comparisons show that the computational results are close to the experimentally observed film features and averaged effectiveness, while also demonstrating that mesh convergence becomes challenging for RANS modelling, especially for the fan-shaped hole.

A curated dataset for data-driven turbulence modelling

The recent surge in machine learning augmented turbulence modelling is a promising approach for addressing the limitations of Reynolds-averaged Navier-Stokes (RANS) models. This work presents the development of the first open-source dataset, curated and structured for immediate use in machine learning augmented corrective turbulence closure modelling. The dataset features a variety of RANS simulations with matching direct numerical simulation (DNS) and large-eddy simulation (LES) data. Four turbulence models are selected to form the initial dataset: k-ε, k-ε-ϕt-f, k-ω, and k-ω SST. The dataset consists of 29 cases per turbulence model, for several parametrically sweeping reference DNS/LES cases: periodic hills, square duct, parametric bumps, converging-diverging channel, and a curved backward-facing step. At each of the 895,640 points, various RANS features with DNS/LES labels are available. The feature set includes quantities used in current state-of-the-art models, and additional fields which enable the generation of new feature sets. The dataset reduces effort required to train, test, and benchmark new corrective RANS models. The dataset is available at https://doi.org/10.34740/kaggle/dsv/2637500.

On the generality of tensor basis neural networks for turbulent scalar flux modeling

This study focuses on Reynolds-averaged Navier Stokes (RANS) models for passive scalar transport. A recently developed machine learning turbulence closure, the tensor basis neural network for scalar flux modeling (TBNN-s), is applied on two different datasets, an inclined jet in crossflow and a wall mounted cube in crossflow. The TBNN-s consists of a deep neural network that predicts a turbulent diffusivity matrix after being trained on high fidelity datasets. The present work addresses the generality question (i.e. will a model trained in one turbulent flow perform well in a different turbulent flow?) by training and testing across different classes of flows. We show that when trained and tested in the same class of flow, the TBNN-s obtains significant improvements in the mean scalar field over a common baseline model. When trained and tested across different classes of flows, the TBNN-s has a similar performance to the baseline model, which is encouraging since training set extrapolation does not produce the nonphysical results some fear. Further analysis suggests that varying training sets for the same test case produces similar results in terms of alignment of the scalar flux vector, but different turbulent diffusivity matrix magnitudes.

S-frame discrepancy correction models for data-informed Reynolds stress closure

Despite their well-known limitations, Reynolds averaged Navier-Stokes (RANS) models remain the most commonly employed tool for modeling turbulent flows in engineering practice. RANS models are predicated on the solution of the RANS equations, but the RANS equations involve an unclosed term, the Reynolds stress tensor, which must be modeled. The Reynolds stress tensor is often modeled as an algebraic function of mean flow field variables and turbulence variables. This, however, introduces a discrepancy between the Reynolds stress tensor predicted by the Reynolds stress model and the exact Reynolds stress tensor. This discrepancy can result in inaccurate mean flow field predictions for complex flows of industrial relevance. In this paper, we introduce a data-informed approach for arriving at Reynolds stress models with improved predictive performance. Our approach relies on learning the components of the Reynolds stress discrepancy tensor associated with a given Reynolds stress model in the mean strain-rate tensor eigenframe. These components are typically smooth and hence simple to learn using state-of-the-art machine learning strategies and regression techniques. Our approach automatically yields Reynolds stress models that are symmetric, and it yields Reynolds stress models that are both Galilean and frame invariant provided the inputs are themselves Galilean and frame invariant. To arrive at computable models of the discrepancy tensor, we employ feed-forward neural networks and an input space spanning the integrity basis of the mean strain-rate tensor, the mean rotation-rate tensor, the mean pressure gradient, and the turbulent kinetic energy gradient, and we introduce a framework for dimensional reduction of the input space to further reduce computational cost. Numerical results illustrate the effectiveness of the proposed approach for data-informed Reynolds stress closure for a suite of turbulent flow problems of increasing complexity.

Numerical prediction of cavitation erosion to investigate the effect of wake on marine propellers

This paper presents validation studies to simulate performance, cavitation and cavitation erosion properties of the King's College (KCD) KCD-193 model propeller. The simulations are performed to investigate the effect of wake on the cavitation formation and the cavitation erosion intensity. The numerical (CFD) study is achieved by using unsteady Reynolds Averaged Navier-Stokes (RANS) and Detached Eddy Simulation (DES) techniques with SST (Menter) k-ω turbulence model. The Schnerr-Sauer cavitation model and Eulerian approach with the Volume of Fluid (VOF) multiphase model are used to model a two-phase cavitation flow. The sliding mesh technique is applied to simulate the rotation of the propeller, and particular mesh refinement is performed on the regions where cavitation development is observed. Erosion intensity values on the propeller surface are obtained by Erosive Power Method (EPM) approach, and erosion formation on the blades is illustrated by utilizing the erosion intensity dataset. Numerical comparison for the erosion intensity values using RANS and DES models is carried out to evaluate the effect of wake in behind condition. Numerical results are evaluated against cavitation tunnel measurements at the Emerson Cavitation Tunnel (ECT) of Newcastle University. The main objectives of the study are to obtain the cavitation and cavitation erosion formation on the propeller in good agreement with experimental data obtained in ECT and to investigate the characteristics of the erosion intensity subject to cavitation conditions. In view of the flow conditions investigated, the RANS model is also found appropriate and required less computing efforts to predict the erosion intensity than the DES model. The results confirm the EPM method is to be compatible in predicting the erosion intensity. The results of the simulations indicate that a decrease in cavitation number does not necessarily cause an increase in the erosion intensity, requiring further investigation.

Safety of a high-speed train passing by a windbreak breach under different wind speeds

A derailment phenomenon could take place on the windward side of a 120 km/h high-speed train when it runs by a breach, between two windbreak walls, subjected to a normal wind speed of 32 m/s. To study the safety of a high-speed train under different normal speeds of crosswind, six wind speeds are investigated; 32 m/s, 28 m/s, 25 m/s, 20 m/s, 15 m/s, and 10 m/s. The wind forces and moments of the moving train are calculated using the Unsteady Reynolds-averaged Navier-Stokes (URANS) model, which are then applied to the train multi-body dynamics. The pressure fields around the train passing by the breach are analysed, which gives a reasonable explanation for the fluctuation of the wind loads. After an analysis on the response of the train, it is apparent that the risk of derailment on the windward side is much greater than the risk of overturning. The lateral distance of the first wheelset increases towards the windward side as along with the yaw angle of the wheelset, which increases as well with wind speeds of higher than 20 m/s.

Parameter Estimation for Reynolds-Averaged Navier–Stokes Models Using Approximate Bayesian Computation

Approximate Bayesian computation (ABC) is used to estimate unknown parameter values, as well as their uncertainties, in Reynolds-averaged Navier–Stokes (RANS) simulations of turbulent flows. The ABC method approximates posterior distributions of model parameters, but does not require the direct computation, or estimation, of a likelihood function. This method thus enables relatively simple and flexible parameter estimation for complex models and a wide range of reference data. This paper describes the ABC approach, including the use of a calibration step, adaptive proposal, and Markov chain Monte Carlo (MCMC) technique to accelerate the parameter estimation, resulting in an ABC approach with improved MCMC, denoted ABC-IMCMC. As a test of the classic ABC rejection algorithm and ABC-IMCMC, parameters are estimated in a nonequilibrium RANS model using reference data from direct numerical simulations of periodically sheared homogeneous turbulence. The use of ABC-IMCMC is then demonstrated to estimate parameters in the Menter shear-stress-transport (SST) model using experimental reference data for an axisymmetric transonic bump. This paper shows that the accuracy of the SST model for this test case can be improved using ABC-IMCMC, indicating that ABC-IMCMC is a promising method for the calibration of RANS models using a wide range of reference data.

An Analytic Study of the Effect of a Vane on the Hydraulic Field around a Cylinder

The flow pattern around the cylinder body is a very serious problem and this problem may become more serious and sensitive, when we place a vane neighboring to the cylinder. The present paper deals with the vane impact on the flow pattern around the cylinder. To investigate this problem the ANSYS fluent software is employed in order to achieve the two dimensional numerical analysis. Here, Reynolds Average Navier Stokes model is adopted. The investigation comprises the following hydraulic variables, like eddy viscosity, turbulent intensity, turbulent kinetic energy, turbulent dissipation rate, flow velocity profile, static pressure and pressure coefficient. The constant flow velocity, cylinder diameter and vane dimension are adopted in this analysis, while the different certain distance between the vane and the cylinder is considered. The used vane has a rectangular shape. In this analysis, it is clear that the vane plays a sensitive vital role in the hydraulic behavior of the flow pattern around the cylinder. The study has taken up four distances between the vane and the cylinder, these distances is a function of the cylinder diameter, in addition to the direct touch that happens between the vane and the cylinder. The analysis also shows that when the cylinder has direct touch with the vane, the dramatic reduction will occur in hydraulic variables.

O-008. Influence of Hypertension in Pregnancy on Hematological Phenotypes of Preterm Infants with Gestational Age Inferior to 32 Weeks

This paper presents validation studies to simulate performance, cavitation and cavitation erosion properties of the King's College (KCD) KCD-193 model propeller. The simulations are performed to investigate the effect of wake on the cavitation formation and the cavitation erosion intensity. The numerical (CFD) study is achieved by using unsteady Reynolds Averaged Navier-Stokes (RANS) and Detached Eddy Simulation (DES) techniques with SST (Menter) k-ω turbulence model. The Schnerr-Sauer cavitation model and Eulerian approach with the Volume of Fluid (VOF) multiphase model are used to model a two-phase cavitation flow. The sliding mesh technique is applied to simulate the rotation of the propeller, and particular mesh refinement is performed on the regions where cavitation development is observed. Erosion intensity values on the propeller surface are obtained by Erosive Power Method (EPM) approach, and erosion formation on the blades is illustrated by utilizing the erosion intensity dataset. Numerical comparison for the erosion intensity values using RANS and DES models is carried out to evaluate the effect of wake in behind condition. Numerical results are evaluated against cavitation tunnel measurements at the Emerson Cavitation Tunnel (ECT) of Newcastle University. The main objectives of the study are to obtain the cavitation and cavitation erosion formation on the propeller in good agreement with experimental data obtained in ECT and to investigate the characteristics of the erosion intensity subject to cavitation conditions. In view of the flow conditions investigated, the RANS model is also found appropriate and required less computing efforts to predict the erosion intensity than the DES model. The results confirm the EPM method is to be compatible in predicting the erosion intensity. The results of the simulations indicate that a decrease in cavitation number does not necessarily cause an increase in the erosion intensity, requiring further investigation.

Assessment of turbulence modeling for massively-cooled turbulent boundary layer flows with transpiration cooling

We assess Reynolds-averaged Navier–Stokes (RANS) turbulent closures for the prediction of a turbulent boundary layer with transpiration cooling via comparison with a high-fidelity direct numerical simulation database. This study considers the canonical zero-pressure gradient, flat-plate, turbulent boundary layer over a massively cooled wall, with transpiration cooling. The simulations are conducted at a low-subsonic Mach number and we study two transpiration cooling configurations with uniform and slit injection at various blowing ratios. The DNS and RANS simulation setups are nearly identical. The RANS-based turbulence models perform well in the qualitative estimation of the velocity and thermal boundary layer evolution at low-blowing ratios (F = 0.2 and 0.6%); more significant differences are noted at higher blowing ratios (F=2.0%). The RANS models, especially the Spalart–Allmaras model, over-predict turbulence production near the wall which results in faster growth in the boundary thickness; this error becomes more pronounced at higher blowing ratios. Despite the greater mixing of momentum, the thermal mixing is under-predicted compared to the DNS in the uniform blowing case but over-predicted for the slit case. These results suggest that modeling errors in the temperature distribution due to turbulent thermal flux modeling can be significant even if the velocity is correctly modeled.

Use and Scalability of OpenFOAM for Wind Fields and Pollution Dispersion with Building- and Ground-Resolving Topography

Complex flow and pollutant dispersion simulations in real urban settings were investigated by using computational fluid dynamics (CFD) simulations with the SST k−ω Reynolds-averaged Navier–Stokes (RANS) equation with OpenFOAM. The model was validated with a wind-tunnel experiment using two surface-mounted cubes in tandem, and the flow features were reproduced with the correct qualitative behaviour. The real urban geometry of the Parade Square in Warsaw, Poland was represented with both laser-scanning data for the ground geometry and the CityGML standard to describe the buildings as an example. The Eulerian dispersion of a passive scalar and the flow behaviour could be resolved within minutes over a computational domain with a size of 958 × 758 m2 and a height of 300 m with over 2 M cells due to the good and strong parallel scalability in OpenFOAM. This implies that RANS modelling with parallel computing in OpenFOAM can potentially be used as a tool for situational awareness on a local urban scale; however, entire cities would be too large.

Comparison of Turbulence Modeling for a Compressor Rotor at Different Tip Clearances

Due to the high Reynolds numbers together with the complex turbulent motions in turbomachinery, selecting a proper turbulence-modeling method is of vital significance in interpreting the real flow physics. In this paper, shear stress transport (SST) as a Reynolds-averaged Navier–Stokes model and scale-adaptive simulation (SAS) and zonal large eddy simulation (ZLES) as two scale-resolving simulation approaches were chosen to simulate the flow in a low-speed axial compressor rotor at three different blade tip clearances (1.3, 2.6, and 4.3% of chord length). Results of unsteady simulations at the design point were compared to the experimental data. The results prove that the ZLES model performs best in characterizing the experimental data regardless of the blade tip clearance size. The performance of the SST model in capturing the tip flow is close to the measurement data when the blade tip clearance is small, while the deviation from the experimental results increases with the clearance size enlarged. In contrast, the use of the SAS model does not result in any benefits over the SST model for the small tip clearance but demonstrates similar performance with the ZLES model for the configurations with middle and large tip clearance sizes.

Wind-Induced Vibration of Traditional Chinese Citygate Buildings in the Ming-Qing Dynasties – A Case Study of the Nanjing Drum Tower

ABSTRACT Compared to normal traditional Chinese timber buildings, citygate buildings are relatively high, lightweight, and quite vulnerable to wind. The present work aims to carry out a wind-induced vibration study on a traditional Chinese citygate building, the Nanjing Drum Tower, with the method of computational fluid dynamics (CFD) simulation to determine the wind loads and finite element (FE) analysis to assess its structural response. The architectural and structural characteristics of traditional Chinese timber buildings in the Ming-Qing dynasties are analysed. The Nanjing Drum Tower has typical characteristics and is taken as the representative building in this study. The wind pressure coefficients of the drum tower are obtained by CFD simulation with the Reynolds-averaged Navier-Stokes (RANS) model. The structural response of the drum tower under wind loads is analysed based on FE analysis. The obtained results can provide a basis for assessing the wind-induced vibration of traditional Chinese citygate buildings.