58 Table E.1 — Benchmark data for aerodynamic coefficients of ICE 3 endcar on flat ground with gap, measured by DB AG on a 1:7-scale model at 80 m/s in DNW wind tunnel .... 63 Table E.2
General
This clause presents various methods to assess the cross wind stability of railway vehicles
The cross wind stability of rolling stock is determined by characteristic wind speeds that indicate the maximum wind forces the rolling stock can endure without exceeding wheel unloading limits By analyzing various input parameters, including train speeds, uncompensated lateral accelerations, and wind angles of attack, we can derive a set of characteristic wind speeds known as characteristic wind curves (CWC).
The cross wind stability of a train is given by the cross wind stability of the most cross wind sensitive vehicle in the train consist
The CWC reflects the crosswind stability of a train in relation to specific wheel vertical forces; however, the characteristic wind curves do not define a threshold for overturning.
The CWC establishes the maximum natural wind speed that a train can endure without exceeding the critical limit for wheel unloading, defined by the average wheel unloading value, ∆Q, of the most critical running gear This average is calculated across the wheel sets of the bogie For assessing the cross wind stability of vehicles, a reference air density of 1.225 kg/m³ is used, while the vehicle mass is determined as the "operational mass in working order" in accordance with EN 15663.
The assessment of cross wind stability of vehicles separates into evaluations of the aerodynamic characteristics (i.e the aerodynamic coefficients) and the vehicle dynamic characteristics
Subclause 5.2 states the applicability of the various methods for the purpose of rolling stock assessment
Subclause 5.3 provides various methods for the determination of aerodynamic coefficients of passenger and freight vehicles and locomotives
Subclause 5.4 provides various methods for the determination of wheel unloading
Subclause 5.5 gives information on the required presentation form of CWC of passenger and freight vehicles and locomotives.
Applicability of cross wind methodologies for rolling stock assessment purposes
Subclauses 5.3 and 5.4 outline different approaches for evaluating aerodynamic and vehicle dynamic characteristics, categorized into simple and complex methods While the simpler methods are more straightforward to implement, they introduce an additional uncertainty factor as they are designed to be more conservative compared to their complex counterparts.
Table 2 specifies which method shall be applied for rolling stock assessment purposes depending on type of rolling stock and its maximum speed v max
When selecting a method for a problem, the required level of accuracy is crucial It is advisable to begin with the simplest suitable techniques and progress to more complex methods only if needed.
All vehicles shall be assessed using any of the methods in Table 2
For a fixed train composition, demonstrating the cross wind stability of the most sensitive vehicle is adequate However, in other scenarios, it is essential to establish the cross wind stability for each individual vehicle in the train.
Table 2 — Application of cross wind methodologies for rolling stock assessment
Passenger rolling stock and locomotives Freight wagons
Speed range v max ≤ 140 km/h 140 km/h < v max
Simplified proof of cross wind stability Yes,
Regulated by national require- ments (see Annex K for more information)
Yes, Subclause 5.3.2 + 5.4.2 or 5.3.3 + 5.4.2 or 5.3.3 + 5.4.3 or 5.3.4 + 5.4.2
5.3.2 not appli- cable for trains with active tilting mode
5.4.2 not appli- cable for articulated trains
5.4.2 not appli- cable for articulated trains
Regulated by national requirements (see Annex K for more information)
Full proof of cross wind stability No Yes,
Subclause 5.3.4 + 5.4.2 or 5.3.4 + 5.4.3 a Above 360 km/h the methods shall be adapted taking into account also compressibility effects
Annex A indicates how these approaches are combined for practical cross wind stability analysis in various European countries.
Determination of aerodynamic coefficients
General
There are several methods to assess aerodynamic loading on passenger and freight vehicles, including static wind tunnel tests, which provide proof of compliance with specifications for rolling stock evaluation Computational Fluid Dynamics (CFD) simulations are utilized for feasibility studies during the initial design phase, while predictive equations offer preliminary analysis The subsequent subclauses outline the application of these methods, and Table 2 illustrates their relevance for rolling stock assessment.
Predictive equations
The predictive equations can only be used subject to the following restrictions:
The method may not be valid for vehicles which differ much in their shape from the shape of current vehicles
The method is not valid for leading end cars with a length L greater than 28 m or less than 10 m
Provided by IHS under license with BSI - Uncontrolled Copy
The method is not valid if the car(s) leading the investigated intermediate car have an overall total length of less than 10 m
The method is valid only for standard track gauge of 1,435 m
The selected equation is conservative, ensuring it does not underestimate wind loading It predicts wind load on a vehicle by utilizing the coefficient for the rolling moment around the lee rail, which differs from the definition in EN 14067-1 that uses the track's midpoint as the moment reference center The rolling moment coefficient around the lee rail is estimated based on the vehicle's main physical dimensions through predictive equations.
The predictive equations allow the wind load to be calculated as follows:
~ β β β β (1) with β~= arctan[tan(β⋅z 5 ) ] when the parameters z 0 , z 1 , z 2 , z 3 , z 4 , z 5 , f L, andf VEH are given (see Table 3) c M x, lee is based on A 0 and d 0 , this normalization differs from EN 14067-1
Table 3 — Parameter set for the standard ground configuration (standard gauge)
Parameter Passenger vehicle (leading endcar) and locomotive Passenger vehicle (midcar) Freight vehicle z 0 0
L VEH Length of the vehicle car body without buffer and inter car gap Length over buffer f L 0
A 0 10 m 2 d 0 3 m ò Yaw angle in degrees h VEH Height of the vehicle from top of rail to roof according to EN 14067-1
Simulations by Computational Fluid Dynamics (CFD)
The primary goal of Computational Fluid Dynamics (CFD) is to identify the aerodynamic loads essential for assessing the mechanical stability of critical vehicles This analysis typically involves evaluating side and lift forces, as well as roll, pitch, and yaw moments, in accordance with EN 14067-1 standards Furthermore, CFD offers predictions of velocity and pressure fields, facilitating the consideration of various configurations.
NOTE Those undertaking CFD simulations and analysis of the results should be able to demonstrate competency through a proven record of railway aerodynamic simulations and have knowledge of vehicle aerodynamics
This method resembles the one suggested for reduced-scale wind tunnel measurements, as detailed in section 5.3.4, utilizing a stationary train and a low turbulence block profile for the onset flow.
The flow around a vehicle exhibits key characteristics such as three-dimensionality, high Reynolds number, turbulence, and variations in acceleration and deceleration It also involves complex factors like curved boundaries, flow separation, potential reattachment, recirculation, and swirling properties Accurately solving this aerodynamic challenge typically requires Direct Numerical Simulation (DNS) of the Navier-Stokes equations; however, practical limitations in computational resources often hinder this approach Alternative solutions can be obtained through turbulence modeling techniques, including Large Eddy Simulation (LES), Detached Eddy Simulation (DES), and Reynolds Averaged Navier Stokes (RANS).
The flow around a vehicle is naturally unsteady and influenced by its geometry and yaw angle Nevertheless, steady methods can be employed when the flow demonstrates sufficient steadiness A widely used criterion for this is to ensure that the residuals decrease by three to four orders of magnitude and that the relevant loads have converged with minimal solution cycling.
The primary challenge in Computational Fluid Dynamics (CFD) lies in selecting the right combination of computational mesh, method, and turbulence modeling A key benefit of CFD is its ability to reduce blockage effects by utilizing a sufficiently large domain size, ensuring that the Reynolds number and Mach number remain unaffected.
To validate the CFD approach, calculations will be performed for at least one specified benchmark vehicle, following the methodology outlined in section 5.3.4.2 for wind tunnel measurement validation The results will be compared to benchmark values obtained from wind tunnel measurements, ensuring adherence to the quality criteria established in section 5.3.4.2.
The vehicle model representation must adhere to the specifications for wind tunnel measurements, particularly regarding intercar gaps, vehicle lengths, and modeling accuracy for both the test vehicle and adjacent vehicles, as outlined in sections 5.3.4.8 and 5.3.4.9.
Contact between the wheels and rail or ground may be simplified to avoid numerical singularities in the contact point
Model dimensions can be either full scale or model scale, with the latter offering practical benefits, especially when comparing results with wind tunnel measurements.
The domain boundaries shall not interfere with the flow around the vehicle in a physically incorrect way
The requirements on blockage ratio shall be taken from 5.3.4.7
Provided by IHS under license with BSI - Uncontrolled Copy
The computational domain must extend at least 8 characteristic heights upstream and 16 characteristic heights downstream of the vehicle of interest The characteristic height is defined as the distance from the top of the train to the ground, which, in the case of modeling an embankment, includes the height of the embankment, ballast, rail, and the train itself.
The mesh must satisfy essential criteria for wall units next to no-slip walls, aligning with the chosen computational method and turbulence model For RANS simulations, the dimensionless wall distance y+ for the first cell layer is generally around 1 for low-Reynolds number turbulence models, while it typically ranges from 30 to 150 for high-Reynolds number turbulence models.
The mesh should capture regions of high pressure/velocity gradient such as: boundary layers, vortices, stagnation zones, recirculation cells, wakes, flow acceleration
For accurate yaw angle calculations, it is essential to maintain a consistent design of the mesh regarding the size and distribution of computational cells Adjustments may be necessary for the location and size of high-resolution regions Additionally, it is important to ensure an adequate number of mesh nodes between the vehicle and the boundaries.
To ensure mesh independency for the vehicle under investigation, it is essential to vary the mesh resolution in regions with high gradients, utilizing at least two levels of mesh refinement The mesh resolution should differ by a factor of 1.5 in each spatial direction where significant velocity gradients are present The results from the three mesh configurations must demonstrate strong agreement in flow topology and force coefficients Specifically, the rolling moment coefficient about the lee rail should maintain an accuracy of 3% across the three refinements for each relevant yaw angle, with the highest value obtained at each yaw angle being reported as the final result.
The computational method must effectively model viscous, turbulent, unsteady, three-dimensional, and strongly separated flows However, if the flow is clearly steady, steady methods may be applied.
Most methods rely on continuum theory, specifically the momentum equations known as the Navier-Stokes equations; however, approaches based on kinetic gas theory, such as the Lattice-Boltzmann equations, can also be utilized.
Turbulence models are essential engineering tools used to predict turbulent stresses, which arise from the averaging or filtering of non-linear convection terms in flow equations These stresses can be viewed as an additional viscosity, often significantly greater than molecular viscosity in turbulent flows Despite their importance, there is currently no universal turbulence model that applies to all scenarios.
Calculations shall involve a sensitivity test, using the most appropriate turbulence models If ambiguous results are obtained, wind tunnel measurements shall be used instead
Reduced-scale wind tunnel measurements
Wind tunnel tests are conducted to determine the aerodynamic force and moment coefficients of a vehicle, including the rolling moment coefficient related to the lee rail This coefficient represents a non-dimensional version of the wind-induced moment acting at the contact point between the wheels and the lee rail It is crucial as it primarily unloads the upwind wheels during strong winds, potentially leading to vehicle rollover in extreme conditions.
Individuals conducting wind tunnel testing and analyzing the results must demonstrate their expertise through a verified history of aerodynamic testing in wind tunnels, along with a solid understanding of vehicle aerodynamics.
The type of wind tunnel test that is described here uses static models of the train in a uniform, low turbulence onset flow
In wind tunnel testing, achieving a block profile and low turbulence intensity is essential for reliable and repeatable results, even though these conditions are rarely found in full-scale environments A simple block profile, with the exception of a thin boundary layer at the floor, promotes consistency across different operators To maintain these conditions, various boundary layer techniques such as moving belts, boundary layer suction, wall jet blowing, and splitter plates have been successfully implemented in wind tunnels.
This wind tunnel test simplifies more complex methods that utilize moving models to replicate train movement over ground with an atmospheric boundary layer Before these advanced tests can be standardized, additional research is necessary Annex G outlines the requirements for conducting wind tunnel tests with a static model in an atmospheric boundary layer.
It is permissible to use different types of wind tunnels Details of the wind tunnel used shall be described according to EN 14067-2
Provided by IHS under license with BSI - Uncontrolled Copy
Wind tunnel tests for rolling stock assessment must include results from specified benchmark vehicles It is advisable to incorporate tests of these benchmark vehicles in other wind tunnel evaluations as well When conducting benchmark tests, at least one of the following vehicle models should be utilized: ICE 3 endcar, TGV Duplex powercar, or ETR 500 powercar, as detailed in Annex C.
Wind tunnel benchmark tests must be conducted under the same conditions as those used for the vehicle being investigated, including techniques, configurations, and onset wind conditions If there are any changes to the wind tunnel setup, such as modifications to the wind tunnel, ground configuration, or boundary layer properties, the benchmark tests must be repeated to ensure accuracy.
The test results for the benchmark vehicle c Mx,lee,test will be compared to the reference values for c Mx,lee,bmk to assess the accuracy of the wind tunnel tests within the specified tolerances This comparison will also identify any systematic biases in the coefficients relative to the reference results While exact agreement is not anticipated due to variations in wind tunnels and experimental setups, the required tolerances for low turbulence tests are defined as: bmk max lee, x bmk lee, x, test lee, max x, |