Bridge Aerodynamics

Bridge Aerodynamics
COWI Expertise COWI Services

Aerodynamic phenomena
Cable oscillations

COWI is an international design consultant with a market leading position in bridge, tunnel and marine engineering. COWI possesses a wide range of expertise within core disciplines of bridge aerodynamics: Establishment of design basis Planning, design and interpretation of wind and turbulence site measurements Wind climate measurements Aeroelastic analysis and computer simulation of wind effects Design, supervision and interpretation of wind tunnel tests Structural modelling of static and dynamic wind loads Optimisation of bridge design with respect to wind Design of countermeasures such as tuned mass dampers Troubleshooting Stay cable vibration assessment and damping. COWIs ISO 9001 certification covers bridge aerodynamics.

COWI has been working with aerodynamics of structures for more than 40 years. We have gained considerable experience in design and prediction methods, balancing theory, experiments and computer simulations to obtain the best results with regard to structural safety and human comfort. COWI works with a large number of internationally established wind tunnel laboratories worldwide on testing of bridges and bridge members. We also operate relevant computer codes developed in-house and thoroughly calibrated against wind tunnel test and field data. Examples of COWIs services within the field of bridge aerodynamics are presented in the following pages. COWIs services are relevant for completed bridges and bridges under construction.

Cable oscillation amplitudes should be kept at a minimum to avoid fatigue problems and problems related to user comfort.

Rain/wind testing of stay cables

-25 -20 -15 -10 -5 0 5 10 15 20 25

10

20

30

40

50

60

70

80

90

100

Flutter instability

Pitch(deg)

Aeroelastic instability (divergent motion of the deck) must be confirmed not to occur at wind speeds foreseen within the design life of the bridge.
Divergent motion (twist) of bridge girder

F low

Vortex excitation

Vortex shedding excitation of the girder or the pylons is important for human comfort and fatigue life and can furthermore induce detrimental large-amplitude cable oscillations due to internal resonance.

F low

Pressure distribution at vortex shedding frequency, w/o and with guide vanes

Vertical vortex excitation response. Note car partly hidden by undulating roadway

Stonecutters Bridge in Hong Kong with a main span of 1018 m is a bridge where wind effects has been very important for the design

Buffeting

Wind turbulence gives rise to a dynamic wind load - the buffeting wind load - that forms a significant part of the design wind loading on the bridge.
Ratio of Propeller Wind Speeds versus Mean Wind Speed

Spectrum

Frequency Response Intensity

2

1.5 1.2 1 0.9 0.8 0.7 0.6 0.5

Wind speed

10

15

20

Wind speed

Measurement mast, power spectrum, turbulence intensities, buffeting response

Computer modelling and analysis

Static wind load

Static wind load coefficients for girders and pylons are important input parameters to the structural design of a bridge. They can be computed very quickly based only on the two-dimensional cross section geometry (including rails, barriers etc.) using COWIs in-house computer code DVMFLOW. This is a very useful tool in the early design phase before wind tunnel tests are performed, for smaller bridges where tests are not carried out, and when evaluating wind tunnel test results. In addition, DVMFLOW simulations allows for quick sensitivity analyses when design changes are considered. DVMFLOW is developed specifically for computation of flow around two-dimensional bluff cross sections and is based on the discrete vortex method. The grid-free nature of the computational scheme allows fast and easy computation of flows around stationary or moving bodies.

Example of DVMFLOW geometric model representation, with crash barriers

1
0.75 0.5 0.25

CD CM

0
-0.25 -0.5 -0.75 -1 -1.25 -1.5 -15

Static wind load coefficients CD, CL and CM as functions of wind incidence angle. DVMFLOW results (points) compared to wind tunnel test results (lines)

CL

-10

-5

10

15

Stability analysis

Simulated flow and time trace of aeroelastic response of the 1st Tacoma Narrows Bridge during the catastrophic event (19 m/s)

Assessment of flutter stability of a bridge girder and the associated critical flutter wind speed can easily be carried out once the motion dependent aerodynamic coefficients for the girder cross section are known. Again, DVMFLOW offers a fast and efficient way of obtaining these coefficients at an early stage in the design process. The girder cross section is subjected to a number of forced vertical bending and twisting motions from which the aerodynamic coefficients are derived. Apart from the cross-section geometry, the input to these simulations consist of modal mass and stiffness of the bridge girder and the centre of rotation. For the subsequent analysis, also the eigenmodes and eigenfrequencies must be known. These can be obtained from a structural dynamic analysis using COWIs in-house bridge modelling software IBDAS, or similar FE models.

Moment Angular response

time

Vortex shedding excitation

The vortex shedding performance of a bridge girder or a pylon can also be assessed computationally using DVMFLOW. By FFT analysis of the static lift coefficient time trace, the dominant vortex shedding frequency can be found. A simulation with the cross section elastically suspended in the wind flow is then carried out assuming lock-in between the vortex shedding frequency and the structural frequency. The output is the response time trace from which the peak response is determined.

Modeshape, simulated flow and vertical response time trace at lock-in

100 50 0 -50 -100 time

Response amplitude (mm)

Buffeting

IBDAS analysis of buffeting

The static and dynamic wind load on a bridge is calculated using COWIs in-house integrated bridge analysis and design software IBDAS. Turbulence intensities, the spectral distribution of the turbulence, the coherence of turbulence along the bridge structure and the mean wind speed profile are all used in the dynamic buffeting calculations. The aerodynamic admittance of the deck cross section can also be included in case it is known from wind tunnel tests. The basic output from a wind load simulation are deflections and sectional forces.

CFD

For detailed analysis of the flow and pressure fields around bridge members, computational Fluid Dynamics can be applied. 3D geometries can be studied and turbulence effects included. COWI uses Star CCM+ for this purpose.
Simulated pressure field and stream lines on girder surface

Wind tunnel testing

Wind tunnel testing forms an integral part of the design and analysis of most long span bridges, and is often a requirement in many codes and national standards. In order to obtain the best possible results based on available resources, it is necessary to carefully plan and execute the tests and be aware of the inherent shortcomings and pitfalls of physical modelling. COWI can offer a full range of services within the field of wind tunnel testing from planning over participation and supervision to interpretation of results in relation to design.

Stonecutters Bridge girder. At NRC Canada

Tower model, Stonecutters Bridge, Hong Kong. At VELUX, Denmark

Section model, Osteroy Bridge, Norway. At VELUX, Denmark

Full bridge model (cantilevered), Stonecutters Bridge, Hong Kong. At FORCE, Denmark

Section model, Xihoumen Bridge, PR China. At BMT, UK

Full bridge model, Chacao Bridge, Chile. At FORCE, Denmark

Section model construction, Sutong Bridge, China. At FORCE, Denmark

Shenzen Western Corridor Bridge. Full bridge model, at BLWTL, Canada

Design basis
International Codes of Practice often gives an insufficient guide to design loads for bridges at specific sites. For long span bridges it is common practice to establish a dedicated Design Basis reflecting the specific environmental conditions at the bridge site. This procedure is based on available data from the area and new measurements designed specifically for the purpose combined with theoretical knowledge and past experience. Single point site measurements are extreapolated to the bridge line by means at terrain model wind tunnel tests.

Turbulence intensity
0.25 0.2 Iu (N) 0.15 0.1 0.05 0 -1000 Iw (N) Iu (land) Iw (land) Iu (NE) Iw (NE)

Terrain model and selected measurements

Iu,Iw

-800

-600

-400

-200

200

400

600

800

1000

distance along span

Assessment of site specific wind climate and synthesis for designers application

Design optimisation and trouble-shooting

An important element of bridge aerodynamics is optimisation of bridge design with respect to wind effects. The optimisation is carried out in close collaboration with the bridge designers and structural analysts in order to achieve the best over-all bridge design.

Three designs proposed to achieve better flutter stability of a plate girder. Question: Which design will achieve the best aerodynamic performance taking into account stability, wind loading and vortex shedding? Vortex shedding, Osteroy Norway. Short sectional guide vanes were mounted to suppress vortex shedding oscillations (see inlaid photo). These were found to work better than continuous guide vanes

Vertical vortex shedding oscillations of the girder of Storeblt Bridge were mitigated by mounting of guide vanes at locations for flow separation. The efficiency of guide vanes were confirmed through wind tunnel testing, and proven through operational experience.
Displacement y (m)
0.6 0.4 0.2 0 -0.2 -0.4 -0.6 33 33.1 33.2 33.3 33.4 33.5 33.6 33.7 33.8 33.9 34 Data recorded : 04-May-1998 Sec. 130/0 S

Guide vanes mounted at locations for flow separation

RMS displacement / deck height
0.08 0.06 0.04 0.02 0.00 0 0.5 1 1:60, Re = 1.5 .105 No guide vanes Guide vanes

130

U/fB

1.5

2.5

Measured vertical displacement, 3rd vertical mode

The guide vanes eliminates the vortex shedding response

10

Damping of cable structures

Cable systems see a wide application in civil engineering structures such as bridges, guyed masts, suspended roofs and power transmission lines. Owing to their long spans and extremely low damping, cable systems are easily set into vibration by the wind often in combination with rain or ice/snow. Stranded cables used in transmission lines and for bridge parapets may encounter excitation by the wind because of cross wind aerodynamic forces created by the stranded surface texture. As it is impractical or virtually impossible to eliminate the excitation caused by the wind, cable vibrations are most readily mitigated by introducing some form of mechanical damping to the cable system. Based on 20 years of experience, COWI offers a full range of services including diagnostics of cable vibrations, design and analysis of damping systems and acceptance tests.

resund Bridge: Stay mounted TMD for mitigation of rain/ice vibrations

Modal damping, log. dec ., Stroke (m)

0.08

A
0.06

0.04

0.02

Storeblt Bridge: Stockbridge dampers for mitigation of hand rope vibrations

2 .10

5 5 5 4 .10 6 .10 8 .10 Damping coefficient C [Ns/m]

1 .10

Measurement of damper characteristics for cable damping diagramme

COWI in bridge aerodynamics science

Tacoma

11

For nearly six decades the underlying aerodynamic mechanisms for the Tacoma collapse were not understood by the bridge engineering community. In year 2000 COWI engineers unveiled the migration of a large vortex structure across the deck as the source of the instability. This instability mechanism was proven in water tunnel experiments as well as in numerical simulations. Today the migrating Tacoma vortex is recognised as being the source of torsion flutter well known in plate girder bridges. The Tacoma vortex was acknowledged in ASCEs 2001 publication (In the Wake of Tacoma).

Tacoma vortex. Water tunnel test (left) and DVMFLOW simulation (right)

Aerodynamic stability often determines the maximum achievable span length for cable supported bridges and suspension bridges in particular.

3500 m spans for the proposed Gibraltar Fixed Link

Gibraltar

Contemporary designs for box girder and truss suspension bridges may be built to span lengths of 1500 m 2000 m without encountering aerodynamic instability at typical design wind speeds, but in case longer spans are called for, special deck and cable designs are needed to tackle the stability problem. A design study for a Gibraltar Strait crossing called for suspended span lengths in the range 3500 m - 5000 m. For these structures a COWI research project developed a twin deck structure which were able to fulfill the requirement to aerodynamic stability while keeping aerodynamic induced deck twist to a minimum.

Hga Kusten Bridge, Sweden

www.cowi.com

021-1700-016e-09a

Osteroy Bridge, Norway

Printed in Denmark by Kailow

COWI is a leading northern European consulting group. We provide state-of-theart services within the fields of engineering, environmental science and economics with due consideration for the environment and society. COWI is a leader within its fields because COWIs 4000 employees are leaders within theirs.

COWI Group Head office COWI A/S Parallelvej 2 DK-2800 Kongens Lyngby Denmark Tel.: +45 45 97 22 11 Fax: +45 45 97 22 12 E-mail: cowi@cowi.com Internet: www.cowi.com Contact: Allan Larsen Senior Specialist, Aerodynamics and Structural Dynamics Major Bridges aln@cowi.com Sanne Poulin Senior Engineer, Aerodynamics Major Bridges sami@cowi.com