Knut Andreas Kvåle

Modal Analysis of a Floating Bridge Without Side-Mooring

The Norwegian Public Roads Administration is currently planning a ferry-free Coastal Highway Route E39. Floating bridges represent feasible options in this project with already two long span floating bridges in function, i.e. the Bergsoysund and Nordhordaland Bridges. In connection with this project, one of the main objectives is to quantify the accuracy of the numerical methods used to predict dynamic behaviour of floating bridges. An extensive monitoring system is installed to measure structural response as well as environmental actions from wind and waves on an existing floating bridge: the Bergsoysund Bridge. These measurements are used to estimate the modal system properties of the structure. The system identification is performed using a parametric time-domain Stochastic Subspace Identification method as well as the Frequency Domain Decomposition method. Challenges of system identification for highly damped structural systems, such as a floating bridge, are especially emphasized. The results are also compared with numerical predictions from a two part combined linear frequency-domain model set-up. The first part consists of a hydrodynamic model, including wave excitation as well as fluid-structure interaction, and relies on linearized potential theory. The results from this are thereafter introduced into a finite element model, for a complete structural dynamic analysis.

Covariance-Driven Stochastic Subspace Identification of an End-Supported Pontoon Bridge Under Varying Environmental Conditions

The Bergsoysund Bridge is currently being extensively monitored with accelerometers, anemometers, wave radars and GNSS sensors. By applying Covariance-driven Stochastic Subspace Identification (Cov-SSI), the modal parameters of the bridge are estimated. The results are interpreted in the context of the environment, represented by significant wave heights. The problem is characterized by the fact that modes are closely spaced in frequency and have high damping. Two weighting algorithms for the Cov-SSI are applied, to assess their performance for application on structures with these characteristics.

Monitoring Wind Velocities and Dynamic Response of the Hardanger Bridge

The Hardanger Bridge is the longest suspension Bridge in Norway and among the top 10 longest suspension bridges in the world. A comprehensive monitoring system was installed after it was completed in August 2013. The monitoring system is designed to provide data that can be used to verify the numerical methods used to predict wind induced dynamic response of slender bridges located in complex terrain. The monitoring system is outlined in this paper together with preliminary analysis of the accuracy of the model used to describe the self-excited forces acting on the bridge deck. Extensive wind tunnel testing was performed in the design of the Hardanger Bridge to achieve an excellent aerodynamic behaviour of the cross-section of the bridge deck. The experimental results of the aerodynamic derivatives that describe the self-excited forces have been combined with a finite element model of the bridge to predict the in-wind natural frequencies and damping ratios of the combined structure and flow system. The numerical predictions have been compared to results obtained from measured data using data-driven and covariance-driven stochastic subspace identification. It is concluded that the model for the self-excited forces provides in-wind frequencies and damping ratios that corresponds well to the observations from measured data.

Experiences from the Five-Year Monitoring of a Long-Span Pontoon Bridge: What Went Right, What Went Wrong and What’s Next?

The Bergsoysund Bridge is a 930-m-long end-supported pontoon bridge located in Norway, and has been the target of a 5-year-long, extensive monitoring program. Herein, we will describe the unique structural characteristics of the bridge. The monitoring system has been under continuous expansion and revision, and consists of sensors monitoring both the excitation and the response of the bridge. Quantification of the uncertainties of the modelling methodology for structures of this nature has been the main goal, for which purpose modal analysis has been an indispensable tool. Modal analysis has also been used to study the effects the environment has on the structure’s dynamic behaviour. We discuss the limitations of the results from modal analyses. Furthermore, we rise the question of how long monitoring campaigns may continue to provide useful information of this bridge and similar civil structures.

Characterization of the Wave Field Around an Existing End-Supported Pontoon Bridge from Simulated Data

The environmental excitation and the dynamic response are currently being monitored on the Bergsoysund Bridge, an existing end-supported pontoon bridge. Wave radars are monitoring the one-point sea surface elevation at six different locations. As the wave excitation is considered the main concern for vibration-based design of similar bridges, an appropriate description of the sea state characterizing the wave excitation is crucial. Furthermore, it is considered a necessity for an assessment of the quality of response predictions by comparison with measurements. In the current paper, time simulations of wave elevation are used to identify the already-known sea states. The Fourier Expansion Method (FEM) and Extended Maximum Entropy Principle (EMEP) are applied for this purpose. The results provide valuable insights about both the identification methods and the sensor layout.

Simulation and Monitoring of Floating Bridge Behaviour

Firstly, this paper presents a review of the main steps for simulating floating bridge behaviour. Both time-domain and frequency-domain approaches are presented. An exemplified model setup and simulation results from the selected case study are presented. Secondly, data obtained from extensive structural and environmental monitoring of the studied bridge are presented. Data analyses attempting to visualize the correlations between excitation sources and response quantities are discussed. Finally, an operational modal analysis is carried out, to attempt to identify the modal parameters from response measurements only.