Modal Testing and Structural Identification

Bridges have been built around the world for at least 3000 years and have made it possible for man to travel over obstacles like valleys and rivers, which has contributed to the rapidly growing infrastructure we have today. Efforts made to save materials and reduce production costs have led to lighter and slender structures, which in turn can cause dynamic problems where the bridge is sensitive to vibrations. Vibrations from traffic, wind, wave and seismic loads may cause major problems for a lightweight design as the risk of fatigue of structural elements is very large. Therefore, it is important to always study the dynamic parameters: natural frequencies, mode shapes and damping ratios using both measurement techniques and computer models. The classical measuring method is called Experimental Modal Analysis (EMA), where both the applied force on the structure and the structure's response are measured. The force can be generated by either an impact hammer or a shaker and recorded using an attached force sensor. The structure’s response is then measured with accelerometers. Since EMA is difficult to perform on larger structures and is sensitive to disturbing vibrations, it has become more common to use so called Operational Modal Analysis (OMA). The OMA technique is based on measuring only the structure’s response from random unknown vibrations. This is very advantageous when performing measurements on for example a bridge, because it can be fully operational during the measurements without disturbing the traffic. The aim of this thesis is to study and compare different types of measurement methods used to identify structures’ dynamic parameters and to compare them with numerical calculations. To facilitate the comparisons between the methods and to avoid environmental interference, measurements were carried out in a lab on a well-defined steel structure. A computer model of the structure was made by using the finite element software ABAQUS, where both the natural frequencies and the mode shapes were predicted. Measurements were performed by using equipment and software from Brüel & Kjær. Measured and calculated natural frequencies and mode shapes of the steel structure have good compliance for all of the different measurement methods. The modal damping was difficult to estimate since it varied a lot between measurements. The measurement with the impact hammer was quick and flexible to perform since the procedure does not require much equipment. The shaker measurement required more equipment but gave much more accurate and clearer results than the measurement with the impact hammer. The results obtained by the OMA were very similar to the results from the EMA with the shaker. The OMA technique has proven to work well even with poor time signals that are far from ideal white noise. The major drawback with the OMA technique is that the measurements require long time signals, but with today’s low storage costs and powerful CPUs that process large amounts of data, this is no longer an issue. The conclusion is that the OMA technique seems to be the most appropriate method to use for measurements on large structures such as towers and bridges that are difficult to excite artificially. The EMA technique could possibly be used for lighter and smaller structures such as footbridges made of wood or steel