16 S. Allahdadian et al. Threshold 1 a c d b Threshold 2 Reference Sessions Safe Zone Critical Zone Unsafe Zone System Order Ref. State Subspace Dynamic Subspace Null Space Selected System Order System Order Ref. State Subspace Dynamic Subspace Null Space Selected System Order log(x2) log(x2) Fig. 2.3 2 values and thresholds of the reference state in (a) structure with fixed supports and (b) with bearings; validation of the reference state in (c) structure with fixed supports and (d) with bearings changed and the major change happens in mass matrix of the structure, which has a direct effect on dynamic vibration of the structure and can be captured effectively with this technique. The damage in column can be captured only for the severe case, i.e. damage ratio equal to 80 %, as shown in Fig. 2.4e. The cap beam in the structure is located between the columns and connects the girders by crossing them. However, the girders are directly located on top of the columns and therefore cap beams do not have much effect in dynamic behaviour of the model. In all cases the damage in cap beams is not captured suggesting that damage in the elements not affecting the dynamic behaviour of structure may not be recognized by this monitoring technique. The SSDD technique is used to evaluate damage in bearings as shown in following figure. The resultant 2 values are compared to the reference state computed for the undamaged structure with bearings. It can be seen that in all cases the damage detection technique can successfully identify the damage in the structure (Fig. 2.5). It can be seen that in all cases except the cap beam case, the severe damage in the structure could be captured by the SSDD technique. In most of the cases the damage for intermediate damage ratio could be also identified using this technique. For the cap beam element type, as this model was investigated, this element has negligible effect on the dynamic response of the structure. For this reason detecting the damage for that types of elements should be investigated in more details. 2.5 Conclusion In this paper an analytical finite element model of a bridge structure namely, S101, located at Reibersdorf, Austria, is created and calibrated. The calibration was performed by measured data from the undamaged bridge. This model was employed as a realistic base model to simulate different damage scenarios. The damage was simulated in different element types, including girder, column, bearing, deck and cap beam. In each element type the damage was simulated separately and for different
RkJQdWJsaXNoZXIy MTMzNzEzMQ==