PSI - Issue 64

Pavel Ryjáček et al. / Procedia Structural Integrity 64 (2024) 228 – 237 Pavel Ryjacek/ Structural Integrity Procedia 00 (2019) 000 – 000

234

7

a b Fig. 10 Second derivative of the change in the diag. members of the modal flexibility matrix [  ] for the support stiffness chang, for 86 (a) and 24 (b) sections 9. Effect of corrosion of the main supporting cables (Alfa) Models in which the area of the main load-bearing cables (Alpha) was reduced in a large part of the main span in the joints and the joints were allowed to open are also investigated and compared with the intact reference model. When comparing the model deviations with the uncertainty of the natural frequency measurements, it is clear that these defects are very difficult to detect based on the natural frequency change. Only 2 values of higher natural frequencies exceed the measurement uncertainty, but not in a detectable way. For detection by advanced methods, comparisons of the eigenmodes of the reference undamaged model with the eigenmodes from the damaged model are shown. The corrosion of the main load-bearing cables was manifested by an increase in ductility, i.e. a decrease in the stiffness of the whole structure. The second derivative method of changing the diag. terms of the modal compliance matrix [  ] localized the failure of the structure in the vicinity of sections 46 and 47 for the model with 86 transverse sections, where a slight increase in its values occurred. Using a model with 24 sections, corresponding to the test performed, it is not possible to localize the failure. For detection by advanced methods, comparisons of the eigenmodes of the reference undamaged model with the eigenmodes from the damaged model are shown. The corrosion of the main load-bearing cables was observed by an increase in flexibility of the whole structure. The second derivative method of changing the diag. members of the modal flexibility matrix [  ] localized the failure of the structure close to the sections 46 and 47 for the model with 86 sections. For the 24 sections, corresponding to the real tests, it is not possible to localize the defect, see Fig. 11.

a

b

Fig. 11 Second derivative of the change in the diag. members of the modal flexibility matrix [  ] for the main cables Alfa corrosion, for 86 (a) and 24 (b) sections In the case of the change in the curvature of the CAMOSUC(j),x, very low manifestation can be found for the 7 and 8 eigenmodes. For 24 sections, the manifestation is significantly lower. 10. Effect of absence of grouting in the main span and corrosion of prestressing cables (Beta) The natural frequencies from the models in which the mass and the stiffness of the grouting of the prestressing cables in a significant part of the main field span were removed are shown. Corrosion weakening of the Beta cables up to 80% was also simulated over the same length in the model. When comparing the resulting model deviations, it