PSI - Issue 64

Bartosz Piątek et al. / Procedia Structural Integrity 64 (2024) 1581–1588 Piątek , B., Howiacki, T., Kulpa, M., Siwowski, T./ Structural Integrity Procedia 00 (2019) 000 – 000

1586

6

Currently developed algorithms allow for calculating crack width with an accuracy of up to 0.05 mm (Howiacki T. et al., (2023)), which is sufficient from an engineering perspective. Exemplary plots of strain distribution along beams under load of 350 kN are presented in Figs 7a and 7b for in-groove sensors (EG) in the PT1 beam and embedded sensors (E) in the PT2 beam, respectively. Estimation of crack widths consists of the strain integration within the close vicinity of the crack over the entire effective length, which was taken in this case as 14 cm (Howiacki, (2022)). The widths of selected cracks are directly indicated in millimeters and are depicted in pink and blue boxes, while reference measurements are shown in white boxes. Reference readings were taken on the lateral surface of the beam using a plastic plate crackmeter with an accuracy of ±0.05 mm. Due to variations in sensor locations and the unpredictable course and width of cracks, comparisons within a single crack can only be considered approximately. However, statistical analysis of a larger number of cracks will provide a more reliable indication of the potential range of errors.

a)

b)

Fig. 7. Crack analysis in PT1 beam based on the strain profiles measured by monolithic sensor EG13 installed in the near-to-surface groove (a) and in PT2 beam based on the strain profiles measured by embedded ES13 monolithic sensor (b) under force of 350 kN.

3.4. Damage detection

The damages were simulated only in the 1.4 m long support areas of the PT beams, i.e. in the areas where no cracking occurred. In these areas, significant differences between the two beams can be noticed. Due to the loss of adhesion to the concrete of the ED2 sensor in the areas of simulated voids, there is no characteristic, monolithic increase in deformation. On the other hand, in the location of the injection with simulated corrosion (the last 1.0-1.2 m of the beam’s length ), the strain increment is accelerated compared to the undamaged beam. In turn, in the injected duct with simulated corrosion (the last 100-120 cm of the beam ’s length ), the increase in deformations is higher as compared to the beam without damage. Details of the damage analysis are presented in Fig. 8 (the y-scale was limited to 250 µ e in order to provide better visibility of strain changes in the ending zones of the beam). The analysis of local phenomena in the support areas of the beams allowed for the diagnosis of damages. There is also the possibility of analyzing the global performance of the beam involving the calculation of statistics (including average values) for selected load phases and comparing them with the results for the undamaged beam at the same loads. For each sensor pair, an increase in average strains was recorded. The smallest increases (of the order of a few percent) were observed for sensors in the compressed and tensioned zones, not located in the immediate vicinity of the simulated damage. On the other hand, the largest increments (of the order of several tens of percent) were shown by ED2 sensors inside the duct and ES2 sensors installed on the duct.