ارائه روش جدید سلامت‌سنجی پایه‌های بتنی پل‌ها با استفاده از همبستگی بین داده سنسورها

Piers are vital design elements for a bridge under seismic loading; ensuring their stability and health is crucial for the overall safety of the bridge. The most common methods available for detecting damage, primarily used in the bridge deck, are modal methods. These methods have errors in detecting damage in the structure for various reasons. On the other hand, these methods require healthy pier information. In addition to modal methods, other methods based on energy and data analysis using wavelet transform also exist, the shortcomings of which are mentioned in this article. The fundamental problem that most methods face is the existence of healthy pier information, and furthermore, parameters that should be used to detect damage must be computable or measurable. In this study, a new method is presented, using the concept of correlation, to detect the presence and location of damage with minimal error and without having healthy pier information. To this end, the tallest pier of Ramp A of the Shahid Bakri Bridge complex in Tehran was selected as the case study. Before modeling, the accuracy of the modeling method was validated, and then a precise nonlinear model was built in the OpenSees software using as-built sheets. For the first time in this article, damage in the model was created in different scenarios: reduction of stiffness in concrete cover materials, reduction of cross-sectional area of rebars at various points of the section. The location of the damage was also separately applied at three heights of the pier: 15, 25, and 35 percent of the height. To generate data on the pier, it was considered to load it with an impact load on the top of the pier. This load was applied as an impact on the deck of the pier for a very short time, and the data was collected using accelerometers at the pier height, and the correlation between each pair of consecutive sensors was calculated. Due to the high velocity of the compressive wave in concrete and the overlap that occurs in the return of the wave at both ends of the pier, data collection was performed only for about 0.001 seconds. Finally, due to the weakness of the correlation coefficient in magnifying the location of damage, using the concept of correlation, a damage index based on this concept was presented, and the capability of the presented index in detecting the presence and location of damage in various damage scenarios was evaluated. The results obtained indicate the proper performance of this index in detecting damage in various scenarios and damage intensities between 10 and 30 percent. The presented index only made an error in detecting the location of damage in cases where damage occurred in the concrete cover and rebars in small areas, but even in these cases, the presence of damage was well demonstrated. Due to limitations in sensor placement at the pier height, increasing the distance between sensors and reducing their number was also investigated, and it was observed that even with greater distances between sensors, the presented index has the capability to detect the presence and location of damage, and after finding the approximate location of damage, it is possible to identify the exact location of damage by re-sensoring in probable areas.