ABSTRACT
The monitoring of bridges is important to ensure structural safety throughout the life-cycle of the enterprise. Environmental effects such as increased water volume in rivers can cause considerable changes in the data provided by instrumentation, sometimes making it challenging to identify whether the structure is exhibiting a damage pattern or just indicating a change in structural behavior due to an environmental component. In some regions of the state of Pará/Brazil, rivers can rise up to 10 meters during flood periods. This characteristic makes bridges in these regions good candidates for testing classification and damage pattern recognition algorithms. Techniques for detecting damage in large structures have been successfully applied in several case studies. Works based on principal component analysis (PCA) have been widely performed due to the ease of implementation of the algorithm and the efficiency in extracting patterns that are important in structural response. In this work, PCA, T²-statistic and Q-statistic were used to evaluate the influence of the effect of floods on the modal properties of the bridge over the Itacaiúnas River. The acceleration data for this structure was obtained at two times: flood period in 2018 and dry period in 2021. The results indicate that the increase in river level causes a decrease in frequencies in the vibration modes of the structure. Furthermore, the use of PCA proved capable of discriminating characteristics from environmental effects.
