Determination of structural behavior and long-term structural health monitoring in Deriner and Yusufeli Arch Dams by considering structure-water-soil interaction

Arch dams have slender sections thanks to their curvature structures. The fact that these dams are large in volume, costly and take many years to build increases the importance of the structure. Huge losses of life and property may occur as a result of errors and/or accidents that may occur during the construction of such structures. Considering the volume, mass and slenderness of arch dams, they can be significantly affected by dynamic forces such as changing water pressure, temperature, wind and earthquakes. The level of these effects and the extent to which they affect the dam can be determined by analysis in the finite element model created with the help of computers. Since Finite Element Analysis is based on certain assumptions, the results obtained must be compared and/or verified with the results obtained by experimental methods. One of the most suitable methods preferred in many engineering structures is the experimental modal analysis method. Within the scope of the thesis, long-term Structural Health Monitoring studies were carried out with sensitive accelerometers placed on the structures in order to experimentally determine the structural behavior of Deriner and Yusufeli Dams, the two highest dams in Turkey. Modal analyses of the finite element models prepared according to the projects of both dams were carried out and verified with reference to experimental dynamic characteristics. In order to determine the effectiveness of three different reservoir modeling approaches and the static/dynamic performances of the dams, analyzes were carried out for the maximum water level in Deriner Dam, and for the empty reservoir in Yusufeli Dam and for eight different water level conditions selected throughout the water retention process. In addition, long-term Structural Health Monitoring studies were carried out with the sensors installed in Deriner and Yusufeli Dams, and the results obtained were trained with Artificial Neural Networks, and estimated frequency maps were created and examined in detail within the scope of the thesis.