Akademik Veri Yönetim Sistemi
Studies in Systems, Decision and Control, Springer International Publishing Ag, ss.169-179, 2026
Seismic base isolators are effective vibration controllers used to control structural vibrations. These devices, which are designed according to the seismicity of the region in which they are located, may suffer from environmental impacts such as various construction and excavation works, strong explosions, natural disasters or mechanical component problems, wear on the isolator bearing, pollution that will affect friction and similar problems, resulting in a decrease in their damping capacity and maintenance and repair needs. In the detection of such situations, it is possible to predict the reduction in damping capacity on a class basis with artificial intelligence techniques. In this study, the Catboost algorithm, a machine learning algorithm capable of extracting meaning from categorical data, is used to create a class prediction model of damping capacity, and the algorithm parameters are optimized to obtain optimum efficiency from the algorithm. In the study, the behavior of isolators with damping capacities ranging between 10 and 50% under earthquake records was recorded and converted into a data set and used in the training of the Catboost machine learning model. To improve the accuracy performance of the developed model, Catboost parameters were optimized with the Optuna tool. The Catboost learning model with optimized parameters showed a prediction performance exceeding 96.3%.