Estimating Dynamic Properties of RC Frame Structures with Various Infill Materials for Quick Assessment

The effect of infill walls on frame rigidity is usually not considered in the performance assessment of structures nor the fundamental period, where infill walls would affect both the performance level and the structure's period negatively or positively depending on their plan and elevation distribution. Moreover, the need for a quick assessment tool to predict the performance level of structures is essential, especially considering infill wall effects. Therefore, buildings were designed according to the Turkish Seismic Code for Buildings 2018 (TSCB-2018) and the Requirements for Design and Construction of Reinforced Concrete Structures Code (TS-500). The number of analyzed models was 336, each with different combinations of peak ground acceleration, soil type, infill material, and infill walls layouts. The data in the Artificial Neural Network (ANN) models used analysis outcomes collected, such as the fundamental period, target displacement, and performance level, as a target, training, and validation data. Furthermore, for performance assessment, the commercial program Seismostruct 2022 was used for nonlinear static analysis (pushover analysis), while Matlab 2018 was used for the creation of neural network models. Bayesian regularization backpropagation was used for all ANN models. ANN models showed excellent accuracy in all three models: period, performance, and target displacement. The models' accuracy obtained was as follows: in the performance level model at 96.15%, while for the fundamental period of the structure, it was 99.99% and lastly, the target displacement was 99.66%. Keywords: ANN, Fundamental Period, Infill Wall, Performance, Target Displacement.