| Summary: | Corrosion damage, which can be considered as a construction or usage defect during the life of the structure, is an important parameter for the structural elements. Strength loss is observed in reinforced concrete (RC) elements due to corrosion, which is an important parameter affecting the performance of the building. Determining the shear strength of beams with corroded reinforcement is important in terms of strength loss, design, and reinforcement criteria in the structural element. In this context, the data of the corroded RC beam experimental tests carried out in the literature were collected and the ultimate shear strength values of the beams were determined depending on the test parameters. Strength estimation was performed with machine learning regression algorithms XGBoost and AdaBoost. The results obtained were evaluated with R2, RMSE and MAE performance metrics and high estimation success was achieved. The study shows that with these systems, which can perform learning based on experimental data, it is possible to estimate the shear strength values of corroded beams with known production parameters without the need for experimental measurements.
|
|---|