PSI - Issue 68

6

Hussam Safieh et al. / Procedia Structural Integrity 68 (2025) 245 – 251 H. Safieh et al / Structural Integrity Procedia 00 (2025) 000–000

250

70

60

50

40

30

20

Predicted Values

Training Data Test Data

10

0

0

10

20

30

40

50

60

70

80

Actual Values

Fig. 2. Training versus test values of Random forest ML model.

4. Conclusions This study explored the application of machine learning techniques to predict the compressive strength of Ordinary Portland Cement (OPC) mixed with varying percentages of fly ash. Among the various models tested, Random Forest emerged as the most effective, significantly outperforming simpler models like Linear Regression, Support Vector Machine, and Ridge Regression. The Random Forest model's superior performance, with an R² of 0.7893 and a low RMSE of 6.5392, underscores its ability to capture complex, nonlinear relationships within the dataset, particularly in the context of variable fly ash content and curing periods. The study demonstrated that the inclusion of fly ash, a supplementary cementitious material, introduces variability in compressive strength that requires advanced modeling techniques to predict accurately. Random Forest's ensemble approach effectively mitigated overfitting, handling both the linear and nonlinear relationships between the input variables and the target output. Ultimately, this work contributes to the optimization of concrete mixes, promoting the development of more sustainable and efficient construction materials Acknowledgements The support presented in this paper had been provided by the American University of Sharjah and Riad T. Al Sadek Endowed Chair in Civil Engineering. The support is gratefully acknowledged and appreciated. The views and conclusions expressed or implied are those of the authors and should not be interpreted as those of the donor or the

institution. References

Abdalla, J. A., Hawileh, R., 2011. Modeling and simulation of low-cycle fatigue life of steel reinforcing bars using artificial neural network. Journal of the Franklin Institute , 348 (7), 1393–1403. doi: 10.1016/j.jfranklin.2010.04.005 Abdalla, J. A., Hawileh, R. A., 2013. Artificial Neural Network Predictions of Fatigue Life of Steel Bars Based on Hysteretic Energy. Journal of Computing in Civil Engineering , 27 (5), 489–496. doi: 10.1061/(asce)cp.1943-5487.0000185 Abuodeh, O. R., Abdalla, J. A., Hawileh, R. A., 2020. Prediction of shear strength and behavior of RC beams strengthened with externally bonded FRP sheets using machine learning techniques. Composite Structures , 234 . doi: 10.1016/j.compstruct.2019.111698 Barcelo, L., Kline, J., Walenta, G., Gartner, E., 2014. Cement and carbon emissions. Materials and Structures/Materiaux et Constructions , 47 (6), 1055–1065. doi: 10.1617/s11527-013-0114-5