Engineered Cementitious Composite (ECC) has emerged as a promising solution with which to address the longstanding challenge of cracking in the tensile zone of reinforced concrete beams. This study conducts an experimental and analytical exploration of the flexural performance of ECC-concrete composite beams reinforced with hot-rolled ribbed steel bars. Sixteen beams, featuring diverse reinforcement ratios and ECC layer thicknesses, underwent rigorous testing through a four-point bending setup. The experimental findings underscore a substantial improvement in crack resistance and flexural bearing capacity of ECC-concrete composite beams reinforced with steel bars. Building on these results, a theoretical model was formulated to predict the moment-deflection responses of ECC-concrete composite beams incorporating steel bars. Furthermore, practical and simplified methods were introduced to predict flexural bearing capacity and effective moment of inertia, as well as anticipate failure modes, offering a user-friendly approach for engineering applications. Validation of the proposed approaches was achieved through simulation results, demonstrating a high degree of accuracy when compared with the experimental outcomes. Moreover, the average crack width at serviceability limit states of composite beams was sensitive to specimen size and the yield strength of steel bars, and a size effect was also observed for ductility expressed as deflection.

Machine learning techniques have become a popular solution to prediction problems. These approaches show excellent performance without being explicitly programmed. In this paper, 448 sets of data were collected to predict the neutralization depth of concrete bridges in China. Random forest was used for parameter selection. Besides this, four machine learning methods, such as support vector machine (SVM), k-nearest neighbor (KNN) and XGBoost, were adopted to develop models. The results show that machine learning models obtain a high accuracy (>80%) and an acceptable macro recall rate (>80%) even with only four parameters. For SVM models, the radial basis function has a better performance than other kernel functions. The radial basis kernel SVM method has the highest verification accuracy (91%) and the highest macro recall rate (86%). Besides this, the preference of different methods is revealed in this study.