Glulam is a product that was engineered to make use of timber in a more efficient way. Bychoosing timber of similar quality and discarding natural defects during production, thedevelopment of a stronger cross-section is achieved.Carbon fiber is a relatively new material with a high tension capacity. This feature is used toexamine how the bending capacity of the beams improve by adhering carbon fiber laminateson the lower edge of the beamsThe strength of the material is tested with three experiments: carbon fiber on the bottom of thebeam (a), carbon fiber attached to the lower sides of the beam (b) and carbon fiber in thebeam, covered with a layer of wood (c) The results show that the first case, where the carbon fiber is attached to the bottom of thebeam, gave the best result with an increase in capacity of 59 % compared to the nonreinforcedcontrol. The other two cases also show an improvement in capacity, beam-type 3had a capacity increase of 47% and beam-type 4 increased with 25 %Tests were also made with glulam beams reinforced with fiberglass, but these tests were notanalyzed in depth because the purpose was to compare the capacity to carbon fiber. Thisbeam improved its capacity by 40.3%.The tests show that carbon fiber as a reinforcement material for glulam is a good choice whenthere is a requirement for stronger cross-sections in both new production and renovation ofold buildings. However there are some disadvantages to carbon fiber, for example costs andincreased demands on work environment, which makes steel a cheaper option.

Yes / This paper explores the development of the most suitable machine learning models for predicting the bending capacity of steel and FRP (Fiber Reinforced Ploymer) bars hybrid reinforced ECC (Engineered Cementitious Composites)-concrete composite beams. Five different machine learning models, namely Support Vector Regression (SVR), Extreme Gradient Boosting (XGBoost), Multilayer Perceptron (MLP), Random Forest (RF), and Extremely Randomized Trees (ERT), were employed. To train and evaluate these predictive models, the study utilized a database comprising 150 experimental data points from the literature on steel and FRP bars hybrid reinforced ECC-concrete composite beams. Additionally, Shapley Additive Explanations (SHAP) analysis was employed to assess the impact of input features on the prediction outcomes. Furthermore, based on the optimal model identified in the research, a graphical user interface (GUI) was designed to facilitate the analysis of the bending capacity of hybrid reinforced ECC-concrete composite beams in practical applications. The results indicate that the XGBoost algorithm exhibits high accuracy in predicting bending capacity, demonstrating the lowest root mean square error, mean absolute error, and mean absolute percentage error, as well as the highest coefficient of determination on the testing dataset among all models. SHAP analysis indicates that the equivalent reinforcement ratio, design strength of FRP bars, and height of beam cross-section are significant feature parameters, while the influence of the compressive strength of concrete is minimal. The predictive models and graphical user interface (GUI) developed can offer engineers and researchers with a reliable predictive method for the bending capacity of steel and FRP bars hybrid reinforced ECC-concrete composite beams.