Continuous beams are often used within RC structures, which are exposed to aggressive environmental impact. The use of the fiber-reinforced polymer (FRP) reinforcement in these objects and environments has a big significance, taking into account tendency of steel reinforcement to corrode. The main aim of these research studies is to estimate ability of continuous beams with glass FRP (GFRP) reinforcement to redistribute internal forces, as a certain way of ductility and desirable behaviour of RC structures. This paper gives the results of experimental research of seven continuous beams, over two spans of 1850 mm length, cross-section of 150 × 250 mm, that are imposed to concentrated forces in the middle of spans until failure. Six beams were reinforced with different longitudinal GFRP and same transverse GFRP reinforcements, and one steel-reinforced beam was adopted as a control beam. The main varied parameters represent the type of GFRP reinforcement and ratio of longitudinal reinforcement at the midspan and at the middle support, i.e., design moment redistribution. The results of the research have shown that moment redistribution in continuous beams of GFRP reinforcement is possible, without decreasing the load-carrying capacity, compared to elastic analysis. The test results have also been compared to current code provisions, and they have shown that the American Concrete Institute (ACI) 440.1R-15 well predicted the failure load for continuous beams with GFRP reinforcement. On the contrary, current design codes underestimate deflection of continuous beams with GFRP reinforcement, especially for higher load levels. Consequently, a modified model for calculation of deflection is proposed.

Fiber-reinforced polymers (FRP) are commonly used as internal reinforcement in RC structures in aggressive environments. The design of concrete elements reinforced with FRP bars is usually ruled by serviceability criteria rather than the ultimate limit state. Six continuous concrete beams over two spans with longitudinal and transverse glass FRP (GFRP) reinforcement were investigated until failure to estimate the effects of different reinforcement arrangements on the limit states of continuous beams. The ratio of longitudinal reinforcement between the midspan and middle support sections (i.e., the design moment redistribution) and the type of GFRP reinforcement were the main parameters. The experimental results were compared to prediction models and other code formulations under serviceability and ultimate limit states. The bond-dependent coefficient kb was investigated to assess adhesion conditions for GFRP reinforcement and concrete. The results showed that moment redistribution in continuous beams with GFRP reinforcement happens with slippage between the reinforcement and concrete in the middle support without the load capacity being reduced. A modified model was suggested for better deflection prediction of continuous beams reinforced with GFRP bars. Based on deformability factors, the tested continuous beams, although containing GFRP reinforcement that has brittle behavior, showed a certain kind of ductile behavior.

Deflections on continuous beams with glass fiber-reinforced polymer (GFRP) reinforcement are calculated in accordance with the appropriate standards (ACI 440.1R-15, CSA S806-12). However, experimental research provides results which differ from the values calculated pursuant to the standards, particularly when it comes to continuous beams. Machine learning methods can be applied for predicting a deflection level on continuous beams with GFRP (glass fiber-reinforced polymer) reinforcement and loaded with a concentrated load. This paper presents research on using artificial neural networks for deflection estimation and an optimal prediction model choice. It was necessary to first develop a database, in order to train the neural network. The database was formed based on the results of the experimental research on continuous beams with GFRP reinforcement. Using the best trained neural network model, high accuracy was obtained in estimating deflection, expressed over the mean absolute percentage error, 9.0%. This result indicates a high level of reliability in the prediction of deflection with the help of artificial neural networks.