Machine learning-based models for FRP shear contribution of wrapped reinforced concrete beams

Determining the shear contribution of fiber-reinforced polymers (FRP) in wrapped reinforced concrete (RC) beams can highly help civil engineers in the design and construction of concrete structures. It is worth noting that since lab works require time and money, and preparing the research test equipment is not quite easy, using predictive models with high ability and accuracy in estimating the target parameter based on the available databanks can reduce the cost and time. This research has used the “artificial neural network (ANN)” and “gene expression programming (GEP), two robust predictive models, and considered such parameters as the beam-section width and height, concrete compressive strength and the number, thickness, width, elastic modulus, center-to-center distance and orientation angle of the FRP layers as the input data to predict the shear contribution of the FRP. According to the results, although both models predicted the mentioned contribution with better accuracy and less error than the models presented in such codes as the ACI, FIB, CSA, TR-55, the ANN model had a higher accuracy/precision and a better correlation coefficient than the GEP model in estimating the target parameter. However, as GEP provides a user-friendly formula with a suitable estimation power, it can satisfy both the manufacturers and researchers.