Evaluating Accuracy of Ethiopian Building Code's Prediction for flexure of RC Beam Using Neural Networks

Abstract: The Ethiopian building code standard is a theoretically based design process that was created to address restrictions, however it is not fully consistent with fundamental material attributes. The goal of this research was to evaluate the accuracy of Ethiopian building code techniques for the design of simply supported RC beams in order to perform flexural capacity tests. The design prediction of the flexural capacity of a simply supported RC beam of the compulsory Ethiopian standard (CES-2015) was compared to available experimental prediction, compressive-force path method based on assumption, different from code adapted, and artificial neural networks that were calibrated based on the available test result. This research shows that artificial neural network predictions of the flexural capacity of RC beams are very close to the experimental flexural capacity of RC beams. In addition, while the compressive force-path method is closer to the experimental than the code, it is not as effective as ANNs. When compared to other approaches, the CES-2015 method was a poor predictor of the RC beam's flexural capacity. As a result, using the input-based systematic technique and the compressive–force path method as RC structure design methods reduces the incompatibility between the CES-2015 methodologies and fundamental concrete material properties. The findings are expected to aid structural engineers who work on high-rise building analysis and design.