Shear capacity estimation of fully grouted reinforced concrete masonry walls using neural network and adaptive neuro-fuzzy inference system models

Abstract

In recent years, fully grouted reinforced concrete block masonry shear walls have been widely used as key structural elements for seismic resistance in medium- and high-rise buildings. However, accurately estimating their shear strength is truly challenging owing to the complex behavior of masonry walls under in-plane loads. This paper proposes the application of artificial neural network and adaptive neuro-fuzzy inference system models for predicting the shear strength of grouted reinforced concrete block masonry walls. To construct these models, an experiment was conducted and additional experimental data were gathered from published literature. Eleven main parameters were considered to be input parameters: compressive strength of grouted concrete block masonry, wall height, wall length, wall thickness, effective wall length, axial load, longitudinal and transverse reinforcement ratios, horizontal reinforcement spacing, and yield strength of longitudinal and transverse reinforcements. The prediction values of the well-trained artificial neural network and adaptive neuro-fuzzy inference system models agreed well with the experimental data. In addition, the comparison results showed that the two proposed models perform better than the existing empirical models. Therefore, they can be considered accurate and reliable models for estimating the shear strength of grouted reinforced concrete block masonry walls.