Development of multi-strut macro models for masonry infilled RC frames using machine learning techniques

The multi-strut model normally has two or more struts located concentric and eccentric to the infill wall corners. The local shear and moment in the frame owing to wall-frame contact can be represented by this model. It is important to estimate the strength of the multi-strut model for accurately capturing the influence of the infill wall on the surrounding frame. However, the prediction of struts’ properties is still a problem. In this paper, we identify efficient features for estimating the strength of the multi-strut model through advanced data science methods. To this end, various feature selection techniques are applied: Stepwise regression and Lasso regression. A comprehensive database of experimental studies from the literature is employed. The findings show that the compressive strength of masonry is the most important feature while wall and column failure modes are found as the second.