Value of deformation index of composite shear walls with parallel circular concrete-filled steel tubular columns combined with machine learning

A steel-plate-provided concrete shear wall with parallel circular concrete-filled steel tube columns (P-CFST-SCSW) was investigated for engineering applications and deformation performance, and this study combines the characteristics of a P-CFST-SCSW to propose material strain limit values for various performance states of this shear wall. Firstly, 378 finite element models are designed using accurate finite element models, which include the ratio of axial compression and the ratio of shear-to-span, steel tube distance, and steel tube thickness. Furthermore, performance points are extracted from all the models, and the effect of each parameter on the limits of the deformation index of the P-CFST-SCSW is investigated. Finally, the machine learning algorithm is used to build models for predicting the deformation of the P-CFST-SCSW and analyze the importance of the feature. The results show that the range of performance point 5 is between 1/25 and 1/12, reflecting the improvement of the P-CFST-SCSW̕s ductility. Performance points are best predicted by the gradient-boosting decision tree model, with an R-squared value exceeding 0.9 across the test set. The importance of steel tube thickness increases gradually with the performance points. The shear-to-span ratio increases and subsequently decreases, unlike the axial compression ratio.