Novel hybrid machine leaning model for predicting shear strength of reinforced concrete shear walls

Accurate prediction of the ultimate shear capacity of reinforced concrete shear walls (RCSWs) is essential for robust design of buildings under seismic and wind loads. However, the shear capacity of RCSWs depends on multiple complex design variables characterized by diverse geometric and materials properties. Thus, a powerful modeling framework is required. In this paper, a hybrid artificial intelligence model is proposed for predicting the ultimate shear capacity of RCSWs named artificial neural network (ANN) coupled with adaptive harmony search optimization (AHS) algorithm. Different statistical metrics were used to compare the performances of the ANN model coupled with AHS (ANN-AHS) to three existing empirical relations and two ANN models combined with harmony search (ANN-HS) and global-best harmony search (ANN-GHS). Results show that the proposed ANN-AHS achieved superior performance in modelling the shear strength of RCSWs compared to ANN-HS and ANN-GHS models. The soft-computing models have proven to be more accurate than existing empirical relations.

Automated summary

A hybrid artificial intelligence model, combining artificial neural network with adaptive harmony search optimization algorithm, demonstrated superior performance in predicting the ultimate shear capacity of reinforced concrete shear walls. The soft-computing model was proven to be more accurate than existing empirical relations.