Failure Mode Detection of Reinforced Concrete Shear Walls Using Ensemble Deep Neural Networks

Abstract

Reinforced concrete structural walls (RCSWs) are one of the most efficient lateral force-resisting systems used in buildings, providing sufficient strength, stiffness, and deformation capacities to withstand the forces generated during earthquake ground motions. Identifying the failure mode of the RCSWs is a critical task that can assist engineers and designers in choosing appropriate retrofitting solutions. This study evaluates the efficiency of three ensemble deep neural network models, including the model averaging ensemble, weighted average ensemble, and integrated stacking ensemble for predicting the failure mode of the RCSWs. The ensemble deep neural network models are compared against previous studies that used traditional well-known ensemble models (AdaBoost, XGBoost, LightGBM, CatBoost) and traditional machine learning methods (Naïve Bayes, K-Nearest Neighbors, Decision Tree, and Random Forest). The weighted average ensemble model is proposed as the best-suited prediction model for identifying the failure mode since it has the highest accuracy, precision, and recall among the alternative models. In addition, since complex and advanced machine learning-based models are commonly referred to as black-box, the SHapley Additive exPlanation method is also used to interpret the model workflow and illustrate the importance and contribution of the components that impact determining the failure mode of the RCSWs.

Hybrid Wavelet Scattering Network-Based Model for Failure Identification of Reinforced Concrete Members

“…The WSN-BT algorithm and WSN-ANN algorithm are ranked second and third, after the WSN-SVM algorithm, in terms of damage classification accuracy. While the recognition of CD damage is challenging in other research studies [40,41], the WSN-SVM algorithm herein achieved an almost 99.1% recall and 99.5% precision in the CD failure mode prediction in the test phase. However, it seems that the identification of the CD damage is difficult for the other proposed hybrid models.…”

Machine learning based prediction model for plastic hinge length calculation of reinforced concrete structural walls

“…The hue in this diagram denotes the variable value, which ranges from low (blue) to high (red), and the dots in it stand for the SHAP values of the specimens (in the test set) for each input variable. Figure 12(b) displays that the model is more likely to anticipate larger PHL as the value of the ratio of shear span rises, which relates to a higher possibility of RCSW yielding in flexure before meeting the RCSW’s nominal shear strength (Barkhordari and Massone 2022a, 2022b). Also, the higher values of wall length (lw) will increase the length of the plastic hinge.…”

“…With the emergence of machine learning (ML) methodologies, popular tools such as artificial neural networks (ANNs) and ensemble models have become increasingly used in material science and structural engineering research and applications. As an example, Barkhordari and Massone (2022a) developed ensemble deep neural network algorithms to predict the failure mode of RCSWs. They suggested that the weighted average ensemble model, which has the highest accuracy of the competing models, be used for determining the failure mechanism.…”

Study the Efficiency of the XGBoost Algorithm for Squat RC Wall Shear Strength Prediction and Parametric Analysis

“…Higher R 2 , lower RMSE, and lower MAPE values, overall, demonstrate that the model is more accurate and highly efficient. According to the SI parameter, a model performs poorly when SI > 0.3, fairly well when 0.2 < SI < 0.3, good performance when 0.1< SI < 0.2, and excellent performance when SI < 0.1 [27].…”