Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Reliability analysis of frame structures under top-floor lateral load using artificial intelligence
The portal frame is an important component of the framed structure building, and researchers in earlier studies have shown the significant ambiguities in determining the displacement under lateral loading, using a deterministic approach. Artificial intelligence (AI) has been applied in the current studies to improve performance and remove associated uncertainty. This study’s primary goal is to get more efficient machine learning (ML) model for calculating displacement on the top floor under lateral load. To estimate the displacement on the top floor under lateral loading, four hybrid models were developed in this study: random forest (RF), gradient boost machine (GBM), extreme gradient boosting (XGBoost) and deep neural network (DNN). All the proposed models could efficiently predict displacement under top-floor lateral loading. The total performance of the models has been revealed using various statistical metrics and RF outperformed among all four models having some testing outputs as NS = 0.9933, R2 = 0.9867, AdjR2 = 0.9857, PI = 1.9758, VAF = 99.3572, LMI = 0.9529, WI = 0.9983, RMSE = 0.0013, WMAPE = 0.0187, SI = 0.0365 and MAE = 0.0006. Also, the RF model outperformed the XGBoost, DNN and GBM models, as per the results of the evaluations utilizing rank analysis and regression curve. For all model types, the reliability index has been determined utilizing the first-order second-moment approach, as the probability of failure. Among all models, the XGBoost model has a higher value of β as well as lower value of Pf, while the RF model has a slightly lower value of β and a little higher value of Pf. Overall, we find that the RF model can be utilized to determine the displacement under top-floor lateral loading, while the XGBoost model can be utilized as a reliable artificial intelligence technique to determine the failure.
Reliability analysis of frame structures under top-floor lateral load using artificial intelligence
The portal frame is an important component of the framed structure building, and researchers in earlier studies have shown the significant ambiguities in determining the displacement under lateral loading, using a deterministic approach. Artificial intelligence (AI) has been applied in the current studies to improve performance and remove associated uncertainty. This study’s primary goal is to get more efficient machine learning (ML) model for calculating displacement on the top floor under lateral load. To estimate the displacement on the top floor under lateral loading, four hybrid models were developed in this study: random forest (RF), gradient boost machine (GBM), extreme gradient boosting (XGBoost) and deep neural network (DNN). All the proposed models could efficiently predict displacement under top-floor lateral loading. The total performance of the models has been revealed using various statistical metrics and RF outperformed among all four models having some testing outputs as NS = 0.9933, R2 = 0.9867, AdjR2 = 0.9857, PI = 1.9758, VAF = 99.3572, LMI = 0.9529, WI = 0.9983, RMSE = 0.0013, WMAPE = 0.0187, SI = 0.0365 and MAE = 0.0006. Also, the RF model outperformed the XGBoost, DNN and GBM models, as per the results of the evaluations utilizing rank analysis and regression curve. For all model types, the reliability index has been determined utilizing the first-order second-moment approach, as the probability of failure. Among all models, the XGBoost model has a higher value of β as well as lower value of Pf, while the RF model has a slightly lower value of β and a little higher value of Pf. Overall, we find that the RF model can be utilized to determine the displacement under top-floor lateral loading, while the XGBoost model can be utilized as a reliable artificial intelligence technique to determine the failure.
Reliability analysis of frame structures under top-floor lateral load using artificial intelligence
Asian J Civ Eng
Sufyan, Md Saeb (Autor:in) / Samui, Pijush (Autor:in) / Mishra, Shambhu Sharan (Autor:in)
Asian Journal of Civil Engineering ; 24 ; 3653-3665
01.12.2023
13 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Lateral Load Path Analysis: Practical Methods for Light-Frame Modular Structures
| British Library Online Contents | 2018
Interaction Analysis of Infilled Frame Structures under Lateral Loads
| British Library Conference Proceedings | 2014
Interaction Analysis of Infilled Frame Structures under Lateral Loads
| Trans Tech Publications | 2014