Probabilistic identification method for seismic failure modes of reinforced concrete beam-column joints using Gaussian process with deep kernel基于深度核 Gauss 过程的钢筋混凝土梁柱节点抗震失效模式概率识别方法
![]()
Yu ZC, Yu B, Li B, 2024. Probabilistic identification method for seismic failure modes of reinforced concrete beam-column joints using Gaussian process with deep kernel. Probabilistic Engineering Mechanics, 76: 103610.DOI: 10.1016/j.probengmech.2024.103610
摘要 | Abstract
识别梁柱节点 (beam-column joint, BCJ) 在地震中的失效模式对于钢筋混凝土 (reinforced concrete, RC) 建筑或结构的抗震安全性和完整性至关重要。然而，传统识别方法无法有效预测不确定性信息，而这些信息有助于评估、识别和改进预测。本研究提出了一种基于深度核 Gauss 过程 (Gaussian process, GP) 的梁柱节点抗震失效模式概率识别方法，该方法将深度神经网络的表征能力与核函数的灵活结构相结合，以准确表征梁柱节点抗震失效模式的演变特征。分析结果表明，所提方法能够提高传统 Gauss 过程的分类精度，并且预测精度优于传统抗剪设计方法和机器学习技术。此外，该方法还为估计梁柱节点抗震失效模式预测中的不确定性提供了有效途径。关键词: 钢筋混凝土, 梁柱节点, 抗震失效模式, 概率识别方法, Gauss 过程, 深度核Identifying the seismic failure modes of beam-column joints (BCJs) is crucial for the safety and integrity of reinforced concrete (RC) buildings or structures withstanding seismic forces. However, traditional identification methods fail to provide any indication about the uncertainties within their predictions, which is beneficial for evaluating, interpreting and improving these predictions. This study develops a probabilistic identification method for seismic failure modes of BCJs using Gaussian process (GP) with a deep kernel, which integrates the representational power of deep neural networks with the flexible structure of kernel functions to accurately represent the evolution characteristics of seismic failure modes of BCJs. Analysis results demonstrated the potential of the proposed method for improving the classification accuracy of traditional GPs, as well as its superiority over the prediction accuracy of traditional shear-resistance design methods and machine learning techniques. Furthermore, the proposed method also provides an efficient approach to estimate the uncertainties within their predictions for seismic failure modes of BCJs.Keywords: Reinforced concrete; Beam-column joints; Seismic failure modes; Probabilistic identification method; Gaussian process; Deep kernel创新点 | Highlights
- 深度核结构集成到 Gauss 过程中，以提高分类精度
评估了传统抗剪设计方法在失效模式分类中的表现
- Probabilistic identification method for failure modes of RC joints was proposed
- Deep kernel architecture was integrated into Gaussian process for improving classification accuracy
- Influences of mechanical-based features on probabilistic evolution of failure modes were analyzed
- Traditional shear-resistance design methods for classifying failure modes were evaluated
![]()
Fig. 1. Three types of seismic failure modes of BCJs
图 2: 基于深度核 Gauss 过程的梁柱节点抗震失效模式概率识别框架Fig. 2. Probabilistic identification framework for seismic failure modes of BCJs based on DGP
图 3: 不同特征参数下梁柱节点的抗震失效模式演化Fig. 3. Seismic failure modes evolution of BCJs with different feature parameters
Fig. 4. Classification performance of Gaussian process model
Fig. 5. Classification performance of deep learning model
图 6: 所提深度核 Gauss 过程模型的分类表现Fig. 6. Classification performance of the proposed DGP model
Fig. 7. Comparison of classification performance metrics of different ML models
图 8: 传统抗剪设计方法与深度核 Gauss 过程的混淆矩阵Fig. 8. Confusion matrices between traditional shear-resistance design method and the DGP)
Fig. 9. Misclassifications for seismic failure modes by different models
图 10: 基于深度核 Gauss 过程的抗震失效模式概率识别Fig. 10. Probabilistic identifications of seismic failure modes based on DGP
图 11: 特征参数对梁柱节点抗震失效模式概率演化的影响Fig. 11. Probabilistic evolution of seismic failure modes of BCJs influenced by feature parameters
作者信息 | Authors
新加坡南洋理工大学 (Nanyang Technological University) 土木与环境工程学院
余波 Bo Yu, 通讯作者 (Corresp.) 广西大学 (Guangxi University) 土木建筑工程学院Email: gxuyubo@gxu.edu.cn
新加坡南洋理工大学 (Nanyang Technological University) 土木与环境工程学院
律梦泽 M.Z. Lyu | 编辑 (Ed)
P.D. Spanos | 审校 (Rev)
陈建兵 J.B. Chen | 审校 (Rev)
彭勇波 Y.B. Peng | 审校 (Rev)
