基于模态数据的八层正交胶合木建筑 Bayes 模型更新
Fig. 1. General information about the building: (a) Dimensions; (b) A typical layout; (c) A typical cross-section; (d) A photo from August 2022
Fig. 2. AVT of an eight-storey CLT building: (a) Layout of the sensor locations on the roof; (b) A photo of typical sensor location; (c) Stabilisation diagram of the first 90 min interval; (d) Stabilisation diagram of a typical 9 min interval; (e) Histograms of the natural frequencies identified from the 80 windowed 9 min intervals (blue lines show estimation with normal distribution using mean and COV from Tab. 1); (f) Mode shapes identified from the first 90 min interval
Fig. 3. Breakdown of the contributions to the total mass of the building
Fig. 4. Modes of vibration of the initial FE model
Fig. 5. Correlation of the experimental and FE modes presented: (a) MAC matrix; (b) FMAC plot
Fig. 6. Effect of different mass distribution assumptions on the natural frequencies
Fig. 7. Effect of different geometry modelling assumptions on the natural frequencies
Fig. 8. Cross-validation of the surrogate model
Fig. 9. Sobol’ indices of the input parameters: coefficient of vertical stiffness of the walls (P_1), coefficient of the in-plane shear stiffness of the walls (P_2), coefficient of the in-plane shear stiffness of the external walls (P_3), the stiffness of the springs modelling the foundation (P_4), and coefficient of the total mass of the building (P_5)
Fig. 10. Pairwise scatter plots of posterior distribution sampled with MCMC are shown in off-diagonal plots
Fig. 11. Posterior distribution of the in-plane shear modulus: (a) Internal walls G_{wall,in}; (b) External walls G_{wall,ex}; (c) A scatter plot of samples projected in the 2D parameter plane (G_{wall,in} and G_{wall,ex})
Fig. 12. KDEs of posterior parameter distributions when different sets of modes are included in the Bayesian updating
Fig. 13. A scatter plot of posterior samples in P_1–P_4 parameter plane for iterations with 3 and 4 modes included
Fig. 14. Prior, posterior and measured frequencies
Fig. 15. One-at-a-time sensitivity analysis for parameter P_4 while other parameters are set to the updated values
作者信息 | Authors
Email: bkurent@fgg.uni-lj.si
匈牙利计算机科学与控制研究所 (Institute for Computer Science & Control) 信息学实验室
意大利拉奎拉大学 (Università degli Studi dell’Aquila) 土木建筑与环境工程系
挪威木材技术研究所 (Norsk Treteknisk Institutt)
挪威生命科学大学 (Norwegian University of Life Sciences) 科学技术学院
斯洛文尼亚卢布尔雅那大学 (University of Ljubljana) 土木与测绘工程学院
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)
