•A deep physics-guided convolutional neural network (PhyCNN) is developed for structural seismic response estimation.•Available physics can provide constraints to the network outputs, alleviate ...overfitting issues, reduce the need of big training datasets, and thus improve the robustness of the trained model for more reliable prediction.•The proposed approach was successfully demonstrated by both numerical and experimental case studies.
Accurate prediction of building’s response subjected to earthquakes makes possible to evaluate building performance. To this end, we leverage the recent advances in deep learning and develop a physics-guided convolutional neural network (PhyCNN) for data-driven structural seismic response modeling. The concept is to train a deep PhyCNN model based on limited seismic input–output datasets (e.g., from simulation or sensing) and physics constraints, and thus establish a surrogate model for structural response prediction. Available physics (e.g., the law of dynamics) can provide constraints to the network outputs, alleviate overfitting issues, reduce the need of big training datasets, and thus improve the robustness of the trained model for more reliable prediction. The surrogate model is then utilized for fragility analysis given certain limit state criteria. In addition, an unsupervised learning algorithm based on K-means clustering is also proposed to partition the datasets to training, validation and prediction categories, so as to maximize the use of limited datasets. The performance of PhyCNN is demonstrated through both numerical and experimental examples. Convincing results illustrate that PhyCNN is capable of accurately predicting building’s seismic response in a data-driven fashion without the need of a physics-based analytical/numerical model. The PhyCNN paradigm also outperforms non-physics-guided neural networks.
•A deep long short-term memory (LSTM) network is developed for nonlinear structural response modeling.•Two input-output schemes (LSTM-s and LSTM-f) are presented.•The deep learning model is capable ...of modeling both elastic and inelastic response of buildings.•An unsupervised learning algorithm is used to cluster the seismic inputs for training enhancement.•The approach was successfully verified by both numerical and experimental examples.
This paper presents a comprehensive study on developing advanced deep learning approaches for nonlinear structural response modeling and prediction. Two schemes of the long short-term memory (LSTM) network are proposed for data-driven structural seismic response modeling. The proposed deep learning model, trained on available datasets, is capable of accurately predicting both elastic and inelastic response of building structures in a data-driven fashion as opposed to the classical physics-based nonlinear time history analysis using numerical methods. In addition, an unsupervised learning algorithm based on a proposed dynamic K-means clustering approach is established to cluster the seismic inputs in order to (1) generate the least but the most informative datasets for training the LSTM and (2) improve the prediction accuracy and robustness of the model trained with limited data. The performance of the proposed approach is successfully demonstrated through three proof-of-concept studies that include a nonlinear hysteretic system, a real-world building with field sensing data, and a steel moment resisting frame. The results show that the proposed LSTM network is a promising, reliable and computationally efficient approach for nonlinear structural response prediction, and offers significant potential in seismic fragility analysis of buildings for reliability assessment.
The work presents the course and results of tests to determine the coefficient of transmission of vibrations from a building sub-base to a building structure. Vibration of structure is caused by ...quasi-seismic excitation; in presented example it is vibrations caused by tram passage. The factor of transmission is relationship between vibrations recorded on the object and the level of vibrations on the ground without an existing object. The value of the transmission degree will enable better design of the building.
•A semi-active tuned mass damper (STMD) with variable stiffness and damping ratio is proposed with a new combined control algorithm.•An eight-story linear/nonlinear full-scale base-isolated structure ...is studied.•For displacement responses, STMD can mitigate the structural first-mode response effectively.•As for acceleration responses of the top story, STMD has an excellent performance in the structural second-mode response mitigation.•STMD, thus, can improve both displacement and acceleration performances of both linear and nonlinear base-isolated structures effectively.
Base isolation can achieve a reduction in floor acceleration and inter-story drift. However, it may suffer from excessive displacements under near-fault and far-field earthquakes. To improve the aseismic performance of passive base-isolated structures, in this paper, a semi-active tuned mass damper (STMD) with variable stiffness and damping is presented. A combined control algorithm based on output signals only is developed for the STMD first. Then, the STMD is applied to an eight-story linear base-isolated structure and also a nonlinear one. As for the linear model, which represents a theoretical benchmark, an optimized passive TMD (PTMD) is used for comparison. As for the nonlinear model, lead rubber bearing (LRB) is considered and simulated using the well-known Bouc-Wen model. Eight earthquakes with different spectral characteristics and peak ground amplitudes are chosen, and two PTMDs are optimized for comparison, while one is tuned to the pre-yield period of the base-isolated model and the other is tuned to the post-yield period of the base-isolated model. Numerical results show that, generally, STMD has the best control effect in both linear and nonlinear models. For displacement responses, because STMD can vary its stiffness and damping, it can mitigate the structural first-mode response effectively, and can achieve both top story and isolated level responses reduction. As for acceleration responses of the top story, STMD achieves excellent performance in the structural second-mode acceleration response mitigation. Therefore, STMD can improve both displacement and acceleration performances of both linear and nonlinear base-isolated structures effectively.
•A friction pendulum inerter system (FPIS) is proposed to reduce seismic response.•The FPIS uses an inerter without adding significantly more mass to the system.•The impact of the FPIS’s mechanical ...layout on response mitigation is investigated.•An optimal design method is developed to meet the target structural performance.•FPIS shows the robustness and outperforms FPS-TMD at response mitigation.
Introducing a tuned mass damper (TMD) into the friction pendulum system (FPS) has been proven to be an effective approach for improving the seismic performances of base-isolated structures. However, its seismic response mitigation effect is related to the quality of the mass employed, which unavoidably requires a necessary large mass under the requirement of a high seismic performance level. To avoid introducing the extra mass of the tuned mass damper (TMD) in the friction pendulum system (FPS) of a base-isolated structure, we herein introduce a lightweight inerter subsystem that has a series-parallel layout; comprises an inerter, a spring, and a damping element; and adds almost no mass. A structure isolated by the proposed friction pendulum inerter system (FPIS) was studied by nonlinear stochastic response analysis within a probabilistic framework, and an optimal design method for a structure with the FPIS was developed to simultaneously reduce the base shear force and the base isolation floor displacement. Based on the stochastic analysis results, parametric studies and a robustness analysis were conducted, and the impact of the FPIS’s mechanical layout on the seismic response mitigation effect was investigated. The analysis results demonstrated that the FPIS significantly reduced structural responses under different types of seismic excitations. Using the proposed optimal design method, target base shear force can be achieved at a minimized cost to the base isolation floor displacement. Compared to the FPS system with a TMD, the proposed FPIS enhances the seismic response mitigation effect by avoiding the extra mass that increases the seismic energy input to the base isolation floor.
Research on the mechanics performance of Song-style Dou-gong joints is absent currently, especially considering Xiaang and Angshuan in this type of joints. Two different types of Dou-gong joints were ...experimentally studied by making models in this paper. The Dou-gong joints can maintain good integrity and rotate around the bottom member in the longitudinal plane under horizontal loads, which shows poor seismic performance. The displacement of layers is less, Xiaang and Angshuan can limit slippage between components to some extent, but they had little effect on the seismic performance of the Dou-gong joints.
In this paper, a new model for predicting seismic responses of buildings based on the correlation of ground motion (GM) and the structure is presented by simulating numerous artificial earthquakes ...(AEQs). In the model, neural network (NN) configurations representing the relationships between GM characteristics and seismic responses of a structure are developed to predict responses of the structure with only GM data measured by monitoring system in future seismic events. To extract the GM characteristics, multiple AEQs corresponding to the design response spectrum are generated based on probabilistic vibration theory, instead of using historical earthquakes. In the presented NN configurations, GM characteristics including mean and predominant period, significant duration, and peak ground acceleration are established as the input layer and the maximum inter-story drift ratio and maximum displacement are established as the output layer. In addition, a new parameter called resonance area is proposed to represent the relationship between a GM and a target structure in the frequency domain and utilized in the NN input layer. By employing the new parameter, dynamic characteristics of the structure are considered in the response estimation of the model with related to GM. The model is applied to seismic response prediction for four multi-degrees-of-freedom (MDOF) structures with different natural periods using 2700 AEQs. The validities of the presented NN models are confirmed by investigating the performance of response prediction. The effectiveness of the resonance area parameter in the NN for predicting the seismic responses is assessed and discussed. Furthermore, the effects of the constitution of NNs and computational costs of those NNs on estimation were investigated. Finally, the presented model is employed for prediction of seismic responses for a structural model of a planar reinforced concrete building structure.
•A new seismic response prediction model is presented with big data concept.•Numerous artificial earthquakes are used in the neural network training of the model.•To reflect a relationship between structure and load, a resonance area is proposed.•The validity of the presented model is confirmed through numerical simulations.•The effectiveness of the newly proposed resonance area is verified.
On the cover: The cover image is based on the Research Article Optimization and performance of metafoundations for seismic isolation of small modular reactors by Tugberk Guner et al., ...https://doi.org/10.1111/mice.12902.
•A novel method is proposed to predict the seismic responses of subway stations.•1D-CNN and LSTM are adopted.•The free-field deformation is taken as the input of the surrogate model.•The performance ...of the surrogate model is tested in different types of waves.•The proposed method reduces the computational cost in stochastic analysis.
A novel and computationally inexpensive method for predicting the nonlinear seismic response of subway stations using deep learning approaches is developed to reduce the computational cost in stochastic seismic responses analysis. The proposed method takes the deformation of the free field where the subway station is located as the input to predict seismic responses of the subway station according to the characteristic of seismic responses of underground structures. One-dimensional convolutional neural network (1D-CNN) and long-short term memory (LSTM) network are adopted for seismic responses modeling of a two-story and three-span subway station in a data-driven fashion as opposed to the computational expensive physics-based finite element model. The prediction performance and extrapolating ability of both models are evaluated and compared with a baseline multi-layer perceptron (MLP) model. With the same training samples, the 1D-CNN has better prediction performance and extrapolating ability than both LSTM and the baseline MLP model and the LSTM model has the worst performance among the three models. The good prediction performance of 1D-CNN makes it suitable to be applied in the stochastic seismic responses analysis using the probability density evolution method (PDEM) which is solved by the finite-difference method (FDM). The evolution characteristics of the probability density function of the layer drift and the distribution characteristics of the peak value of layer drift can be captured by a low computational cost with the proposed method.