Multiple-Surrogate Models for Probabilistic Performance Assessment of Wind-Excited Tall Buildings under Uncertainties

Micheli, Laura; Alipour, Alice; Laflamme, Simon

doi:10.1061/ajrua6.0001091

Cited by 15 publications

(3 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Step 3 of Algorithm 1, multiple surrogate models are employed as replacement of the original numerical simulation model to reduce the computational demand of the optimization process. [ 38 ] A Kriging surrogate model is defined for each floor of the structure. Each surrogate model includes the floor wind speed, structural characteristics, and damping device properties as inputs and the corresponding EDPs as output.…”

Section: Lcc Optimization Methodologymentioning

confidence: 99%

Life‐cycle cost optimization of wind‐excited tall buildings using surrogate models

Micheli

Alipour

Laflamme

2021

Structural Design Tall Build

Self Cite

View full text Add to dashboard Cite

As buildings become taller and slender, they become more sensitive to wind-induced vibrations. Commonly used solutions to mitigate wind-induced vibrations are structural system modifications and integration of supplemental damping devices. This paper presents a procedure to optimize the structural system and the damping device configuration with the goal of reducing wind-induced vibrations. The performance of the building is expressed in terms of life-cycle cost (LCC), allowing to consider not only the initial costs associated with the integration of wind-induced vibration mitigation system but also the lifetime savings due to vibration suppression. The proposed procedure employs a set of Kriging surrogate models to analyze a large number of different structural properties/damping device characteristics. The combination that minimizes the LCC is taken as the optimal configuration. The procedure is demonstrated on a wind-sensitive 39-story building equipped with passive dampers. Results demonstrated that the accuracy of the Kriging surrogate models depends on the number of input variables considered, with an average root mean square error of 2.5% for the floors without dampers and 5% for the floors equipped with damping devices, respectively. It was also demonstrated that optimal stiffness-damping device configuration and LCC depend on the assumed cost of the structural system modifications.

show abstract

Section: Lcc Optimization Methodologymentioning

confidence: 99%

Life‐cycle cost optimization of wind‐excited tall buildings using surrogate models

Micheli

Alipour

Laflamme

2021

Structural Design Tall Build

Self Cite

View full text Add to dashboard Cite

show abstract

“…The authors exploited the metamodel to construct fragility functions of a 4-story building under stochastic ground motions and uncertain structural parameters. Recently, Micheli et al [15] proposed a multiple-surrogate models framework for risk assessment of wind-excited tall buildings. The proposed framework exploits a set of Kriging surrogates to reduce the computational demand of risk assessment of high-rise structures subjected to wind load time histories.…”

Section: Introductionmentioning

confidence: 99%

“…Synthetic input/output observations are usually derived from numerical simulation models, which might lead to approximate responses. Furthermore, when the response of the structure is highly nonlinear, a large data pool might be required to train a reliable surrogate model [15][16][17], yielding a significant computational burden. A solution to this drawback could be to extract input/output observations directly from field data [18].…”

Section: Introductionmentioning

confidence: 99%

A Neural Network-Based Approach for Seismic-Induced Damage Assessment of Steel Liquid Storage Tanks

Micheli¹,

Faytarouni²

2021

Proceedings of the 8th International Conference on Computational Methods in Structural Dynamics and Earthquake Engineering (COM

Self Cite

View full text Add to dashboard Cite

Surrogate models, including neural network (NN), machine learning, and Kriging, are used in various fields to reduce the computational demand of risk assessment and uncertainty analysis. In civil engineering applications, surrogate models are usually trained on synthetic data generated with numerical simulation models, which might yield approximate responses and significant computational burdens. Post-disaster reconnaissance observations represent an alternative source of data that could be used to train a surrogate model without the need for numerical models. However, the limited number of reconnaissance observations available in the literature might yield challenges, such as unbalanced data distributions. This paper presents a surrogate-based approach for seismic-induced damage prediction, where postearthquake reconnaissance data are exploited to train a NN model. The approach is demonstrated on steel liquid storage tanks. Field data from past earthquake reconnaissance reports are first collected. Then, features representative of tank characteristics, seismic hazard parameters, and seismic-induced damage are extracted. A traditional two-layers NN model is built to map the relationship between tank characteristics, seismic hazard parameters, and seismic-induced damage. To tackle the challenge of unbalanced dataset, a cascade NN approach is proposed. In the cascade approach, two NN models are employed to predict the damage level of the steel tanks. The first NN returns a binary classification of the tank damage (i.e., damage, no damage). If the tank results damaged, a second NN identifies the level of damage (i.e., minor, severe, collapse). The performance of traditional and cascade NN approaches is compared using different metrics. Results demonstrate that the cascade NN strategy leads to more accurate damage predictions than the traditional NN approach.

show abstract

A new method for earthquake-induced damage identification in historic masonry towers combining OMA and IDA

Kita

Cavalagli

Venanzi

et al. 2021

Bull Earthquake Eng

View full text Add to dashboard Cite

This paper presents a novel method for rapidly addressing the earthquake-induced damage identification task in historic masonry towers. The proposed method, termed DORI, combines operational modal analysis (OMA), FE modeling, rapid surrogate modeling (SM) and non-linear Incremental dynamic analysis (IDA). While OMA-based Structural Health Monitoring methods using statistical pattern recognition are known to allow the detection of small structural damages due to earthquakes, even far-field ones of moderate intensity, the combination of SM and IDA-based methods for damage localization and quantification is here proposed. The monumental bell tower of the Basilica of San Pietro located in Perugia, Italy, is considered for the validation of the method. While being continuously monitored since 2014, the bell tower experienced the main shocks of the 2016 Central Italy seismic sequence and the on-site vibration-based monitoring system detected changes in global dynamic behavior after the earthquakes. In the paper, experimental vibration data (continuous and seismic records), FE models and surrogate models of the structure are used for post-earthquake damage localization and quantification exploiting an ideal subdivision of the structure into meaningful macroelements. Results of linear and non-linear numerical modeling (SM and IDA, respectively) are successfully combined to this aim and the continuous exchange of information between the physical reality (monitoring data) and the virtual models (FE models and surrogate models) effectively enforces the Digital Twin paradigm. The earthquake-induced damage identified by both data-driven and model-based strategies is finally confirmed by in-situ visual inspections.

show abstract

Multiple-Surrogate Models for Probabilistic Performance Assessment of Wind-Excited Tall Buildings under Uncertainties

Cited by 15 publications

References 35 publications

Life‐cycle cost optimization of wind‐excited tall buildings using surrogate models

Life‐cycle cost optimization of wind‐excited tall buildings using surrogate models

A Neural Network-Based Approach for Seismic-Induced Damage Assessment of Steel Liquid Storage Tanks

A new method for earthquake-induced damage identification in historic masonry towers combining OMA and IDA

Contact Info

Product

Resources

About