Adaptive Online One-Class Support Vector Machines with Applications in Structural Health Monitoring

2018 ◽  
Vol 9 (6) ◽  
pp. 1-20 ◽  
Author(s):  
Ali Anaissi ◽  
Nguyen Lu Dang Khoa ◽  
Thierry Rakotoarivelo ◽  
Mehrisadat Makki Alamdari ◽  
Yang Wang
Keyword(s):  
Support Vector Machines ◽  
Structural Health Monitoring ◽  
Health Monitoring ◽  
Support Vector ◽  
Structural Health ◽  
Vector Machines

A new pinched nonlinear hysteretic structural model for automated creation of digital clones in structural health monitoring

2020 ◽  
pp. 147592172092064 ◽  
Author(s):  
Cong Zhou ◽  
J Geoffrey Chase
Keyword(s):  
Decision Making ◽  
Structural Health Monitoring ◽  
Hysteresis Loop ◽  
Health Monitoring ◽  
Average Error ◽  
Support Vector ◽  
Model Parameters ◽  
Loop Analysis ◽  
Structural Health ◽  
Vector Machines

Optimizing risk treatment of structures in post-event decision-making is extremely difficult due to the lack of information on building damage/status after an event, particularly for nonlinear structures. This work develops an automated, no human intervention, modeling approach using structural health monitoring results to create accurate digital building clones of nonlinear structures for collapse prediction assessment and optimized decision-making. Model-free hysteresis loop analysis structural health monitoring method provides accurate structural health monitoring results from which model parameters of a nonlinear computational foundation model are identified. A new identifiable nonlinear smooth hysteretic model capturing essential structural dynamics and deterioration is developed to ensure robust parameter identification using support vector machines. Method performance is validated against both numerical and experimental data of a scaled 12-story reinforced concrete nonlinear structure. Results of numerical validation show an average error of 1.5% across 18 structural parameters from hysteresis loop analysis and an average error of 2.0% over 30 identified model parameters from support vector machines in the presence of 10% added root-mean-square noise. Validation using experimental data of the scale test reinforced concrete structure also shows a good match of identified hysteresis loop analysis and predicted nonlinear stiffness changes using the digital clones created with an average difference of 1.4%. More importantly, the predicted response using the digital clones for the highly nonlinear pinched hysteretic behavior matches the measured response well, with the average correlation coefficient Rcoeff = 0.92 and average root-mean-square error of 4.6% across all cases. The overall approach takes structural health monitoring from a tool providing retrospective damage data into automated prospective prediction analysis by “cloning” the structure using computational modeling, which in turn allows optimized decision-making using existing risk analyses and tools.

scholarly journals Structural Health Monitoring Using Machine Learning and Cumulative Absolute Velocity Features

2021 ◽  
Vol 11 (12) ◽  
pp. 5727
Author(s):  
Sifat Muin ◽  
Khalid M. Mosalam
Keyword(s):  
Machine Learning ◽  
Logistic Regression ◽  
Structural Health Monitoring ◽  
Health Monitoring ◽  
Degree Of Freedom ◽  
Absolute Velocity ◽  
Support Vector ◽  
Damage State ◽  
Structural Health ◽  
Cumulative Absolute Velocity

Machine learning (ML)-aided structural health monitoring (SHM) can rapidly evaluate the safety and integrity of the aging infrastructure following an earthquake. The conventional damage features used in ML-based SHM methodologies face the curse of dimensionality. This paper introduces low dimensional, namely, cumulative absolute velocity (CAV)-based features, to enable the use of ML for rapid damage assessment. A computer experiment is performed to identify the appropriate features and the ML algorithm using data from a simulated single-degree-of-freedom system. A comparative analysis of five ML models (logistic regression (LR), ordinal logistic regression (OLR), artificial neural networks with 10 and 100 neurons (ANN10 and ANN100), and support vector machines (SVM)) is performed. Two test sets were used where Set-1 originated from the same distribution as the training set and Set-2 came from a different distribution. The results showed that the combination of the CAV and the relative CAV with respect to the linear response, i.e., RCAV, performed the best among the different feature combinations. Among the ML models, OLR showed good generalization capabilities when compared to SVM and ANN models. Subsequently, OLR is successfully applied to assess the damage of two numerical multi-degree of freedom (MDOF) models and an instrumented building with CAV and RCAV as features. For the MDOF models, the damage state was identified with accuracy ranging from 84% to 97% and the damage location was identified with accuracy ranging from 93% to 97.5%. The features and the OLR models successfully captured the damage information for the instrumented structure as well. The proposed methodology is capable of ensuring rapid decision-making and improving community resiliency.

Sign in / Sign up

Export Citation Format

Share Document