A neural network approach for structural identification and diagnosis of a building from seismic response data

Huang, Chiung‐Shiann; Hung, Shan Ling; Wen, Cheng-Hao; Tu, Tse-Yi

doi:10.1002/eqe.219

Cited by 111 publications

(51 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Learning mechanisms of the brain are based primarily on experience, and the extraordinary power of the brain to absorb these experiences originates from the presence of a tremendous number of neurons and their natural connections. The core principles of ANN models are based on a similar logic [6][7][8][9][10][11]. A schematic representation of neurons in a network is shown in Fig.…”

Section: Artificial Neural Networkmentioning

confidence: 99%

Soft Computing-based Approach on Capacity Prediction of FRP Strengthened RC Joints

2018

View full text Add to dashboard Cite

Abstract. Shear failure of the RC beam-column joints is a brittle failure which has no prior warning and can induce tremendous damages because of collapse of column and joint before the connected beam. This paper is focused on one particular method of strengthening the RC joints, that is, the use of FRP composites as confining element. The results of previous studies have shown that strengthening the RC beam-column joints with FRP composites can improve their shear capacity. In this study, the data collected from the existing standards and studies regarding the FRP strengthened RC joints were used to develop an artificial neural network model for predicting the shear strength contribution of FRP jacket. The developed model was then used to evaluate the role of different parameters on this contribution, and finally derive a formula for contribution of FRP jacket to the shear strength of the RC beam-column joints.

show abstract

Section: Artificial Neural Networkmentioning

confidence: 99%

Soft Computing-based Approach on Capacity Prediction of FRP Strengthened RC Joints

2018

View full text Add to dashboard Cite

show abstract

“…Then, they calculated the dynamic characteristics of the system directly from the coefficient matrices of the ARV model by adopting the concept behind Ibrahim's system identification technique. Huang, Hung, Wen, and Tu (2003) presented a procedure for identifying the dynamic characteristics of a steel frame using a back-propagation neural network. The dynamic characteristics were directly evaluated from the weighting matrices of the neural network trained by the observed acceleration responses and base excitation input.…”

Section: Introductionmentioning

confidence: 99%

Model updating of multistory shear buildings for simultaneous identification of mass, stiffness and damping matrices using two different soft-computing methods

Khanmirza¹,

Khaji²,

Majd³

2011

Expert Systems with Applications

View full text Add to dashboard Cite

“…The SI process is only for the structural parameters, it is called parameter identification. A great amount of researches with various methodologies have been conducted in parameter identification, such as artificial neural network approach (ANN) [1], wavelet analysis method [2], Fourier transform based method [3], finite element-based Iterative Least-squares methods (ILS) [4,5,6], frequency domain decomposition (FDD) [7], natural excitation technique coupled with eigen-system realization algorithm (NExT-ERA) [8], Random decrement technique (RDT) [9], extended Kalman filter technique (EKF) [10,11], and so on.…”

Section: Introductionmentioning

confidence: 99%

Structural Parameter Identification in Time Domain using Generalized Hybrid Inversion Method

Wang¹,

Cui²

2016

Proceedings of the 2015 4th International Conference on Sensors, Measurement and Intelligent Materials

View full text Add to dashboard Cite

A structural parameter identification method without input information and with limited observations is proposed to identify the stiffness parameters and damping coefficient, to inverse the input information and to estimate the unmeasured responses in time domain. The unique feature of this technique is that it requires only the displacement responses of substructure. This method has two steps: 1) structural physical parameter identification of substructure and input inversion using hybrid inversion method (HIM) with introducing the combination method of the modified LevenbergMarquadt and the singular value decomposition (SVD-mLM) and rectangular window method, and 2) structural physical parameter identification of whole structure and estimation of unmeasured responses using the extended Kalman filter method with weighted global iteration (EKF-WGI) coupled with memory fading filter (MFF) technique and Runge-Kutta (R-K) method. The improved method is called as generalized hybrid inversion method (GHIM) by the authors. The parameter identification and the response estimation are carried out using the proposed method by a numerical example. Three noise contaminated cases of the displacement responses are considered. The robustness, efficiency and accuracy of the new technique are verified by the results.

show abstract

A neural network approach for structural identification and diagnosis of a building from seismic response data

Cited by 111 publications

References 17 publications

Soft Computing-based Approach on Capacity Prediction of FRP Strengthened RC Joints

Soft Computing-based Approach on Capacity Prediction of FRP Strengthened RC Joints

Model updating of multistory shear buildings for simultaneous identification of mass, stiffness and damping matrices using two different soft-computing methods

Structural Parameter Identification in Time Domain using Generalized Hybrid Inversion Method

Contact Info

Product

Resources

About