EFFICIENCY OF 1D CNNS IN FINITE ELEMENT MODEL PARAMETER ESTIMATION USING SYNTHETIC DYNAMIC RESPONSES

. A critical step of building a reliable Finite Element (FE) model is closing the gap between the actual and the simulated structural behaviour by utilizing FE model updating methods. Current updating methods typically require hand-crafted features such as natural frequencies and mode shapes. As a result, limitations in identifying these features are becoming also limitations in the updating process. With the ability to utilize raw data, Deep-Learning (DL) algorithms have the potential to avoid such limitations. This paper aims to validate the efficiency of 1D Convolutional Neural Networks (CNNs) in estimating FE model parameters using synthetic acceleration data. The performance of 1D CNNs was tested on a simple FE beam model. The influence of training pair numbers, and acceleration signal duration on the algorithm performance was evaluated. The results indicate that the 1D CNNs have the ability to estimate FE model parameters with high accuracy. Final results indicate that the standard damage detection and the more novel model parameter estimation, through using 1D CNNs, can be practised synchronously with equally good results.

[1]  L. Hao,et al.  Partial discharge source discrimination using a support vector machine , 2010, IEEE Transactions on Dielectrics and Electrical Insulation.

[2]  Hyeonjoon Moon,et al.  Background Information of Deep Learning for Structural Engineering , 2017 .

[3]  Harmesh Kumar,et al.  Structural Dynamic Model Updating Techniques: A State of the Art Review , 2016 .

[4]  Chih-Chen Chang,et al.  Adaptive neural networks for model updating of structures , 2000 .

[5]  Yasunori Miyamori,et al.  Vibration‐based structural state identification by a 1‐dimensional convolutional neural network , 2019, Comput. Aided Civ. Infrastructure Eng..

[6]  Boualem Boashash,et al.  1-D CNNs for structural damage detection: Verification on a structural health monitoring benchmark data , 2018, Neurocomputing.

[7]  Wei Cai,et al.  Lightweight Deep Residual CNN for Fault Diagnosis of Rotating Machinery Based on Depthwise Separable Convolutions , 2019, IEEE Access.

[8]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[9]  John E. Mottershead,et al.  Model Updating In Structural Dynamics: A Survey , 1993 .

[10]  T. M. Wang.  Natural frequencies of continuous Timoshenko beams , 1970 .

[11]  Josh Patterson,et al.  Deep Learning: A Practitioner's Approach , 2017 .

[12]  José L. Zapico,et al.  Finite element model updating of a small steel frame using neural networks , 2008 .

[13]  Izzuddin Zaman,et al.  A review on model updating in structural dynamics , 2015 .

[14]  Yixin Chen,et al.  Multi-Scale Convolutional Neural Networks for Time Series Classification , 2016, ArXiv.

[15]  Ashutosh Bagchi,et al.  System identification and model updating of highway bridges using ambient vibration tests , 2018, Journal of Civil Structural Health Monitoring.

[16]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[17]  M. H. Sadr,et al.  Improving the neural network method for finite element model updating using homogenous distribution of design points , 2007 .

[18]  Pengcheng Jiang,et al.  Intelligent fault diagnosis of rotating machinery based on one-dimensional convolutional neural network , 2019, Comput. Ind..

[19]  Ri Levin,et al.  DYNAMIC FINITE ELEMENT MODEL UPDATING USING NEURAL NETWORKS , 1998 .

[20]  Onur Avci,et al.  Structural Damage Detection in Real Time: Implementation of 1D Convolutional Neural Networks for SHM Applications , 2017 .

[21]  Debahuti Mishra,et al.  Empirical Analysis of Activation Functions and Pooling Layers in CNN for Classification of Diabetic Retinopathy , 2019, 2019 International Conference on Applied Machine Learning (ICAML).

[22]  Guido De Roeck,et al.  Dealing with uncertainty in model updating for damage assessment: A review , 2015 .

[23]  Yi Zheng,et al.  Time Series Classification Using Multi-Channels Deep Convolutional Neural Networks , 2014, WAIM.

[24]  Jing Lin,et al.  A multivariate encoder information based convolutional neural network for intelligent fault diagnosis of planetary gearboxes , 2018, Knowl. Based Syst..

[25]  Yong Lu,et al.  A two-level neural network approach for dynamic FE model updating including damping , 2004 .

[26]  Gokhan Pekcan,et al.  Vibration‐based structural condition assessment using convolution neural networks , 2018, Structural Control and Health Monitoring.

[27]  Namgyu Kim,et al.  Finite element model updating considering boundary conditions using neural networks , 2017 .

[28]  Omer W. Blodgett,et al.  Design of Welded Structures , 2016 .

[29]  Francesco Visin,et al.  A guide to convolution arithmetic for deep learning , 2016, ArXiv.

[30]  Tshilidzi Marwala,et al.  Finite Element Model Updating Using Computational Intelligence Techniques: Applications to Structural Dynamics , 2010 .

[31]  Moncef Gabbouj,et al.  Real-time vibration-based structural damage detection using one-dimensional convolutional neural networks , 2017 .