Unsupervised deep learning framework for ultrasonic-based distributed damage detection in concrete: integration of a deep auto-encoder and Isolation Forest for anomaly detection
Restricted accessResearch articleFirst published online May, 2024
Unsupervised deep learning framework for ultrasonic-based distributed damage detection in concrete: integration of a deep auto-encoder and Isolation Forest for anomaly detection
This study presented an unsupervised anomaly detection-based framework for distributed damage detection in concrete using ultrasonic response signals. A deep fully connected auto-encoder was employed to reconstruct the ultrasonic response signals. This model was trained on the intact specimen’s responses. The auto-encoder demonstrated a relatively high prediction error encountering the damaged specimen’s responses. Two time-domain features (mean squared error and reconstructed-to-original signal ratio) and one frequency-domain feature (fundamental amplitude ratio) were defined to measure the reconstruction error of the auto-encoder (the damage-sensitive features). Finally, the Isolation Forest algorithm was implemented for anomaly (damage) detection. The beauty of this framework is that it requires a few numbers of data only from the intact specimen for training the auto-encoder and collecting the binary decision trees of the Isolation Forest. The framework was successfully implemented for damage detection in five geopolymer concrete specimens with different mix proportions. Using all three introduced damage-sensitive features, the framework demonstrated an average prediction accuracy of 95.0% and 93.0% for damaged and intact stages, respectively.
ZhangSJiaZXiongY, et al. Wave amplitude of embedded ultrasonic transducer-based damage monitoring of concrete due to steel bar corrosion. Struct Health Monitor2022; 21: 1694–1709.
2.
AliRChaY-J. Attention-based generative adversarial network with internal damage segmentation using thermography. Autom Constr2022; 141: 104412.
3.
FröjdPUlriksenP. Detecting damage events in concrete using diffuse ultrasound structural health monitoring during strong environmental variations. Struct Health Monitor2018; 17: 410–419.
4.
OstachowiczWSomanRMalinowskiP. Optimization of sensor placement for structural health monitoring: a review. Struct Health Monitor2019; 18: 963–988.
5.
DongC-ZCatbasFN. A review of computer vision–based structural health monitoring at local and global levels. Struct Health Monitor2021; 20; 692–743.
6.
RautelaMGopalakrishnanS. Ultrasonic guided wave based structural damage detection and localization using model assisted convolutional and recurrent neural networks. Exp Syst Appl2021; 167: 114189.
7.
RautelaMSenthilnathJMollJ, et al. Combined two-level damage identification strategy using ultrasonic guided waves and physical knowledge assisted machine learning. Ultrasonics2021; 115: 106451.
8.
SoleimanianEToufighV. Assessment of plain and glass fiber-reinforced concrete under impact loading: a new approach via ultrasound evaluation. J Nondestr Eval38 (2019) 1–8.
9.
GhoshRSagarSPKumarA, et al. Estimation of geopolymer concrete strength from ultrasonic pulse velocity (UPV) using high power pulser. J Build Eng2018; 16: 39–44.
10.
AshrafianAAmiriMJTRezaie-BalfM, et al. Prediction of compressive strength and ultrasonic pulse velocity of fiber reinforced concrete incorporating nano silica using heuristic regression methods. Constr Build Mater2018; 190: 479–494.
11.
AshkarMToufighVGhaemianM, et al. Application of ultrasonic models for investigating the properties of the interface between roller compacted concrete (RCC) layers. NDT & E Int2021; 124: 102516.
12.
KamadaTNagatakiSIwanamiM. Evaluation of material deterioration in concrete by nondestructive testing methods. In: International conference on engineering materials, Thessaloniki, Greece, JSCE, vol. 95, 1997, pp. 453–466.
13.
AntonaciPBrunoCLEGliozziA, et al. Monitoring evolution of compressive damage in concrete with linear and nonlinear ultrasonic methods. Cem Concr Res2010; 40: 1106–1113.
14.
ShahARibakovY. Effectiveness of nonlinear ultrasonic and acoustic emission evaluation of concrete with distributed damages. Mater Des2010; 31: 3777–3784.
15.
EirasJKunduTBonillaM, et al. Nondestructive monitoring of ageing of alkali resistant glass fiber reinforced cement (GRC). J Nondestr Eval2013; 32: 300–314.
16.
BasuSThirumalaiselviASasmalS, et al. Nonlinear ultrasonics-based technique for monitoring damage progression in reinforced concrete structures. Ultrasonics2021; 115: 106472.
17.
ShahARibakovYZhangC. Efficiency and sensitivity of linear and non-linear ultrasonics to identifying micro and macro-scale defects in concrete. Mater Des2013; 50: 905–916.
18.
KimGLoretoGKimJ-Y, et al. In situ nonlinear ultrasonic technique for monitoring microcracking in concrete subjected to creep and cyclic loading. Ultrasonics2018; 88: 64–71.
19.
ZhaoMNieZWangK, et al. Nonlinear ultrasonic test of concrete cubes with induced crack. Ultrasonics2019; 97: 1–10.
20.
ChenJWuYYangC, Damage assessment of concrete using a non-contact nonlinear wave modulation technique. NDT & E Int2019; 106: 1–9.
21.
SepehrinezhadAToufighV. The evaluation of distributed damage in concrete based on sinusoidal modeling of the ultrasonic response. Ultrasonics2018; 89: 195–205.
22.
LeCunYBengioYHintonG. Deep learning. Nature2015; 521: 436–444.
23.
YeJToyamaN. Automatic defect detection for ultrasonic wave propagation imaging method using spatio-temporal convolution neural networks. Struct Health Monitor2022; 21(6): 2750–2767.
24.
RanjbarIToufighVBoroushakiM. A combination of deep learning and genetic algorithm for predicting the compressive strength of high-performance concrete. Struct Concr2022; 23(4): 2405–2418.
25.
SonodaYLuCYinY. Basic research on usefulness of convolutional autoencoders in detecting defects in concrete using hammering sound. Struct Health Monitor2022: 14759217221122296.
26.
XuYJinR. Measurement of reinforcement corrosion in concrete adopting ultrasonic tests and artificial neural network. Constr Build Mater2018; 177: 125–133.
27.
Tenza-AbrilAJVillacampaYSolakAM, et al. Prediction and sensitivity analysis of compressive strength in segregated lightweight concrete based on artificial neural network using ultrasonic pulse velocity. Constr Build Mater2018; 189: 1173–1183.
28.
ShahAAAlsayedSHAbbasH, et al. Predicting residual strength of non-linear ultrasonically evaluated damaged concrete using artificial neural network. Constr Build Mater2012; 29: 42–50.
29.
RanjbarIToufighV. Deep long short-term memory (LSTM) networks for ultrasonic-based distributed damage assessment in concrete. Cem Concr Res2022; 162: 107003.
TzengEHoffmanJSaenkoK, et al. Adversarial discriminative domain adaptation. In: Proceedings of the IEEE conference on computer vision and pattern recognition, Honolulu, HI, USA2017, pp. 7167–7176.
32.
DiezAKhoaNLDMakki AlamdariM, et al. A clustering approach for structural health monitoring on bridges. J Civil Struct Health Monitor2016; 6: 429–445.
33.
RafieiMHAdeliH. A novel unsupervised deep learning model for global and local health condition assessment of structures. Eng Struct2018; 156: 598–607.
34.
MaoJWangHSpencer JrBF. Toward data anomaly detection for automated structural health monitoring: exploiting generative adversarial nets and autoencoders. Struct Health Monitor2021; 20: 1609–1626.
35.
WangZChaY-J. Unsupervised deep learning approach using a deep auto-encoder with a one-class support vector machine to detect damage. Struct Health Monitor2021; 20: 406–425.
36.
RautelaMSenthilnathJMonacoE, et al. Delamination prediction in composite panels using unsupervised-feature learning methods with wavelet-enhanced guided wave representations. Compos Struct2022; 291: 115579.
37.
EybpooshMBergesMNohHY. An energy-based sparse representation of ultrasonic guided-waves for online damage detection of pipelines under varying environmental and operational conditions. Mech Syst Signal Process2017; 82: 260–278.
38.
ShinTSunHZhuJ, et al. Nondestructive damage detection of concrete with alkali-silica reactions using coda wave and anomaly detection. IEEE Sens J2022; 22: 6124–6135.
39.
KramerMA. Nonlinear principal component analysis using autoassociative neural networks. AIChE J1991; 37: 233–243.
40.
HintonGESalakhutdinovRR. Reducing the dimensionality of data with neural networks. Science2006; 313: 504–507.
41.
VincentPLarochelleHLajoieI, et al. Stacked denoising autoencoders: learning useful representations in a deep network with a local denoising criterion. J Mach Learn Res2010; 11: 3371–3408.
42.
BengioYLamblinPPopoviciD, et al. Greedy layer-wise training of deep networks. Adv Neural Inform Processing systems2006; 19: 153–160.
43.
ChowJKSuZWuJ, et al. Anomaly detection of defects on concrete structures with the convolutional autoencoder. Adv Eng Inform2020; 45: 101105.
44.
NguyenHTranKPThomasseyS, et al. Forecasting and anomaly detection approaches using lstm and lstm autoencoder techniques with the applications in supply chain management. Int J Inform Manage2021; 57: 102282.
45.
HaJMSeungHMChoiW. Autoencoder-based detection of near-surface defects in ultrasonic testing. Ultrasonics2022; 119: 106637.
46.
RuderS. An overview of gradient descent optimization algorithms, arXiv preprint arXiv:1609.04747, 2016, https://arxiv.org/abs/1609.04747
47.
KingmaDPBaJ. Adam: a method for stochastic optimization, arXiv preprint arXiv:1412.6980, 2014, https://arxiv.org/abs/1412.6980
48.
LiuFTTingKMZhouZ-H. Isolation forest, in: 2008 eighth ieee international conference on data mining. Pisa, Italy: IEEE, 2008, pp. 413–422.
49.
HeZXuXDengS. Discovering cluster-based local outliers. Pattern Recogn Lett2003; 24: 1641–1650.
50.
AbeNZadroznyBLangfordJ. Outlier detection by active learning. In: Proceedings of the 12th ACM SIGKDD international conference on Knowledge discovery and data mining, Philadelphia, PA, USA, 2006, pp. 504–509.
51.
RousseeuwPJDriessenKV. A fast algorithm for the minimum covariance determinant estimator. Technometrics1999; 41: 212–223.
52.
PreissBR. Data structure and algorithms: with object-oriented design patterns in Java. New York, NY: John Wiley & Sons, 1999.
53.
SchubertFKöhlerB. Three-dimensional time domain modeling of ultrasonic wave propagation in concrete in explicit consideration of aggregates and porosity. J Comput Acoust2001; 9: 1543–1560.
54.
PetitjeanFKetterlinAGançarskiP. A global averaging method for dynamic time warping, with applications to clustering. Pattern Recogn2011; 44: 678–693.
SchölkopfBPlattJCShawe-TaylorJ, et al. Estimating the support of a high-dimensional distribution. Neural Comput2001; 13: 1443–1471.
59.
PearsonK. Liii on lines and planes of closest fit to systems of points in space. Lond. Edinb. Dublin Philos Mag J Sci1901; 2: 559–572.
60.
ShangZSunLXiaY, et al. Vibration-based damage detection for bridges by deep convolutional denoising autoencoder. Struct Health Monitor2021; 20: 1880–1903.
61.
WuXZhangYChengC, et al. A hybrid classification autoencoder for semi-supervised fault diagnosis in rotating machinery. Mech Syst Signal Process2021; 149: 107327.
62.
ZhengSZhaoJ. A new unsupervised data mining method based on the stacked autoencoder for chemical process fault diagnosis. Comput Chem Eng2020; 135: 106755.
