Unsupervised deep learning approach using a deep auto-encoder with a one-class support vector machine to detect damage

Abstract

<jats:p> This article proposes an unsupervised deep learning–based approach to detect structural damage. Supervised deep learning methods have been proposed in recent years, but they require data from an intact structure and various damage scenarios of monitored structures for their training processes. However, the labeling work on the training data is typically time-consuming and costly, and sometimes collecting sufficient training data from various damage scenarios of infrastructures in service is impractical. In this article, the proposed unsupervised deep learning method based on a deep auto-encoder with an one-class support vector machine only uses the measured acceleration response data acquired from intact or baseline structures as training data, which enables future structural damage to be detected. The major contributions and novelties of the proposed method are as follows. First, an appropriate deep auto-encoder is carefully designed through comparative studies on the depth of neural networks. Second, the designed deep auto-encoder is taken as an extractor to obtain damage-sensitive features from the measured acceleration response data, and an one-class support vector machine is used as a damage detector. Third, experimental and numerical studies validate the high accuracy of the proposed method for damage detection: a 97.4% mean average for a 12-story numerical building model and a 91.0% accuracy for a laboratory-scaled steel bridge. Fourth, the proposed method also detects light damage (i.e. a 10% reduction in stiffness) with 96.9% to 99.0% accuracy, which shows its superior performance compared with the current state of the art. Fifth, it provides stable and more robust damage detection performance with reduced tuning parameters. </jats:p>

A Deep Learning-Based Framework for Damage Detection With Time Series

“…A one-class support vector machine (OC-SVM) is then served as the damage detector on the basis of the extracted features. Wang and Cha [24] also propose a damage detection method integrated with an autoencoder and an OC-SVM. The autoencoder is designed to acquire features related to the damage information from the acceleration signals, while the OC-SVM is a classifier that determines damage states of samples.…”

“…After each excitation, the building is exposed to the same random white noise to record its ambient response, which is supposed to significantly vary as excitation intensity increases. [24]. Given that the frequency is proportional to √ [33]- [34], where is the stiffness of building, therefore, the stiffness of building degrades by 23.1%, 34.5%, and 44.3% after excitations with 25%, 50%, and 100% intensities compared with the undamaged state.…”

Development of FE Model Updating for Three-Story Building considering Soil-Structure Interaction

“…이러한 대형구조물의 사용연한 증가는 구조물의 유지 보수 및 성능 평가에 대한 중요성을 부각시켰 으며, 이에 따라 성능 평가를 위한 정확한 모델 개발의 필요성 이 증가하고 있다 (Gong and Park, 2019). 2015년 서해대교에 서 낙뢰로 인해 케이블이 파단되었을 때, 현재 상태를 반영하 는 사용 가능한 유한요소모델의 개발에 72시간 이상의 시간 이 소요된 것으로 파악되었으며 (Gil et al, 2016), 소록대교의 구조식별 작업에 준비 단계를 제외한 순수 계측 및 분석에 약 2일이 소요되었고, 모델 업데이팅 작업의 최적화 프로그 램 실행에 약 2일이 소요된 것으로 알려져 있다 (Sung, 2018 (Lin et al, 2017;Tran-Ngoc et al, 2018;Yoon et al, 2018;Lim and Yoon, 2019;Wang and Cha, 2020).…”

Damage‐sensitive feature extraction with stacked autoencoders for unsupervised damage detection

“…In this context, focusing on the feature classification for damage detection, AE models have been used for automatic and semi‐automatic feature extraction. In Wang and Cha, 36 Wang et al utilized a seven‐layer autoencoder trained with undamaged acceleration data as an automatic extractor of damage‐sensitive features and an one‐class support vector machine algorithm (OC‐SVM) for statistical modeling and feature classification. After the training, the test step is carried out with data from unknown health states.…”