Search

This paper investigates the use of deep transfer learning based on convolutional neural networks (CNNs) to monitor the condition of bolted joints using acoustic emissions. Bolted structures are critical components in many mechanical systems, and the ability to monitor their condition status is crucial for effective structural health monitoring. We evaluated the performance of our methodology using the ORION-AE benchmark, a structure composed of two thin beams connected by three bolts, where highly noisy acoustic emission measurements were taken to detect changes in the applied tightening torque of the bolts. The data used from this structure is derived from the transformation of acoustic emission data streams into images using continuous wavelet transform, and leveraging pretrained CNNs for feature extraction and denoising. Our experiments compared single-sensor versus multiple-sensor fusion for estimating the tightening level (loosening) of bolts and evaluated the use of raw versus prefiltered data on the performance. We particularly focused on the generalization capabilities of CNN-based transfer learning across different measurement campaigns and we studied ordinal loss functions to penalize incorrect predictions less severely when close to the ground truth, thereby encouraging misclassification errors to be in adjacent classes. Network configurations as well as learning rate schedulers are also investigated, and super-convergence is obtained, i.e., high classification accuracy is achieved in a few number of iterations with different networks. Furthermore, results demonstrate the generalization capabilities of CNN-based transfer learning for monitoring bolted structures by acoustic emission with varying amounts of prior information required during training.

Structural Health Monitoring (SHM) relies on non-destructive techniques such as Acoustic Emission (AE) which provide a large amount of data over the life of the systems. The analysis of these data is often based on clustering in order to get insights about damage evolution. In order to evaluate clustering results, current approaches include Clustering Validity Indices (CVI) which favor compact and separable clusters. However, these shape-based criteria are not specific to AE data and SHM. This paper proposes a new approach based on the sequentiality of clusters onsets. For monitoring purposes, onsets indicate when potential damage occurs for the first time and allows to detect the inititation of the defects. The proposed CVI relies on the Kullback-Leibler divergence and enables to incorporate prior on damage onsets when available. Three experiments on real-world data sets demonstrate the relevance of the proposed approach. The first benchmark concerns the detection of the loosening of bolted plates under vibration. The proposed onset-based CVI outperforms the standard approach in terms of both cluster quality and accuracy in detecting changes in loosening. The second application involves micro-drilling of hard materials using Electrical Discharge Machining. In this industrial application, it is demonstrated that the proposed CVI can be used to evaluate the electrode progression until the reference depth which is essential to ensure structural integrity. Lastly, the third application is about the damage monitoring in a composite/metal hybrid joint structure. As an important result, the timeline of clusters generated by the proposed CVI is used to draw a scenario that accounts for the occurrence of slippage leading to a critical failure., Comment: Submitted in May 2023 to a journal

The data set presented in this work, called ORION-AE, is made of raw AE data streams collected by three different AE sensors and a laser vibrometer during five campaigns of measurements by varying the tightening conditions of two bolted plates submitted to harmonic vibration tests. With seven different operating conditions, this data set was designed to challenge supervised and unsupervised machine/deep learning as well as signal processing methods which are developed for material characterization or Structural Health Monitoring (SHM). One motivation of this work was to create a common benchmark for comparing data-driven methods dedicated to AE data interpretation. The data set is made of time-series collected during an experiment designed to reproduce the loosening phenomenon observed in aeronautics, automotive or civil engineering structures where parts are assembled together by means of bolted joints. Monitoring loosening in jointed structures during operation remains challenging because contact and friction in bolted joints induce a nonlinear stochastic behavior. ORION-AE data set is available on a shared directory for Research Data, Harvard Dataverse (https://doi.org/10.7910/DVN/FBRDU0). A Matlab code is provided to extract the data stream from each sensor. This article is a preprint, the published version is available at https://doi.org/10.3390/data7030031., Comment: Version 2

The interpretation of unlabeled acoustic emission (AE) data classically relies on general-purpose clustering methods. While several external criteria have been used in the past to select the hyperparameters of those algorithms, few studies have paid attention to the development of dedicated objective functions in clustering methods able to cope with the specificities of AE data. We investigate how to explicitly represent clusters onsets in mixture models in general, and in Gaussian Mixture Models (GMM) in particular. By modifying the internal criterion of such models, we propose the first clustering method able to provide, through parameters estimated by an expectation-maximization procedure, information about when clusters occur (onsets), how they grow (kinetics) and their level of activation through time. This new objective function accommodates continuous timestamps of AE signals and, thus, their order of occurrence. The method, called GMMSEQ, is experimentally validated to characterize the loosening phenomenon in bolted structure under vibrations. A comparison with three standard clustering methods on raw streaming data from five experimental campaigns shows that GMMSEQ not only provides useful qualitative information about the timeline of clusters, but also shows better performance in terms of cluster characterization. In view of developing an open acoustic emission initiative and according to the FAIR principles, the datasets and the codes are made available to reproduce the research of this paper.

[This paper was initially published in PHME conference in 2016, selected for further publication in International Journal of Prognostics and Health Management.] This paper describes an Autoregressive Partially-hidden Markov model (ARPHMM) for fault detection and prognostics of equipments based on sensors' data. It is a particular dynamic Bayesian network that allows to represent the dynamics of a system by means of a Hidden Markov Model (HMM) and an autoregressive (AR) process. The Markov chain assumes that the system is switching back and forth between internal states while the AR process ensures a temporal coherence on sensor measurements. A sound learning procedure of standard ARHMM based on maximum likelihood allows to iteratively estimate all parameters simultaneously. This paper suggests a modification of the learning procedure considering that one may have prior knowledge about the structure which becomes partially hidden. The integration of the prior is based on the Theory of Weighted Distributions which is compatible with the Expectation-Maximization algorithm in the sense that the convergence properties are still satisfied. We show how to apply this model to estimate the remaining useful life based on health indicators. The autoregressive parameters can indeed be used for prediction while the latent structure can be used to get information about the degradation level. The interest of the proposed method for prognostics and health assessment is demonstrated on CMAPSS datasets.

In the context of human action recognition in video sequences, a temporal belief filter is presented. It allows to cope with human action disparity and low quality videos. The whole system of action recognition is based on the Transferable Belief Model (TBM) proposed by P. Smets. The TBM allows to explicitly model the doubt between actions. Furthermore, the TBM emphasizes the conflict which is exploited for action recognition. The filtering performance is assessed on real video sequences acquired by a moving camera and under several unknown view angles., Comment: 8 pages