Classifying the type of damage occurring within a structure using a structural health monitoring system can allow the end user to assess what kind of repairs, if any, that a component requires. This paper investigates the use of acoustic emission (AE) to locate and classify the type of damage occurring in a composite, carbon fibre panel during buckling. The damage was first located using a bespoke location algorithm developed at Cardiff University, called delta-T mapping. Signals identified as coming from the regions of damage were then analysed using three AE classification techniques; Artificial Neural Network (ANN) analysis, Unsupervised Waveform Clustering (UWC) and corrected Measured Amplitude Ratio (MAR). A comparison of results yielded by these techniques shows a strong agreement regarding the nature of the damage present in the panel, with the signals assigned to two different damage mechanisms, believed to be delamination and matrix cracking. Ultrasonic C-scan images and a digital image correlation (DIC) analysis of the buckled panel were used as validation. MAR’s ability to reveal the orientation of recorded signals greatly assisted the identification of the delamination region, however, ANN and UWC have the ability to group signals into several different classes, which would prove useful in instances where several damage mechanisms were generated. Combining each technique’s individual merits in a multi-technique analysis dramatically improved the reliability of the AE investigation and it is thought that this cross-correlation between techniques will also be the key to developing a reliable SHM system

Acoustic emission is widely used for mechanical diagnostics and to characterise damage in composite materials. Distinction between different damage mechanisms is still one of the major challenges and remains an unresolved issue. The objective of cluster analysis is to separate an acoustic emission data set into multiple classes that reflect different acoustic emission sources. This article is concerned with the implementation of unsupervised clustering techniques to classify acoustic emission transients from a carbon fibre laminate buckling test. A new approach to signal feature extraction was utilised, whereby principal components provide signal features that represent the greatest data variance while remaining linearly uncorrelated with each other; feature selection was undertaken using a hierarchical clustering method and finally a cluster analysis was performed using k-means and Fuzzy C-means techniques. The aim of the work is to reduce the data required in the classification process, thereby reducing the processing time and computational power required, without significantly affecting the classification result. Thus, an approach which is more suited to online processing, allowing fast and efficient processing and storage of data is provided. The proposed unsupervised clustering analysis was able to separate acoustic emission signals into two different clusters that were correlated to the damage mechanisms observed. The results show that the clustering groups have a good fit with ultrasonic C-scan and digital image correlation strain data. The application of a clustering process that uses the most effective acoustic emission features as input data is an objective method, and this investigation shows that it may be a useful complement in the field of non-destructive evaluation.

Structural health monitoring has gained wide appeal for applications with high inspection costs, such as aircraft and wind turbines. As the structures and materials used in these industries evolve, so too must the technologies used to monitor them. Acoustic emission is a passive method of detecting damage which lends itself well to structural health monitoring. One form of acoustic emission monitoring, known as wavestreaming, involves intermittently recording data for set periods of time and using the sequential recordings to detect changes in the state of the structure. However, at present, there is no standard method for selecting appropriate wavestream recording parameters, such as their length or their interval of collection. This article investigates a method of optimising acoustic emission wavestreaming for structural health monitoring purposes by introducing the novel concept of adjoining consecutive discrete acoustic emission hit signals to create synthetic wavestreams. To this end, a pre-notched 492 mm × 67.5 mm × 20 mm, 300M grade steel cantilever specimen was subject to cyclic loading and both acoustic emission hit data and conventional wavestreams were collected as a crack grew in the notched region; crack growth activity was also monitored using digital image correlation for comparison. To demonstrate the proposed optimisation process, four sets of synthetic wavestreams were created from the hit data, 0.25, 0.5, 1.0 and 1.5 s in length, and compared with the 1.5-s-long conventional wavestreams. The activity of the peak frequency and frequency centroid bands of interest within the conventional and synthetic wavestreams were examined to determine whether or not cracking activity could be inferred through them. Across comparisons of all data, it was found that the 0.5-s-long synthetic wavestreams contained enough information to identify the same trends as the conventional wavestreams for this application; thus, the use of synthetic wavestreams as a tool for selecting an appropriate wavestream recording length was demonstrated.

Thermoelastic stress analysis (TSA) is a non-contact technique for measuring the distribution of stress in the surface of a component subject to cyclic loading by using a sensitive infrared camera. The stress concentrations indicative of a crack can be located and tracked using an optical flow method, allowing the position of the crack-tip to be identified at a given time. Acoustic emission (AE) has been used to validate the TSA algorithm. AE events from cracking, located using the Delta-T Mapping method, were detected several seconds before the TSA algorithm first detected cracking; however, TSA provided significantly more accurate location information.