The use of pseudo-faults for damage location in SHM: An experimental investigation on a Piper Tomahawk aircraft wing

Abstract

The application of pattern recognition-based approaches in damage localisation and quantification will eventually require the use of some kind of supervised learning algorithm. The use, and most importantly, the success of such algorithms will depend critically on the availability of data from all possible damage states for training. It is perhaps well known that the availability of damage data through destructive means cannot generally be afforded in the case of high value engineering structures outside laboratory conditions. This paper presents the attempt to use added masses in order to identify features suitable for training supervised learning algorithms and then subsequently test the trained classifiers with damage data, with the ultimate purpose of damage localisation. In order to test the approach of adding masses, two separate cases of a dual-class classification problem, representing two distinct locations, and a three-class problem representing three distinct locations, are examined with the help of a full-scale aircraft wing. It was found that an excellent rate of correct classification could be achieved in both the dual-class and threeclass cases. However, it was also found that the rate of correct classification was sensitive to the choices made in training the supervised learning algorithm. The results for the dual-class problem demonstrated a comparatively high level of robustness to these choices with a substantially lower robustness found in the three-class case.

A security risk plan search assistant decision algorithm using deep neural network combined with two-stage similarity calculation

“…These three pictures are trained in three identical and shared weights and biases networks through these three channels. In the training process, the network is trained by Btelling^which two samples belong to the same type and which two samples belong to the different type in the three images input by the network [15,16]. When the recognition result of the network is similar to that of the same sample, the network will be Bpunished^to train the weights and biases in the network structure.…”

Optimal statistical damage detection and classification in an experimental wind turbine blade using minimum instrumentation

“…Therefore, damage was simulated nondestructively by attaching small masses at these locations, see Figure 2. Papatheou et al introduced such pseudo-faults in SHM studies on a simple plate structure 26 as well as on an aircraft wing 27 . The masses were varied between 10 g and 100 g in the different locations, as detailed in Table 1.…”

Structural Health Monitoring Suitable for Airborne Components Using the Speckle Pattern in Plastic Optical Fibers

“…Thus, to ensure a safe operation it is essential to conduct periodic inspections to the structure integrity, using external equipment's, and, if necessary, replace damaged components. Future tendency will be to replace completely the inspections by continuous structural health monitoring using on-board systems (implemented within the aircraft), increasing security and confidence in the systems and reducing the periodic inspection cycles, eliminating the costs of unplanned maintenance on new aircrafts and ensuring greater longevity of the components [6], [7]. For these SHM systems, the inspections will be conducted using non-destructive techniques, being the sensors and actuators permanently attached to the structure, collecting data continuously during the entire service life of the structure, without operator intervention, so that an integrity evaluation could be performed in real time and if necessary carrying out repairs in a timely manner [4].…”