Citation: Heesch, M.;
Dziendzikowski, M.; Mendrok, K.;
Dworakowski, Z. Diagnostic-Quality
Guided Wave Signals Synthesized
Using Generative Adversarial Neural
Networks. Sensors 2022, 22, 3848.
https://doi.org/10.3390/s22103848
Academic Editors: Kim Phuc Tran,
Athanasios Rakitzis and Khanh T. P.
Nguyen
Received: 26 April 2022
Accepted: 18 May 2022
Published: 19 May 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Diagnostic-Quality Guided Wave Signals Synthesized Using
Generative Adversarial Neural Networks
Mateusz Heesch
1
, Michał Dziendzikowski
2
, Krzysztof Mendrok
1
and Ziemowit Dworakowski
1,
*
1
Department of Robotics and Mechatronics, AGH University of Science and Technology, Al. A. Mickiewicza 30,
30-059 Krakow, Poland; heesch@agh.edu.pl (M.H.); mendrok@agh.edu.pl (K.M.)
2
Airworthiness Division, Air Force Institute of Technology, ul. Ks. Boleslawa 6, 01-496 Warsaw, Poland;
michal.dziendzikowski@itwl.pl
* Correspondence: zdw@agh.edu.pl
Abstract:
Guided waves are a potent tool in structural health monitoring, with promising machine
learning algorithm applications due to the complexity of their signals. However, these algorithms
usually require copious amounts of data to be trained. Collecting the correct amount and distribution
of data is costly and time-consuming, and sometimes even borderline impossible due to the necessity
of introducing damage to vital machinery to collect signals for various damaged scenarios. This
data scarcity problem is not unique to guided waves or structural health monitoring, and has been
partly addressed in the field of computer vision using generative adversarial neural networks. These
networks generate synthetic data samples based on the distribution of the data they were trained on.
Though there are multiple researched methods for simulating guided wave signals, the problem is
not yet solved. This work presents a generative adversarial network architecture for guided waves
generation and showcases its capabilities when working with a series of pitch-catch experiments
from the OpenGuidedWaves database. The network correctly generates random signals and can
accurately reconstruct signals it has not seen during training. The potential of synthetic data to be
used for training other algorithms was confirmed in a simple damage detection scenario, with the
classifiers trained exclusively on synthetic data and evaluated on real signals. As a side effect of the
signal reconstruction process, the network can also compress the signals by 98.44% while retaining
the damage index information they carry.
Keywords: guided waves; structural health monitoring; neural networks
1. Introduction
1.1. Guided Waves in Structural Health Monitoring
Structural health monitoring (SHM) is widely researched [
1
–
3
], as the viability of
continuously monitoring an object offers many potential benefits, such as cutting down
maintenance costs by ensuring the maintenance schedule is based on the actual condition of
the elements and early detection of material deterioration and damage in various structures.
One such widely-researched method is based on guided wave (GW) signal analysis
[4–8]
,
in which the propagation of transducer-induced vibration along natural boundaries of
a structure is investigated for signs of fatigue or damage. This method has numerous
strong points, such as being able to monitor large structures with relatively few transducers
and its sensitivity to small changes [
4
]. Unfortunately, analyzing these output signals is
a non-trivial task due to the complexity of the output. This naturally motivates research
in applying the latest advances in machine learning for processing and interpreting these
signals [
9
–
14
]. However, machine learning generally requires vast amounts of data to be
appropriately trained.
Sensors 2022, 22, 3848. https://doi.org/10.3390/s22103848 https://www.mdpi.com/journal/sensors
