Case Study: Development of a Fibre-Optic Impact Detection

and Location System for Aerospace Structures

1. Introduction

The increasing use of composite materials in aerospace structures offers numerous advantages,
including low mass and adaptable properties. However, these materials are susceptible to damage
from low-velocity impacts that may leave little visible sign, such as barely visible impact
damage (BVID). Such damage can significantly affect the structural integrity of an aircraft.
Consequently, there is a critical need for effective structural health monitoring (SHM) systems
capable of detecting and assessing impact events in aerospace structures. Traditional
non-destructive evaluation (NDE) and non-destructive inspection (NDI) methods can be slow
and labour-intensive. This has driven the development of embedded or surface-mounted sensor
technologies for continuous monitoring.

Optical fibre sensors, particularly fibre Bragg gratings (FBGs), hold great promise for SHM in
the aerospace and defence industries due to their lightweight nature, immunity to
electromagnetic interference, and potential for multiplexing. This case study examines the
development of a demonstrator system presented in this thesis, which integrates various research
aspects to create a functional FBG-based impact detection and location system for a curved
composite leading edge structure, mimicking a section of an aircraft wing. The aim was to
produce a scalable optical system capable of interrogating multiplexed sensors with sufficient
sensitivity to measure the acoustic waves generated during an impact event and accurately locate
the impact.

2. Background

2.1. Composite Materials and Impact Damage

Composite materials used in aerospace are formed by combining two or more distinct materials,
typically long fibres embedded in a matrix, to achieve enhanced properties. Unlike metals,
composites exhibit anisotropic and inhomogeneous properties, and their damage mechanisms
under impact differ significantly. Low-velocity impacts can cause various forms of damage in
composites, including matrix cracking, delamination (separation between layers), fibre breakage,
and fibre pull-out. Delamination, in particular, can severely reduce structural integrity without
always being visually apparent.

Impact events generate dynamic strain waves that propagate through the structure. Monitoring
these strain waves using embedded or bonded sensors can provide valuable information about
the occurrence, location, and severity of an impact.
2.2. Fibre Bragg Grating (FBG) Sensors

Fig. Diagram of a fibre Bragg grating showing the multiple reflections from the periodic refractive index
modulation of period

FBGs are wavelength-selective optical filters created by a periodic modulation of the refractive
index within the core of an optical fibre. When a broadband light source illuminates an FBG, it
reflects a narrow band of wavelengths centred at the Bragg wavelength ($\lambda_B$), which is
sensitive to changes in strain and temperature. This wavelength shift can be measured to
determine the strain or temperature experienced by the sensor. FBGs are advantageous for SHM
due to their small size, lightweight nature, immunity to electromagnetic interference, and
capability for wavelength division multiplexing (WDM), allowing multiple sensors with
different Bragg wavelengths to be interrogated along a single fibre.

2.3. FBG Interrogation Techniques

To extract the strain or temperature information from FBG sensors, an interrogation system is
required to accurately measure the shift in the reflected Bragg wavelength. Various interrogation
techniques exist, including using tunable lasers, optical spectrum analysers, and interferometric
methods. For dynamic strain measurements, techniques offering high temporal bandwidth are
necessary.

Fig. Arrayed Waveguide Grating

One promising interrogation approach involves using an arrayed waveguide grating (AWG). An
AWG is a passive optical device that diffracts incoming light based on its wavelength, spatially
separating different wavelengths into distinct output channels. By aligning the reflected spectrum
of an FBG with the transmission spectrum of one or more AWG channels, a shift in the Bragg
wavelength due to strain or temperature will cause a change in the light intensity transmitted
through these channels. By monitoring the intensity changes in adjacent AWG channels, a
ratiometric measurement can be performed, which can reduce the sensitivity to overall intensity
fluctuations.

3. The Challenge: Developing a Scalable and Sensitive Impact Detection and

Location System

The development of an effective impact detection and location system for aerospace structures
using FBG sensors faces several challenges:

●​ Sensitivity to Dynamic Strain: Impact events generate rapid, transient strain waves with
frequency content ranging from tens of Hz to approximately 200 kHz for low-velocity
impacts. The interrogation system must have sufficient temporal bandwidth to capture
these dynamic signals accurately.
●​ Multiplexing Capability: Large aerospace structures require monitoring over extensive
areas, necessitating the use of numerous sensors. The interrogation system should be
capable of simultaneously interrogating multiple FBG sensors to provide comprehensive
coverage without adding excessive complexity or weight.
●​ Accurate Impact Location: Determining the location of an impact is crucial for targeted
inspection and repair. Algorithms are needed to process the signals from multiple sensors
to estimate the impact location with sufficient accuracy.
●​ Environmental Factors: Aerospace environments are harsh, with large temperature
ranges, vibration, and mechanical shock. The sensor system and interrogation unit must
be robust and stable under these conditions.
●​ Scalability and Practicality: The system should be scalable for deployment on various
aircraft sizes and types. It should also be practical in terms of weight, power
consumption, cost, and ease of integration and maintenance.
4. Proposed Solution: The Demonstrator System

To address these challenges, the thesis developed and implemented a demonstrator system that
combined an AWG-based FBG interrogator with neural network-based impact location
algorithms. The system was designed to detect and locate impacts on a curved composite leading
edge structure.

Fig. Optical diagram for the demonstrator system featuring an amplified spontaneous emission broadband
source (ASE), four multiplexed FBG sensors, semiconductor optical amplifier (SOA), arrayed waveguide
grating (AWG) and set of eight fibre coupled photodetectors (PDs).

4.1. System Components

The demonstrator system comprised the following key components:

●​ FBG Sensors: Four new FBG sensors were bonded onto the curved composite leading
edge structure.
●​ AWG Interrogator: An Alcatel A1936AWG with a 100 GHz channel spacing (0.81 nm)
and 32 channels covering the wavelength range of 1536.1 – 1560.1 nm was used for
interrogation. The AWG was temperature controlled to ensure stability.
●​ Broadband Light Source: A broadband amplified spontaneous emission (ASE) source
was used to illuminate the FBG sensors.
●​ Optical Coupler: A 3 dB directional coupler was used to direct the reflected light from
the FBGs to the AWG (an optical circulator would have been preferable to reduce light
loss).
●​ Photodetectors: Four photodetectors, constructed from pigtailed InGaAs photodiodes,
were used to measure the light intensity at specific AWG channels flanking the reflected
FBG spectra. The photodetectors had a 3 dB bandwidth of 175 kHz.
 ●​ Data Acquisition System: A LabVIEW program was developed to acquire signals from
the eight detector channels (two channels per FBG sensor) and reconstruct the FBG
signal based on the ratiometric interrogation function.
●​ Neural Network for Impact Location: A neural network algorithm, building on the
work in Chapter 3, was implemented in LabVIEW to determine the impact location based
on the dynamic strain signals captured by the FBG sensors.

4.2. Impact Location using Neural Networks

Previous work in the thesis compared time-of-flight triangulation and neural network techniques
for impact location using piezoelectric sensors on the same curved composite structure. The
neural network method demonstrated superior location accuracy compared to triangulation. For
the demonstrator system utilizing FBG sensors, a neural network was again chosen for impact
location.

The neural network was trained using a dataset of 283 valid impact events at twenty marked
locations on the structure. The input to the neural network consisted of features extracted from
the dynamic strain signals captured by the four FBG sensors. The output of the network provided
an estimate of the impact location on the curved composite structure. The use of a neural
network allowed the system to learn the complex relationship between the sensor signals and the
impact location on the inhomogeneous and anisotropic composite structure.

4.3. Multiplexed Interrogation

The AWG-based interrogation scheme inherently supports wavelength division multiplexing

(WDM). By assigning different Bragg wavelengths to the four FBG sensors, their reflected
spectra would be centred at different positions within the operating wavelength range of the
AWG. The AWG then spatially separated these different wavelengths onto different sets of
output channels, allowing for the simultaneous monitoring of all four FBG sensors using the
corresponding photodetectors. This multiplexing capability is crucial for creating scalable SHM
systems for large aerospace structures.

The ratiometric interrogation function, defined as ρ=log⁡(In+1In), where $I_n$ and $I_{n+1}$
are the light intensities captured by adjacent AWG channels for each FBG, was used to extract
the strain information. This ratiometric approach helps to mitigate the effects of broadband
intensity fluctuations from the light source or losses in the optical path.
5. Implementation and Results

5.1. System Implementation and Testing

Four new FBG sensors were bonded to the curved composite leading edge structure. Impact tests
were performed at twenty different locations on the structure, with fifteen impacts at each
location, resulting in 300 impact events. From these, 283 valid data sets were captured and used
for training and testing the neural network.

The LabVIEW program was extended to acquire data from all eight photodetector channels and
calculate the interrogation function for each of the four FBG sensors. This provided real-time
information about the dynamic strain experienced by each sensor during an impact event.

5.2. Neural Network Training and Impact Location Accuracy

A neural network was trained using the captured impact data to learn the relationship between
the FBG sensor signals and the impact location. The performance of the neural network-based
impact location system was evaluated by comparing the estimated locations with the actual
impact locations.

Fig. Comparison of impact location techniques by comparing the average location error from many impacts
at 20 locations.

The demonstrator system, using FBG sensors and a neural network, achieved an average location
error of 105.9 mm with an average standard deviation of 44.7 mm. This was a significant
improvement over a previously implemented system using chirped fibre Bragg grating sensors
(CFBG) and time-of-flight triangulation, which had an average error of 177.8 mm. While the
piezoelectric sensor-based neural network system from Chapter 3 achieved a higher accuracy
(89.8 mm average error), the FBG-based demonstrator offered the advantages of multiplexing
and suitability for aerospace environments. The increased complexity in the location problem
with FBG sensors was attributed to their potential directionality, requiring a different neural
network architecture compared to the piezoelectric sensor system.

5.3. Simultaneous Quasi-Static and Dynamic Strain Measurement

The demonstrator system also demonstrated the capability of simultaneously measuring

quasi-static and dynamic strain. By placing a 3.8 kg weight near one of the FBG sensors, a
change in the D.C. signal level, representing the quasi-static strain due to the load, was observed.
When an impact was then applied to the loaded structure, the dynamic strain response was
superimposed on the quasi-static strain offset and successfully captured by the system. This
capability is important for SHM applications where structures experience both static loads and
dynamic events. Furthermore, with the addition of a strain-isolated FBG, temperature could also
be easily determined simultaneously.

6. Discussion and Advantages

The development of the FBG-based impact detection and location demonstrator system
showcases the potential of optical fibre sensor technology for SHM in aerospace structures. The
system offers several advantages:

●​ Scalability and Multiplexing: The use of an AWG allows for the simultaneous
interrogation of multiple FBG sensors, making the system scalable for monitoring larger
and more complex structures.
●​ Environmental Suitability: Optical fibres are lightweight, immune to electromagnetic
interference, and do not pose an ignition risk in fuel-filled structures, making them
well-suited for aerospace applications.
●​ Improved Location Accuracy: Compared to the previously implemented CFBG-based
triangulation method, the AWG-interrogated FBG system with a neural network achieved
significantly better impact location accuracy on the curved composite structure.
●​ Simultaneous Measurement Capabilities: The system can simultaneously measure both
quasi-static and dynamic strain, providing comprehensive structural health information.

Despite these advantages, further work is needed to optimize the system's performance and
address potential limitations. The location accuracy, while improved, could be further enhanced
with more sophisticated neural network architectures and larger training datasets, potentially
accounting for the directional sensitivity of the FBG sensors. Furthermore, for real-world
aerospace deployment, issues such as long-term stability, robustness in harsh environments
(temperature variations, vibrations), and integration with existing aircraft systems would need to
be thoroughly addressed.
7. Conclusion

This case study has presented the development of a demonstrator system for FBG-based impact
detection and location on a curved composite structure relevant to aerospace applications. By
integrating an AWG interrogation unit with a neural network-based location algorithm, the
system demonstrated the ability to simultaneously interrogate multiple sensors, accurately locate
impact events, and measure both static and dynamic strain. The findings highlight the significant
potential of fibre-optic sensor technology for creating scalable and effective SHM systems for
the aerospace industry, contributing towards enhanced safety and reduced maintenance costs.
Future work should focus on further improving the system's accuracy, robustness, and readiness
for integration into real aircraft platforms.
 References

[1] Y.-J. Rao, "In-fibre Bragg grating sensors," Measurement Science and Technology, vol. 8, pp.
355–375, 1996.

[2] A. D. Kersey et al., "Fiber Grating Sensors," Journal of Lightwave Technology, vol. 15, no. 8,
1997.

[3] S. P. Joshi and C. T. Sun, "Impact-induced microcracking in laminated composites - a finite

element model," Composites: Part A, vol. 33, pp. 1055–1062, 2002.

[4] Y. Sano and T. Yoshino, "Fast Optical Wavelength Interrogator Employing Arrayed
Waveguide Grating for Distributed Fiber Bragg Grating Sensors," Journal of Lightwave
Technology, vol. 21, no. 1, 2003.

[5] W. Staszewski, C. Boller, and G. Tomlinson, Health Monitoring of Aerospace Structures,

Chichester, UK: Wiley, 2004, ch. 1.

[6] W. Staszewski, C. Boller, and G. Tomlinson, Health Monitoring of Aerospace Structures,

Chichester, UK: Wiley, 2004, ch. 1.

[7] C. Y. Park, B.-W. Jang, J. H. Kim, C.-G. Kim, and S.-M. Jun, "Bird strike event monitoring in
a composite UAV wing using high-speed optical fiber sensing system," Composites Science and
Technology, vol. 72, pp. 498–505, 2012.