Page 85 - IJAMD-2-3

International Journal of AI for
Materials and Design SHM using improved CNT-BP and LSTM-NN

its low density, high electrical conductivity, and significant nanocomposite films, as cracks develop and propagate
load sensitivity. Wang et al. correlated the electrical in the metallic host structure. Bian et al. proposed a
19
5
resistance of BP sensor with the strain for monitoring the facile technique to enhance performance of carbon fiber
health of composite structures. It can be seen that the BP composites through interlaminar insertion of aligned CNT
sensor has very high strain sensitivity in the static tensile sheets. The inserted CNT sheets also provide electrical
test. Lu et al. employed an omnidirectional, nanomaterial- conductivity in the composites even at a low CNT loading
6
based sensor for impact damage detection in composite below the electrical percolation threshold established for
structures. The influence of both tensile and low-velocity CNT-filled composites. Lin et al. presented a flexible
20
impact on the BP sensors was also further investigated. Lu CNT-based strain sensor that has a significant potential
et al. proposed a real-time monitoring of resin infiltration for applications in human motion monitoring systems and
7
process in vacuum-assisted molding of composites with electronic skins in water. Olson et al. investigated a CNT-
21
CNT-BP sensors. Her et al. employed CNT-BP sensor on a based sensor to detect crack propagation in aluminum
8
complex surface for monitoring the strain and temperature structures underneath composite patching. Initial tests
at the critical areas of the sample. Yee et al. developed an were conducted to determine the correct procedure and
9
improved fabrication process for enhancing the sensitivity materials to properly fabricate a highly sensitive CNT-
of the BP sensor for strain monitoring applications. based sensor.
Yang et al. proposed a flexible, lightweight, and low-
10
thickness BP sensor for high-performance electromagnetic The present work proposes an improved fabrication
interference shielding materials for catering to the demand process for the CNT-BP for measuring the fatigue crack
for smart and wearable electronic devices. De Paula Santos propagation in a metallic specimen under uniaxial loading.
et al. investigated the effect of incorporating CNT-BP on First, an optimal composition for different ingredients
11
the interlaminar fracture toughness by testing under cyclic required for fabricating CNT-BP was obtained and further
loading in Mode I and II. Ahmed et al. investigated the optimized for its sensing architectures. This includes the
12
efficiency of a CNT sensor to detect and monitor fatigue CNT-BP sensor size, sensing locations, and number of
crack initiation and propagation in metal structures. The sensing points, and the silver ink electrode placement. It
sensor consisted of a non-woven carrier fabric with a thin is found that the fabricated CNT-BP sensor has enhanced
film of CNT that is superglued to the structure’s surface sensitivity and has adequate ductility so that it can be
using an epoxy adhesive. Jiang et al. employed CNT-BP easily superglued to the test structure. Moreover, a strong
13
sensors to monitor the structural health status of composite correlation was observed among the actual crack lengths
structures subjected to ambient vibration condition. Hehr measured by a high-speed optical camera, the crack
et al. employed CNT thread in unidirectional glass fiber estimates based on CNT-BP resistance, and the predictions
14
composites to identify the onset of track crack growth, from the long short-term memory (LSTM) neural network.
matrix cracking, and differentiate between crack breathing This shows that CNT-BP has promising potential for fatigue
and closing states. This information is obtained by crack prognosis and diagnosis without the need for complex
analyzing the resistance response of the thread with a simple equipment, signal processing, and measurement units.
Wheatstone bridge circuit and a low-speed data acquisition The rest of the paper is organized as follows. The
system. Ribeiro et al. reviewed the general characteristics, theoretical model for prediction of crack length is presented
15
physical properties, and processing conditions of CNT-BP in Section 2, while the CNT-BP fabrication process is
and its polymer composites. Wan et al. investigate the in detailed in Section 3. Section 4 discusses the LSTM neural
16
situ monitoring of Mode I interlaminar crack propagation network for prognosis of a metallic plate with fatigue crack
behavior in woven glass fiber-reinforced epoxy (WGF/ propagation, and the experimental study is discussed in
epoxy) composite laminates. Lecompte et al. employed Section 5. The results obtained are discussed in Section
17
two different optical measurement techniques for the 6, and lastly, the concluding remarks and future study are
detection of cracks at the surface of a concrete beam being given in Section 7. The following section discusses the
subjected to flexural loading. The primary aim of the theoretical model for crack length prediction based on
study was to examine which of the two methods appears the measurement of resistance change of CNT-BP as a
to be the most suitable for crack detection. Moreover, it function of fatigue cycles.
was shown that it is possible to detect the appearance and
evolution of cracks, even before the cracks become visually 2. Theoretical model for prediction of crack
detectable. Ashrafi et al. focused on the application of length
18
epoxy nanocomposite thin film sensors for continuous
monitoring of crack evolution in metallic structures. The A measurement model to correlate between the fatigue
primary aim was to monitor the resistance change in these crack at any instant of time and the electrical resistance

Volume 2 Issue 3 (2025) 79 doi: 10.36922/IJAMD025310028