Cover page
Title: Strain Monitoring of Filament Wound Composite Tank Using Fiber Bragg Grating Sensors
Authors: Hyun-Kyu Kang1 Dong-Hoon Kang2 Jae-Sung Park3 Chang-Sun Hong2 Chun-Gon Kim2§
1 Senior Engineer, Samsung Electronics Co., LTD., Nongseo-Ri, Giheung-Eup, Yongin-City, Gyeonggi-Do, 449-711, KOREA
2 Division of Aerospace Engineering, Korea Advanced Institute of Science and Technology, 373-1, Guseong-dong, Yuseong-gu, Daejeon, 305-701, KOREA
3 Senior Engineer, Korea Aerospace Research Institute, 45, Eoeun-dong, Yuseong-gu, Daejeon, 305-333, KOREA
§ : To whom correspondence should be addressed. Tel:+82-42-869-3719, Fax:+82-42-869-3710, E-mail:cgkim@kaist.ac.kr
ABSTRACT
In this paper, we present strain monitoring of a filament wound composite tank using fiber Bragg grating sensors during hydrostatic pressurization. 20 fiber Bragg grating (FBG) sensors and 20 strain gauges were attached on the domes and cylinder of the composite tank. A wavelength-swept fiber laser (WSFL) was used as a light source to supply high signal power. From the experimental results, it was successfully demonstrated that the FBG sensor system is useful for large structures that require a large number of sensor arrays.
1. INTRODUCTION
Composite materials are increasingly being used as engineering materials in aircrafts, buildings, containers, and other structures. In particular, the use of filament wound composite tanks is increasingly prevalent because of high specific strength and specific stiffness over their metal counterparts, as well as excellent corrosion and fatigue resistance. Filament wound composite tanks are finding their use in the applications such as fuel tanks, pressure tanks, and motor cases of aerospace structures.
The complexity in damage mechanisms and failure modes makes the use of composite materials difficult. Most of the conventional damage assessment and nondestructive inspection methods are time-consuming and are often difficult to implement on hard-to-reach-parts of the structure. For these reasons, a built-in assessment system must be developed to constantly monitor the structural integrity of critical components.
_____________
Hyun-Kyu Kang, Samsung Electronics Co., LTD., Nongseo-Ri, Giheung-Eup, Yongin-City, Gyeonggi-Do, 449-711, Korea
Dong-Hoon Kang, Chang-Sun Hong, Chun-Gon Kim, Division of Aerospace Engineering, Korea Advanced Institute of Science and Technology, 373-1, Guseong-dong, Yuseong-gu, Daejeon, 305-701, Korea - cgkim@kaist.ac.kr
Jae-Sung Park, Korea Aerospace Research Institute, 45, Eoeun-dong, Yuseong-gu, Daejeon, 305-333, Korea
Fiber optic sensors (FOSs) have shown a potential to serve real time health monitoring of the structures. They can be easily embedded or attached to the structures and are not affected by the electro-magnetic field. Also, they have not only the flexibility in the selection of the sensor size but also high sensitiveness. Recently, fiber optic sensors have been introduced into composite structures[1-4]. FBG sensors based on WDM (wavelength division multiplexing) technology are attracting considerable research interest and appear to be ideally suitable for structural health monitoring of smart composite structures. FBG sensors are easily multiplexed and have many advantages such as linear response, absolute measurement, etc.
In this paper, the strains of a filament wound composite tank were monitored using a FBG sensor system in real time. 20 FBG sensors and 20 strain gauges were attached on the surface of the dome and cylinder part of the composite tank. The strains measured by FBG sensors were compared with those measured by strain gauges.
2. FIBER BRAGG GRATING SENSOR
A fiber Bragg grating is a periodic refractive index change that is formed in the core of an optical fiber by exposure to an intense UV interference pattern. This grating structure results in the reflection of the light at a specific narrowband wavelength, called the Bragg wavelength. The Bragg condition is given by λB =2neΛ, where λB is the Bragg wavelength of FBG, ne is the effective refractive index of the fiber core, and Λ is the grating period. The shift of Bragg wavelength due to strain and temperature can be expressed as
∆λB =λB
[ (
αf +ξf)
∆T +(
1− pe)
ε]
. (1)where α is the coefficient of thermal expansion (CTE), f ξ is the thermo-optic f coefficient, and p is the strain-optic coefficient of the optical fiber. The value of e
227 . 0
e =
p was measured experimentally and used for this study. If there is no temperature change, we can measure the strain from the wavelength shift as
B B
pe λ
ε ∆λ
= − 1
1 . (2)
In this experiment, we used a dummy FBG, which was not attached on the composite tank, for the temperature compensation.
3. EXPERIMENTAL DETAILS
3.1. Filament wound composite tank
Figure 1 shows the filament wound composite tank, which is to be used as the 3rd stage kick motor case of a 3-stage Korean scientific rocket(KSR). This composite tank
(a) Front dome (b) Aft dome (c) Side view (d) Assembly with nozzle Figure 1. Filament Wound Composite Tank.
consists of a front dome, an aft dome, a cylinder, and a skirt for joining with a satellite.
Epoxy filler filled the gap between the skirt and domes to prevent stress concentration due to sudden changes of curvature and thickness in the joint part of the cylinder and domes. The composite tank was fabricated by a dry winding process using tow prepreg. The tow prepreg is the composite prepreg tape with wide bandwidth of ~ mm. The dome parts consist of 12 layers (helical winding layers) with ±22° winding angle and the cylinder part is made of helical layers and several hoop layers(90°) inserted between the helical layers. The helical winding layer is 0.198 mm thick per 2 layers (+22°/-22°) and the hoop winding layer is 0.168 mm thick per single layer. The 1st skirt and 2nd skirt have [±30°4 / 90°2 / ±30°4 / 90°2] and [90°2 / ±15°2 / 90°2 / ±15°2 / 90°6] lay-up, respectively. The dimensions of the filament wound composite tank are shown in Table 1.
3.2. Wavelength-swept fiber laser
In this study, a wavelength-swept fiber laser (WSFL)[5], a broadband source, was used to interrogate the FBG sensors. The WSFL was in a unidirectional ring configuration with isolators, a 3-dB output coupler, and an Er3+-doped fiber pumped by a laser diode at 980 nm. An F-P tunable filter was used as the intracavity scanning filter and had a 3-dB bandwidth of 0.27 nm and a free spectral range of 58 nm. We modulated the F-P filter with a triangular waveform to produce a wavelength sweep over 40 nm from 1525 to 1565 nm at a 200 Hz repetition rate. The output power of the WSFL was over 1000 times as large as that of the amplified spontaneous emission (ASE) of an LD-pumped Er3+-doped fiber (EDF), which is usually used for the light source in a FBG sensor system [6].
Table 1. Dimensions of filament wound composite tank.
Forward Dome Aft Dome
Cylinder radius, Rc 250 mm
Boss radius, Rb 50 mm 85 mm
Cylinder thickness, tc 2.92 mm Thickness of 1st skirt 4.16 mm Thickness of 2nd skirt 3.68 mm
P ressure Device Steel Shield
Pressure Transducer
D igital Indicator
4 O ptical lines
D ata A cquisition &
S ignal Processing U nit
W SFL 4 channel detector 20 S train gage lines
S train M easurem ent U nit
W ater injection
…
FBG peak detection circuit SC XI-1001 P ressure
Device Steel Shield
Pressure Transducer
D igital Indicator
4 O ptical lines
D ata A cquisition &
S ignal Processing U nit
W SFL 4 channel detector 20 S train gage lines
S train M easurem ent U nit
W ater injection
…
FBG peak detection circuit SC XI-1001
Figure 2. Experimental set-up.
3.3. Experimental apparatus and method
Figure 2 depicts the experimental set-up for the strain monitoring of a filament wound composite tank during hydrostatic pressurization. The strain measurement was performed at intervals of 100 psi up to 1000 psi(6.895 MPa). The data from the FBG sensors, strain gauges, and a pressure transducer were acquired and processed by computers in real time. In total, 20 FBG sensors and 20 strain gauges were attached at the same longitudinal locations on the composite tank. The array of FBG sensors consisted of 5 gratings along each of four separate fiber optic lines. The 5 sensors of FBG line A were attached on the front dome and other 5 sensors of FBG line B on the aft dome. The 5 sensors of FBG line C and the 5 sensors of FBG line D were attached on the cylinder and the skirt. Since most of the loading applied to composite tank is sustained by reinforcing fibers, it is important to measure the strains along the fiber direction. Hence, all sensors were aligned to fiber directions of the composite tank except two sensors of FBG line D attached perpendicular to the fiber direction on the cylinder. On the domes, the pairs of FBG sensors and strain gauges were attached on the same tow prepreg in the helical winding direction (±22°) of fibers and at the same longitudinal location. On the cylinder and the skirt, also, 8 pairs of two sensors were adhered in the hoop winding direction, i.e., the circumferential direction of cylinder, and at the same longitudinal location two pairs were adhered perpendicular to the winding direction, i.e., the longitudinal direction.
The signals of the FBG sensors, strain gauges and pressure transducer were acquired simultaneously by computers, and processed and displayed by a signal-processing program written in LabVIEW software. Pressure level was held for checking the pressure drop occurred due to leaking water.
4. RESULTS AND DISCUSSIONS
Figures 3 and 4 illustrate the results of strain measurements using the FBG sensors and electrical strain gauges (ESG). As shown in these results, the strains measured by both types of sensors increased linearly with increasing pressure.
0 200 400 600 800 1000 0.00
0.05 0.10 0.15 0.20 0.25 0.30
Strain (%)
Pressure (psi) FBG1
ESG1 FBG2 ESG2 FBG3 ESG3 FBG4 ESG4 FBG5 ESG5
S1 S2 S3 S4 S5
180224 150 107
88 S1 S2 S3 S4 S5
180224 150 107
88
0 200 400 600 800 1000
0.00 0.05 0.10 0.15 0.20 0.25 0.30
Strain (%)
Pressure (psi) FBG1
ESG1 FBG2 ESG2 FBG3 ESG3 FBG4 ESG4 FBG5 ESG5
S1 S2 S3 S4 S5
180224 150 107
88 S1 S2 S3 S4 S5
180224 150 107
88
0 200 400 600 800 1000
0.0 0.1 0.2 0.3 0.4 0.5
Strain (%)
Pressure (psi) FBG6
ESG6 - N/A FBG7 ESG7 FBG8 ESG8 FBG9 ESG9 FBG10 ESG10
S6 S7 S8 S9
S10 225 200 169 137
119 S6 S7 S8 S9
S10 225 200 169 137
119
0 200 400 600 800 1000
0.0 0.1 0.2 0.3 0.4 0.5
Strain (%)
Pressure (psi) FBG6
ESG6 - N/A FBG7 ESG7 FBG8 ESG8 FBG9 ESG9 FBG10 ESG10
S6 S7 S8 S9
S10 225 200 169 137
119 S6 S7 S8 S9
S10 225 200 169 137
119
(a) Front dome – FBG channel 1. (b) Aft dome – FBG channel 2.
Figure 3. Strains measured by FBG and ESG attached on domes.
The signal peaks of FBG5 of line A, FBG9 and FBG10 of line B attached on domes split when the pressure level was over 600 psi. The signal peaks of FBG19 and FBG20 of line D attached on the cylinder split at the pressure level over 300 psi. Hence, these FBG sensors could not measure strains over those pressure levels. Also, since ESG6
was in an abnormal condition, it could not measure the strains.
In figure 3(a), the strains measured by FBG2 and FBG3 were much larger than those by other FBG sensors on the front dome. In figure 3(b), also, the strains measured by FBG7 and FBG8 were much larger than those by other FBG sensors on the aft dome.
From these results, we can find that the larger deformations occurred by the expansion in the middle of the domes and that more reinforcements in the circumferential direction of the domes are needed. In figures 4(a) and 4(b), the strains measured by FBG12, FBG13, FBG16 and FBG17 were much larger than those by other FBG sensors on the cylinder and the skirt. From these results, we also find that the large deformations by the expansion in the middle of the cylinder can be reduced by the more reinforcements of the cylinder. Moreover, we can predict that the small deformations might occur in the skirt and the skirt does not play an important role in the load bearing.
The differences of strains measured by FBG sensors and strain gauges may be occurred by a mismatch of attaching angles and locations between them. Also, these differences may be caused by the large variation of the fiber angle at points of the same meridian line in a tow prepreg of the dome.
Jeusette et al [7] considered the variation of fiber angle along the width of the same winding tow in the analysis of filament wound motor case. It is pointed out that the relative difference between fiber directional strains can reach up to 40 % in the polar region due to the winding angle difference along the width of a winding tow. In our experiment, the width of a winding tow was about 20 mm. Although each pair which consisted of a FBG sensor and a strain gauge was attached at the same longitudinal location on a tow prepreg of domes, distances between FBG sensors and strain gauges were about 6~8 mm. Hence, winding angles of two points where FBG and strain gauge were attached could be different with each other. As a result, fiber directional strains measured by FBG sensors and strain gauges showed remarkable discrepancies.
0 200 400 600 800 1000 0.00
0.05 0.10 0.15 0.20 0.25 0.30
Strain (%)
Pressure (psi) FBG11
ESG11 FBG12 ESG12 FBG13 ESG13
FBG14 ESG14 FBG15 ESG15
4648 614642 92
S11S12S13S14S15 Cylinder Skirt Skirt
Filler Filler
4648 614642 92
S11S12S13S14S15 Cylinder Skirt Skirt
Filler Filler
0 200 400 600 800 1000
0.00 0.05 0.10 0.15 0.20 0.25
Strain (%)
Pressure (psi) FBG16
ESG16 FBG17 ESG17 FBG18 ESG18 FBG19 ESG19 FBG20 ESG20
Cylinder Skirt Skirt
Filler Filler
84 72 64 115
126 81 133
S16 S20
S17 S19
S18 Cylinder Skirt Skirt
Filler Filler
84 72 64 115
126 81 133
S16 S20
S17 S19
S18
(a) Cylinder – FBG channel 3. (b) Cylinder – FBG channel 4.
Figure 4. Strains measured by FBG and ESG attached on cylinder.
The signal peaks of 5 FBG sensors among the 20 FBG sensors split during pressurization. Figure 5 shows the variations of FBG signal peaks of FBG line D attached on the cylinder with increasing pressure. Since strains in the circumferential direction of the cylinder are uniform, peaks of FBG sensor signals attached in the fiber direction did not split. However, since strains in the axial direction of the cylinder vary with high strain gradient, the peak splitting of FBG sensor signals attached in the axial direction became larger with increasing pressure. For the prevention of the peak splitting, the FBG must be protected by an external tube such as a glass capillary tube, as shown in figure 6.
From the point of view of the measurement equipment and method, a FBG sensor system is superior to the strain gauge system. Since each strain gauge requires lead wires and an amplifier, increasing the number of strain gauge results in a proportionally larger measurement system. Also, since the strain gauge system is interfered by electro-magnetic field, remote sensing is difficult for monitoring of hazardous structures. Conversely, since only one source is needed for numerous FBG sensors by using WDM technique, the FBG sensor system is simple. In addition, the FBG sensor system can be used for remote sensing from long distances because optical fiber is immune from electro-magnetic interference. Hence, the FBG sensor system is more efficient for multiple point strain measurements of large and hazardous structures.
1 5 30 1 5 3 5 1 5 4 0 1 54 5 1 5 5 0 1 5 5 5
1E -6 1E -5 1E -4 1E -3 0 .0 1
Intensity (dBm)
W a v e le n g th (n m ) 0 p s i 5 0 0 p s i
1 0 0 0 p s i S 1 9 S 2 0 S 1 6
S 1 7
S 1 8
1 5 30 1 5 3 5 1 5 4 0 1 54 5 1 5 5 0 1 5 5 5
1E -6 1E -5 1E -4 1E -3 0 .0 1
Intensity (dBm)
W a v e le n g th (n m ) 0 p s i 5 0 0 p s i
1 0 0 0 p s i S 1 9 S 2 0 S 1 6
S 1 7
S 1 8
Figure 5. Variation of FBGs spectrum of FBG line D according to pressure elevation.
Fiber Bragg grating Optical fiber
Glass tube
Adhesive
Gage length Fiber Bragg grating Optical fiber
Glass tube
Adhesive
Gage length
Figure 6. Configuration of protected fiber Bragg grating sensor.
5. CONCLUSIONS
20 FBG sensors were attached on the domes and cylinder parts of a filament wound composite tank to measure the strains in real time during hydrostatic pressurization.
FBG sensors attached on the cylinder and the skirt showed close agreement with strain gauges, while the agreement of the two types of sensors on domes was not good. Since the signal peak splitting of some FBG sensors during pressurization makes the strain measurement impossible, the compensatory manufacturing methods of the sensor heads are needed to prevent the peak splitting of the FBG sensor signal. From the experiment, it was successfully demonstrated that the FBG sensor system could be useful for the strain monitoring of large structures that require a large number of sensor arrays.
Acknowledgements
The authors would like to thank the Ministry of Science and Technology, Korea, for the financial support by a grant from the Critical Technology 21 project.
REFERENCES
1. I. B. Kwon, C. G. Kim, and C. S. Hong. 1997. “Simultaneous sensing of the strain and failure instants of composite beam using fiber optic Michelson sensor,” Composite Science and Technology, 57(12):1639-1651.
2. R. C. Foedinger, D. L. Rea, J. S. Sirkis, C. S. Baldwin, J. R. Troll, R. Grande, C. S. Davis, and T. L.
VanDiver. 1999. “Embedded fiber optic sensor arrays for structural health monitoring of filament wound composite pressure vessels,” Proc. Of SPIE, 3670: 289-301.
3. H. K. Kang, J. W. Park, C. Y. Ryu, C. S. Hong, and C. G. Kim. 2000. “Development of fibre optic ingress/egress methods for smart composite structures,” Smart Materials and Structures, 9(2):
149-156.
4. J. W. Park, C. Y. Ryu, H. K. Kang, and C. S. Hong. 2000. “Detection of buckling and crack growth in the delaminated composites using fiber optic sensor,” Journal of Composite Materials, 34(19):
1602-1623.
5. S. H. Yun, D. J. Richardson and B. Y. Kim. 1998. “Interrogation of fiber grating sensor arrays with a wavelength-swept fiber laser,” Optics Letters, 23(11):843-845.
6. C. S. Hong, C. Y. Ryu, B. Y. Koo, C. G. Kim and S. H. Yun. 2000. “Strain monitoring of smart bridge using fiber Bragg grating sensor system with wavelength-swept fiber laser,” Proc. Of SPIE, 3988:371-379.
7. J. P. Jeusette, G. Laschet, P. Charpentier, and P. H. Deloo. 1987. “Finite element analysis of composite revolution structures wound by wide plies,” Composite Structures, 8:221-237.
