Hierarchical structural health monitoring system combining a fiber optic spinal cord network and distributed nerve cell devices Shu Minakuchi*a, Haruka Tsukamotob, Nobuo Takedaa a
Department of Advanced Energy, Graduate School of Frontier Sciences, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa-shi, Chiba, 277-8561, Japan b Department of Aeronautics and Astronautics, School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656, Japan ABSTRACT This study proposes novel hierarchical sensing concept for detecting damages in composite structures. In the hierarchical system, numerous three-dimensionally structured sensor devices are distributed throughout the whole structural area and connected with the optical fiber network through transducing mechanisms. The distributed “sensory nerve cell” devices detect the damage, and the fiber optic “spinal cord” network gathers damage signals and transmits the information to a measuring instrument. This study began by discussing the basic concept of the hierarchical sensing system thorough comparison with existing fiber optic based systems and nerve systems in the animal kingdom. Then, in order to validate the proposed sensing concept, impact damage detection system for the composite structure was proposed. The sensor devices were developed based on Comparative Vacuum Monitoring (CVM) system and the Brillouin based distributed strain sensing was utilized to gather the damage signals from the distributed devices. Finally a verification test was conducted using prototype devices. Occurrence of barely visible impact damage was successfully detected and it was clearly indicated that the hierarchical system has better repairability, higher robustness, and wider monitorable area compared to existing systems utilizing embedded optical fiber sensors. Keywords: hierarchical sensing concept, repairability, robustness, monitorable area, nerve systems in animal kingdom, optical fiber, Brillouin based distributed sensing , Comparative Vacuum Monitoring (CVM), composite material, impact damage
1. INTRODUCTION Use of advanced composite materials represented by carbon fiber reinforced plastic (CFRP) has been increasing rapidly in wide range of industrial applications since they have excellent mechanical properties. Especially, in civil aviation aircraft, the weight percentage of the composite materials to the whole weight of the airplane has significantly increased for past 30 years. In the composite materials, however, centimeter internal damages (i.e., impact damage) are critical [1]. The invisible internal cracks and damages significantly degrade the strength of the structures, restricting design space of airplanes with the composite materials. In order to improve the reliability of the composite materials and save the maintenance cost, smart technologies to detect these internal damages in the large-scale composite structures have been established [2]. Among the developed techniques, optical fiber sensors have attracted a considerable amount of attention since they are small size, lightweight, immune to electromagnetic interference, environmentally stable, and having very little signal loss over extremely long distances. The optical fiber sensors can be embedded in the advanced composite materials and have been applied to detect wide-ranging damages [3, 4]. Recently, many innovative techniques enhanced distributed strain and temperature sensing in the aspect of the spatial resolution [5-7], and Brillouin based strain measurement was successfully utilized to detect barely visible impact damage in large-scale composite sandwich structures [8]. However the existing systems with the embedded optical fiber network are unsatisfactory in the following three properties: repairability, robustness, and monitorable area. After the critical damage is detected, the damaged region needs to be repaired. In the case where the optical fiber sensor is embedded in the damaged area, the optical fiber may be broken. However, it is difficult to repair and reconnect the broken optical fiber embedded in the composite materials. And a failure at only one point on the optical fiber may lead to a breakdown of the entire sensing network. *
[email protected]; phone/fax +81-4-7136-4032; http://www.smart.k.u-tokyo.ac.jp/index.html Smart Sensor Phenomena, Technology, Networks, and Systems 2009, edited by Norbert G. Meyendorf, Kara J. Peters, Wolfgang Ecke, Proc. of SPIE Vol. 7293, 72930G · © 2009 SPIE · CCC code: 0277-786X/09/$18 · doi: 10.1117/12.815581 Proc. of SPIE Vol. 7293 72930G-1
Furthermore, since the fiber optic sensing obtains one-dimensional strain and temperature information along the thin optical fiber, the damage far from the sensing fiber can not be detected. In this context, this study proposes an innovative hierarchical sensing system excelling in those three aspects. First, the basic concept is discussed through comparison with the existing system and a nerve system in the animal kingdom. Then, in order to validate the hierarchical sensing concept, impact damage detection system for the composite structure is proposed. Finally, the verification test is conducted using the prototype devices, clarifying the advantages and disadvantages of the hierarchical concept. A future research plan is also mentioned.
2. HIERARCHICAL SENSING CONCEPT The proposed hierarchical system and the existing system are compared in Fig. 1. In the existing system, the optical fiber network is formed within the structural material and the embedded sensors detect the damage from the strain or temperature changes. It is important to note that the optical fiber is further utilized to transmit the “damage information” from the damage region to the measuring instrument. This is why the existing system lacks robustness. A failure at only one point on the optical fiber disables monitoring the other area. Even though the embedded optical fiber can break by introducing damages in the structure and/or is difficult to repair after the breakage, the optical fiber are fully utilized to both detect damage and transmit damage information. Furthermore, since the fiber optic sensing obtains basically onedimensional strain and temperature distribution, the existing system cannot monitor the whole structural area. In the
Composite Structure Embedded optical fibers
Measurable area
Optical fibers
amages
Measuring instrument
Damage in formation
flow
(a) Existing sensing system utilizing embedded optical fiber.
Optical fibers attached on surface
MM..'.,-
MMMMM
flistrihi itp.d
sensor devices
Sensor evices
\ mmrr
eas u ring
instrument
-. Damage in formation flow I
mechanism
(b) Hierarchical sensing system combining distributed sensor devices and optical fiber. Fig. 1 Schematic of sensing system.
Proc. of SPIE Vol. 7293 72930G-2
Brain
Spinal cord
Sensory nerve cells
II
Fig. 2 Schematic of hierarchical nerve system in human body.
Fig. 3 Schematic of diffuse nerve system in hydra.
hierarchical system, on the other hand, numerous three-dimensionally structured sensor devices are distributed throughout the whole structure and connected with the fiber optic network through transducing mechanisms. The devices are bonded on the structural back surface or embedded in the material to detect the damage. Meanwhile, the optical fiber network attached on the back surface of the structure gathers the damage signals from the distributed sensor devices and transmits the damage information to a measuring instrument. If the hierarchical sensing system is compared to a nerve system in vertebrates (Fig.2), the distributed sensor devices, the optical fiber network, and the measuring instrument may correspond to the sensory nerve cells, the spinal cord, and the brain, respectively. Within the animal kingdom, hydra (Fig.3) or coelenterates in general are the first organisms which contain cells specialized for nerve function [9]. Hydra is a very simple fresh-water animal having diffuse nervous system, which is primitive in so far as it has no specialized nervous organs. The nerve cells receive sensory stimuli, make synapses with other nerve cells or muscle fibers, and initiate and propagate action potentials. That is to say, the first nerve system was composed of “jack-of-all-trades” nerve cells. Then, in evolutionary history, the nerve cells were progressively specialized and the organisms began to have hierarchical nervous system consisting of central nerves and peripheral nerves. It is thought that the explosive increase in the variety and the number of the nerve cells enabled the higher organisms to dramatically improve their abilities to sense stimuli, process information, and behave. In this context, the proposed hierarchical sensing system must be an evolved concept. Unlike the existing sensing system, the hierarchical system utilizes two specialized devices. The distributed “sensory nerve cell” devices are specialized in detecting damage. The three-dimensionally structured devices can monitor the whole structure and, moreover, the broken devices can be easily replaced during the material repairing process as the peripheral nerves regenerate in the damaged body. On the other hand, the fiber optic “spinal cord” network is specialized in gathering and transmitting the damage signals. Brillouin based distributed strain and temperature sensing techniques can be utilized to monitor the huge number of the distributed devices. Furthermore, since the optical fiber is attached on the back surface of the structure, the fiber hardly breaks when the structure is damaged. And, if by any chance the optical fiber is broken, the broken parts can be
Proc. of SPIE Vol. 7293 72930G-3
readily reconnected by using a fusion splicer since the optical fiber is not embedded. Finally, in the hierarchical system, several devices, which are connected to different optical fibers, are placed in the same area. Hence the failure at only one point on the devices or the fiber optic network does not affect the monitoring performance. The hierarchical system has high redundancy and robustness. Thus, the hierarchical concept suitably combines strengths of the distributed sensor devices and the optical fiber. The evolved system may have better repairability, higher robustness, and wider monitorable area compared to the existing system using embedded optical fiber. In the next section, in order to validate the new concept, impact damage detection system for the composite structure is established. A verification test using prototype devices is conducted to clarify advantages and disadvantages of the proposed concept. Finally, a future research plan is also mentioned.
3. HIERARCHICAL IMPACT DAMAGE DETECTION SYSTEM 3.1 System overview The schematic of the proposed system is presented in Fig. 4. The distributed sensor devices are based on Comparative
Back surface
Foreign object
11111111111111111
Optical Sensor devices vLuJLuJLuJv fiber I I_
III
tll
irrmin rrnr
o 11111 Delamination I Bending crack
jjiii.
pact damage
_LLJULJU O
II,
Ii I Damage signal transducing mechanism II
iliilliil
A Sensor device
B
Before damage Channel
-
Polymer tube
Optical fiber Bending crack
After damage
Ad hesive face Fig. 4 Schematic of impact damage detection system.
Proc. of SPIE Vol. 7293 72930G-4
Vacuum Monitoring (CVM) system [10]. The CVM system, which has been developed by Structural Monitoring Systems in Australia, uses an elastomeric surface sensor with fine channels and can detect surface cracks by monitoring the internal pressure variation of the sensor. Since the system is extremely simple, sensitive, and reliable, several aircraft manufacturing companies have been verifying and improving the sensor technologies and equipments for future aircraft applications. Basically, the CVM technology has been applied to monitor localized areas such as joining parts. In this study, however, tightly-arranged devices are bonded on a back surface of the composite structure. The thin polymer device has fine channels on the adhesive face and the internal pressure of the channels is kept lower than atmospheric pressure. Then the devices are connected to the optical fiber via damage signal transducing mechanisms. The mechanism converts the internal pressure variation of the distributed sensor device, which is the damage signal, into the axial strain change of the optical fiber. In this study, soft polymer tubes are connected to the channels of the distributed sensor devices and the optical fiber is wound and fixed on the tubes. The internal pressure change deforms the polymer tubes in the radial direction and, consequently, the optical fiber is axially strained. The strain variation of the whole devices is monitored by using a distributed strain measurement system having high spatial resolution. When the impact damage is induced, a bending crack occurs along the reinforcing fiber direction on the back surface just below the impact point. The surface crack forms a leakage path between the atmosphere and the channels, increasing the internal pressure of the distributed sensor devices. Finally, the fixed optical fiber is strained and the distributed strain measurement system detects it. It is important to note that the polymer sensor developed in the CVM system is lightweight, flexible, and involves no electrical excitation. Thus the sensor device is highly suitable to be combined with the optical fiber in the hierarchical sensing concept. 3.2 Verification test Figure 5 presents a specimen viewed from directly above. One prototype device was attached on the back surface of the
Unit: mm
Transducing nechanism
pact point
230
Fig. 5 Back surface of specimen for impact damage detection test.
Proc. of SPIE Vol. 7293 72930G-5
Fixed area
400
AAft impact
300
200
impact
:1 100
kBefore
I -200 8
8.5
9
9.5
Position (m) Fig. 6 Strain distribution along optical fiber wound on tube. CFRP laminates (T700S/2500, Toray Industries, Inc., [02/904/02], 230×230×1.1 mm3) using a silicone sealant. The sensor device was made of an polystyrene thin plate (thickness: 0.3 mm) and two independent sensor channels were formed. The internal pressure of the channels was kept lower than atmospheric pressure by sealing the channels in a vacuum environment. A vinyl chloride tube (internal diameter: 4 mm, external diameter: 6 mm) was connected to each channel using the silicone sealant and an optical fiber (Heatop 300, Totoku Electric Co., Ltd.) was wound and fixed on the tubes. The length of the fixed parts was about 20 cm. Then, low-velocity impact loading (6.1 J) was applied to the center of the specimen using a drop-weight impact tester (Dynatup Mini-Tower, Instron Corporation, impactor diameter: 16 mm). The distributed strain measurement was conducted before and after the impact loading. Pre-pump pulse Brillouin optical time domain analysis (PPP-BOTDA, spatial resolution: 10 cm, sampling interval: 5 cm, strain measurement accuracy: ±0.0025%) was utilized [6]. Figure 6 shows the obtained strain distribution from the optical fiber fixed on one tube. Before the impact loading, tensile strain was already induced in the fixed area, since the optical fiber was tightly wound on the tube. Hence the optical fiber could monitor the internal pressure variation sensitively. After the impact loading, the strain at the fixed part significantly increased. A small dent (depth: 0.25 mm, diameter: 3 mm) was visibly observed at the impact point and a 20 mm tiny bending crack appeared on the back surface of the specimen, creating the leakage path between the atmosphere and the channels of the sensory devices. Hence, it would appear that air entered to the channels, which had lower internal pressure, and the increased internal pressure deformed the tubes, straining the wound optical fiber. A comparable result was obtained from the optical fiber fixed on the other tube. Thus, the barely visible impact damage was successfully detected. The strain increase at the fixed area clearly indicated the impact damage occurrence. Even though the distributed strain measurement had only two measure points (i.e. the two tubes), the hierarchical sensing system could monitor the wide plane area by utilizing the three-dimensionally structured sensor devices. In practical application of CFRP structures, the severely damaged region is hollowed out and/or repaired with new materials. The damaged device can be easily repaired and, if necessary, replaced due to its simple structure. Furthermore, since the two channels connected to different optical fibers monitored the same plane area, the system has high robustness, as described above. Meanwhile, the hierarchical system may have following two drawbacks. First, since
Proc. of SPIE Vol. 7293 72930G-6
the several devices (i.e. the distributed sensor devices, the transducing mechanisms, and the optical fibers) are attached, the structural back surface is touchy. It is vital for the devices to be easily-handled and acceptable in the manufacturing and assembly process and operations. Second, the sensing system has several interfaces because of its hierarchical nature. In the impact damage detection system, the sensor device is firstly connected to the transducing mechanism, and then the mechanism is coupled with the optical fiber. Hence in order to develop highly reliable system, it is crucial to utilize simply-structured and high-integrity devices. In the near future, the proposed impact damage system is applied to large-scale CFRP structures. The feasibility of the hierarchical sensing system will be clearly verified by addressing its above-mentioned disadvantages. Since the sensing capability of the hierarchical system is determined by the characteristics of the distributed devices, development of new sensor devices is also planned. Specifically, embeddable microstructured devices, which can be physically and/or chemically-based, are expected for detecting microscopic damages in composite materials. Moreover, the impact damage detection technique proposed in this study determined the damage occurrence just based on the strain increase. Quantitative damage detection system will be established by utilizing the strain values.
4. CONCLUSIONS This study proposed a novel hierarchical sensing concept for detecting damages in composite structures. First, the basic concept of the hierarchical sensing system was discussed thorough comparison with existing fiber optic based systems and nerve systems in the animal kingdom. Then, in order to validate the proposed sensing concept, impact damage detection system based on CVM system was established. In the verification test, impact damage was successfully detected and it was clearly indicated that the hierarchical system has better repairability, higher robustness, and wider monitorable area compared to existing systems utilizing embedded optical fiber sensors. Furthermore, the disadvantages of the hierarchical system were also discussed. In the near future, by developing embeddable microstructured sensor devices, highly reliable sensing system will be developed for quantitative detection of microscopic damages in composite structures.
ACKNOWLEDGEMENT This work was financially supported by Japan Society for the Promotion of Science under Grant No. 20860027.
REFERENCES [1] [2] [3] [4] [5] [6] [7]
Abrate, S., [Impact on composite structures], Cambridge University Press (1998). Staszewski, W. J., Tomlinson, G. R. and Boller, C., editors, [Health Monitoring of Aerospace Structures: Smart Sensor Technologies and Signal Processing], John Wiley & Sons, Inc. (2004). Kuang, K. S. C. and Cantwell, W. J., “Use of conventional optical fibers and fiber Bragg gratings for damage detection in advanced composite structures: A review,” Applied Mechanics Reviews 56(5), 493-513 (2003). Takeda, N., Okabe, Y. and T. Mizutani, “Damage detection in composites using optical fibre sensors,” Proceedings of the Institution of Mechanical Enigineers Part G - Journal of Aerospace 221(G4), 497-508 (2007). Bao, X., Brown, A., DeMerchant, M. and Smith J., “Characterization of the Brillouin-loss spectrum of single-mode fibers by use of very short (< 10-ns) pulses,” Optics Letters 24(8):510-512 (1999). Kishida, K., Li, C. H. and Nishiguchi, K., “Pulse pre-pump method for cm-order spatial resolution of BOTDA,” Proceeding of the SPIE 5855, 559-562 (2005). Song, K. Y., He, Z. Y. and Hotate, K., “Distributed strain measurement with millimeter-order spatial resolution based on Brillouin optical correlation domain analysis,” Optics Letters 31(17), 2526-2528 (2006).
Proc. of SPIE Vol. 7293 72930G-7
[8]
[9] [10]
Minakuchi, S., Mizutani, T., Okabe, Y. and Takeda, N., “Barely visible impact damage detection by distributed strain measurement along embedded optical fiber with 10 cm spatial resolution,” Proceedings of the 8th International Conference on Sandwich Structures, 357-368 (2008). Agata, K. and Koizumi, O., editors, [Neuronal diversity: its origin and evolution (in Japanese)], Baifukan Co., Ltd, (2007). Barton, D. P., “Comparative vacuum monitoring: a new method of in-situ real-time crack detection and monitoring,” Structural Monitoring Systems Ltd., Perth, WA (2004). Available from: .
Proc. of SPIE Vol. 7293 72930G-8
