Skip to main content

Full text of "Sensor Technology for Integrated Vehicle Health Management of Aerospace Vehicles"

See other formats



W. H. Prosser, T. L. Brown, S. E. Woodard, G. A. Fleming, and E. G. Cooper 
NASA Langley Research Center, Hampton, VA 23681 

ABSTRACT. NASA is focusing considerable efforts on technology development for Integrated 
Vehicle Health Management systems. The research in this area is targeted toward increasing 
aerospace vehicle safety and reliability, while reducing vehicle operating and maintenance costs. On- 
board, real-time sensing technologies that can provide detailed information on structural integrity are 
central to such a health management system. This paper describes a number of sensor technologies 
currently under development for integrated vehicle health management. The capabilities, current 
limitations, and future research needs of these technologies are addressed. 


The application of traditional NDE methods for on-ground inspection of aerospace 
vehicles contributes greatly to their safety and reliability. However, periodic inspections 
significantly increase operating expense and vehicle processing time. Further, the need to 
disassemble and reassemble structural components to allow inspections can lead to damage 
or degradation of the structure or auxiliary systems (e.g., electrical wiring and hydraulic 
lines). NASA is focusing on technology development for Integrated Vehicle Health 
Management (IVHM) systems to address these issues, and to meet demanding goals in 
increasing aerospace vehicle safety and reliability while reducing vehicle operating costs. 
On-board, real-time sensing systems for structural integrity assessment are central to the 
IVHM approach. Such sensing systems will minimize the need for periodic NDE 
inspections, or at least focus these inspections to specific vehicle areas where damage was 
indicated. Sensors comprising an IVHM system must be able to withstand harsh aerospace 
operating environments, while having minimal size, weight, and power requirements. 
Several candidate sensor technologies for use in an IVHM system are discussed in this 
paper. These include fiber-optic sensors, active and passive ultrasonic methods, remote 
wireless technologies, and remote non-contact sensing. Additionally, a brief discussion on 
IVHM system architecture is provided to illustrate the considerations given to establishing 
architectures capable of handing the data acquisition, processing, analysis, and storage for 
massive numbers of multiple sensor types. 


Considering the large acreage of aerospace vehicle structural elements, it is a given 
that extremely large numbers of sensors will be required for on-board structural integrity 
assessment. Fiber optic sensors have been identified as the leading candidate technology 

for meeting this requirement with minimal weight penalty. Numerous sensor sites can be 
multiplexed along a single optical fiber, mitigating the complexity and weight inherent 
with the wiring required for a large number of single ended sensors. Fiber optic sensors 
also provide other advantages such as the ability to measure many different structural 
parameters of interest, immunity to electromagnetic interference (EMI), and the ability to 
operate over very large temperature environments. 

Fiber optic sensors can be separated into two classes for discrete strain and 
temperature measurement: cavity-based designs and grating-based designs [1]. Cavity- 
based designs utilize an interferometric cavity in the fiber to create the sensor. Examples 
include the extrinsic Fabry-Perot interferometer (EFPI), the intrinsic or fiber Fabry-Perot 
interferometer (IFPI or FFPI), and all other etalon-type devices. Although such sensor 
designs have been utilized in a wide variety of applications such as in high temperature 
and EMI environments, they do not allow for multiplexing capability in a single fiber, and 
thus may be limited for applications requiring large number of sensors. 

Grating-based designs utilize a photo- or heat-induced periodicity in the fiber core 
refractive index to create a sensor whose reflected or transmitted wavelength is a function 
of this periodicity. Grating-based sensors (e.g., Bragg gratings) can be easily multiplexed 
by using gratings of different wavelength as in the case of wavelength division 
multiplexing (WDM). Factors limiting the number of sensors in a single fiber include the 
limited bandwidth of the source as well as that supported by the fiber, and the range over 
which the physical parameter of interest is being measured. 

Another grating-based system developed at NASA Langley [2,3] has the ability to 
multiplex hundreds or thousands of Bragg gratings (with the same wavelength) in a single 
fiber. The system is based on the principle of optical frequency domain reflectometry 
(OFDR) and essentially eliminates the bandwidth limitations imposed by the WDM 
technique. Figure 1 shows results obtained when the system was used to measure strains in 
composite coupons subjected to combined thermal (elevated and cryogenic) and 
mechanical loading. The plots compare measurements obtained using the fiber optic 
sensors versus traditional resistance strain gages. Three optical fibers, each containing 
Bragg gratings, were bonded to the surface adjacent to two conventional foil strain gages. 
A single resistive thermal measurement device was in close proximity to the strain gages. 


Load (lbs) 


-12000 -6000 

Load (lbs) 

a) IM7/977-2, Tensile, -320°F. b) IM7/977-2, Compressive, -320°F. 

Figure 1 . Comparison of strain measurements from Bragg gratings and foil strain gages. 

The graphs in Figure 1 contain tension/compression data for IM7/977-2 composite 
specimens subjected to -320°F temperatures. The dashed line trace in each of the graphs 
represents the data from the fiber Bragg gratings while the solid trace represents data from 
the conventional foil strain gages. The horizontal axis displays load values in pounds- 
force while the vertical axis displays strain in microinches per inch. The fiber optic 
sensors show excellent agreement with the foil gages. At elevated temperatures (500°F), 

in tests of T650-35/PMR-15 composite specimens, a similarly good agreement between 
foil gage and fiber optic sensors was seen in compressive loading. Some small 
discrepancies were observed for tensile testing, possibly due to fiber-optic thermal 
compensation effects, and are under further investigation. 

The preliminary findings from these thermo-mechanical tests indicate that fiber 
Bragg gratings are capable of accurately measuring tensile and compressive strain at 
elevated and cryogenic temperatures, although compensation for the effect of temperature 
on the optical parameters of the Bragg gratings may be necessary, particularly at elevated 
temperatures. Additional research areas that are of concern include: 1) adhesive selection 
and bonding procedures for surface mounting the fiber optic sensors, 2) embedded fiber 
optic sensor characterization at elevated and cryogenic temperatures, and 3) transverse 
sensitivity of fiber optic sensors. These areas will continue to be explored in future 
research to support aerospace vehicle requirements. 


Ultrasonic sensing, applied in both active and passive modes, is another sensor 
technology area receiving considerable attention. Analysis of actively transmitted 
ultrasonic signals is a conventional NDE methodology that has long been used to detect 
and assess damage. However, such approaches use sensors that are scanned over the 
structure to provide a point-by-point representation of material properties and/or damage 
locations. Such scanning probe approaches are not currently feasible for continuous, on- 
board monitoring. Therefore, the use of arrays of permanently attached or embedded 
ultrasonic transducers, which act dually as transmitters and receivers, is being researched. 
Ultrasonic signals generated by one transducer are detected by neighboring transducers 
within an array. Damage along paths between the transducers can be detected, and with 
more complex analysis methods, material along secondary propagation paths that include 
reflections from structural boundaries can also be evaluated. The development of the 
Stanford Multi-Actuator Receiver Transduction (SMART) layer is an excellent example of 
recent efforts in this area [4]. Ongoing areas of research in active ultrasonic sensing 
technology for structural health monitoring include 1) the further improvement and 
characterization of miniaturized, rugged, embeddable sensors, 2) analysis methodologies 
for optimized sensor placement to enable characterization of damage throughout the entire 
structure rather than just along direct propagation paths, and 3) modeling of ultrasonic 
guided wave propagation that occurs when such sensors are attached or embedded on thin- 
walled aerospace structures. 

Passive ultrasonic monitoring, also known as acoustic emission (AE), also utilizes 
an array of ultrasonic sensors. The sensor array is used to passively monitor acoustic 
signals generated by damage mechanisms such as crack growth. AE is widely used as a 
conventional method for off-line structural assessment, and can also be implemented in- 
situ to monitor a structure while in service. This capability makes it well suited for on- 
board structural health monitoring of aerospace vehicles. However, considerably more 
research and development is required to make AE a more viable technology for IVHM. 
Successful implementation of AE requires sensors having lighter weight, increased 
sensitivity, and increased ruggedness over those currently available. Additionally, 
reductions in size, weight, and power requirements of the associated AE monitoring 
instrumentation are also needed. Advances in AE analysis methodologies are required to 
more accurately locate and identify damage, while intelligently discriminating extraneous 
noise from signals indicating actual damage. Ongoing efforts in this field include the 
development of AE multiplexing instrumentation that can miniaturize AE flight systems, 

the development of fiber optic AE sensors [5] and the development of Modal AE based 
analysis methods [6]. Another significant development is that of modeling approaches to 
better understand and predict AE propagation phenomena [7]. Such models are of benefit 
for a number of reasons to include the characterization of AE transducers, optimization of 
sensor placement on a structure, scaling of AE results from laboratory test coupons to full 
scale structures, and the development of new and automated AE data analysis methods. 


Conventional sensors such as strain gages, thermocouples, and accelerometers will 
also be used for structural health monitoring. One major issue for such sensors is the need 
to route large numbers of wires to provide power and data communication. This is an 
especially difficult problem when retrofitting these sensors into existing structures, such as 
the aging aircraft fleet. To address this concern, a prototype adaptable vehicle health- 
monitoring architecture has been developed [8] and flight tested. The architecture is self- 
contained and requires limited integration intrusion into existing systems, having "bolt- 
on/bolt-off ' simplicity. There are three operational levels to the architecture: one or more 
remote data acquisition units (RDAU) located throughout the vehicle; a command and 
control unit (CCU) located within the vehicle; and, a terminal collection unit (TCU) to 
collect analysis results from all vehicles. 

The RDAUs are multi-sensor interfaces with an on-board miniature computer, 
programmable digital interface, nonvolatile solid-state memory and a wireless transceiver 
for communication with the command and control unit. Communication is achieved by 
using wireless radio frequency transceivers operating at 433MHz. The RDAUs were 
designed to withstand impact during aircraft landing while mounted on the main landing 
gear, and have been vibration tested up to an acceleration amplitude of 20g at 2000 Hz. It 
was also designed to operate in non-environmentally controlled locations of the plane. 
The RDAU was thermally tested for temperatures ranging from -50°C to 55°C and 

pressure tested to simulate 50,000 ft altitude. Vibration tests verified that the remote data 
acquisition unit could operate at vibration levels representative of those experienced by 
commercial aircraft. During vibration testing, the final acceleration amplitude was 20g at 
2000 Hz. The remote data acquisition unit has an eight channel programmable digital 
interface, which allows the user discretion in choosing type of sensors, number of sensors, 
sensor sampling rate and sampling duration for each sensor. Programmable data 
acquisition circuitry and expert systems trained to performance baselines in each RDAU 
allow the architecture to be adaptable for many types of vehicles and structures. Once a 
suite of sensors has been chosen for each RDAU and installed on the vehicle, a baseline of 
acceptable vehicle performance is established from measurements acquired when the 
vehicle is performing correctly. Each RDAU uses an embedded expert system trained to 
its respective baseline 

The CCU is a computer-based subsystem that provides the communications, 
analysis repository, and user interface functions for the RDAUs. The CCU can also serve 
as a power management tool by regulating when individual or combinations of RDAUs 
are powered. A simple radio frequency (RF) wireless network of RDAUs can be 
controlled from a single CCU. The TCU provides the means to autonomously retrieve 
vehicle analysis results from all vehicle CCUs. The TCU performs analysis on results 
collected from all vehicles to identify any fleet-wide anomalies (e.g., all aircraft have the 
same faulty bearing at a similar location). The TCU develops the final summary of the 
vehicle health monitoring results that gets routed to the appropriate users (e.g., 
maintenance workers, airlines operations, etc.). 

This architecture system has been flight tested on NASA Langley's Airborne 
Research Integrated Experiments System (ARIES). There were 13 flight tests of the 
RDAU and CCU. The flight tests were performed to validate the following: the wireless 
radio frequency communication capabilities of the system, the hardware design, command 
and control, software operation, and, data acquisition, storage, and retrieval. A very 
rigorous test of the mechanical design was achieved by mounting the device on the left 
main landing gear. During the initial flight tests, none of the autonomous features had 
been installed. The system functioned as a remotely controlled data acquisition device. 
Measurements acquired during flights included take-offs, landings, vibration while gear 
was fully retracted, taxiing, and, touch and go landings. The flight tests demonstrated that 
the remotely controlled data acquisition capability worked correctly. 


Although most current visions of structural health monitoring systems are based on 
sensors that are attached to or embedded within the structure, the adaptation of non- 
contacting measurement systems should not be ruled out. Methods such as laser 
vibrometry [9], shearography [10], laser ultrasound [11], and infrared thermography [12] 
are examples of these techniques. These methods are typically applied externally to a 
structure to interrogate specific vehicle components where damage may have occurred. As 
such, they satisfy a critical role as part of an integrated vehicle health management system 
by providing enhanced ground-based diagnostic capabilities. These techniques can thus be 
used to validate fault indicators or damage sites identified by the on-board sensor systems. 
Further, there is potential that in the future such non-contact sensor systems could be 
incorporated into some aerospace structural systems, such as large space platforms. 

The use of structural vibration signatures as an indicator for airframe integrity is a 
growing field [13]. Vibration signatures are typically acquired at a single or small set of 
points on the aircraft surface using either a scanning laser vibrometer or accelerometers. 
The surface vibration data are acquired in response to an impulse force or frequency chirp 
applied by an excitation source at several locations about the vehicle. Comparison of the 
time-frequency and/or wavelet analyses of the signals obtained in baseline and aged 
conditions can lead to the identification of airframe cracks, disbonds, or fatigue [14-15]. 

Measuring the vibration signatures at only a few select points can often cause 
difficulties in determining the locus of damage. Therefore, it is desirable to acquire 
measurements at multiple points simultaneously. This enables spatial-temporal cross 
correlations between the data obtained at each measurement site, yielding improved 
accuracy in determining the location of airframe flaws. These capabilities are currently 
being pursued by the development of a multi-point laser vibrometer. 

Figure 2 shows the multi-point laser vibrometer configuration currently being 
pursued for the acquisition of vibration signatures over a two-dimensional array of 
measurement sites. Contrary to conventional scanning laser Doppler vibrometry (SLDV), 
the laser beam is not scanned from measurement site to measurement site. The time- 
dependent surface vibration is measured at each measurement site simultaneously so that 
vibration transients are preserved. Measurement site locations are generated by passing 
the output laser beam through a diffractive optical element, which can be fabricated to split 
the beam into any desired pattern with better than 90% efficiency and uniformity. 
Doppler-shifted scattered light is collected from each measurement location on the 
vibrating surface using a standard video camera lens, and mixed with reference light 
derived from the fundamental laser beam. The resulting light energy is intensity- 
modulated at the Doppler shift frequency experienced at each measurement location. The 

intensity-modulated light is subsequently focused to discrete sensor locations and digitized 
to obtain the time-dependent vibration signature. The data are further processed off-line to 
examine spatial-temporal cross correlation patterns for NDE purposes. 


optic ^ 

Test Specimen 


Fiber coupled 
reference light 

Interferometer and 
detector assembly 

Figure 2. Schematic of multi-point laser vibrometer for simultaneous measurements over 
a two-dimensional area 

Figure 3 shows a prototype multi-point vibrometer designed to measure the 
propagation of structural vibrations along a line. The object under test was a 4.9-meter 
long, 1.2-meter diameter aluminum cylinder fabricated in the same manner as an aircraft 
fuselage. Surface vibration measurements in response to an applied impulse force were 
obtained at 512 individual sensor sites along the 0.4-meter interrogation line. An example 
plot showing the time-dependent surface wave propagation is shown in figure 4. 

Figure 3. A 1 -dimensional multi-point 

u — nggassp 


Surf it h 


Vclocit; inm/s 


- 4.0 


- 2.0 


- 0.0 




III"" 4 - 


200 r 

1 1 1 1 1 1 1 

5 150 125 100 75 50 25 
Pixel Position 

Figure 4. Vibration measurements across the 
interrogation line after impulse force 


In addition to considering the types of sensors required to characterize structural 
integrity as part of an IVHM system, the data systems and processing architectures 
necessary to support such large numbers of heterogeneous sensors must also be 
considered. The architecture will be highly complex, as it must provide for the 

interrogation, digitizing, pre-processing, and archiving of massive amounts of raw signal 
information for consumption by modeling and analysis modules that will assess the 
integrity of the affected structural elements. Furthermore, since it is anticipated that a 
portion of the cost benefits gained through the deployment of on-board SHM systems is 
achieved through the elimination of certain maintenance and inspection procedures, the 
architecture's level of reliability must be commensurate with current regulatory guidance 
for assuring continued airworthiness [16, 17]. The magnitude of raw signal data, coupled 
with the complexity of the network interconnections and evolving diagnostic and 
prognostics methodologies, necessitate key architecture characteristics of scalability, 
robustness, flexibility, and maintainability [18]. 

Recognizing that architecture cannot be completely separated from application, 
work is underway to define a methodology that will aid in designing architectures for 
IVHM environments, and a layered reference architecture that facilitates scalability, 
robustness, flexibility, and maintainability [18, 19] is being developed. It is anticipated 
that a SHM architecture will support a data flow that includes real-time flight data (e.g., 
altitude, airspeed, accelerations, etc.) and sensor data (e.g., acoustic emission, strain, 
vibration, corrosion, etc.) that is tagged and conditioned (e.g., when, where, amount, rate, 
etc.), archived for trend analysis and usage history, then forwarded to a flight profiler for 
determination of phase-of-flight and maneuver. The tagged and conditioned data, coupled 
with flight profile, usage history, certification load data, and archived maintenance data, is 
then made available to the diagnostics/prognostics modules for degradation assessment. 

NASA is currently giving specific emphasis to architectures supporting the 
deployment of Langley's OFDR fiber optic Bragg grating sensor system technology as a 
key sensor suite component for on-board structural health and usage monitoring. A series 
of simulated axial fuselage lap joints have been instrumented with Bragg gratings and 
tested at NASA Langley for purposes of developing an architecture concept as well as 
building a proof-of-concept diagnostic inference model that can infer the presence of 
growing fatigue cracks at affected and adjacent fasteners [20]. As a result of these 
preliminary tests, several key architecture areas were identified as needing further 
investigation including (1) reduction, representation, and archival of large data sets 
suitable for retrieval by current degradation and damage assessment modules, (2) optimal 
techniques for increasing timeliness in demodulating the waveform, including dedicated 
distributed processors and analog techniques, (3) automatic identification and location of 
Bragg gratings within each fiber string, (4) miniaturization of components for sub-system 
distribution throughout the airframe, (5) fusion of fiber optic strain sensor data with other 
pertinent sensor information, and (6) architecture compatibility between laboratory test 
environments and flight- worthy avionics [21]. 


Extremely large numbers of a variety of sensor types will be necessary to provide real-time, 
on-board structural integrity assessment as part of an IVHM system for aerospace vehicles. 
These sensors will measure a multitude of parameters including strain, temperature, load, 
pressure, vibration, ultrasonic waves, and local chemistry. For flight applications, such 
sensors will need to be extremely lightweight, as well as be able to survive rugged 
environments. At present, fiber optic sensing is the leading candidate for such applications 
because of the ability to multiplex hundreds to thousands of sensors in a single fiber. 
Ultrasonic sensors, utilized in both active and passive modes, are also being studied for on- 
board structural health monitoring. For retrofit onto existing vehicles, a remote wireless 
sensor architecture is being developed that can support a variety of conventional sensor 

types, and be bolted into locations on vehicles without having to route wires to provide 
communication and power. Remote, non-contacting sensor technologies are being 
developed for complimentary ground inspections, and possibly for on-vehicle deployment. 
Further, the data systems and processing architectures that will be required to support these 
massive numbers of diverse sensors are being considered, with special emphasis on the 
integration of fiber optic sensors with more conventional sensor types. 


1. Steenkiste, R. J. and Springer, G. S., Strain and Tempertaure Measurement with Fiber 
Optic Sensors, Technomic Publishing (Lancaster, PA) 1997, pp. 3-4. 

2. Froggatt, M. and Moore, J., in Applied Optics, vol. 37, No. 10, 1998, pp. 1741-1746. 

3. Childers, B. A., Froggatt M. E., Allison, S. G., Moore, T. C, Hare, D. A., Batten, C. 
F., and Jegley, D. C, in Smart Structures and Materials Conference, SPIE Vol. 4332, 
2001, pp. 133-142. 

4. M. Lin, 2nd International Workshop on Structural Health Monitoring, ed. by F. K. 
Chang, Stanford Univ., Technomic, 1999, pp. 603-611. 

5. Duke, J. C, Cassino, C. D., Childers, B. A., and Prosser, W. FL, "Characterization of 
an EFPI AE Sensor," submitted to Research in Nondestructive Evaluation (2002). 

6. Gorman, M.R., in Review of Progress in QNDE, Vol. 17, eds. D. O. Thompson and D. 
E. Chimenti, Plenum Press, New York, 1998, pp. 557-564. 

7. Prosser, W. FL, Hamstad, M. A., Gary, J, and O'Gallagher, A, Journal of Acoustic 
Emission,Yo\. 17(1-2), 1999, pp. 37-47. 

8. Woodard, S. E., Coffey, N. C, Taylor, B. D., Brett, R. R. and Woodman, K. L., 
"Autonomous Health Monitoring Architecture and Hardware," United States Patent 
Provisional Application and NASA Langley Research Center Invention Disclosure 
LAR 16407-1, Feb 2002. 

9. Castellini, P., and Revel, G. M., "Damage Detection by Laser Vibration 
Measurement", 15 th World Conference on Non-Destructive Testing, Rome, Italy, 
October 15-21, 2000. 

10. Maji A. K., Experimental Mechanics J., Vol. 34, No. 2, 1997, pp. 197-204. 

11. Dewherst, R. J., Applied Optics and Optoelectronics IoPP, Bristol, 1996, pp. 289-298 

12. Cramer, K. E., and Winfree, W. P., Thermosense XVIII, Orlando, April 8-12, 1996. 

13. U. Qidwai, A. H. Costa, and C. H. Chen, Insight, vol. 41, no. 11, 1999, pp.700-703 

14. Deobling, S.W.; Farrar, C.R.; Prime, M.B., The Shock and Vibration Digest, Vol. 30, 
(1998) 2, pp. 91-105. 

15. Krohn, N., Stoessel, R., and Busse, G., in Review of Progress in QNDE, Vol. 20, eds. 
D. O. Thompson and D. E. Chimenti, AIP, New York, 2000, pp. 1666-1672. 

16. Kent, R. M., Murphy, D.A., NASA/CR-2000-209848, January 2000. 

17. Munns, T.E., Beard, R.E., Culp, A.M., Murphy, D.A., Kent, R.M., NASA/CR-2000- 
2 1 064 3, December 2000. 

18. Mukkamala, R., Proceedings of the IEEE/AIAA 1 9th Digital Avionics Systems 
Conference; 2000, pp. 6.C.4.1 - 6.C.4.8. 

19. Mukkamala, R., Bhoopalam, K., Dammalpati, S., Proceedings of the IEEE/AIAA 20th 
Digital Avionics Systems Conference; 2001, pp. 3.D.2.1 - 3.D.2.10 

20. Munns, T.E., Kent, R.M., Bartolini, A., Gause, C.B., Borinski, J.W., Dietz, J., Elster, 
J.L., Boyd, C, Vicari, L., Cooper, K., Ray, A., Keller, E., Venkata, V., Sastry, S.C., 
NASA/CR-2002-2 11428 February 2002. 

21. Doggett, W., Vazquez, S., Proceedings of the IEEE/AIAA 19th Digital Avionics 
Systems Conference; 2000, pp. 6.D.2.1 - 6.D.2.8