INITIAL INTEGRATION OF ULTRASONIC STRUCTURAL HEALTH MONITORING OF SELF-HEALING MATERIALS

Shvetashva SURI 1, Ameralys CORREA 2,

Pradeep ROHATGI 2, Nathan P. SALOWITZ 1

1 Department of Mechanical Engineering, University of Wisconsin-Milwaukee 3200 N Cramer St, Milwaukee, Wisconsin, 53211 U.S.A. Salowitz@UWM.edu

2 Department of Materials Science and Engineering, University of Wisconsin-Milwaukee 3200 N Cramer St, Milwaukee, Wisconsin, 53211 U.S.A.

Key words: Ultrasonic, SHM, Self-Healing Materials, Smart Structures, Repair. Abstract Structural Health Monitoring (SHM) and self-healing materials are both areas of current,

ongoing scientific research toward intelligent, maintenance-free materials. SHM has developed techniques to detect, locate, characterize, and quantify damage in simple and complex structures. Simultaneously, self-healing materials have been developed with an intrinsic capability to mend themselves, restoring mechanical properties after damage occurs. These technologies have largely been developed independently, but together could lead to the use of lighter weight structures that can incur, detect, and automatically heal damage without catastrophic failure as well as inspect the structural integrity of healed structures and estimate residual healing capabilities that could be applied if further damage occurred in the future. However, there are significant challenges to integrating these technologies.

Each technology has developed multiple approaches but has been developed with

fundamental incompatibilities. For example, samples of self-healing materials are often relatively thick (order centimeters) and composed of materials with low strain wave propagation speeds, or even containing encapsulated liquid, leading to short attenuated ultrasonic waveforms propagating in thick structures and inhibiting electromagnetic SHM.

This paper presents initial studies, experiments, results, and analyses integrating structural

health monitoring with self-healing materials. This preliminary study focuses on ultrasonic SHM with metal matrix self-healing composites containing shape memory alloys to enable crack closing capabilities. Initial techniques, data, and capabilities of detecting damaged and healed states of a self-healing structure using ultrasonic SHM are presented demonstrating a fundamental potential to identify damaged and healed states in a basic system.

Remaining challenges and potential approaches in both fields are presented including

tuning self-healing materials to support SHM, characterizing healed structures with embedded SHM systems, determining remaining intrinsic healing potential of self-healing materials, and integrating combined SHM techniques with alternate or more complex self-healing materials.

8th European Workshop On Structural Health Monitoring (EWSHM 2016), 5-8 July 2016, Spain, Bilbao

www.ndt.net/app.EWSHM2016M

ore

info

abo

ut th

is a

rtic

le: h

ttp://

ww

w.n

dt.n

et/?

id=

1991

9

mailto:Salowitz@UWM.edu

2

1 INTRODUCTION

Structural Health Monitoring (SHM) and self-healing materials are complementary areas of current, ongoing scientific research with similar goals of developing intelligent, maintenance free materials. These fields seek to improve the safety and efficiency of engineered structures ranging from vehicles to civil infrastructure while reducing maintenance costs. Each field takes a different approach to mitigate the effects of damage and prevent catastrophic failure improving safety. This mitigation will also result in a shift in design with reduced margins, enabling the creation of lighter weight, more fuel efficient vehicles [1, 2, 3]. Real time knowledge of damage will enable a shift from maintenance schedules to condition based maintenance, or modified performance envelopes and self-healing capabilities will enable repairs without disruption to service completing the closed loop damage-repair cycle illustrated in Figure 1. However, to date the fields have developed independently despite their complementary capabilities.

2 BACKGROUND

2.1 Structural health monitoring

SHM takes the approach of permanently integrating sensors into structures to enable real time automated damage detection. Knowledge of the state of damage can be used to refine operational envelopes or plan maintenance and repairs as necessary. The permanent nature of these systems reduces signal variation and potential for user error enabling a greater level of automation and autonomy than traditional Non-Destructive Evaluation (NDE) [4]. Many

Service

Self-Healing

Evaluation of “Healed” Structure

Damage Occurrence

Sim

ula

ted

d

ama

ge

Ind

icat

ed

dam

age

Damage Detection

Figure 1 – The damage cycle of self-healing materials with structural health monitoring capabilities

Damage-repair Cycle

3

approaches have been employed to detect damage with SHM systems including ultrasonic methods, strain mapping, and eddy current analysis [5, 6]. Ultrasonic techniques are among the most common because of their many benefits including: sensitivity to multiple forms of damage, ability to locate damage, broad structural coverage with a sparse array of sensors, and ability to place hardware in locations removed from anticipated damage [7, 8, 9]. Despite these advantages, ultrasonic waveforms can be hard to interpret and detection of damage smaller than one half the wavelength of the ultrasonic signal is an ongoing challenge requiring specific setups, calibration, and complex analysis techniques [10, 11, 12]. Self-healing materials provide an alternative approach where these issues are not as critical.

2.2 Self-healing materials

Self-healing materials avoid catastrophic failure by creating an intrinsic ability for the material to repair itself after a damaging event. Many techniques have been developed to accomplish this including integration of shape memory fibers that bridge a crack and enable crack closing of a metallic or polymer matrix. Encapsulated healing agents, which can be thermoset, polymeric, catalytically activated, or solder based, can also be integrated into a structure.

Self-Healing materials have been classified into two groups: autonomic which heal without actuation, and nonautonomic which require external actuation [13]. Nonautonomic self-healing materials require knowledge of the existence and location of damage so that energy can be applied to actuate healing in the correct location.

The most basic self-healing materials simply have the ability to close a crack. This simplicity has led to the development of multiple systems of shape memory material combined with different matrices in composite structures, often with very low fiber fractions [14, 15, 16].

Much of the work on self-healing materials has focused on empirically demonstrating the ability to close and/or seal a crack. The thermodynamics and kinetics of self-healing are currently not well understood, and a broad framework for design of self-healing microstructure has yet to be developed.

The unique capabilities and remaining challenges in these technologies complement each

other. For example, in nonautonomous self-healing materials with crack closing capabilities employing shape memory fibers in a matrix, it is typical to heat the entire structure to actuate healing. Knowledge of the existence of damage is assumed and location is ignored. Structural health monitoring can solve these issues by detecting the damage existence and identifying damage location, addressing the real issue of damage existence and enabling localized heating to actuate healing, reducing energy requirements.

3 APPROACH

This preliminary research seeks to integrate two of the most common and fundamental technologies from each field to demonstrate initial compatibility and identify challenges and areas of future research.

4 RESULTS

As an initial proof of concept, piezoelectric transducers were adhered to two distinct

samples to inspect ultrasonic wave propagation capabilities in soft metals and across closed

4

cracks in self-healing materials. In each case, two piezoceramic transducers measuring 6 mm in diameter and 1 mm thick, composed of APC-850 Lead Zirconate Titanate (PZT) material were adhered to each sample [17]. In both cases, the system was actuated with a Hann windowed 5 peak tone burst centered at 350 kHz using a Keysight 33500B waveform generator amplified to ±40 Volts using a Krohn-Hite 7602M amplifier. A Tektronix MDO-3014 oscilloscope triggered on the actuation signal and then collected data on the actuation input and sensed signal at a sampling rate of 50 MHz. An example test setup is shown in Figure 2. The results of these experiments are discussed in the following subsections.

4.1) Experimental ultrasonic testing of metals with low melting temperature and slow wave propagation

Metals of low melting temperature are of interest for this research such as solders, Tin and Zinc. While not structural, many current basic self-healing materials are composed of them because they allow for casting of the self-healing materials with crack closing capabilities without exceeding the training temperature of NiTi shape memory alloy [18, 19]. It is anticipated that future development will enable the use of stiffer structural materials perhaps in conjunction with more advanced shape memory fibers possessing higher training and transition temperatures. However, the current technology is potentially limiting in this study because the metals currently in use typically have: low stiffness, low speed of sound, and high damping compared to the aluminum or steel often inspected with ultrasonics as illustrated by the data in Table 1 [20, 21].

Longitudinal

(m/s)

Transversal

(m/s)

Extensional

(m/s)

Steel 5790 3100 5000

Aluminum

(Al)

6420 3040 5000

Tin (Sn) 3320 1670 2730

Zinc (ZN) 4210 2440 3850

Table 1 – Wave speeds in various materials of interest

Keysight 33500B waveform generator

Krohn-Hite 7602M amplifier

0.5 1 1.5 2 2.5 3 3.5 4 4.5

x 10-4

-0.1

-0.05

0

0.05

0.1

time (s)

Volta

ge (V

)

Sensed propagated signal

Sample

Actuation Signal

Tektronix MDO3014 Oscilloscope

Figure 2 – Example test setup

1 1.1 1.2 1.3

x 104

-40

-20

0

20

40

time (s)

Volta

ge (V

)

5

To ensure experimental capabilities, a simple plate of 99% pure Tin plate measuring 30.6 cm by 30.7 cm by 4 mm thick was equipped with piezoceramics 20.6 cm apart and a signal was propagated between the piezoceramic transducers. The actuation signal is shown in blue in Figure 3. The sensed signal is shown in green in Figure 3. The sensed signal was very strong with peak voltages near ±0.2V and much of the signal around ±0.050 V, well above the noise floor.

A travel time of approximately 0.000112 seconds was measured for the first arrival which corresponds to the signal traveling 0.200 m at 1785 m/s.

4.2) Experimental ultrasonic testing of wave propagation across a crack in a thick self-healing, crack closing structure

A sample of a zinc alloy with crack closing capabilities, shown in Figure 4, was also tested for basic compatibility. Piezoceramic transducers were adhered to the specimen 79 mm apart (on center, in the closed state). The specimen had an initial crack before the piezoceramics were adhered to it. The crack was created by etching a stress concentration into the specimen and then mechanically overloading it in tension. The self-healing, crack closing zinc alloy was subjected to cycling with data taken in: a ‘crack closed’ state, like that shown in Figure 5, an open state with a 2 mm gap like that shown in Figure 6, and another closed state like that shown in Figure 5 through 3 crack opening events.

The, far from ideal, sample measured 9.7 cm by 3.4 cm by 1.4.cm in the closed state. As a result, the wave propagation is likely better modeled as a surface or Rayleigh wave, not a Lamb wave. Surface and Rayleigh waves are most sensitive to surface conditions with decreasing sensitivity through the depth of the structure. In this specific work, that is not critical because the cracking is through the entire thickness. However, in future work thinner samples better suited to Lamb wave propagation will be developed. While the propagated strain signal was weak, the signal showed unique traits in the open and closed states. Conditioning of the signals involved averaging them over 4 consecutive datasets taken in the same condition and smoothed to reduce high frequency noise. The data was also centered to a mean Voltage of 0 V, removing any DC offset.

The first observed signal to arrive, above the noise floor, for this series of tests is shown in Figure 7. The left hand column of plots in Figure 7 is data collected in ‘crack closed’ states, corresponding to Figure 5, and the right hand column is data collected in ‘crack open’ states corresponding to Figure 6, with each row being data from an initial ‘crack closed’ state and

Figure 4 – Thick self-healing Zinc alloy for initial testing.

0 1 2

x 10-4

-50

0

50

0 1 2

x 10-4

-0.2

0

0.2

Figure 3 – Signal propagated 20.0 cm through a 4mm thick simply supported Tin plate

Time (S)

Vol

tage

(V

)

6

1.4 1.5 1.6 1.7

x 10-4

-0.02

-0.01

0

0.01

0.02Initial crack closed (healed) state

Time (S)

Vol

tage

(V

)

1.4 1.5 1.6 1.7

x 10-4

-0.02

-0.01

0

0.01

0.02First crack open state

Time (S)

Vol

tage

(V

)

1.4 1.5 1.6 1.7

x 10-4

-0.02

-0.01

0

0.01

0.02Second crack closed state

Time (S)

Vol

tage

(V

)

1.4 1.5 1.6 1.7

x 10-4

-0.02

-0.01

0

0.01

0.02Second crack open state

Time (S)

Vol

tage

(V

)

1.4 1.5 1.6 1.7

x 10-4

-0.02

-0.01

0

0.01

0.02Third crack closed state

Time (S)

Vol

tage

(V

)

1.4 1.5 1.6 1.7

x 10-4

-0.02

-0.01

0

0.01

0.02Third crack open state

Time (S)

Vol

tage

(V

)

Figure 7 – First arrival of signals propagated in small thick samples of self-healing materials with crack closing capabilities in the closed (left column) and open (right column) states.

Figure 5 – Specimen in the ‘crack closed’ state

97 mm

Figure 6 – Specimen in the ‘crack open’ state with a 2 mm gap (enlarged to the

right showing bridging NiTi wires)

99 mm 2 mm

Crack Bridging

NiTi wires

7

then the subsequent ‘crack open’ state, and the row below being data from the next ‘crack (re)closed’ state, and following crack (re)opened state. The consistency of the signals in the left column (‘crack closed’ state) combined with the differences compared to the right column (‘crack open’ state), indicates the potential to automate and identify these two states in the structure. Variation exists between signals in the different ‘crack closed’ cases and increases over cycling. This variation is likely due to increasing damage through cycling, changing the structure, combined with changes in contact caused in part because the means of opening the crack was poorly controlled. Signal variation will need to be analyzed and compensated in order to automate systems for SHM.

The fact that there are signals in the open state, though weaker, also indicates that some wave propagation occurs through the bridging structure. Analysis of this phenomena could provide insights into identifying signal traits of the two structural states. Additional significant signal differences between the two states occurred later in the signal, however, that region of the signal also contains reflections and constructive and destructive interference which is prohibitively complex to analyze and therefore is regularly discarded.

5 CONCLUSION & FUTURE WORK

This experiment found potential to integrate ultrasonic SHM with self-healing materials with crack closing capabilities, as indicated by signal consistency in ‘crack open’ states as opposed to ‘crack closed’ states. Significant work remains to understand these signals, signal properties, and signal variation. A major step to support this research would be to produce stronger signals, clearly above the noise floor, and reduce or delay the arrival of reflections. To that end, self-healing samples are under development which are intended to be better tuned to the integration of ultrasonic structural health monitoring systems. Specific goals are to make the samples:

Thinner to promote lamb wave propagation that is a full plate wave and contains multiple wave modes. [22, 23] Larger in lateral dimensions to reduce the effect of reflections. Generate significant crack closing forces to better promote strain wave propagation across the closed crack. [24]

Modifying the form of the specimens will support collection of better data, but analysis and

theory remain to be developed to enable automated SHM in self-healing materials. Ongoing work is seeking to:

Compensate for signal changes between ‘crack closed’ states through cycles and accumulated damage. Identify signal traits indicative of generic ‘crack closed’ and ‘crack open’ states. Model ultrasonic wave propagation across crack bridging fibers and accounting for disbonding between the shape memory fibers and damaged matrix. This may be particularly complicated due to the complex nature of shape memory materials which may have varying properties due to the strain exposed to and temperature cycling

Integrating Structural Health Monitoring systems into self-healing materials has the

potential to advance both technologies and the broader field of intelligent structures leading to improved safety and efficiency of engineered systems. These systems take two complementary approaches toward the same goal and simultaneous deployment can make a system with greater

8

capabilities than the sum of the parts. Structural health monitoring has limitations on accurate location, characterization, and quantification of small scale damage that self-healing materials can tolerate and mend autonomically. Conversely, larger scale damage can be detected by SHM that could then automatically actuate non-autonomic self-healing systems.

Future development may also enable SHM systems to analyze the mechanical properties of the healed structure and estimate residual healing capability in a self-healing structure to mend possible future damage. These capabilities would serve to verify the quality of a mend in a structure after healing and provide information about the operational envelope and tolerable future damage. Estimation of remaining healing agents could be accomplished based on the extent of damage repaired or more direct measurement of the presence of liquid agents that would attenuate ultrasonic signals or alter electrical conductivity. These goals present significant challenges that are not new to SHM, however, the automated nature of repair in self-healing materials will help to reduce structural variations supporting automated SHM analysis at a level not previously possible [25, 26].

The inherent complexity of self-healing materials that may include combinations of shape memory materials, encapsulated liquid mending agents, and chemical catalysts also complicates analysis. SHM has demonstrated the capability to inspect complex structures, using different SHM technologies (e.g. ultrasonics, eddy currents, strain mapping) [27]. Integration of multiple heterogeneous sensing systems enabled by microfabricated sensor networks provides another path to address structural complexity [28, 29, 30, 31]. The inherent composite nature of self-healing materials also supports embedding sensor networks [32].

Prudent harmonious system design and integration of SHM with self-healing materials can also improve capabilities. SHM capabilities could guide the development of a design framework for self-healing microstructure that simultaneously addresses electrical resistance, ultrasonic attenuation, generation of acoustic emissions, and repair activation in pristine, damaged, and healed states [33]. Initial studies, like this one, must focus on simplified systems to build a foundation on which to build more complex and complete systems.

REFERENCES

[1] M. M. Derriso, S. E. Olson, M. P. Desimio and D. M. Pratt, "Why are there so Few Fielded SHM Systems for Aerospace Structures," in Proceedings of the 6th International Workshop on Structural Health Monitoring, Stanford, CA, 2007.

[2] J. Telgkamp and H.-J. Schmidt, "Benefits by the Application of Structureal Health Monitoring (SHM) Systems on Civil Transport Aircraft," in Structural Health Monitoring 2013 - From Diagnostics & Prognostics to Structural Health Management, Stanford, CA, 2003.

[3] G. Li, Self-Healing Composites: Shape Memory Polymer Based Structures, John Wiley & Sons, 2014.

[4] M. M. Derriso, M. P. Desimio, C. D. McCurry, C. M. S. Kabban and S. E. Olson, "Industrial Age NDE to Information Age SHM," in Structural Health Monitoring 2013, Stanford, CA, 2013.

[5] F.-K. Chang, "Structural Health Monitoring 2013 - Roadmap to Intelligent Structures," in Proceedings of the 9th International Workshop on Structural Health Monitoring, Stanford, CA, 2013.

9

[6] F.-K. Chang and F. Kopsaftopoulos, "Structural Health Monitoring 2015 - System Reliability for Verification and Implementation," Stanford, 2015.

[7] V. Giurgiutiu, "Recient Advances in Structural Health Monitoring with Piezoelectric Wafer Active Sensors," Coumbia, SC, 2006.

[8] V. Giurgiutiu, A. Zagrai and J. J. Bao, "Piezoelectric wafer embedded active sensors for aging aircraft structural health monioring," Structural Health Monitoring, vol. 1, no. 1, pp. 41-61, July 2002.

[9] V. Giurgiutiu, A. Zagrai and J. Bao, "Damage Identification in Aging Aircraft Structures with Piezoelectric Wafer Active Sensors," Journal of Intelligent Material Systems and Structures, vol. 15, pp. 673-687, 2004.

[10] Y. Lu, L. Ye, Z. Su and C. Yang, "Quantitative assessment of through-thickness crack size based on Lamb wave scattering in aluminium plates," NDT&E International, vol. 41, no. 1, pp. 59-68, January 2008.

[11] B. C. Lee and W. J. Staszewski, "Modelling of Lamb waves for damage detection in metallic structures: Part II. Wave intearactions with damage," Smart Materials and Structures, vol. 12, no. 5, 25 September 2003.

[12] D. Rehbein, R. Thompson and O. Buck, "CRACK LENGTH DETERMINATION BY ULTRASONIC METHODS," in Review of Progress in Quantitative Nondestructive Evailuation, Vol. 12, New York, Plenum Press, 1993, pp. 2159-2165.

[13] S. K. Gosh, Self-healing Materials: Fundamentals, Design Strategies, and Applications, WILEY-VCH Verlag GmbH & Co, 2009.

[14] A. D. Moghadam, B. F. Schultz, J. Ferguson, E. Omrani, P. K. Roahtgi and N. Gupta, "Functional metal matrix composites: self-lubricating, self-healing, and nanocomposites-an outlook," Journal of Materials, vol. 6, pp. 872-881, 2014.

[15] M. Nosonovski and P. Rohatgi, Biomimetics in Materials Science: Self-healing, self-lubricating, and self-cleaning materials, Springer, 2011.

[16] J. M. Lucci, R. S. Amano, P. Rohatgi and B. Schultz, "Experiment and computational analysis of self-healing in an aluminum alloy," in Proceedings of the ASME 2008 International Mechanical Engineering Congress and Exposition, 2008.

[17] American Piezoelectric Corperation, February 2011. [Online]. Available: http://www.americanpiezo.com/materials/apc_properties.html.

[18] W. P. Leser, Embedded Shape Memory Alloy Particles for the Self-Sensing of Fatigue Crack Growth in an Aluminum Alloy, Raleigh, North Carolina: North Carolina State University, 2014.

[19] J. Martinez-Lucci, A. Ruzek, S. Misra, P. Rohatgi and R. Amano, "Self Healing in Metal Castings," in American Foundary Society Proceedings 2011, Schaumberg, IL, 2011.

[20] J. Aldred, Manual of Sound Recording, London: Fountain Press, 1972.

[21] Wikipedia, "Speed of sound of the elements (Data Page)," Wikimedia Foundation, Inc., 27 12 2015. [Online]. Available: https://en.wikipedia.org/wiki/Speeds_of_sound_of_the_elements_(data_page). [Accessed 01 02 2016].

[22] A. Cuc and V. Giurgiutiu, "Propagation of Guided Lamb Waves in Bonded Specimens Using Piezoelectric Wafer Active Sensors," in SPIE Smart Structures and

10

Materials 2006: Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems, 2006.

[23] V. Giurgiutiu, "Lamb Wave Generation with Piezoelectric Wafer Active Sensors for Structural Health Monitoring," in SPIE Smart Structures and Materials 2003: Smart Structures and Integrated Systems, 2003.

[24] J. Yang, STRUCTURAL HEALTH MONITORING TECHNOLOGY FOR BOLTED CARBON-CARBON THERMAL PROTECTION PANELS, Stanford: Stanford University, 2005.

[25] J.-B. Ihn and F.-K. Chang, "Detection and monitoring of hidden fatigue crack growth using a built-in piezoelectric sensor/actuator network: II. Validation using riveted joints and repair patches," Smart Materials and Structures, vol. 13, pp. 621-630, 2004.

[26] I. Mueller, V. JANAPATI, S. BANERJEE, K. LONKAR, S. ROY and F.-K. Chang, "On the Performance Quantification of Active Sensing SHM Systems Using Model-Assisted POD Methods," in Structural Health Monitoring 2011: Condition Based Maintenance and Intelligent Structures, Stanford, CA, 2011.

[27] N. Salowitz, Y.-H. Li, S.-J. Kim, S. Roy and F.-K. Chang, "Structural Health Monitoring of High Temperature Composites," in Proceedings of the ASME 2011 International Mechanical Engineering Congress & Exposition, Denver, Colorado, U.S.A., 2011.

[28] N. Salowitz, Z. Guo, S. Roy, R. Nardari, Y.-H. Li, S.-J. Kim, F. Kopsaftopoulos and F.-K. Chang, "Recent advancements and vision toward stretchable bio-inspired networks for intelligent structures," Structural Health Monitoring - An International Journal, vol. 13, no. 6, pp. 609-620, November 2014.

[29] N. Salowitz, Z. Guo, Y.-H. Li, K. Kim, G. Lanzara, K. Kim, Y. Chen and F.-K. Chang, "Development of a bio-inspired stretchable network for intelligent composites," in International Conference on Composite Materials, Jeju, South Korea, 2011.

[30] N. Salowitz, Z. Guo, S.-J. Kim, Y.-H. Li, G. Lanzara and F.-K. Chang, "Microfabricated Expandable Sensor Networks for Intelligent Sensing Materials," IEEE Sensors Journal, vol. 14, no. 7, pp. 2138-2144, 2014.

[31] N. Salowitz, Z. Guo, S.-J. Kim, Y.-H. Li, G. Lanzara and F.-K. Chang, "Screen Printed Piezoceramic Actuators/Sensors Microfabricated on Organic Films and Stretchable Networks," in Structural Health Monitoring 2013 - A Roadmap to Intelligent Structures, Stanford, CA, 2013.

[32] N. Salowitz, Z. Guo, Y.-H. Li, K. Kim, G. Lanzara and F.-K. Chang, "Bio-inspired stretchable network-based intelligent composites," Journal of Composite Materials, vol. 47, no. 1, pp. 97-106, January 2013.

[33] R. Oishi and H. Nagai, "Strain sensors of shape memory alloys using acoustic emissions," Sensors and Actuators A Physical, vol. 122, pp. 39-44, 2005.