River Rapids Conference Proceedings of the Society for Experimental Mechanics Series Experimental and Applied Mechanics, Volume 6 Nancy Sottos Robert Rowlands Kathryn Dannemann Proceedings of the 2014 Annual Conference on Experimental and Applied Mechanics River Publishers
Conference Proceedings of the Society for Experimental Mechanics Series Series Editor Tom Proulx Society for Experimental Mechanics, Inc. Bethel, CT, USA
River Publishers Nancy Sottos • Robert Rowlands • Kathryn Dannemann Editors Experimental and Applied Mechanics, Volume 6 Proceedings of the 2014 Annual Conference on Experimental and Applied Mechanics
Published, sold and distributed by: River Publishers Broagervej 10 9260 Gistrup Denmark www.riverpublishers.com ISBN 978-87-7004-902-3 (eBook) Conference Proceedings of the Society for Experimental Mechanics An imprint of River Publishers © The Society for Experimental Mechanics, Inc. 2015 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, or reproduction in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Preface Experimental and Applied Mechanics, Volume 6: Proceedings of the 2014 Annual Conference on Experimental and Applied Mechanics represents one of eight volumes of technical papers presented at the 2014 SEM Annual Conference & Exposition on Experimental and Applied Mechanics organized by the Society for Experimental Mechanics and held in Greenville, SC, June 2–5, 2014. The complete Proceedings also includes volumes on: Dynamic Behavior of Materials; Challenges in Mechanics of Time-Dependent Materials; Advancement of Optical Methods in Experimental Mechanics; Mechanics of Biological Systems and Materials; MEMS and Nanotechnology; Experimental Mechanics of Composite, Hybrid, and Multifunctional Materials; Fracture, Fatigue, Failure and Damage Evolution. Each collection presents early findings from experimental and computational investigations of an important topic within the field of Experimental Mechanics. This volume includes papers on Residual Stress, Thermomechanics and Infrared Imaging, Hybrid Techniques and Inverse Problems. Residual stresses are especially important in engineering systems and design. The insidious nature of residual stresses often causes them to be underrated or overlooked. However, they profoundly influence structural design and substantially affect strength, fatigue life, and dimensional stability. Since residual stresses are induced during most materials processing procedures, for example, welding/joining, casting, thermal conditioning, and forming, they must be included and addressed in engineering design and applications. In recent years the applications of infrared imaging techniques to the mechanics of materials and structures have grown considerably. The expansion is marked by the increased spatial and temporal resolution of the infrared detectors, faster processing times, and much greater temperature resolution. The improved sensitivity and more reliable temperature calibrations of the devices have meant that more accurate data can be obtained than were previously available. Advances in inverse identification techniques have been coupled with optical methods that provide surface deformation measurements and volumetric measurements of materials. In particular, inverse methodology was developed to more fully use the dense spatial data provided by optical methods to identify mechanical constitutive parameters of materials. Since its beginnings during the 1980s, creativity in inverse methods has led to applications for a wide range of materials, with different constitutive behavior, across heterogeneous material interfaces. Complex test fixtures have been implemented to produce the necessary strain fields for identification. Force reconstruction has been developed for high strain rate testing. As developments in optical methods improve for both very large and very small length scales, applications of inverse identification methods have expanded to include geological and atomistic events. Urbana, IL, USA Nancy Sottos Madison, WI, USA Robert Rowlands San Antonio, TX, USA Kathryn Dannemann v
Contents 1 Electrical Impedance Spectroscopy for Structural Health Monitoring........................... 1 Geoffrey A. Slipher, Robert A. Haynes, and Jaret C. Riddick 2 In situ Observation of NiTi Transformation Behaviour: A Micro–Macro Approach................. 13 Kasun S. Wickramasinghe, Rachel A. Tomlinson, and Jem A. Rongong 3 Bio-Inspired Design of a Multi-scale Pass Band Frequency Sensor Using Local Resonance Phenomena...................................................... 21 Riaz Ahmed and Sourav Banerjee 4 Dynamic Equations for an Isotropic Spherical Shell Using Power Series Method and Surface Differential Operators ...................................................... 29 Reza Okhovat and Anders Bostr€om 5 Hydrogen Embrittlement and “Cold Fusion” Effects in Palladium During Electrolysis Experiments. . . . . . 37 A. Carpinteri, O. Borla, A. Goi, S. Guastella, A. Manuello, and D. Veneziano 6 Torque Arm Actuated Bi-Stable Buckled Energy Harvester Characterization..................... 49 D.A. Porter and T.A. Berfield 7 Validating FSI Simulations in LS-DYNA 971 R7............................................ 55 Kevin A. Gardner, Jeremy D. Seidt, and Amos Gilat 8 Fundamental Frequencies of Slender Beams Subject to Imposed Axial End Displacements ........... 59 G. Piana, A. Manuello, R. Malvano, and A. Carpinteri 9 Characterization of a Heating and Quenching Apparatus for Microgravity Testing................. 67 Anthony S. Torres, Jeff Ganley, and Arup Maji 10 Phase Unwrapping Work Done via Graphic Processing Unit .................................. 75 M.J. Huang and Y.C. Liu 11 Classification of Low Velocity Impact Using Spiral Sensing Technique........................... 79 Chijioke Agbasi and Sourav Banerjee 12 Residual Stress Measurements in Finite-Thickness Materials by Hole-Drilling..................... 89 Gary S. Schajer and Colin Abraham 13 Residual Stress Response to Peening in Metallic Glass ....................................... 99 B. Jayakumar, M. Allahkarami, and J.C. Hanan 14 Residual Stress Modeling and Measurement in Aluminum Wrought Alloys ....................... 105 Bowang Xiao, Qigui Wang, Cherng-Chi Chang, and Josie E. Rewald 15 Notch Fatigue Behaviour of Shot Peened High-Strength Aluminium Alloys: Role of the Residual Stress Field Ahead of the Notch Root .................................... 113 M. Benedetti, V. Fontanari, M. Allahkarami, and J.C. Hanan vii
16 Residual Stress of Individual Aluminum Grains from Three Dimensional X-Ray Diffraction.......... 123 M. Allahkarami, B. Jayakumar, and J.C. Hanan 17 Incremental Ring Core by Optical Methods: Preliminary Results ............................... 131 Antonio Baldi and Filippo Bertolino 18 Uncertainty Quantification in VFM Identification........................................... 137 P. Wang, F. Pierron, O.T. Thomsen, M. Rossi, and P. Lava 19 Modal Identification of Over-Damped Structural Systems Using Extended Ibrahim Time-Domain Method......................................................... 143 Chang-Sheng Lin and Tse-Chuan Tseng 20 Structural Health Monitoring by Laser Shearography: Experimental and Numerical Investigations ........................................................... 149 Xiaoran Chen, Morteza Khaleghi, Ivo Dobrev, Weiyuan Tie, and Cosme Furlong 21 On Improving Thermoelastic Stress Analysis Data Near Edges of Discontinuities ................... 157 W.A. Samad and R.E. Rowlands 22 Measurement of Stress Network in Granular Materials from Infrared Measurement ................ 163 Pawarut Jongchansitto, Xavier Balandraud, Michel Gre´diac, and Ittichai Preechawuttipong 23 Influence of Relative Humidity on the Thermomechanical Behavior of PA6.6...................... 167 Adil Benaarbia, Andre´ Chrysochoos, and Gilles Robert 24 Temperature Field in FSW Process: Experimental Measurement and Numerical Simulation.......... 177 C. Casavola, A. Cazzato, V. Moramarco, and C. Pappalettere 25 Dynamics of Strain Localization Associated with Lu¨ders Deformation: An Insight .................. 187 Srinivasan Nagarajan, Raghu Narayanaswamy, and Venkatraman Balasubramaniam 26 Raman Spectroscopy-Enhanced IIT: In Situ Analysis of Mechanically Stressed Polycrystalline Si Thin Films ........................................................... 195 Yvonne B. Gerbig, Chris A. Michaels, and Robert F. Cook viii Contents
Chapter 1 Electrical Impedance Spectroscopy for Structural Health Monitoring Geoffrey A. Slipher, Robert A. Haynes, and Jaret C. Riddick Abstract Structural heath monitoring (SHM) can provide an estimate of the state of damage in a structure, and of the remaining useful life of that structure. The work presented here is an investigation of a proposed new SHM technique for composite structures composed of carbon-fiber-reinforced-polymers (CFRP). Electrical impedance spectroscopy (EIS) is employed to estimate the damage state of the composite. No modification to current CFRP processing methods is required, nor is the proposed technique invasive or destructive. EIS interfaces can be either permanently attached or temporarily connected. We hypothesize that EIS has the potential to be more sensitive and selective for damage detection by using a full complex-plane analysis, considering both impedance magnitude and phase angle. This is in contrast to electrical SHM approaches employing resistance measurement, the real component of impedance, which ignores phase angle and reactance information. In order to test our hypothesis we implemented three different experiments to evaluate the effectiveness of the EIS technique: (1) specimen load sensitivity; (2) specimen damage sensitivity; and (3) specimen fatigue sensitivity. Multiple electrical interrogation paths through the specimen are considered. Keywords Electrical impedance spectroscopy • Structural health monitoring • Carbon fiber • Fatigue state • Damage state detection • Load state detection 1.1 Introduction Carbon-fiber reinforced composite materials are rapidly gaining wide use for a variety of applications pertaining to aircraft, spacecraft and civil infrastructure. The recent trend toward condition-based maintenance of advanced structural platforms has given rise to sensor-based monitoring of critical structural components. Structural health monitoring refers to the use of sensor data as a means to provide an estimate of the state of damage in a structure, and of the remaining useful life of that structure. Pervasive use of SHM in composite structures in the future will enable new structural applications that exploit the presence of sensors to enable novel lightweight designs and reduced maintenance burdens. Damage detection by structural self-sensing in carbon fiber composites has been investigated thoroughly in the literature. In particular concepts for the use of electrical resistance measurements to detect damage in fiber-reinforced composites have been investigated, showing that resistance changes irreversibly upon damage inflicted by flexure, tension, fatigue, and impact [1]. Resistance change can be associated with the mode of damage. For example, fiber breakage increases the longitudinal resistance, and delamination increases the through-thickness resistance. Embedding of carbon nanotubes (CNTs) into glass-epoxy composites to create percolating electrical networks for sensing damage has been shown to work based on measuring the change in resistance due to cracking [2–5]. Electrical resistance measurements have shown some promise in the detection of damage in fiber composites, particularly in those systems where embedded sensors or electrical networks were used. However, electrical impedance measurements have the potential to be more sensitive and selective for damage detection by using a full complex-plane analysis, considering both impedance magnitude and phase angle. Electrical resistance can be considered zero-phase angle impedance. Impedancebased methods have been employed for qualitative health monitoring by correlating variations in mechanical impedance to G.A. Slipher (*) • R.A. Haynes • J.C. Riddick Vehicle Technology Directorate, U.S. Army Research Laboratory, Aberdeen Proving Ground, MD 21005, USA e-mail: geoffrey.a.slipher.civ@mail.mil N. Sottos et al. (eds.), Experimental and Applied Mechanics, Volume 6: Proceedings of the 2014 Annual Conference on Experimental and Applied Mechanics, Conference Proceedings of the Society for Experimental Mechanics Series, DOI 10.1007/978-3-319-06989-0_1, #The Society for Experimental Mechanics, Inc. 2015 1
electrical impedance measurements of a piezoelectric patch [6]. Experimental measurements of carbon fiber composites with embedded piezoceramic patches have shown a close connection between mechanical properties and impedance [7]. Electrical conductivity measurements of embedded CNT thin films have been used to detect damage from low velocity impacts in glass fiber reinforced polymer composites [8]. The electrical impedance tomography response of the composite with the electrical conductive strain sensitive embedded thin film was measured by recording current-voltage measurements at the periphery of the composite. More recent work has shown the effectiveness of using piezoelectric patches to monitor the health of adhesive joints in composites subjected to various environmental conditions using electro-mechanical impedance [9]. Other recent approaches that rely on impedance and frequency related responses have been presented in the literature. An approach for probing embedded sensors wirelessly using inductive coupling depends on electrical impedance, acoustic response, and pulse-echo response of the system [10]. Another study shows the promise of using metamaterial electromagnetic lenses to detect delaminations in composite materials by extracting effective magnetic permeability and frequency measurements [11]. Carbon fiber composites have been shown to be excellent electromagnetic interference (EMI) shielding materials with low surface impedance and high reflectivity [12]. Several models have been presented for electromagnetic characterization of fiber composites [13]. Models for effective properties of fiber composites commonly used in aircraft EMI shielding often employ various levels of detail, and therefore capture various physical aspects in the estimation values such as the reflection and transmission coefficients. The approach of embedding conductive networks or sensor/actuator patches has proven to be effective in exploiting the electromechanical impedance of composite material systems. The work presented here employs electrical impedance spectroscopy (EIS) to estimate the damage state of the carbon fiber reinforced composites without the modifications introduced by embedding conductive networks or sensors. EIS interfaces can be either permanently attached or temporarily connected. Multiple interrogation paths through the specimen are considered. If successful, the present method would offer the potential of an inspection method for SHM that depends wholly on the intrinsic properties of the material. This paper contains a description of the composite material considered, along with a description of the experimental procedure. 1.2 Motivation We are seeking structural health monitoring techniques that are simple and readily implementable, involve minimal modification to existing CFRP component manufacturing techniques, can operate in real-time, have minimal cost, and have the potential to spatially resolve damage and/or damage precursor locations. Our hypothesis that we are testing, in part with the methods described in this paper, is that by moving beyond resistance measurement techniques to EIS we can make a significant step toward achieving these objectives, in particular minimal cost and impact to existing manufacturing techniques. Electrical interrogation techniques, such as resistance or impedance measurement, require only simple electrical connection to an electrically conductive specimen, such as carbon fiber reinforced components. For this reason they are attractive as minimally invasive techniques that require little or no modification to the way carbon fiber components are manufactured. As damage is introduced to a carbon fiber reinforced component, electrical properties change, and that change can be detected and correlated with the type and location of the damage. Fiber breakage is expected to manifest predominantly as variations in component resistivity as well as in higher frequency response characteristics due to variations in the effective antenna length paths within a component. Delamination is expected to manifest predominantly as variations in the capacitive reactivity of the specimen as the physical spacing between parallel layers varies. A method with the capability to simultaneously detect changes in resistivity, frequency response, and reactivity would thus be required to resolve as many damage modes as possible. Resistance measurement techniques provide only one dimension of information along the real axis of the complex plane to resolve damage in CFRP components. Impedance analysis techniques add two additional dimensions of information: reactance and phase, as shown in Fig. 1.1. We therefore propose the EIS method as a potential improvement in NDE methods for CFRP structural health monitoring in Army systems such as rotorcraft. This paper describes initial experimental efforts to validate these hypotheses. Fig. 1.1 Impedance complex plane 2 G.A. Slipher et al.
1.3 Experimental Procedure Fifteen specimens were manufactured from a woven graphite/epoxy material system with a thickness of 0.125 in. and cut to dimensions of 1 in. by 12 in. Tabs of dimension 1.25 in. by 3 in. were attached to both sides and both ends of the specimen. The tabs extended 0.125 in. beyond the specimen on both sides and the end to reduce the likelihood that the wire leads would touch the testing machine; see Fig. 1.2. Five of the specimens were used to estimate the material’s static strength as 68 kN, five specimens were used to estimate the material’s fatigue life, and five specimens underwent cyclic loading with predefined pauses to record impedance measurements. The cyclic loading was performed in an MTS 22-kip load frame with an MTS FlexTest 40 controller. An R-ratio of 0.1 was used with a maximum load of 55 kN and a frequency of 5 Hz. Silver epoxy was used to attach wire leads to the specimen in the configuration shown in Fig. 1.2. Five different electrical paths were interrogated through the specimen: A1-B1, A2-B2, A3-B2, A4-B3, and A5-B3. Electrical impedance was measured using an Agilent E5061B-LF network analyzer (NA) using a port 1–2 thru series method. Prior to initiating sample characterization, a calibration procedure was executed on the NA that pushed the calibration plane out to the specimen. The A1-A5 connections were used to inject the electrical signal through port 1, and the B1-3 connections were connected to port 2 for the NA return. For the fatigue tests, cyclic loading was applied as described above. Cyclic loading was stopped periodically and logarithmic frequency scans were performed with the NA from 1 kHz to 100 MHz for both statically loaded (55 kN) and unloaded states through each of the five electrical paths. Impedance magnitude and phase angle data were collected and saved. The electrical interrogation signal was a sine wave of varying frequency with power of 10 dBm (707 mV @ 50Ω). During the first sample run, a manual testing protocol for running the impedance scans was used, and was identified as undesirable for two key reasons: (1) it consumed too much personnel time; (2) it introduced error in the data in the form of jump discontinuities in the cycle/failure domain as shown in Fig. 1.3. We believe the discontinuities arose from small variations in the calibration introduced by shutting down the equipment at night and recalibrating in the morning. For both of the above stated reasons an around-the-clock automated testing routine was developed and implemented for the subsequent four samples described in this paper. The key component of the automated testing routine was the switching electronics that used magnetic reed relays triggered by a custom-programmed IC to switch between each of the five conductive paths through the specimen. Implementation of the automated routine completely eliminated the jump discontinuities in the cycle/failure domain. The automated experimental setup is shown in Fig. 1.4. 1.4 Results and Discussion In order to test our hypotheses outlined above we implemented three different experiments to evaluate the effectiveness of the EIS technique: (1) specimen load sensitivity; (2) specimen damage sensitivity; and (3) specimen fatigue sensitivity. The specimen load sensitivity experiment involved incrementing the load applied to the specimen to identify if the EIS technique could detect changes in load levels. The damage sensitivity experiment involved scanning an undamaged specimen and subsequently introducing gross damage to the gage length and rescanning to determine if the EIS technique could detect the damage. The fatigue sensitivity experiment involved scanning the specimen at various numbers of loading cycles to identify whether or not the EIS technique can detect fatigue-related damage or damage precursors. The results of these three experiments are presented and discussed below. Fig. 1.2 Specimen configuration 1 Electrical Impedance Spectroscopy for Structural Health Monitoring 3
1.4.1 Detection of Specimen Load State Load on specimen ‘S6’ was incremented from 0 to 50 kN, and all five electrical paths were interrogated using the NA at each load increment. The results for phase angle can be seen in Fig. 1.5 and impedance magnitude in Fig. 1.6. Specimen load resulting in strain can be detected using both impedance magnitude and phase information from the EIS scans. Fig. 1.3 Jump discontinuities in the data arising from manual data collection procedure, eliminated by implementing an automated procedure Fig. 1.4 Automated experimental setup: (a) wide view; (b) close-in specimen view with switching electronics 4 G.A. Slipher et al.
Fig. 1.5 Phase angle resolution of specimen S6 load state for different electrical paths (a–e) 1 Electrical Impedance Spectroscopy for Structural Health Monitoring 5
Fig. 1.6 Impedance magnitude resolution of specimen S6 load state for different electrical paths (a–e) 6 G.A. Slipher et al.
Comparing the results from phase and impedance we conclude that phase is better at resolving sample load state at higher frequencies, whereas impedance magnitude is more sensitive at lower frequencies. For the specimens tested, the approximate cutoff frequency at which the phase becomes more effective at resolving specimen load state is about 500 kHz. Electrical path through the specimens also influences sensitivity to the load state, with paths A2B2 and A3B2 showing the highest level of sensitivity for both phase and impedance magnitude. 1.4.2 Detection of Specimen Damage State Specimen ‘S6’ was scanned undamaged and unloaded. Damage was introduced to the specimen at the middle of the gage section without removing it from the grips using a spring-loaded center punch, as shown in Fig. 1.7. The specimen was then rescanned to determine if the EIS technique is capable of detecting the known presence of damage. The specimen was then subjected to 10,000 cycles at an r-ratio of 0.1 and a peak load of 50 kN to identify whether or not the EIS technique is able to detect changes in the specimen arising from a change in the damage due to the cyclic loading. All five electrical paths were interrogated using the NA at the undamaged, damaged, and damaged-cycled states. The results for phase angle can be seen in Fig. 1.8 and impedance magnitude in Fig. 1.9. The specimen damage state can be detected using both impedance magnitude and phase information from the EIS scans. The EIS technique is also capable of resolving changes in the damage state arising from the cyclic loading of the specimen. Comparing the results from phase and impedance we conclude that, as in the case with load state above, phase information is better at resolving damage state at higher frequencies, whereas impedance magnitude is more sensitive at lower frequencies. For the specimens tested, the approximate cutoff frequency at which the phase becomes more effective at resolving specimen load state is about 500 kHz, again the same result as for the load state. Electrical path through the specimens also influences sensitivity to the damage state, with paths A2B2, A3B2, and A4B3 showing the highest level of sensitivity for both phase and impedance magnitude. 1.4.3 Detection of Specimen Fatigue State Five specimens were prepared for the fatigue testing (S1, S2, S3, S4, S5). The data for specimen S1 were inconclusive due to the aforementioned issues arising from the initial manual protocol. Samples S2 and S3 for the fatigue evaluation failed at very low numbers of cycles, and so are not included in the data presented in this paper. Therefore, the data for the fatigue sensitivity results presented in this paper are based on only two specimens, S4, and S5, and conclusions based on a sample size of two are thereby limited. Due to limitations in the way the initial experiments were performed we currently have no way to confidently discriminate between specimen fatigue damage and fatigue damage to the electrical lead bonds. Figure 1.10 shows an example of how obvious lead damage is in the phase and impedance data. Examination of the data was used to identify which leads failed, and when. We have evidence to suggest that both types of damage are being detected. For sample S4, lead damage is apparent for the electrical path A2B2 in the impedance magnitude data shown in Fig. 1.10 at approximately 48 % cycle/failure. Similarly, lead damage was detected in S4 in electrical path A4B3 at approximately 25 % cycle/failure (see plot for S4 in Fig. 1.11). In both of these cases, the lead damage for paths A4B3 and A2B2 showed up in the scans along path A1B1 as shown in Fig. 1.7 Photograph of intentionally introduced damage to the specimen 1 Electrical Impedance Spectroscopy for Structural Health Monitoring 7
Fig. 1.8 Phase angle resolution of specimen S6 damage state for different electrical paths (a–e) 8 G.A. Slipher et al.
Fig. 1.9 Impedance magnitude resolution of specimen S6 damage state for different electrical paths (a–e) 1 Electrical Impedance Spectroscopy for Structural Health Monitoring 9
the left side of Fig. 1.11. Likewise, for sample ‘S5,’ lead damage in path A2B2 and A4B3 were detected at 10 and 62 % cycles/failure respectively, and showed up in the A1B1 electrical path scans as shown on the right side of Fig. 1.11. As also illustrated in Fig. 1.11, in the case of sample S4, the damage is correlated with the center of the contour regions, versus in sample S5, the damage is correlated with the onset (left edge) of the contour regions. Lead damage is not correlated with the contour region at the right side of the S4 scan in Fig. 1.11, thus leading us to believe that this region may be indicative of damage manifesting in the sample itself, and leading to eventual specimen failure. We need to further refine our experimental techniques to allow for reliable and confident discrimination between specimen damage and damage to the electrical leads. We have multiple strategies to realize this: (1) remove the electrical leads from the gage length of the specimens to the extent possible in order to eliminate the potential for electrical lead damage; (2) analyze specimens with single pairs of electrical leads to eliminate the possibility of electrical paths into/out of the specimens via the passive non-scanning leads. Fig. 1.10 Electrical lead failure indication in both phase angle (a) and impedance magnitude (b) data Fig. 1.11 Comparison of delta-phase contours for two specimens, S4 (left) and S5 (right), with electrical lead damage indicated in the cycle/ failure domain 10 G.A. Slipher et al.
1.5 Conclusions We have succeeded in an initial laboratory implementation of an EIS technique for damage detection in CFRPs. We have shown that EIS can be used to detect changes in both the loading and the gross damage states of a CFRP specimen. We found that impedance magnitude data has more resolution on specimen damage and loading state at lower frequencies, and that phase data has better resolution on specimen damage and loading state at higher frequencies. For the samples tested the cross-over point between when one would choose impedance magnitude versus phase angle to resolve specimen states occurred around 500 kHz. Electrical path through the specimen also needs to be carefully considered since we found that it also strongly influences sensitivity of the EIS technique to both damage and loading state. The principal conclusions based on our initial fatigue cycling results center mainly around refinements for our experimental methodology. At this point we cannot, with confidence, state that the EIS technique is able to resolve either damage precursors or internal damage to CFRP specimens caused by fatigue loading. This lack of confidence arises from the inability to confidently discriminate electrical lead damage from specimen damage, which, in turn, comes from flaws in the initial experimental methodology that we employed, especially electrical lead placement in the gage section of the evaluated specimens. However, our initial results clearly indicate directions of experimental refinement that we believe will allow us to discriminate between electrical contact and specimen damage in future experiments. Acknowledgements The authors would like to acknowledge the expert and valuable assistance of Mr. Wosen Wolde with the design, production, assembly, and trouble shooting of the electronics switching board. References 1. Chung DDL (2007) Damage detection using self-sensing concepts. Proc IME G J Aero Eng 221(4):509–520 2. Liu A, Wang KW, Bakis CE (2010) Damage detection of epoxy polymer via carbon nanotube fillers and external circuitry. In: 18th international conference on composites or nano engineering, Anchorage, Alaska 3. Gao L, Thostenson ET, Zhang Z, Chou T-W (2009) Coupled carbon nanotube network and acoustic emission monitoring for sensing of damage development in composites. Carbon 47(5):1381–1388 4. Thostenson ET, Chou T‐W (2006) Carbon nanotube networks: sensing of distributed strain and damage for life prediction and self healing. Adv Mater 18(21):2837–2841 5. Chung DDL (2012) Carbon materials for structural self-sensing, electromagnetic shielding and thermal interfacing. Carbon 50(9):3342–3353 6. Ayres JW, Lalande F, Chaudhry Z, Rogers CA (1998) Qualitative impedance-based health monitoring of civil infrastructures. Smart Mater Struct 7(5):599 7. Pohl J, Herold S, Mook G, Michel F (2001) Damage detection in smart CFRP composites using impedance spectroscopy. Smart Mater Struct 10(4):834 8. Loyola BR, Briggs TM, Arronche L, Loh KJ, La Saponara V, O’Bryan G, Skinner JL (2013) Detection of spatially distributed damage in fiberreinforced polymer composites. Struct Health Monitor 12(3):225–239 9. Na S, Tawie R, Lee HK (2011) Impedance-based non-destructive evaluation of the FRP adhesive joints in corrosive environment with re-usable technique. In: SPIE smart structures and materials + nondestructive evaluation and health monitoring. International Society for Optics and Photonics, pp 79811B–79811B 10. Zhong CH, Croxford AJ, Wilcox PD (2012) Inductively coupled transducer system for damage detection in composites. In: SPIE smart structures and materials + nondestructive evaluation and health monitoring. International Society for Optics and Photonics, pp 83480H–83480H 11. Grimberg R, Premel D, Lemistre MB, Balageas DL, Placko D (2001) Compared NDE of damages in graphite/epoxy composites by electromagnetic methods. In: 6th annual international symposium on NDE for health monitoring and diagnostics. International Society for Optics and Photonics, pp 65–72 12. Luo X, Chung DDL (1999) Electromagnetic interference shielding using continuous carbon-fiber carbon-matrix and polymer-matrix composites. Comp Part B Eng 30(3):227–231 13. Holloway CL, Sabrina Sarto M, Martin J (2005) Analyzing carbon-fiber composite materials with equivalent-layer models. IEEE Trans Electromagn Compat 47(4):833–844 1 Electrical Impedance Spectroscopy for Structural Health Monitoring 11
Chapter 2 In situ Observation of NiTi Transformation Behaviour: A Micro–Macro Approach Kasun S. Wickramasinghe, Rachel A. Tomlinson, and Jem A. Rongong Abstract A novel experimental investigation is presented of thermally and stress induced transformation behaviour of a Polycrystalline NiTi Shape Memory Alloy (SMA) plate for flexural-type applications: In situ techniques are employed to allow simultaneous macroscopic and microstructural observation of the SMA in a 4-point flexural test. Forming part of a wider research towards realising a NiTi SMA Variable Stator Vane assembly for the gas turbine engine, the study explores variables critical to flexural-type morphing NiTi structures: (1) temperature; (2) strain; and (3) cyclic loading. It builds a relationship between the macro and micro response of the SMA under these key variables and lends critical implications for the future understanding and modelling of shape memory alloy behaviour for all morphing applications. This paper presents the methodological aspects of this study. Keywords Shape memory • NiTi • In situ • Phase transformations • Micro–macro approach • Cyclic loading 2.1 Introduction Gas turbine performance development has been governed traditionally by the “worst case” deterioration and operating condition [1]. This leads to severe compromises and large safety margins. Active control of the engine operation using smart materials could potentially improve engine efficiency. Shape Memory Alloys (SMAs), a class of smart materials, exhibit several desirable characteristics exploitable for this purpose. NiTi, based on an equiatomic compound of nickel and titanium is the most widely used SMA in commercial applications [2]. Besides the ability of tolerating relatively large amounts of shape memory strain, NiTi shows high stability in cyclic applications, possesses an elevated electrical resistivity, and is corrosion resistant [3]. An exciting possibility is the incorporation of SMA in plate form into blade structures of the Gas Turbine compressor. Airflow control is introduced into compressor designs through the use of Variable Inlet Guide Vanes and a number of stages incorporating Variable Stator Vanes. They operate by progressively closing as the compressor speed is reduced from the original design value to maintain an acceptable air angle value into the following rotor blades. Traditionally, such a system encompasses a complex structure employing control levers that are actuated through an electrical or bleed air system. Switching this system to a NiTi plate based actuation mechanism is intrinsically very attractive due to its high power density, solid-state actuation, high damping capacity, durability and fatigue resistance. A concept actuator design using NiTi plates to form a solid-state actuator that replicates the behaviour of the VSV system is depicted in Fig. 2.1. In this concept, the activation of each NiTi plate translates to a deflection at the tip of the actuator. To achieve this, the system exploits a flexural type strain application/recovery using the Shape Memory Effect. K.S. Wickramasinghe • R.A. Tomlinson (*) • J.A. Rongong Department of Mechanical Engineering, University of Sheffield, Mappin St., Sheffield S11 3JD, UK e-mail: r.a.tomlinson@sheffield.ac.uk N. Sottos et al. (eds.), Experimental and Applied Mechanics, Volume 6: Proceedings of the 2014 Annual Conference on Experimental and Applied Mechanics, Conference Proceedings of the Society for Experimental Mechanics Series, DOI 10.1007/978-3-319-06989-0_2, #The Society for Experimental Mechanics, Inc. 2015 13
2.2 Previous Studies A thorough understanding of the flexural performance of NiTi plate is required for the implementation of such a system. Previous studies have primarily evaluated material behaviour under tension employing mostly uni-axial loads. Of particular interest are studies that employ in-situ loading and microscopy [4–12]: they offer a rich understanding of SMA response that addresses both microstructural and macromechanical response simultaneously. In their micro–macro study, Brinson et al. [12] clarified several aspects of the transformation behaviour under loading at different strain rates and fatigue. This included observations of low-cycle localized plastic deformation, correlations between latent heat release and stress relaxation, and a refined definition of “full transformation” for polycrystalline materials. However, only a few experimental studies have addressed SMA behaviour under compression or bending, despite the prevalence of these deformation modes in applications. Most studies have restrained from using traditional point-bending fixture tests; instead, they have opted for custom-built pure-bending apparatus. Studies on small diameter NiTi wire naturally restrict point-bending fixtures as a result of small radii of curvature and large displacements—conditions where undesirable axial loads can develop due to the support constraints. The studies conducted by Berg [13] and Bundara et al. [14] have employed moment-controlled experiments to assess the constitutive relationship between applied pure bending moments and the resulting curvatures. These studies, however, suffer a major limitation due to their lack of control over strain rate, leading to a significant impact on the stress/strain response during phase transformations [12]. More recent work by Rejzner et al. [15] and Reedlunn et al. [16] avoid this problem by using displacement-controlled, custom pure bending fixtures integrated into the load frame. The study by Reedlunn et al. [16] is particularly interesting as they used larger diameter NiTi tube instead of wire, which negates the requirement of a small radius of curvature to incur strain. Additionally, the resulting smaller deformations and larger specimen dimensions allow the use of in-situ imaging techniques in the guise of Digital Image Correlation for strain measurement. Their experimental tests were able to discover a significant asymmetry between the compressive and tensile deformation modes. This resulted in a shift of the neutral axis towards the compressive side to equalise the distribution of tensile and compressive stresses in the cross section. One limitation that was noted, however, was the appearance of the Brazier effect; the tendency of thin walled tubes to ovalise in bending due to longitudinal tension and compression [17]. Ovalisation is particularly undesirable if it leads to variant re-orientation along this new deformation path at a granular level as it questions the homogeneity over compressive and tensile responses in their bend tests. This could also result in accelerated development of stress localisations under fatigue, which could lead to premature failure [12]. Another limitation encountered by Reedlunn et al. was through undertaking tests at room temperature without direct measurement of the specimen temperature. This procedure is not ideal because of the high temperature sensitivity of NiTi. A common theme to most bending studies is the exclusive focus on SMA behaviour in its Superelastic phase. This can be attributed to the myriad of applications that already employ Superelastic SMAs in this deformation mode, making it a more lucrative research area. Moreover, most studies on bending have followed a macro level approach to exploring material behaviour. This has left the investigation of its microstructural behaviour largely unexplored, under this deformation mode. Fig. 2.1 Concept sketch 14 K.S. Wickramasinghe et al.
2.3 Experimental Techniques Inspired by the above stream of research, the present study explores NiTi behaviour in bending. What distinguishes this research is the analysis of both thermal and stress induced transformation behaviour of NiTi plate. Additionally, in-situ techniques allow simultaneous observation at both macroscopic and microstructural levels. The inclusion of these different aspects, however, presents the study with a new series of hurdles. The methodology employed to conduct this study is detailed below. 2.3.1 Testing Stage Figure 2.2a illustrates the general experimental set up. The load frame—a tabletop MTS 858 servo-hydraulic test system— uses a cross-head assembly which includes a single moving upper grip, a stationary lower grip and a LVDT position sensor. The system is designed to function in a closed-loop configuration under computer control. The specimen sits within a temperature controlled environmental chamber—a Thermcraft LB-series Box Laboratory oven. A custom 4-point bend fixture is built and integrated into the uniaxial load frame to perform the flexural tests, as depicted in Fig. 2.2b. The external and internal supports are inverted to limit the deflection of the optically investigated central section of the specimen. The fixture is designed with adequate spatial clearances in mind, allowing upper support deflection of up to 12.5 mm (equivalent to a 4 % strain at the outer fibre with our specimens). 2.3.2 Optical Microscopy A digital camera is used to observe changes in the surface transformation characteristics at a macroscopic and fine-scale microstructural level as a function of loading parameters. All images are taken during static holds of constant temperature and strain. Figure 2.3 depicts the unique twin surface preparation employed to conduct this in-situ observation. Digital Image Correlation (DIC) is employed to obtain local strain measurement without the need to rely solely on grip displacements. The specimen is prepared by applying a random speckle pattern to its surface. DIC starts with a reference image and followed by subsequent images recorded during the deformation process. A strain distribution map is created by calculating the correlation between the reference and deformed speckle patterns. In contrast, the microstructural changes are observed from the finely polished surface on the left hand side of the specimen. A polarized light interference filter is used 68 50 25 mm 1 6 Crosshead Load unit control module Force transducer Hydraulic actuator Environmental Chamber Columns a b Fig. 2.2 (a) Testing stage; (b) custom 4-point bend fixture 2 In situ Observation of NiTi Transformation Behaviour: A Micro–Macro Approach 15
for higher detection sensitivity of fine topological details. A maximum zoom which focuses on a surface area approximately 1mm2 is required to investigate observation of the material at a variant level. This allows the simultaneous viewing of both compressive and tensile deformation modes. 2.4 Specimen Preparation Machining—Near-equiatomic 1 mm thick NiTi plate was obtained from Memry Metalle GmbH with manufacturer specified Austenite finish (Af) temperature of 65 C. The plate is machined to dimensions of 66 mm 8 mm using a wire cut electrical discharge machining process. This minimizes heat distortion and cold work induced on the specimen. As the tests optically investigate the specimen surface and relate it to the material behaviour as a whole, the homogeneity over their behaviours is important. Annealing—The specimens are heat treated in an oven at 450 C for 1 h followed by water quench. This annealing condition ensures a large grain size—the amorphous bands start to crystallize without completely annealing the structural defects in the material, as crystallite nucleation is suppressed [18]. This step is essential for the material to recover from any previous cold work or damage it has sustained and to stabilize transformation characteristics [18]. Thermal cycling—Additionally, the specimens are thermally cycled through martensite and austenite phases using a water bath. Miyazaki et al. [6] indicate that Mf andMs temperatures decrease with increasing number of thermal cycles. The decrease is rapid at first but more gradual with increasing cycles as values stabilize. Their TEM observations reveal the introduction of dislocations from the first thermal cycle onwards, which act as obstacles to dislocation movement and stabilizes behaviour through further thermal cycles. However, their thermal cycling had also shown to activate R-Phase transformation in NiTi SMAs that had a previously suppressed R-Phase. In our study, the specimens are subjected to 30 thermal cycles despite possible R-Phase activation, as it is paramount to obtain stable and repeatable transformation characteristics throughout. Micropreparation—The thermally prepared specimens are then subjected to several steps of grinding and polishing using the Bueler Automet 250, to allow observation of variant microstructure at the specimen surface. As the surface was initially rough, all sides of the specimen are grinded using P800 and P1200 abrasive discs. The observational side of the specimen is then sequentially polished using 3 μm MetaDi Supreme Diamond abrasive for coarse polishing and then finally using 0.02–0.06 μm MasterMet Silica for fine polishing. 2.4.1 Specimen Characterisation Differential Scanning Calorimetry (DSC) is used to identify phase transformation temperatures of the NiTi plates. This is a thermoanalytical method commonly used in NiTi SMA studies that determines absolute phase transformation temperatures. It measures the difference between the amount of heat required to increase the temperature of a sample and reference, as a function of temperature. Our DSC tests adopt a temperature profile of 15 C to +100 C with a constant scanning rate of 10 K/min. This is considered the most efficient temperature rate to measure the intrinsic transition quantities Speckle pattern for DIC Finely polished surface for optical microscopy Fig. 2.3 Preparation of observational surface 16 K.S. Wickramasinghe et al.
and also a common value within published literature. The characterisation of the transformation temperatures is essential in establishing temperature ranges for the experimental study. DSC scans of the NiTi plate specimens in the following conditions are presented: as received (grey), annealed (red), and thermally cycled (blue). Figure 2.4a, b depict thermograms of each of these states under heating (M!A) and cooling (A!M), respectively. The transformation temperatures for each of the specimen states are recorded in Table 2.1. 2.5 Test Plan The experimental test study can be subcategorized into two sections: (1) constant temperature tests with load increments, and (2) constant load tests with temperature increments. The tests employ a temperature range between Af 25 C !Af + 25 C and a peak flexural strain of 2 %. Using these load parameters, the study performs a comprehensive evaluation of the material behaviour within the operating envelope of the planned NiTi SMA actuator. The extremities of these parameters are established due to the requirement of the maximum strain falling within the de-twinning phase at the lowest temperature, Af 25 C. However, it must be noted that due to the non-linear response of the material, localised strains could be considerably more, especially during the de-twinning phase where large heterogeneities can exist between grain boundaries. An adequate safety margin is thus kept from the end of the strain plateau observed from previous data from tensile tests on this material: ~4 % at room temperature. This strain magnitude range ensures that thermomechanical cycling effects are minimised and life span of each specimen is improved. Furthermore, the lower strains reflect the behaviour of a high cycle element such as the variable stator vane where uniformity of performance must be established throughout the life cycle. Three NiTi plate specimens are used in this study. The thermal and fatigue history of each of the specimens is recorded and carefully balanced to enhance comparability between results. In each test cycle, the loading is temporarily paused at planned intervals and the specimen is held at constant displacement, while photographs are taken of the observational surface. Constant temperature tests with load increments attempt to replicate the deflection behaviour of the NiTi plate actuator concept over a range of static temperatures. Table 2.2 highlights the parameters used in this study. The tests are sequenced so 15 20 a b 25 30 35 40 45 50 55 Heat Flow Temperature (°C) Annealed Thermally cycled As received 25 30 35 40 45 50 Heat Flow Temperature (°C) Annealed Thermally cycled As received Fig. 2.4 (a) DSC Thermogram for heating (M!A); (b) DSC Thermogram for cooling (A!M) Table 2.1 Phase transformation temperatures Specimen Ms ( C) Mf ( C) As ( C) Af ( C) As received 47.8 30.7 28.5 50.9 Annealed 34.9 30.6 33.1 37.4 Thermally cycled 35.6 31.3 33.5 38.2 2 In situ Observation of NiTi Transformation Behaviour: A Micro–Macro Approach 17
that strain is always recovered after a cycle that is loaded from the martensitic phase: the subsequent cycle runs above Af and activates the SMA. Loading sequences for each of the initial 36 tests are repeated in reverse order to account for the temperature history sensitivity of the material. The strain is regulated by a load input to allow strain relaxation in non-isothermal type loading. Constant load tests with temperature increments attempt to replicate the actuation behaviour of the NiTi plate actuator concept over a range of static loads. Strain recovery characteristics of the NiTi plate specimens at different load magnitudes are analysed to accomplish this. Table 2.3 highlights the parameters used in this study, which allows the evaluation of macroscopically observable material characteristics and microstructural changes over the stated temperature range. The strain is regulated by a load input to allow strain recovery. The specimen is heated past its Austenite finish (Af) temperature prior to each subsequent test to ensure that strains are not carried over between tests. This is executed at zero strain, using an identical temperature ramp rate. 2.6 Predictions and Limitations In this section, we evaluate the possible outcomes of the experimental tests. While the broader aim of the study is to evaluate material behaviour under the possible operating envelope of the NiTi plate actuator, this study also assesses several NiTi SMA behavioural aspects that may have academically significant implications, as discussed below. Asymmetry between tension and compression—The bend tests allow the simultaneous evaluation of the material behaviour under both tensile and compressive deformation modes. While classical Euler–Bernoulli bending kinematics assumes a linear strain profile across the cross section, the heterogeneous nature of the bending strain fields in NiTi result in a much more complex strain profile. In their bend study, Reedlunn et al. [16] report a significant shift of the neutral axis under bending. However, the appearance of the Brazier effect, due to their choice in specimen geometry, could have affected this to a certain extent. In contrast, the use of plate specimens in our study negates such effects. Moreover, this study potentially allows a richer understanding of the two deformation modes as we use broader testing conditions and follow a micro–macro approach. Relationships between macroscopic behaviour and microstructural observations—This study combines macroscopic and microstructural analysis via DIC and optical microscopy. Most previous studies have either focused on a micro or macro level approach: micromechanical studies are often theoretical and not directly applicable to end applications whereas macro level studies do not contain enough depth to explain the observed phenomena. The employment of both approaches allows relationships to be constructed, explaining material behaviour in greater detail. A complete map of transformation behaviour—Phase transformations, temperatures and applied stresses are closely intertwined where one affects the other. As this study varies both temperature and load in its tests, a complete phase diagram can be developed, showing the effect of applied flexural deflection on the phase transformation temperatures for the tested specimens. Effects of loading rates—The constant temperature tests repeat every loading sequence at three loading rates: 1 10 4 s 1, 1 10 3 s 1 and1 10 2 s 1. Uniaxial studies by Brinson et al. [12] report on slower strain rates following an isothermal behaviour where martensite initiates and grows from a single band, while at higher strain rates, the formation of multiple bands is present. To our knowledge, however, the implications on loading rate for bend tests using full-field, in-situ methods such as DIC are largely unexplored. Table 2.2 Constant temperature tests Strain Strain rate Temperature Optical imaging Total tests 2% 1 10 4 s 1 !1 10 2 s 1 (3 steps) Af 25 C!Af + 25 C (12 steps + 12 steps reverse) Static strain holds 0.25%!2%!0 % (16 steps) 72 Table 2.3 Constant load tests Load Temperature Temperature ramp rate Optical imaging Total tests 0.1%!2% (at Af 25 C/ 20 steps) Af 25 C!Af + 25 C!Af 25 C 1 10 1 Cs 1 Static temperature holds (11 steps/every 5 C) 60 18 K.S. Wickramasinghe et al.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==