Acta Polytechnica CTU Proceedings doi:10.14311/APP.2016.3.0035 Acta Polytechnica CTU Proceedings 3:35–38, 2016 © Czech Technical University in Prague, 2016 available online at http://ojs.cvut.cz/ojs/index.php/app IDENTIFICATION OF DAMAGED ZONES IN C/PPS SPECIMENS SUBJECTED TO FATIGUE LOADING BASED ON HIGH-RESOLUTION THERMOGRAPHY Petr Koudelkaa,∗, Daniel Kytýřa,b, Nela Fenclováa,b, Martin Šperlb a Czech Technical University In Prague, Faculty of Transportation Sciences, Konviktská 20, 120 00 Prague 1, Czech Republic b Czech Academy of Sciences, Institute of Theoretical and Applied Mechanics, Prosecká 76, 190 00, Prague 9, Czech Republic ∗ corresponding author: koudelka@fd.cvut.cz Abstract. In this study high resolution thermography is used for identification of damaged zones in Carbon fiber/polyphenylene sulfide (C/PPS) long fiber composite specimens with induced impact damage subjected to tensile fatigue loading. Image processing techniques were applied on thermographs from all loading cases to obtain segmented images of the damaged location that were then used for calculation of the heated area. Results show that the considered method can be used to identify heated area in the vicinity of damage with high confidence at low number of cycles where no significant fatigue effect is present in the material. Keywords: fatigue, damage, long fiber composite, C/PPS, thermography. 1. Introduction Composites with different matrix materials reinforced by short or long fibers have been widely used as struc- tural materials thanks to their favorable mechanical properties (particularly relative stiffness and strength) for several decades. One of very popular composite compounds widely used in aviation industry consists of polyphenylene sulphide (PPS) matrix and carbon fiber reinforcement. The material is commonly used for production of both external and internal aircraft components because of its excellent mechanical prop- erties, chemical resistance to aerospace fluids, low density and good resistance to heat and flames [1]. Micrograph of damaged composite structure after ma- terial failure is depicted in figure 1. Figure 1. Detail of damaged structure obtained by scanning electron microscope, the plies are consisted of yarn with individual 5µm fibres. Still, application of this material in airframes is limited by its relatively low impact resistance and fragile behavior that have to be thoroughly investi- gated before its operational deployment [2]. In the life-cycle of an airframe, damage from impacts may be inflicted during flight or ground maintenance. This in combination with unique nature of composites and high cycle fatigue may lead to unexpected failure due to damage accumulation in the material’s microstruc- ture [3]. In this study high-resolution thermography has been used to observe dynamically loaded damaged specimens in order to quantify thermal characteris- tics of the damaged area and to evaluate suitability of considered approach for NDT inspection/damage tolerance analysis [4–6]. Thus, primary goal of this study was to identify lowest possible stress level for reliable assessment of the damage in low number of loading cycles. 2. Materials and Methods 2.1. Specimen Description The C/PPS material used in this study consists of 8-ply of carbon fabric with volume fraction of fibres approximately 60 %, overall density 1.35 g · cm−3 and quasi-isotropically distributed fibres. Surface of the samples was covered by thin glass fibre cloth protect- ing the core against UV light, surface damage and chemical influences. Set of eight dog bone shaped specimens were prepared using numerical controlled water jet cutter from two plates of C/PPS material (dimension 160×160 mm) with thickness 2.5±0.1 mm. Overall length of specimens was 160 mm. Distal part with width 25 mm was used for best possible fixation in mechanical grips whereas central part with length 105 mm and width 20 mm was subjected to the initial loading and thermographically investigated. 35 http://dx.doi.org/10.14311/APP.2016.3.0035 http://ojs.cvut.cz/ojs/index.php/app P. Koudelka, D. Kytýř, N. Fenclová, M. Šperl Acta Polytechnica CTU Proceedings 2.2. Initial Damage Before dynamic loading the specimens were subjected to impact loading to impose initial damage under controlled conditions using instrumented drop tower. Experimental setup with maximal impact energy 50 J was used in configuration with steel spherical impactor having diameter 15 mm. Impact energy of 15 J was set to inflict initial damage in the central part of the sample leading to imprint diameter in range of mil- limeters and depth in range tens of micrometers. The geometrical characteristics of imprints were identified using laser profilometry method [7]. 2.3. Experimental Setup For the investigation of thermal response to the me- chanical loading Mikrotron (Russenberger Prüfmaschi- nen, Germany) resonant testing machine was em- ployed. Loading was performed in tensile mode at four stress levels with mean load 4.5, 5.0, 5.5, 6.5, 8.0 kN and amplitude 2.0, 3.0, 4.0, 4.5, 5.0, 6.0 kN in sinusoidal cyclic force. Testing frequency was approximately 75 Hz. Highest stress level did not exceed 35 % of ten- sile strength of the material and fully elastic behaviour during loading cycles was assumed. One of damage degradation parameter obtained from resonant excita- tion is cycle dependent decrease of natural frequency. According to our previous study eigenfrequency de- crease lower than 0.05 % indicates no damage inflicted in the material’s microstructure. Over the whole test- ing campaign this level was not exceeded as can be seen from graph of natural frequencies plotted against number of cycles (figure 2). −0.02 −0.01 0 0.01 0.02 0.03 0.04 0.05 0 10000 20000 30000 40000 50000 ∆ na tu ra l f re qu en cy [H z] Loading cycles 6.5 ± 6.0 kN 8.0 ± 6.0 kN 5.5 ± 5.0 kN 5.0 ± 4.5 kN 4.5 ± 4.0 kN 8.0 ± 4.0 kN 5.0 ± 3.0 kN 8.0 ± 2.0 kN Figure 2. Variation of natural frequency as a function of number of cycles. The test was performed at constant room tem- perature 22 ◦C. Loading scene was observed us- ing scientific-grade thermal imaging camera SC7650 (FLIR Systems, USA). The device is equipped with InSb cooled sensor with pixel pitch 15µm and res- olution 640 × 512 px. The device allows acquisition of images in temperature range −20 to 300 ◦C with full frame sampling frequency 100 Hz. In this work full frame resolution and 0.5 Hz acquisition frequency was used. Control of the camera and acquisition of Figure 3. Experimental setup for thermographical observation of dynamic loading. Specimen Initial damage 1 1 1 1 Figure 4. Distribution of ROIs in thermograph for segmentation and area calculation: 1) background tem- perature, 2) specimen temperature, 3) heated area. thermographs was performed using GigE interface. Experimental setup is depicted in figure 3. 2.4. Image Processing Procedure Thermographs were subjected to image processing procedures in order to identify the damaged zone and to calculate the heated area in its vicinity. Adaptive identification technique based on the median value of the background (blue areas in figure 4) was used for differentiation of the sample and background. Median of specimen’s temperature was then found and area with temperature higher than 105 % of the median temperature was subjected to investigation (red area in figure 4). Standard morphological opening and closing procedures with circular structuring element were applied on the identified heated areas to improve accuracy of impacted area calculation. In result all holes in the identified area smaller than 3 px2 were morphologically closed and areas smaller than 10 px2 were neglected from the calculation. Visualisation of a series of thermographs subjected to investigation using image processing techniques is shown in figure 5. 3. Results During the thermographically observed experiments it has been found out that normalized peak temperature (temperature of specimen relative to ambient temper- ature) converges to constant value in less than 50 000 36 vol. 3/2016 Identification of Damaged Zones in C/PPS Based on Thermography loading cycles. In figure 6 dominant influence of the amplitude compared to mean value of applied load to the maximal temperature is clearly visible. In case of the highest loading level temperature of the specimen reached absolute value of 45 ◦C corresponding to less than 50 % of the material’s glass transition tempera- ture. Therefore negative thermal effects influencing the material properties (particularly the assumed elas- tic behavior) during the test were avoided. Figure 5. Example of a series of acquired thermo- graphs showing heating of the damaged zone during cyclic loading at 1 s (left), 32 s (middle) and 200 s (right); temperature scale is in ◦C. 1 1.2 1.4 1.6 1.8 2 2.2 0 10000 20000 30000 40000 50000 M a x im u m t e m p e ra tu re t o b a c k ro u n d r a ti o Loading cycles 6.5 ± 6.0 kN 8.0 ± 6.0 kN 5.5 ± 5.0 kN 5.0 ± 4.5 kN 4.5 ± 4.0 kN 8.0 ± 4.0 kN 5.0 ± 3.0 kN 8.0 ± 2.0 kN Figure 6. Ratio of specimen to ambient temperature as a function of number of cycles. Threshold temperature higher than 105 % of aver- age specimen temperature (calculated from the yellow area in figure 4) was chosen for identification of the local damaged zone. According to acquired results sig- nificant increase of the temperature in damaged zone became observable after only a few hundred of loading cycles. After that peak temperature in the damaged zone rose sharply together with average temperature in the entire volume of the samples. This average tem- perature reached its maximum equal to temperature after approximately 10 000 loading cycles yielding the thermal footprint of the samples homogeneous. Hence higher number of loading cycles was unnecessary for our study. Identified damaged zones for all loading cases are depicted in figure 7. It can be seen that minimal loading level 5 ± 4.5 kN was sufficient for reliable identification of the dam- 0 20 40 60 80 100 120 140 0 10000 20000 30000 40000 50000 Id e n ti fi e d a re a r a ti o 5 x 1 0 − 5 Loading cycles 6.5 ± 6.0 kN 8.0 ± 6.0 kN 5.5 ± 5.0 kN 5.0 ± 4.5 kN 4.5 ± 4.0 kN 8.0 ± 4.0 kN 5.0 ± 3.0 kN 8.0 ± 2.0 kN Figure 7. Variation of identified heated area in the vicinity of damage as a function of number of cycles. aged zone in the vicinity of impact due to its apparent heating. Detailed plot of identified heated area in the vicinity of damage and median temperature of spec- imens for two representative loading cases is shown in figures 8 and 9. Here increase of the loading to 6.5 ± 6 kN shows significant increase of the relative heated area (ratio of identified heated area in the vicinity of damage to total area of specimen) of mag- nitude more than 30 times. This together with the fact that in all cases number of loading cycles lower than 10 000 was sufficient for identification of the influenced zone indicates that the proposed method is suitable for non-destructive inspection of C/PPS long-fiber composites. 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 10000 20000 30000 40000 50000 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 Id e n ti fi e d a re a r a ti o 5 x 1 0 -5 M a x im u m t e m p e ra tu re t o b a c k ro u n d r a ti o Loading cycles Area Temperature Figure 8. Evolution of identified heated area dur- ing mean temperature of specimen during 5 ± 4.5 kN loading. 4. Conclusions It can be seen that in majority of loading cases the material exhibits consolidation and hardening in the first 50 000 cycles. For the two cases where combi- nation of mean force and amplitude causes damage accumulation image processing methods have been applied on thermographs to obtain segmented im- ages of the damaged location. From the segmented 37 P. Koudelka, D. Kytýř, N. Fenclová, M. Šperl Acta Polytechnica CTU Proceedings 0 20 40 60 80 100 120 140 0 10000 20000 30000 40000 50000 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 Id e n ti fi e d a re a r a ti o 5 x 1 0 -5 M a x im u m t e m p e ra tu re t o b a c k ro u n d r a ti o Loading cycles Area Temperature Figure 9. Evolution of identified heated area dur- ing mean temperature of specimen during 6.5 ± 6 kN loading. images heated areas were calculated together with specimen/background temperature ratio and plotted against number of loading cycles. It is apparent that heated area in the vicinity of damage can be identified with high confidence at low number of cycles (10 000) where no significant fatigue effect is present in the material. The considered inspection method is thus suitable for NDT inspections or damage tolerance analysis. Acknowledgements The research was supported by Grant Agency of Czech Technical University in Prague (projects no. SGS15/225/OHK2/3T/16, 1051505F014 RPMT 2015 43e), by Technology Agency of the Czech Republic (project no. TA03010209) and by RVO: 68378297. References [1] M. Favarolo: Proceedings of SAMPE 2010 Conference, 2010. [2] P. Koudelka, T. Fíla, T. Doktor, D. Kytýř, J. Valach, et al.: Inspection of post impact fatigue damage in carbon fibre composite using modulus mapping technique, Key Engineering Materials, 606:245-248, 2014. doi:10.4028/www.scientific.net/KEM.606.245 [3] R. Aoki, J. Heyduck, Developments in the Science and Technology of Composite Materials, 633-642, 1993. [4] J. P. Gallagher, USAF damage tolerant design handbook, Flight Dynamics Laboratory, Air Force Wright Aeronautical Laboratories, 1984. [5] M. Naderi, A. Kahirdeh, M.M. Khonsari: Dissipated thermal energy and damage evolution of Glass/Epoxy using infrared thermography and acoustic emission, Composites Part B: Engineering, 43(3):1613-1620, 2011. doi:10.1016/j.compositesb.2011.12.013 [6] E.Z. Kordatos, D.G. Aggelis, T.E. Matikas: Monitoring mechanical damage in structural materials using complimentary NDE techniques based on thermography and acoustic emission, Composites Part B: Engineering, 43(6):2676-2686, 2012. [7] Kytýř, D., Fíla, T., Šleichrt, J., Doktor, T., Šperl, M. Assessment of Post Impact Damage Propagation in Carbon-Fibre Composite under Cyclic Loading. Materials and Technology, 48(5):777-780, 2014. 38 http://dx.doi.org/10.4028/www.scientific.net/KEM.606.245 http://dx.doi.org/10.1016/j.compositesb.2011.12.013 Acta Polytechnica CTU Proceedings 3:35–38, 2016 1 Introduction 2 Materials and Methods 2.1 Specimen Description 2.2 Initial Damage 2.3 Experimental Setup 2.4 Image Processing Procedure 3 Results 4 Conclusions Acknowledgements References