Experimental investigation of polymer matrix reinforced composite erosion characteristics

Wear 270 (2011) 146–151

Contents lists available at ScienceDirect

Wear journal homepage: www.elsevier.com/locate/wear

Experimental investigation of polymer matrix reinforced composite erosion characteristics G. Drensky, A. Hamed ∗ , W. Tabakoff, J. Abot Department of Aerospace Engineering and Engineering Mechanics, University of Cincinnati, Cincinnati, OH, USA

a r t i c l e

i n f o

Article history: Received 27 November 2009 Received in revised form 27 April 2010 Accepted 19 August 2010 Available online 3 December 2010 Keywords: Polymer matrix Composite Erosion

a b s t r a c t Solid particle erosion of a composite material of polyetheretherketone (PEEK) matrix and unidirectional (AS4) carbon fiber was investigated experimentally. Erosion tests were conducted with 10 ␮m Arizona road dust and 100 ␮m sieved runway sand particles in especially designed erosion tunnel at temperature up to 260 ◦ C (500 ◦ F) and impact velocities up to 152.4 m/s (500 ft/s). Experimental results are presented for the measured erosion rates of the unidirectional (UD) composite material at two perpendicular fiber alignment settings relative to the impacting particle stream. The results indicate a quasi-ductile behavior with peek erosion rate at 45◦ impact angle. Overall the erosion rate was found to increase with impact velocity. The sieved runway sand caused more than double the erosion of the Arizona road dust under the same impact conditions. The erosion rate was also found to increase with temperature except at normal impact. The fiber alignment orientation relative to the impacting particle stream influenced the erosion rate of the (UD) composite material. Higher erosion rates were measured at 90◦ fiber orientation than at 0◦ at ambient test temperatures as reported in prior investigations. However, lower erosion rates were measured at 90◦ fiber orientation than at 0◦ for the erosion at 260 ◦ C. Scanning electron micrographs (SEM) of post-erosion surfaces are presented. © 2010 Published by Elsevier B.V.

1. Introduction Traditional metal alloys are being replaced by lighter composite materials that offer weight saving and strength improvements in propulsion systems [1]. In particular, the introduction of composite fan blades is one of the revolutionary advances in modern high by-pass turbofan engines. There is a need, however, to expand our knowledge of the behavior of aging components made of composite materials. The prognosis and life management of airbreathing propulsion systems requires characterization of foreign object damage and erosion of blade and coating materials [2,3]. However, solid particle erosion of composite materials has not been investigated to the same extent as metal alloys and ceramics. The presence of solid suspended particles in turbofan and gas turbine engine flow fields can have serious consequences on engine performance and life [2,3]. Particle impacts can reduce blade chord, increase surface roughness, and change the leading trailing and edge shapes of the blades in the compression system [4,5]. Ingested particles could impair thermal protection in the hot section through film cooling passage blockage and thermal barrier coating erosion [6]. Characterizing blade surface material erosion resistance under a wide range of particle impact conditions is critical to mitigate the

∗ Corresponding author. Tel.: +1 513 556 3553; fax: +1 513 556 5038. E-mail address: [email protected] (A. Hamed). 0043-1648/$ – see front matter © 2010 Published by Elsevier B.V. doi:10.1016/j.wear.2010.08.017

consequences of these changes. This requires special erosion facilities that can simulate the aero-thermal conditions encountered in turbomachines [7]. Experimental based correlations for particle restitution and surface erosion by impacts are also needed, in combination with blade surface impact statistic from turbomachinery multi-phase flow numerical simulations, to predict the intensity and pattern of blade erosion [8]. It is well known that the erosion resistance of metal alloys and ceramic materials by solid particles depends on the impact velocity, impingement angle, target and erosive particle materials, particle size and shape, and that it is strongly influenced by temperature. The mechanisms of material removal by erosion include crack formation and fracture in brittle materials, and cutting and chipping in ductile materials. Consequently, the effect of impingement angle on the erosion rate differs significantly in ductile and brittle materials. Peak erosion of brittle materials occurs at normal impact while that of ductile material occurs at glancing impingement angles between 20◦ and 30◦ . Composite materials generally consist of two or more phases and their erosion mechanisms are more complex since they are influenced by the matrix and fiber materials, fiber content and reinforcement type [9,10]. Models for conventional material erosion are not applicable to composite materials due to their heterogeneity and anisotropy. Composites with thermosetting matrix were found to erode in a brittle manner with peak erosion at normal impact. On the other hand, those with thermoplastic metrics were found to erode in duc-

G. Drensky et al. / Wear 270 (2011) 146–151

147

Fig. 3. 100 ␮m sieved runway sand particles.

by the fiber orientation relative to the incoming particle stream. The erosion rate was generally higher when the fiber orientation was normal to the particle stream [13–15]. The test conditions in previous polymer matrix composite erosion studies were generally limited to ambient temperatures and to particle impact velocities below 100 m/s. One exception is the experimental investigation [16] at 93 ◦ C of advance polymer

Fig. 1. Schematic of erosion test facility.

tile or quasi-ductile mode [9–12]. It is generally believed that the matrix material is removed first, causing the fibers to be exposed to the erosive environment. The breakage of the exposed fibers is then followed by its detachment and separation. Consequently, composite materials erosion is influenced by fiber ductility, orientation, volume fraction, and by fiber/matrix adhesion [13,14]. A number of erosion studies were conducted for unidirectionally (UD) reinforced fiber composites, since they represent the basic element of more complex structures. Their erosion was found to be influenced

Fig. 2. Arizona road dust particles.

Fig. 4. Erosion rate for 10 ␮m Arizona road dust (T = 21 ◦ C, ˇ = 0◦ ).

Fig. 5. Erosion rate for 100 ␮m sieved runway sand (T = 21 ◦ C, ˇ = 0◦ ).

148

G. Drensky et al. / Wear 270 (2011) 146–151

Fig. 6. Erosion rate for 10 ␮m Arizona road dust (T = 260 ◦ C, ˇ = 0◦ ).

Fig. 9. Erosion rate for 100 ␮m sieved runway sand (T = 21 ◦ C, ˇ = 90◦ ).

matrix composite (PMC) erosion due to normal impact by Arizona road dust including two erosion resistant coatings. In the present work, erosion tests of polymer matrix composite material were conducted in the University of Cincinnati high temperature erosion wind tunnel at high particle impact velocities and temperatures that simulate turbofan flow conditions. The tested composite

Fig. 10. Effect of temperature on peak erosion rate (˛ = 45◦ , ˇ = 0◦ ).

Fig. 7. Erosion rate for 100 ␮m sieved runway sand (T = 260 ◦ C, ˇ = 0◦ ).

Fig. 8. Erosion rate for 10 ␮m Arizona road dust (T = 21 ◦ C, ˇ = 90◦ ).

material was made of (AS4) carbon fibers drawn from polyacrolonitrile (PAN) and polyetheretherketone (PEEK) matrix that were processed in the University of Cincinnati Multi-scale materials Characterization and Composite Structures Lab” [17]. The unidirectional composite material of a thermoplastic polymer with high

Fig. 11. Effect of temperature on erosion rate at normal impact (˛ = 90◦ , ˇ = 0◦ ).

G. Drensky et al. / Wear 270 (2011) 146–151

149

Table 1 Test condition. Tunnel temperature 21 ◦ C 70 ◦ F

260 ◦ C 500 ◦ F

Particle velocity 61.0 m/s 200 ft/s

97.5 m/s 320 ft/s

152.4 m/s 500 ft/s

2. Methodology

Fig. 12. The erosion rate at normal impact normalized by that at 45◦ (ˇ = 0◦ ).

stiffness and high heat distortion temperature, and a carbon fiber with intermediate elastic modulus offered high strength and operating temperature up to 350 ◦ C. The morphology of basic thermo mechanical properties was determined pre- and post-erosion tests, in a special loading stage. The eroded sample surfaces were also examined using a scanning electron microscope (SEM).

The composite material was fabricated at the University of Cincinnati Multi-scale Materials Characterization Laboratory from 20 sheets of carbon fibers stacked unidirectionally using a high temperature and high pressure press mould method. The resulting 20.32 cm × 10.16 cm × 0.02 cm (8.0 × 4.0 × 0.008 ) composite was then cut into 2.54 cm × 2.54 cm × 0.02 cm (1 × 1 × 0.008 ) samples for erosion testing by 10 ␮m Arizona road dust and 100 ␮m sieved runway sand. The University of Cincinnati erosion wind tunnel [7] shown schematically in Fig. 1 consists of the following components: particle feeder (A), main air supply pipe (B), combustor (C), particle pre-heater (D), particle injector (E), acceleration tunnel (F), test section (G), and exhaust tank (H). Abrasive particles of a given constituency and measured weight are placed into the particle

Fig. 13. Eroded surface with 100 ␮m sieved runway sand: (a) 260 ◦ C, ˇ = 0◦ , V = 152.4 m/s, ˛ = 15◦ ; (b) T = 260 ◦ C, ˇ = 0◦ , V = 152.4 m/s, ˛ = 30◦ ; (c) T = 260 ◦ C, ˇ = 0◦ , V = 152.4 m/s, ˛ = 45◦ ; (d) T = 260 ◦ C, ˇ = 0◦ , V = 152.4 m/s, ˛ = 60◦ ; (e) T = 260 ◦ C, ˇ = 0◦ , V = 152.4 m/s, ˛ = 90◦ .

150

G. Drensky et al. / Wear 270 (2011) 146–151

feeder (A). The particles are fed into a secondary air source and blown into the particle pre-heater (D), and then to the injector (E), where they mix with the primary air supply (B), which is heated by the combustor (C). The particles are then accelerated via high velocity air in a constant-area steam-cooled duct (F) and impact the specimen in the test section (G). The particulate flow is then mixed with coolant and directed to the exhaust tank. As can be seen from Fig. 1, the tunnel geometry is uninterrupted from the acceleration tunnel throughout the test section in order to preserve the aerodynamics of the flow passing over the samples. Referring to Table 1, sample erosion tests were conducted with two types of erosive particles over a range of impact velocities (61.0, 97.5, and 152.4 m/s), and temperatures (21 and 260 ◦ C). At each test condition the unidirectional composite material samples were subjected to impacting particles at (˛ = 15◦ , 30◦ , 45◦ , 60◦ and 90◦ ) impingement angles relative to the target sample surface, and at fiber orientations parallel and normal to (ˇ = 0, 90◦ ) the impacting particle stream. The 100 ␮m Sieved runway sand (SRS) and ISO 12103-1,A2 fine 10 ␮m Arizona road dust (ARD), shown in Figs. 2 and 3 were supplied by Power Technology Incorporated. The associated variation in impact velocity with particle size distribution [18] was less than 5.6%.

3. Results and discussions Figs. 4–9 present the measured erosion rates for 10 ␮m road dust and 100 ␮m sieved runway sand. Figs. 4–7 show the results when the composite sample fiber orientation is parallel to the impacting particle stream (ˇ = 0◦ ) while Figs. 8 and 9 present the measured erosion rates at normal fiber orientation (ˇ = 90◦ ). The figures indicate a quasi-ductile erosion behavior with peak erosion at ˛ = 45◦ impingement angle at all tested impact velocities of (61.0, 97.5 and 152.4 m/s) and temperatures (21 and 260 ◦ C). Comparing Figs. 4 and 5 one can see that the 100 ␮m sieved runway sand particles cause more than double the erosion rate of the 10 ␮m Arizona road dust at the same impact conditions. This trend also prevails at 260 ◦ C, as can be seen in Figs. 6 and 7. The same could be observed in the erosion results of Figs. 8 and 9 for ˇ = 90◦ fiber orientation relative to the impacting particles. In general, the erosion rate increases with impact velocity at all impingement angles, fiber orientations, and temperatures. Comparing Fig. 4 with Figs. 8 and 5 with Fig. 9 one can see that the erosion rate is higher when the tested samples’ unidirectional fiber orientation is perpendicular to the impacting particle stream (ˇ = 90◦ ). These findings are consistent with prior polymer matrix composite (PMC) erosion studies [7,9,11–13] at ambient

Fig. 14. Eroded surface with 100 ␮m sieved runway sand: (a) T = 260 ◦ C, ˇ = 90◦ , V = 152.4 m/s, ˛ = 15◦ ; (b) T = 260 ◦ C, ˇ = 90◦ , V = 152.4 m/s, ˛ = 30◦ ; (c) T = 260 ◦ C, ˇ = 90◦ , V = 152.4 m/s, ˛ = 45◦ ; (d) T = 260 ◦ C, ˇ = 90◦ , V = 152.4 m/s, ˛ = 60◦ ; (e) 260 ◦ C, ˇ = 90◦ , V = 152.4 m/s, ˛ = 90◦ .

G. Drensky et al. / Wear 270 (2011) 146–151

temperatures and impact velocities below 100 m/s. Contrasting fiber orientation dependence was observed at 60◦ impact angle in a fiber reinforced plastics erosion study [10]. In most studies, the erosive particles were delivered using nozzles to generate particle jets with significant variation in impact velocities and impingement angles at the tested samples [4]. Furthermore, most test results were reported for total volume [12] or mass [13] removal by erosion. Only few studies [15] presented their test results in the form of erosion rate (ratio of eroded sample to impacting particle mass). Their maximum erosion rate measured at 85 m/s impact velocity and room temperature for parallel and normal unidirectional CF/PEEK, ∼0.8 g/kg for corundum particles of angular shape, which is comparable to the 10 ␮m Arizona dust results at 21 ◦ C and a low 0.28 and 0.47 for round steel particles. The measured erosion rate was higher at 260 ◦ C for all impact velocities and oblique impact angles as can be seen for peak erosion at ˛ = 45◦ in Fig. 10 for both Arizona Dust (ARD) and sieved runway sand (SRS). On the other hand, the erosion rate at normal impact (˛ = 90◦ ) was lower at 260 ◦ C than at 21 ◦ C as seen in Fig. 11. Fig. 12 presents the variation of the ratio between the measured erosion rate at normal impact and the maximum erosion rate at 45◦ impact angle with velocity and temperature for the erosion test results in which the unidirectional fiber orientation was parallel to the impacting particle stream (ˇ = 0◦ ). It shows that above 97.5 m/s, E90/E45 is independent of particle impact velocity, and that it decreases with increased temperature. Finally it is interesting to note that E90/E45 is lower in the case of erosion with the large 100 ␮m sieved runway sand. Figs. 13 and 14 present scanning electron micrographs (SEM) of eroded test surfaces at different impact angles for parallel (ˇ = 0◦ ) and normal (ˇ = 90◦ ) fiber alignment relative to the impacting particle stream. A comparison of the two figures reveals that the higher erosion rate at ˇ = 90◦ alignment is associated with a wholesale fracture and removal of the reinforcing fibers, which can be attributed to the bending stresses on the exposed fibers. Comparing Fig. 14a–c one can observe an increase in the number of fiber layers subjected to fracture and removal with increased impact angles up to ˛ = 45◦ . It is important to point out that the carbon fibers (6.5–9.5 ␮m in diameter) are an order of magnitude smaller than the impacting 100 ␮m sieved runway sand particles shown in Fig. 3. 4. Conclusions An experimental study was conducted to determine the erosion resistance of a composite material of polyaryletherketone (PEEK) matrix and unidirectional (AS4) carbon fiber at temperatures up to 260 ◦ C and velocities up to 152.4 m/s over a range of impact angles. The results indicate a quasi-ductile erosion behavior with peek erosion rate at 45◦ impact angle. The peak erosion rate increased with temperature and particle impact velocity. However, the tempera-

151

ture effect was found to reverse at 90◦ particle impact. The sieved runway sand was found to cause more than double the erosion of the Arizona road dust under the same impact conditions. The effect of unidirectional (UD) composite material fiber alignment orientation relative to the impacting particle stream on the measured erosion rate was influenced by the test temperature. Higher erosion rates were measured at 90◦ fiber orientation than at ambient temperatures as reported in prior investigations. However in the 260 ◦ C tests lower erosion rates were measured at 90◦ fiber orientation than at 0◦ . A comparison of the scanning electron micrographs (SEM) of post-erosion surfaces of tested samples reveals wholesale fracture and removal of the reinforcing UD fibers in the case of higher erosion rates measured at 90◦ fiber orientation relative to the impacting particle stream. References [1] J.K. Sutter, S.K. Naik, R. Horan, K. Miyoshi, C. Bowman, K. Ma, G. Leisser, R. Sinatra, R. Cupp, “Erosion Resistant Coatings for Plymer Matrix Composites in Propulsion Applications,” NASA/TM-2003-212201, March 2003. [2] R.C. Sirs, “The Operation of Gas Turbine Engines in Hot & Sandy ConditionsRoyal Air Force Experiences in the Gulf War,” AGARD-CP-558, 1994, Paper No. 2. [3] H.J. Mitchell, F.R. Gilmore, “Dust-Cloud Effects on Aircraft Engines: Emerging Issues and New Damage Mechanisms,” RDA-TR-120012-001, March 1982. [4] A. Hamed, W. Tabakoff, R. Wenglarz, Erosion and deposition in turbomachinery, Journal of Propulsion and Power 22 (March–April (2)) (2006). [5] A. Hamed, W. Tabakoff, R. Wenglarz, Particulate Flow and Blade Erosion Von Karman Institute for Fluid Dynamics Lecture Series 1988-08, May 1980. [6] R. Swar, A. Hamed, W. Tabakoff, R.A. Miller, Surface Deterioration of Thermal Barrier Coated Turbine Blades by Erosion, ISABE, Montreal, Canada, September 2009. [7] W. Tabakoff, T. Wakeman, Test Facility for Material Erosion at High Temperature ASTM Special Publication 664, 1979, pp. 123–135. [8] A. Hamed, W. Tabakoff, R.B. Rivir, K. Das, P. Arora, Turbine blade surface deterioration by erosion, Journal of Turbomachinery 127 (July 2005). [9] U.S. Tewari, A.M. Harsha, A.M. Hager, K. Friedrich, Solid particle erosion of carbon fibre- and glass fibre-epoxy composites, Composites Science and Technology 63 (2003) 549–557. [10] N.M. Barkoula, J. Karger-Kocsis, Review processes and influencing parameters of solid particle erosion of polymers and their composites, Journal of Materials Science 37 (2002) 3807–3820. [11] K.V. Pool, C.K.H. Dharan, I. Finne, Erosive wear of composite materials, Wear 107 (1986) 1–12. [12] T.H. Tsiang, Sand erosion of fiber composites: testing and evaluation, in: C. Chamis (Ed.), Test Methods for Design Allowables for Fibrous Composites: 2nd Volume, ASTM STP 1003, 1989, pp. 55–74. [13] N. Miyazaki, T. Hamao, Effect of interfacial strength on erosion behavior of FRPs, Journal of Composites Materials 30 (1) (1996) 35–50. [14] J. Zahavi, G.F. Schmitt, Solid particle erosion of reinforced composite materials, Wear 71 (1981) 179–190. [15] A. Hager, K. Friedrich, Y.A. Dzenis, S.A. Paipetis, Study of erosion wear of advanced polymer composites, in: Proceedings of ICCM-10, Whistler, BC, Canada, 1995, pp. IV-155–152. [16] S.K. Naik, J.K. Sutter, W. Tabakoff, R.G. Siefker, H.S. Haller, R.J. Cupp, K. Miyoshi, Wear resistant polymer matrix composites for aerospace applications, in: Proceedings of the ASME Turbo Expo, vol. 2, 2004, pp. 339–346. [17] G.K. Drensky, “Experimental Investigation of Composite Material, Erosion Characteristics under Conditions Encountered in Turbofan Engines”, Ph.D. Dissertation, University of Cincinnati, April 9, 2007. [18] W. Tabakoff, A. Hamed, B. Beacher, Investigation of Gas Particle Flow in an Erosion Wind Tunnel, October 12, 1982.