- Email: [email protected]
Effect of temperature on fatigue strength of vibration welded and unwelded glass reinforced nylon 6
Accepted Manuscript Effect of Temperature on Fatigue Strength of Vibration Welded and Unwelded Glass Reinforced Nylon 6 K.T. Lockwood, Y. Zhang, P.J. Bates, D.L. DuQuesnay PII: DOI: Reference:
S0142-1123(14)00099-1 http://dx.doi.org/10.1016/j.ijfatigue.2014.03.017 JIJF 3351
To appear in:
International Journal of Fatigue
Received Date: Revised Date: Accepted Date:
20 October 2013 15 March 2014 19 March 2014
Please cite this article as: Lockwood, K.T., Zhang, Y., Bates, P.J., DuQuesnay, D.L., Effect of Temperature on Fatigue Strength of Vibration Welded and Unwelded Glass Reinforced Nylon 6, International Journal of Fatigue (2014), doi: http://dx.doi.org/10.1016/j.ijfatigue.2014.03.017
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Effect of Temperature on Fatigue Strength of Vibration Welded and Unwelded Glass Reinforced Nylon 6
K.T. Lockwood 1 Y. Zhang 1 , P.J. Bates 2 , D.L. DuQuesnay* 1 1
Department of Mechanical and Aerospace Engineering, Royal Military College of Canada, Kingston, Ontario, Canada. 2
Department of Chemistry and Chemical Engineering, Royal Military College of Canada, Kingston, Ontario, Canada, K7K 7B4. *Corresponding author, Department of Mechanical and Aerospace Engineering, Royal Military College of Canada, PO Box 17000, Station Forces, Kingston, Ontario CANADA, K7K 7B4 [email protected], Telephone: 1 (613) 541-6000 ext 6483 Abstract The effect of temperature on the tensile and fatigue strength of vibration welded and unwelded 30 wt % glass fiber reinforced nylon 6 (PA6GF) was experimentally examined. Fatigue tests were performed under sinusoidal constant amplitude tension-tension load at a stress ratio of R=0.1 and within the frequency range of 2-10 Hz. Stress levels from just under the tensile strength down to the run-out point, at 5 million cycles, were used. It was found that increasing temperature led to a significant decrease in both tensile strength and fatigue life. However, it was also noted that for both welded and unwelded PA6GF, the endurance ratio, i.e., the ratio of fatigue strength to static tensile strength, was approximately 45% regardless of the temperature. The fatigue notch factor (Kf) lies between 1.5 and 1.75 regardless of test temperature. Keywords Fatigue strength; temperature; nylon 6; vibration welding; glass fiber reinforcement
1
1.
Introduction
Nylons have been the material of choice for vibration welded under-hood automotive parts such as air intake manifolds (AIM) for over a decade [1]. The service conditions for such applications can be harsh, with temperature extremes from -40°C to 130°C, which is why nylon and its composites have become commonly used materials within the automotive industry [1]. The vibration welding (VW) process offers advantages over other joining processes due to its short manufacturing cycle times, relatively simple equipment requirements, no pre-welding surface preparation and low risk of material deterioration due to overheating at the interface [2]. Previous studies have demonstrated that the vibration welding process parameters such as welding frequency and amplitude, weld penetration and weld pressure can affect weld strength [3-8]. Under optimal conditions, the static vibration welding strength of unreinforced nylons can approach that of the unwelded nylon; however for glass fiber reinforced nylons, the static vibration welding strength can only match that of the unreinforced unwelded nylon, because little glass fiber reinforcement occurs at the weld, due to poor fiber orientation [5, 9]. Since nylon materials are used increasingly in applications where they bear considerable stress under cyclic loads, the fatigue properties of vibration welded nylon materials have been investigated [10-12]. Tsang et al. [10, 11] studied the fatigue behavior of vibration welded nylon 6, nylon 66, as well as their 30 wt % glass fiber reinforced composites, under sinusoidal tensiontension load at a loading ratio (R) of 0.1. They studied the effect of a high, 4 MPa, and a low, 0.8 MPa, welding pressure on the static and fatigue strength of vibration welded specimens at room temperature. Regardless of the matrix material and reinforcement, the fatigue behavior of these
2
nylon composites was similar, with a slight improvement of fatigue life when a low welding pressure of 0.8 MPa was applied. For the vibration welded nylon materials studied, the fatigue limit in terms of maximum stress at R=0.1 was reported to be 0.33 of the tensile strength. Glass fiber reinforcement did not improve the fatigue behavior over unreinforced nylon, because the glass fibers were oriented parallel to the weld plane after vibration welding. The effect of inherent anisotropy of the material caused by flow-induced orientation of fibers was also reported by Zhou and Mallick [13] for nylon 66 reinforced with 33 wt % short glass fiber. They found that the fatigue strength of the material was significantly higher in the flow direction than in the transverse direction. Hartmann et. al. [14] showed that fatigue results obtained from simple geometry specimens are transferrable in the design of more complex components. They determined the fatigue life of flat specimens made of short glass fibre reinforced polyamide 6,6 under both constant and variable amplitude loadings. It was verified that that fatigue life behaviour observed from the flat specimens could be transferred to the component related specimen by conducting finite element calculation and internal pressure fatigue tests on the component related specimens. Less attention has been directed towards the fatigue performance of nylon materials at elevated temperatures. Jia and Kagan [15] conducted tension-tension fatigue tests on glass fiber reinforced nylons, at various test temperatures ranging from -40 °C to 121 °C, and observed that fatigue strength decreased monotonically with increasing test temperature. They showed that at room temperature (23°C) nylon 66 had slightly higher fatigue strength than nylon 6, but the trend was reversed at lower and higher temperatures. A previous study by Hahn et al. [16] has shown that the fatigue crack propagation rate in nylon 66 reached a minimum at 0°C and then increased with either increasing or decreasing temperature, corresponding to an optimum combination of 3
storage modulus so as to minimize fatigue damage and loss compliance values that maximize dissipation. However, Hahn et al. [16] studied temperatures up to only 50°C. Since complex nylon components are often welded and required to serve at elevated temperatures, the impact of vibration welding on material properties at elevated temperatures cannot be overlooked. This study experimentally examines the effect of temperature on the tensile and fatigue properties of vibration welded 30 wt % glass fiber reinforced nylon 6, the latter obtained under sinusoidal tension-tension loading at a loading ratio of R=0.1. Non-welded material data are also included for reference. The results are analyzed using stress-life (S-N) curves, and the fatigue notch factors at various test temperatures for VW specimens are evaluated as a guideline for designing purposes.
2.
Experimental Procedures
A 30 wt % short glass fiber reinforced nylon 6 (PA6GF, glass content determined by burning off the organic phase and fillers in a furnace at 900 °C in air) was used for this research. The PA6GF was provided in the form of pellets, which were dried at 80°C for 24 hours prior to injection molding into plaques using an Engel 55 ton injection molding machine. The molded plaque dimensions were 100 mm x 100 mm x 3.2 mm. The orientation of the glass fibers in the molded plaque can have a significant impact on its mechanical properties. The orientation of the fibers is affected by the flow of the material in the injection molding process [9, 17, 18]. Generally, in thin edge-gated plaques during cavity filling, the preferential fiber orientation is parallel to the melt flow direction as shown in Figure 1 [19].
4
The plaques were cut and the edges were milled to create flat plates with the dimensions of 60 mm × 100 mm × 3.2 mm in preparation for manufacturing butt welds via vibration welding. The cuts were made parallel to the flow direction (Figure 2) [11] so that, as is generally found in industry, the fibers are preferentially aligned along the length of the plate. Two 100 mm × 60 mm plates were welded together at the milled edges to form one weldment approximately 100 mm × 120 mm.
The method used to manufacture vibration welded specimens was duplicated from Tsang et al. [10, 11]. In the present study, a low welding pressure of 1 MPa was applied. The top plate was oscillated by a spring-mass system vibrating at 212 Hz, with peak-to-peak amplitude of 1.8 mm, while the bottom plate was held in place. The weld penetration was 1.5 mm which, once reached, the vibration was ended and the weld was allowed to solidify under constant welding pressure. The welded plates were then machined into dog bone specimens with the dimensions shown in Figure 3, with the loading axis perpendicular to the preferential fiber direction. Unwelded plaques were also machined into dog bone specimens with the same gauge dimensions and the same orientation; they were slightly shorter in overall length. Since it has been shown that water content can have a significant effect on the properties of nylon 6 [16, 20], the nylon specimens were tested dry-as-molded or were dried in an oven at 80°C for approximately 50 h before testing. The methodology recommended in ASTM D7791 was adapted for the fatigue testing in this work. For the fatigue tests, the specimens were subjected to constant amplitude sinusoidal tension-tension fatigue. The loading ratio, R was 0.1. Specimens that ran for over 5·(10)6 cycles without breaking were considered to be runouts. The 5
highest stress level at which run out was achieved was taken as the fatigue limit; three specimens were typically taken to runout for each condition. The loading frequency for the fatigue tests was kept between 2 and 10 Hz to be consistent with previous studies available in the literature [11, 21]. Test frequency was decreased with increasing stress level in order to avoid the excessive hysteretic heating that causes the specimens to fail due to permanent plastic deformation rather than fracture. The loading frequency during each test was kept constant as stipulated in ASTM D7791 - Standard Test Method for Uniaxial Fatigue Properties of Plastics. All fatigue tests were performed within an environmental chamber incorporating a furnace that allowed for the temperature to be controlled within ±2°C at the gauge section of the specimen. Figure 4 shows one actual welded fatigue specimen set up in the environmental chamber, the red frame marks the thermocouples used to control and measure the temperature within the chamber. Monotonic tension tests were also performed on the specimens at a constant displacement rate of 5 mm/min using a screw-driven testing frame in accordance to ASTM D 638. These tests were performed within an environmental chamber incorporating a furnace that allowed for the temperature to be controlled within ±2°C at the gauge section of the specimen in a configuration similar to that shown in Figure 4. The temperature range used for tensile and fatigue tests was from room temperature to 120°C. At least five specimens were tested for tensile strength under each condition.
6
3.
Results and Discussion
3.1
Tension Test Results
The results from the monotonic tension tests are shown in Figure 5. Three specimens were tested at each examined temperature and the average values are plotted. The highest standard deviation observed for a single temperature was 6.3 MPa and the average was about 1.5 MPa excluding the highest standard deviation. The lines in the tensile test results curve are only meant to guide the eye and distinguish between welded and unwelded data sets. As expected, the welded specimens are weaker than the unwelded specimens. This is caused by the vibration welding interface that is significantly weaker than the bulk material. During welding, the glass fibers tend to orient themselves in the plane of the weld creating an interface whose properties are largely dictated by the properties of the thermoplastic polymer matrix [5, 9, 17]. As expected, strength is also observed to decrease with increasing temperature. The tension data was then normalized by dividing the tensile strength at a given temperature by the tensile strength at 24°C. The results are shown in Figure 6. As can be seen, the normalized tensile strength falls on a single curve for both welded and unwelded specimens with a significant drop of 60 % in normalized strength as the testing temperature was increased from ambient to 120 °C. This shows that both welded and unwelded specimen strengths follow the same trend as temperature varies, which indicates that the reduction in strength is a result of the diminished ability of the softened matrix material to transfer stress to the reinforcing fibers as well as a loss of matrix-fiber adhesion at elevated
7
temperatures. Similar behavior in tensile strength variation at elevated test temperature has been reported for unidirectional carbon fiber reinforced epoxy [22]. 3.2
Fatigue Results
The results from the fatigue testing of the unwelded specimens are displayed in Figure 7. The solid lines are not curve fits to the data and meant only to demonstrate the trends in the results. For reference, the average tensile strengths were plotted at 1 cycle. As can be seen, increasing the temperature lowers fatigue life and fatigue strength. Also it appears that temperature has much less of an effect on low stress level fatigue tests than on high stress level fatigue tests. The reduction in fatigue life with increasing temperature could be explained by the fact that fatigue crack propagation in nylon 6 is dependent on the materials’ viscoelastic properties. As the temperature increases, the storage modulus decreases while the loss compliance increases, forming a synergy resulting in higher levels of cyclic deformation and accelerating fatigue crack propagation as demonstrated in earlier work by Hahn et al. [16]. It is also useful to note that while increasing temperature does lead to a decrease in fatigue life of the specimens, they still retain approximately half of the fatigue strength at room temperature even at the highest temperature examined. For example, for the unwelded PA6GF, a fatigue strength of 25 MPa was observed at 120 °C, while the value tested at room temperature is around 50 MPa. Figure 8 presents the fatigue life data for the welded specimens tested at various temperatures. The solid lines are not curve fits to the data and meant only to demonstrate the trends in the results. It is shown that the trends for the welded specimens are the same as for unwelded specimens. As expected, the welded specimens are weaker than the unwelded specimens tested at the same temperature. There are two main reasons for this. Firstly, the weld line introduces a geometric stress concentration, and secondly the unwelded specimens derive some strength from 8
the fibres. To elucidate the second reason, SEM fractography of two specimens tested at 70 °C was performed on their failure surfaces. Even though the maximum stress applied to the unwelded specimen was nearly twice as that applied to the welded specimen, both attained similar fatigue lives of about 105 cycles. Fractographs of unwelded specimens show that there are some fibers intersecting the fracture plane at an angle (Figure 8(a)). On the contrary, in welded specimens, fractography in the vicinity of the weld line shows that the glass fibers were oriented parallel to the weld plane after vibration welding (Figure 9(b)). Thus instead of propagating through the polymer matrix and causing debonding between the fibers and the matrix material across the weld plane (Figure 9 (b)), the fatigue crack in the unwelded specimen was deflected by the reinforcing fibers ((Figure 9 (a)); as a result, the fatigue lives of unwelded specimens were prolonged compared to welded specimens tested under the same conditions. All welded specimens failed at the weld line as expected. Tsang et al. [10] observed this in an earlier experiment and proposed that the cause was likely the notches in the vibration welded samples created by the geometry of the weld and inherent weld defects (microstructure, voids, glass, etc).
3.3
Endurance Ratios
The fatigue endurance ratio is defined as the ratio of the fatigue limit (Se) to the tensile strength (Su) for a given set of conditions. The results for both welded and unwelded specimens at each temperature tested are shown in Figure 10. The endurance ratio is consistently in the 0.4 to 0.5 range, with an average of about 0.45. This is in agreement with the ratio found by Tsang et al. [11] and equal to about 0.4 for this material in an earlier study looking at the effect of vibration welding pressure at room temperature (24°C). For unwelded glass fiber reinforced nylon 66, an endurance ratio of about 0.5 was reported by
9
Wyzgoski et al. [12]. This indicates that the reduction of fatigue strength and tensile strength remained at the same rate as temperature was increased for both unwelded and welded PA6GF. Similar rules of thumb have been found for fatigue conditions in metals. For example for fatigue testing at a stress ratio R = -1, steel and aluminium typically have a fatigue endurance ratio of 0.5 and 0.3, respectively [23]. To determine if the same type of rule could be applied also for fatigue strength, S, and not just the fatigue limit, Se, the ratio of fatigue strength to tensile strength (S/Su) was plotted versus lifespan and the results are shown in Figure 11. For the unwelded specimen data set (n=53, r2=0.9256), an empirical equation can be fitted to predict fatigue life at a given maximum fatigue stress (S) from the tensile strength (Suuw): (1)
Similarly for the welded specimen data set (n=53, r2=0.9366), the fatigue life can be predicted by: (2)
where Suw is the tensile strength of welded specimens. However, because the two sets of fitted parameters are numerically close, a universal fit to the entire unwelded and welded data set as well as calculation of a scatter band can be considered. Both welded and unwelded specimens at any temperature all fit within the same scatter band in Figure 11. The least squares best fit line to the entire data set (n=106, r2=0.9256) is shown in the figure and given by:
10
(3)
This suggests that, with only the tensile strength (Su) of this material at any temperature in the range examined, one can predict the average lifespan, N, of a specimen tested at a given maximum stress level, S, at a loading ration R=0.1. Obviously, since there is scatter in the fatigue data, this method should only be used to obtain an approximate fatigue life and a suitable factor of safety must be applied for design purposes. 3.4 Fatigue Notch factors It is known that cyclic loads cause specimens to fail at stress levels lower than their tensile strength. The fatigue notch factor, Kf, is defined in the present study as the ratio of fatigue strength of unwelded (Suw) and welded (Sw) specimens at a given fatigue life, N, and temperature. In practice, the fatigue notch factor curve can be constructed by calculating the ratio of the fatigue strength functions (S(N)) over a fatigue life interval, which for the current study is from 100 to 5 million cycles:
(2)
where Suw(N) and Sw(N) are functions of fatigue life obtained from least squares best fit lines for data presented in Figure 7 and Figure 8. The resulting fatigue notch factor curves for welded specimens as a function of fatigue life at different temperatures are presented in Figure 12. The fatigue notch factor for the welded specimens ranged from 1.5 to 1.75 regardless of the testing temperature. Analogous to the stress concentration factor, the fatigue notch factors have values higher than 1 over the entire fatigue life range. However, aside from weld geometry and
11
possible misalignment of the welded plates, other factors are also responsible for the observed reduction of fatigue strength, Sw, compared to that of unwelded specimens, Suw. Firstly, the change in the microstructure, i.e. the orientation of glass fibers along the weld plane [5, 9, 17], creates a localized weak zone for fatigue that is evident in Figure 9 (a) and 8 (b). For the unwelded specimens, the fatigue cracks could be deflected by fibers intercepting the planes perpendicular to the load direction while for the welded specimens, the fibers oriented parallel to the weld plane along the weld line, leaving the cracks propagating through the polymer matrix unimpeded. Secondly, this weak zone also provide higher density of micro-cracks, which acted as potential starting points for more crack initiation and growth, leading to accelerated fatigue failure of the welded specimens. From Figure 7 and Figure 8, it is shown that fatigue strength decreases over fatigue life at all examined temperatures for both unwelded and welded specimens. Thus a decreasing trend in Kf over fatigue life means only that the fatigue strength reduction of unwelded specimens is less severe than the welded specimens over the same span of fatigue life and vice versa. No underlying physical reason is known.
4.
Conclusions
It was found that temperature has a significant effect on both the tensile strength and fatigue life of welded and unwelded nylon 6, with higher temperatures creating an overall weaker material. It should be noted that the welded nylon specimens were found to be weaker than the unwelded specimens due to the dominant parallel fiber orientation along the welding plane and possible defects within it. Empirical equations that predict fatigue life from the respective tensile strength were proposed for welded and unwelded specimens separately. However, the two sets of
12
parameters in the equations were very close; a single empirical equation was found adequate for both welded and unwelded specimens to be estimated. It was observed that in fatigue testing at R=0.1, there was a consistent ratio of about 0.45 between the fatigue limit stress and ultimate tensile strength of the glass filled nylon 6 regardless of testing temperature for both welded and unwelded specimens. The fatigue notch factor for the welded specimen varied between 1.5 and 1.75. 5.
Acknowledgements
Financial support of this research was provided by AUTO 21, the Ontario Centre of Excellence, OCE Project NM50966, and by the Natural Sciences and Engineering Research Council of Canada, grant 239174. The work was also supported by an industrial partner, MAHLE Filter Systems Canada, ULC, and the technical support and contribution of Jim Vanderveen and Bobbye Baylis are also gratefully acknowledged.
13
REFERENCES
[1] J. Lee, L. Roessler, Vibration Welded Composite Intake Manifolds - Design Considerations and Material Selection Criteria, Proceed. 1997 Int. Cong. Expo.1253 (1997) 27-37. [2] V. K. Stokes, Toward a Weld-strength Data Base for Vibration Welding of Thermoplastics, ANTEC Conf. Proceed. 1 (1995) 1280-1284. [3] V. K. Stokes, Vibration Welding of Thermoplastics. Part I: Phenomenology of the Welding Process, Polym. Eng. Sci. 28 (1988) 718-727. [4] J. J. MacDonald, Vibration Welding of Engineering Plastics, M. Sc. Thesis, Royal Military College of Canada (2001). [5] V. A. Kagan, Vibration Welding of Thermoplastics, Technical Paper - Society of Manufacturing Engineers. AD (1999) 1-21. [6] J. Panaswich, Vibration Welding Joins Thermoplastics, SAE J. Auto. Eng. 92 (1984) 40-44. [7] P. Bates, D. Couzens, J. Kendall, Vibration Welding of Continuously Reinforced Thermoplastic composites, J. Thermoplast. Compos. Mater. 14 (2001) 344-354. [8] B. Patham, P. H. Foss, Thermoplastic Vibration Welding: Review of Process Phenomenology and Processing-structure-property Interrelationships, Polym. Eng. Sci. 51 (2011) 1-22. [9] V. A. Kagan, C. Roth, The Effects of Weld Geometry and Glass-fiber Orientation on the Mechanical Performance of Joints - Part 1: Weld Design Issues, J. Reinf. Plast. Compos. 23 (2004) 167-175. [10] K. Y. Tsang, D. L. DuQuesnay, P. J. Bates, Fatigue Strength of Vibration-welded Unreinforced Nylon Butt Joints, Polym. Eng. Sci. 45 (2005) 935-944. [11] K. Y. Tsang, D. L. DuQuesnay, P. J. Bates, Fatigue Properties of Vibration-welded Nylon 6 and Nylon 66 Reinforced with Glass Fibres, Compos. B: Eng. 39 (2008) 396-404. [12] M. G. Wyzgoski, J. A. Krohn, G. E. Novak, Fatigue of Fiber-reinforced Injection Molded Plastics. I: Stress-lifetime Data, Polym. Compos. 25 (2004) 489-498. [13] Y. Zhou, P.K. Mallick, Fatigue Performance of an Injection-molded Short E-glass Fiberreinforced Polyamide 6,6. I. Effects of Orientation, Holes, and Weld Line, Polym. Compos. 27 (2006) 230-237.
14
[14] J. Hartmann, E. Moosbrugger, A. Buter, Variable Amplitude Loading with Components Made of Short Fiber Reinforced Polyamide 6.6. 10 (2011) 2009-2015. [15] N. Jia, V. A. Kagan, Effects of Time and Temperature on the Tension-tension Fatigue Behavior of Short Fiber Reinforced Polyamides, Polym. Compos. 19 (1998) 408-414. [16] M. T. Hahn, R. W. Hertzberg, J. A. Manson, L. H. Sperling, Influence of Temperature and Absorbed Water on The Fatigue Crack Propagation in Nylon-6,6, Polymer. 27 (1986) 18851888. [17] V. Kagan, S. Lui, G. R. Smith, J. Patry, Optimized Performance of Linear Vibration Welded Nylon 6 and Nylon 66 Butt Joints, ANTEC Conf. Proceed. 1 (1996) 1266-1274. [18] P. F. Bright, M. W. Darlington, Factors Influencing Fibre Orientation and Mechanical Properties in Fibre Reinforced Thermoplastics Injection Mouldings, Plast. Rub. Process. App. 1 (1981) 139-147. [19] N. G. McCrum, C. P. Buckley, C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., Oxford University Press, Oxford, 1997. [20] J. F. Mandell, M. G. Steckel, S. S. Chung, M. C. Kenney, Fatigue and Environmental Resistance of Polyester and Nylon Fibers, Polym. Eng. Sci. 27 (1987) 1121-1127. [21] J. A. Sauer, G. C. Richardson, Fatigue of Polymers, Int. J. Fract. 16 (1980) 499-532. [22] S. Cao, X. Wang, Z. Wu, Evaluation and Prediction of Temperature-dependent Tensile Strength of Unidirectional Carbon Fiber-reinforced Polymer Composites, J. Reinf Plast. Compos. 30 (2011) 799-807. [23] R. I. Stephens, A. A. Fatemi, R. R. Stephens, H. O. Fuchs, Metal Fatigue in Engineering, 2nd ed., Revised ed., Wiley-Interscience, John Wiley & Sons Inc., Hoboken, 2000
15
APPENDIX Table 1 The tensile strength and standard deviation of unwelded PA6GF. Test Temperature (°C)
Tensile Strength (MPa)
Standard Deviation (MPa)
24
113.3
1.5
50
94.4
1.4
70
75.3
1.8
80
62.0
1.8
120
49.2
0.6
Table 2 The tensile strength and standard deviation of welded PA6GF. Test Temperature (°C)
Tensile Strength (MPa)
Standard Deviation (MPa)
24
72.0
6.3
50
64.5
1.5
70
46.5
1.0
120
29.1
2.0
16
Figure 1. Flow direction and predominant glass fiber orientation for a plaque [11].
17
Figure 2. Pre-welding cut and welding alignment for nylon plaques used in butt welds. [11]. The arrows denote the flow direction of the polymer melt during injection molding.
18
Figure 3. Geometry of an unwelded plaque (left), welded plates (middle) and a test specimen (right) [11].
19
Figure 4 A welded PA6GF specimen clamped in the environmental chamber on the fatigue tester; the red frame marks the thermocouples used to control and measure the temperature within the chamber.
20
Figure 5. Effect of temperature on the average tensile strength of welded and unwelded PA6GF. The lines are meant to guide the eye only.
21
Normalized Tensile Strength
1.2 unwelded
1
welded
0.8 0.6 0.4 0.2 0 0
50
100 Temperature (°C)
150
Figure 6. Effect of temperature on normalized tensile strength.
22
Maximum Stress (MPa)
120 110 100 90 80 70 60 50 40 30 20 10 0 1E+0
24 °C 50 °C 70 °C 120 °C
1E+1
1E+2
1E+3 1E+4 1E+5 Cycles to Failure
1E+6
1E+7
Figure 7. Fatigue life data at various temperatures for unwelded PA6GF specimens. Each arrow denotes one (1) run-out specimen. The solid lines are meant to demonstrate the trends in the data.
23
80 24°C
Maximum Stress (MPa)
70
50°C
60
70°C
50
120°C
40 30 20 10 0 1E+0
1E+1
1E+2
1E+3 1E+4 1E+5 Cycles to Failure
1E+6
1E+7
Figure 8. Fatigue life data at various temperatures for welded PA6GF specimens. Each arrow denotes one (1) run-out specimen. The solid lines are meant to demonstrate the trends in the data.
24
(a)
(b)
Figure 9 SEM image of the fatigue failure surface after testing at 70°C for (a) an unwelded specimen tested at 40 MPa and (b) a welded specimen tested at 25 MPa; the fatigue life was 25
about
105
cycles
for
both
specimens.
26
0.6
Endurance Ratio
0.5 0.4 0.3 0.2
Unwelded
0.1
Welded
0 0
50
100 Temperature (°C)
150
Figure 10. Endurance ratios for both unwelded and welded PA6GF specimens as a function of temperature. Error bars denote the scattering of the endurance ratio caused by the error in the static tensile strength measurement.
27
1.2 1
S/Su
0.8 0.6 0.4 0.2 0 1.E+0
24 °C
50 °C
70 °C
120 °C
24 °C
50 °C
70 °C
120 °C
1.E+2
1.E+4 Cycles to Failure
1.E+6
Figure 11. Ratio fatigue strength (S) to their respective tensile strength (Su) for both unwelded (hollow symbols) and welded (filled symbols) specimens at all the tested temperatures as a function of fatigue life. The solid line represents the least squares fit to the data set. The dashed lines represent the scatter band. The difference between the fitted line and the scatter band is the ratio of the standard deviation of the fit to the local predicted value provided by Eqn (1).
28
2 1.8 1.6
Kf = Suw/Sw
1.4 1.2 1 0.8 0.6
24 °C
50 °C
0.4
70 °C
120 °C
0.2 0 1E+1
1E+2
1E+3
1E+4 1E+5 1E+6 Cycles to Failure
1E+7
1E+8
Figure 12 Fatigue notch factor (Kf) as a function of cycles to failure (N) for the welded specimens under different test temperatures.
29
Highlights • The tensile and fatigue strength of glass fiber reinforced nylon 6 were examined • The effects of temperature, from 24°C to 120°C, and vibration welding were studied • Increasing temperature causes a reduction of both tensile and fatigue strengths • For both welded and unwelded PA6GF, the endurance ratio is approximately 45% • The fatigue notch factor lies between 1.5 and 1.75
30
S0142-1123(14)00099-1 http://dx.doi.org/10.1016/j.ijfatigue.2014.03.017 JIJF 3351
To appear in:
International Journal of Fatigue
Received Date: Revised Date: Accepted Date:
20 October 2013 15 March 2014 19 March 2014
Please cite this article as: Lockwood, K.T., Zhang, Y., Bates, P.J., DuQuesnay, D.L., Effect of Temperature on Fatigue Strength of Vibration Welded and Unwelded Glass Reinforced Nylon 6, International Journal of Fatigue (2014), doi: http://dx.doi.org/10.1016/j.ijfatigue.2014.03.017
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Effect of Temperature on Fatigue Strength of Vibration Welded and Unwelded Glass Reinforced Nylon 6
K.T. Lockwood 1 Y. Zhang 1 , P.J. Bates 2 , D.L. DuQuesnay* 1 1
Department of Mechanical and Aerospace Engineering, Royal Military College of Canada, Kingston, Ontario, Canada. 2
Department of Chemistry and Chemical Engineering, Royal Military College of Canada, Kingston, Ontario, Canada, K7K 7B4. *Corresponding author, Department of Mechanical and Aerospace Engineering, Royal Military College of Canada, PO Box 17000, Station Forces, Kingston, Ontario CANADA, K7K 7B4 [email protected], Telephone: 1 (613) 541-6000 ext 6483 Abstract The effect of temperature on the tensile and fatigue strength of vibration welded and unwelded 30 wt % glass fiber reinforced nylon 6 (PA6GF) was experimentally examined. Fatigue tests were performed under sinusoidal constant amplitude tension-tension load at a stress ratio of R=0.1 and within the frequency range of 2-10 Hz. Stress levels from just under the tensile strength down to the run-out point, at 5 million cycles, were used. It was found that increasing temperature led to a significant decrease in both tensile strength and fatigue life. However, it was also noted that for both welded and unwelded PA6GF, the endurance ratio, i.e., the ratio of fatigue strength to static tensile strength, was approximately 45% regardless of the temperature. The fatigue notch factor (Kf) lies between 1.5 and 1.75 regardless of test temperature. Keywords Fatigue strength; temperature; nylon 6; vibration welding; glass fiber reinforcement
1
1.
Introduction
Nylons have been the material of choice for vibration welded under-hood automotive parts such as air intake manifolds (AIM) for over a decade [1]. The service conditions for such applications can be harsh, with temperature extremes from -40°C to 130°C, which is why nylon and its composites have become commonly used materials within the automotive industry [1]. The vibration welding (VW) process offers advantages over other joining processes due to its short manufacturing cycle times, relatively simple equipment requirements, no pre-welding surface preparation and low risk of material deterioration due to overheating at the interface [2]. Previous studies have demonstrated that the vibration welding process parameters such as welding frequency and amplitude, weld penetration and weld pressure can affect weld strength [3-8]. Under optimal conditions, the static vibration welding strength of unreinforced nylons can approach that of the unwelded nylon; however for glass fiber reinforced nylons, the static vibration welding strength can only match that of the unreinforced unwelded nylon, because little glass fiber reinforcement occurs at the weld, due to poor fiber orientation [5, 9]. Since nylon materials are used increasingly in applications where they bear considerable stress under cyclic loads, the fatigue properties of vibration welded nylon materials have been investigated [10-12]. Tsang et al. [10, 11] studied the fatigue behavior of vibration welded nylon 6, nylon 66, as well as their 30 wt % glass fiber reinforced composites, under sinusoidal tensiontension load at a loading ratio (R) of 0.1. They studied the effect of a high, 4 MPa, and a low, 0.8 MPa, welding pressure on the static and fatigue strength of vibration welded specimens at room temperature. Regardless of the matrix material and reinforcement, the fatigue behavior of these
2
nylon composites was similar, with a slight improvement of fatigue life when a low welding pressure of 0.8 MPa was applied. For the vibration welded nylon materials studied, the fatigue limit in terms of maximum stress at R=0.1 was reported to be 0.33 of the tensile strength. Glass fiber reinforcement did not improve the fatigue behavior over unreinforced nylon, because the glass fibers were oriented parallel to the weld plane after vibration welding. The effect of inherent anisotropy of the material caused by flow-induced orientation of fibers was also reported by Zhou and Mallick [13] for nylon 66 reinforced with 33 wt % short glass fiber. They found that the fatigue strength of the material was significantly higher in the flow direction than in the transverse direction. Hartmann et. al. [14] showed that fatigue results obtained from simple geometry specimens are transferrable in the design of more complex components. They determined the fatigue life of flat specimens made of short glass fibre reinforced polyamide 6,6 under both constant and variable amplitude loadings. It was verified that that fatigue life behaviour observed from the flat specimens could be transferred to the component related specimen by conducting finite element calculation and internal pressure fatigue tests on the component related specimens. Less attention has been directed towards the fatigue performance of nylon materials at elevated temperatures. Jia and Kagan [15] conducted tension-tension fatigue tests on glass fiber reinforced nylons, at various test temperatures ranging from -40 °C to 121 °C, and observed that fatigue strength decreased monotonically with increasing test temperature. They showed that at room temperature (23°C) nylon 66 had slightly higher fatigue strength than nylon 6, but the trend was reversed at lower and higher temperatures. A previous study by Hahn et al. [16] has shown that the fatigue crack propagation rate in nylon 66 reached a minimum at 0°C and then increased with either increasing or decreasing temperature, corresponding to an optimum combination of 3
storage modulus so as to minimize fatigue damage and loss compliance values that maximize dissipation. However, Hahn et al. [16] studied temperatures up to only 50°C. Since complex nylon components are often welded and required to serve at elevated temperatures, the impact of vibration welding on material properties at elevated temperatures cannot be overlooked. This study experimentally examines the effect of temperature on the tensile and fatigue properties of vibration welded 30 wt % glass fiber reinforced nylon 6, the latter obtained under sinusoidal tension-tension loading at a loading ratio of R=0.1. Non-welded material data are also included for reference. The results are analyzed using stress-life (S-N) curves, and the fatigue notch factors at various test temperatures for VW specimens are evaluated as a guideline for designing purposes.
2.
Experimental Procedures
A 30 wt % short glass fiber reinforced nylon 6 (PA6GF, glass content determined by burning off the organic phase and fillers in a furnace at 900 °C in air) was used for this research. The PA6GF was provided in the form of pellets, which were dried at 80°C for 24 hours prior to injection molding into plaques using an Engel 55 ton injection molding machine. The molded plaque dimensions were 100 mm x 100 mm x 3.2 mm. The orientation of the glass fibers in the molded plaque can have a significant impact on its mechanical properties. The orientation of the fibers is affected by the flow of the material in the injection molding process [9, 17, 18]. Generally, in thin edge-gated plaques during cavity filling, the preferential fiber orientation is parallel to the melt flow direction as shown in Figure 1 [19].
4
The plaques were cut and the edges were milled to create flat plates with the dimensions of 60 mm × 100 mm × 3.2 mm in preparation for manufacturing butt welds via vibration welding. The cuts were made parallel to the flow direction (Figure 2) [11] so that, as is generally found in industry, the fibers are preferentially aligned along the length of the plate. Two 100 mm × 60 mm plates were welded together at the milled edges to form one weldment approximately 100 mm × 120 mm.
The method used to manufacture vibration welded specimens was duplicated from Tsang et al. [10, 11]. In the present study, a low welding pressure of 1 MPa was applied. The top plate was oscillated by a spring-mass system vibrating at 212 Hz, with peak-to-peak amplitude of 1.8 mm, while the bottom plate was held in place. The weld penetration was 1.5 mm which, once reached, the vibration was ended and the weld was allowed to solidify under constant welding pressure. The welded plates were then machined into dog bone specimens with the dimensions shown in Figure 3, with the loading axis perpendicular to the preferential fiber direction. Unwelded plaques were also machined into dog bone specimens with the same gauge dimensions and the same orientation; they were slightly shorter in overall length. Since it has been shown that water content can have a significant effect on the properties of nylon 6 [16, 20], the nylon specimens were tested dry-as-molded or were dried in an oven at 80°C for approximately 50 h before testing. The methodology recommended in ASTM D7791 was adapted for the fatigue testing in this work. For the fatigue tests, the specimens were subjected to constant amplitude sinusoidal tension-tension fatigue. The loading ratio, R was 0.1. Specimens that ran for over 5·(10)6 cycles without breaking were considered to be runouts. The 5
highest stress level at which run out was achieved was taken as the fatigue limit; three specimens were typically taken to runout for each condition. The loading frequency for the fatigue tests was kept between 2 and 10 Hz to be consistent with previous studies available in the literature [11, 21]. Test frequency was decreased with increasing stress level in order to avoid the excessive hysteretic heating that causes the specimens to fail due to permanent plastic deformation rather than fracture. The loading frequency during each test was kept constant as stipulated in ASTM D7791 - Standard Test Method for Uniaxial Fatigue Properties of Plastics. All fatigue tests were performed within an environmental chamber incorporating a furnace that allowed for the temperature to be controlled within ±2°C at the gauge section of the specimen. Figure 4 shows one actual welded fatigue specimen set up in the environmental chamber, the red frame marks the thermocouples used to control and measure the temperature within the chamber. Monotonic tension tests were also performed on the specimens at a constant displacement rate of 5 mm/min using a screw-driven testing frame in accordance to ASTM D 638. These tests were performed within an environmental chamber incorporating a furnace that allowed for the temperature to be controlled within ±2°C at the gauge section of the specimen in a configuration similar to that shown in Figure 4. The temperature range used for tensile and fatigue tests was from room temperature to 120°C. At least five specimens were tested for tensile strength under each condition.
6
3.
Results and Discussion
3.1
Tension Test Results
The results from the monotonic tension tests are shown in Figure 5. Three specimens were tested at each examined temperature and the average values are plotted. The highest standard deviation observed for a single temperature was 6.3 MPa and the average was about 1.5 MPa excluding the highest standard deviation. The lines in the tensile test results curve are only meant to guide the eye and distinguish between welded and unwelded data sets. As expected, the welded specimens are weaker than the unwelded specimens. This is caused by the vibration welding interface that is significantly weaker than the bulk material. During welding, the glass fibers tend to orient themselves in the plane of the weld creating an interface whose properties are largely dictated by the properties of the thermoplastic polymer matrix [5, 9, 17]. As expected, strength is also observed to decrease with increasing temperature. The tension data was then normalized by dividing the tensile strength at a given temperature by the tensile strength at 24°C. The results are shown in Figure 6. As can be seen, the normalized tensile strength falls on a single curve for both welded and unwelded specimens with a significant drop of 60 % in normalized strength as the testing temperature was increased from ambient to 120 °C. This shows that both welded and unwelded specimen strengths follow the same trend as temperature varies, which indicates that the reduction in strength is a result of the diminished ability of the softened matrix material to transfer stress to the reinforcing fibers as well as a loss of matrix-fiber adhesion at elevated
7
temperatures. Similar behavior in tensile strength variation at elevated test temperature has been reported for unidirectional carbon fiber reinforced epoxy [22]. 3.2
Fatigue Results
The results from the fatigue testing of the unwelded specimens are displayed in Figure 7. The solid lines are not curve fits to the data and meant only to demonstrate the trends in the results. For reference, the average tensile strengths were plotted at 1 cycle. As can be seen, increasing the temperature lowers fatigue life and fatigue strength. Also it appears that temperature has much less of an effect on low stress level fatigue tests than on high stress level fatigue tests. The reduction in fatigue life with increasing temperature could be explained by the fact that fatigue crack propagation in nylon 6 is dependent on the materials’ viscoelastic properties. As the temperature increases, the storage modulus decreases while the loss compliance increases, forming a synergy resulting in higher levels of cyclic deformation and accelerating fatigue crack propagation as demonstrated in earlier work by Hahn et al. [16]. It is also useful to note that while increasing temperature does lead to a decrease in fatigue life of the specimens, they still retain approximately half of the fatigue strength at room temperature even at the highest temperature examined. For example, for the unwelded PA6GF, a fatigue strength of 25 MPa was observed at 120 °C, while the value tested at room temperature is around 50 MPa. Figure 8 presents the fatigue life data for the welded specimens tested at various temperatures. The solid lines are not curve fits to the data and meant only to demonstrate the trends in the results. It is shown that the trends for the welded specimens are the same as for unwelded specimens. As expected, the welded specimens are weaker than the unwelded specimens tested at the same temperature. There are two main reasons for this. Firstly, the weld line introduces a geometric stress concentration, and secondly the unwelded specimens derive some strength from 8
the fibres. To elucidate the second reason, SEM fractography of two specimens tested at 70 °C was performed on their failure surfaces. Even though the maximum stress applied to the unwelded specimen was nearly twice as that applied to the welded specimen, both attained similar fatigue lives of about 105 cycles. Fractographs of unwelded specimens show that there are some fibers intersecting the fracture plane at an angle (Figure 8(a)). On the contrary, in welded specimens, fractography in the vicinity of the weld line shows that the glass fibers were oriented parallel to the weld plane after vibration welding (Figure 9(b)). Thus instead of propagating through the polymer matrix and causing debonding between the fibers and the matrix material across the weld plane (Figure 9 (b)), the fatigue crack in the unwelded specimen was deflected by the reinforcing fibers ((Figure 9 (a)); as a result, the fatigue lives of unwelded specimens were prolonged compared to welded specimens tested under the same conditions. All welded specimens failed at the weld line as expected. Tsang et al. [10] observed this in an earlier experiment and proposed that the cause was likely the notches in the vibration welded samples created by the geometry of the weld and inherent weld defects (microstructure, voids, glass, etc).
3.3
Endurance Ratios
The fatigue endurance ratio is defined as the ratio of the fatigue limit (Se) to the tensile strength (Su) for a given set of conditions. The results for both welded and unwelded specimens at each temperature tested are shown in Figure 10. The endurance ratio is consistently in the 0.4 to 0.5 range, with an average of about 0.45. This is in agreement with the ratio found by Tsang et al. [11] and equal to about 0.4 for this material in an earlier study looking at the effect of vibration welding pressure at room temperature (24°C). For unwelded glass fiber reinforced nylon 66, an endurance ratio of about 0.5 was reported by
9
Wyzgoski et al. [12]. This indicates that the reduction of fatigue strength and tensile strength remained at the same rate as temperature was increased for both unwelded and welded PA6GF. Similar rules of thumb have been found for fatigue conditions in metals. For example for fatigue testing at a stress ratio R = -1, steel and aluminium typically have a fatigue endurance ratio of 0.5 and 0.3, respectively [23]. To determine if the same type of rule could be applied also for fatigue strength, S, and not just the fatigue limit, Se, the ratio of fatigue strength to tensile strength (S/Su) was plotted versus lifespan and the results are shown in Figure 11. For the unwelded specimen data set (n=53, r2=0.9256), an empirical equation can be fitted to predict fatigue life at a given maximum fatigue stress (S) from the tensile strength (Suuw): (1)
Similarly for the welded specimen data set (n=53, r2=0.9366), the fatigue life can be predicted by: (2)
where Suw is the tensile strength of welded specimens. However, because the two sets of fitted parameters are numerically close, a universal fit to the entire unwelded and welded data set as well as calculation of a scatter band can be considered. Both welded and unwelded specimens at any temperature all fit within the same scatter band in Figure 11. The least squares best fit line to the entire data set (n=106, r2=0.9256) is shown in the figure and given by:
10
(3)
This suggests that, with only the tensile strength (Su) of this material at any temperature in the range examined, one can predict the average lifespan, N, of a specimen tested at a given maximum stress level, S, at a loading ration R=0.1. Obviously, since there is scatter in the fatigue data, this method should only be used to obtain an approximate fatigue life and a suitable factor of safety must be applied for design purposes. 3.4 Fatigue Notch factors It is known that cyclic loads cause specimens to fail at stress levels lower than their tensile strength. The fatigue notch factor, Kf, is defined in the present study as the ratio of fatigue strength of unwelded (Suw) and welded (Sw) specimens at a given fatigue life, N, and temperature. In practice, the fatigue notch factor curve can be constructed by calculating the ratio of the fatigue strength functions (S(N)) over a fatigue life interval, which for the current study is from 100 to 5 million cycles:
(2)
where Suw(N) and Sw(N) are functions of fatigue life obtained from least squares best fit lines for data presented in Figure 7 and Figure 8. The resulting fatigue notch factor curves for welded specimens as a function of fatigue life at different temperatures are presented in Figure 12. The fatigue notch factor for the welded specimens ranged from 1.5 to 1.75 regardless of the testing temperature. Analogous to the stress concentration factor, the fatigue notch factors have values higher than 1 over the entire fatigue life range. However, aside from weld geometry and
11
possible misalignment of the welded plates, other factors are also responsible for the observed reduction of fatigue strength, Sw, compared to that of unwelded specimens, Suw. Firstly, the change in the microstructure, i.e. the orientation of glass fibers along the weld plane [5, 9, 17], creates a localized weak zone for fatigue that is evident in Figure 9 (a) and 8 (b). For the unwelded specimens, the fatigue cracks could be deflected by fibers intercepting the planes perpendicular to the load direction while for the welded specimens, the fibers oriented parallel to the weld plane along the weld line, leaving the cracks propagating through the polymer matrix unimpeded. Secondly, this weak zone also provide higher density of micro-cracks, which acted as potential starting points for more crack initiation and growth, leading to accelerated fatigue failure of the welded specimens. From Figure 7 and Figure 8, it is shown that fatigue strength decreases over fatigue life at all examined temperatures for both unwelded and welded specimens. Thus a decreasing trend in Kf over fatigue life means only that the fatigue strength reduction of unwelded specimens is less severe than the welded specimens over the same span of fatigue life and vice versa. No underlying physical reason is known.
4.
Conclusions
It was found that temperature has a significant effect on both the tensile strength and fatigue life of welded and unwelded nylon 6, with higher temperatures creating an overall weaker material. It should be noted that the welded nylon specimens were found to be weaker than the unwelded specimens due to the dominant parallel fiber orientation along the welding plane and possible defects within it. Empirical equations that predict fatigue life from the respective tensile strength were proposed for welded and unwelded specimens separately. However, the two sets of
12
parameters in the equations were very close; a single empirical equation was found adequate for both welded and unwelded specimens to be estimated. It was observed that in fatigue testing at R=0.1, there was a consistent ratio of about 0.45 between the fatigue limit stress and ultimate tensile strength of the glass filled nylon 6 regardless of testing temperature for both welded and unwelded specimens. The fatigue notch factor for the welded specimen varied between 1.5 and 1.75. 5.
Acknowledgements
Financial support of this research was provided by AUTO 21, the Ontario Centre of Excellence, OCE Project NM50966, and by the Natural Sciences and Engineering Research Council of Canada, grant 239174. The work was also supported by an industrial partner, MAHLE Filter Systems Canada, ULC, and the technical support and contribution of Jim Vanderveen and Bobbye Baylis are also gratefully acknowledged.
13
REFERENCES
[1] J. Lee, L. Roessler, Vibration Welded Composite Intake Manifolds - Design Considerations and Material Selection Criteria, Proceed. 1997 Int. Cong. Expo.1253 (1997) 27-37. [2] V. K. Stokes, Toward a Weld-strength Data Base for Vibration Welding of Thermoplastics, ANTEC Conf. Proceed. 1 (1995) 1280-1284. [3] V. K. Stokes, Vibration Welding of Thermoplastics. Part I: Phenomenology of the Welding Process, Polym. Eng. Sci. 28 (1988) 718-727. [4] J. J. MacDonald, Vibration Welding of Engineering Plastics, M. Sc. Thesis, Royal Military College of Canada (2001). [5] V. A. Kagan, Vibration Welding of Thermoplastics, Technical Paper - Society of Manufacturing Engineers. AD (1999) 1-21. [6] J. Panaswich, Vibration Welding Joins Thermoplastics, SAE J. Auto. Eng. 92 (1984) 40-44. [7] P. Bates, D. Couzens, J. Kendall, Vibration Welding of Continuously Reinforced Thermoplastic composites, J. Thermoplast. Compos. Mater. 14 (2001) 344-354. [8] B. Patham, P. H. Foss, Thermoplastic Vibration Welding: Review of Process Phenomenology and Processing-structure-property Interrelationships, Polym. Eng. Sci. 51 (2011) 1-22. [9] V. A. Kagan, C. Roth, The Effects of Weld Geometry and Glass-fiber Orientation on the Mechanical Performance of Joints - Part 1: Weld Design Issues, J. Reinf. Plast. Compos. 23 (2004) 167-175. [10] K. Y. Tsang, D. L. DuQuesnay, P. J. Bates, Fatigue Strength of Vibration-welded Unreinforced Nylon Butt Joints, Polym. Eng. Sci. 45 (2005) 935-944. [11] K. Y. Tsang, D. L. DuQuesnay, P. J. Bates, Fatigue Properties of Vibration-welded Nylon 6 and Nylon 66 Reinforced with Glass Fibres, Compos. B: Eng. 39 (2008) 396-404. [12] M. G. Wyzgoski, J. A. Krohn, G. E. Novak, Fatigue of Fiber-reinforced Injection Molded Plastics. I: Stress-lifetime Data, Polym. Compos. 25 (2004) 489-498. [13] Y. Zhou, P.K. Mallick, Fatigue Performance of an Injection-molded Short E-glass Fiberreinforced Polyamide 6,6. I. Effects of Orientation, Holes, and Weld Line, Polym. Compos. 27 (2006) 230-237.
14
[14] J. Hartmann, E. Moosbrugger, A. Buter, Variable Amplitude Loading with Components Made of Short Fiber Reinforced Polyamide 6.6. 10 (2011) 2009-2015. [15] N. Jia, V. A. Kagan, Effects of Time and Temperature on the Tension-tension Fatigue Behavior of Short Fiber Reinforced Polyamides, Polym. Compos. 19 (1998) 408-414. [16] M. T. Hahn, R. W. Hertzberg, J. A. Manson, L. H. Sperling, Influence of Temperature and Absorbed Water on The Fatigue Crack Propagation in Nylon-6,6, Polymer. 27 (1986) 18851888. [17] V. Kagan, S. Lui, G. R. Smith, J. Patry, Optimized Performance of Linear Vibration Welded Nylon 6 and Nylon 66 Butt Joints, ANTEC Conf. Proceed. 1 (1996) 1266-1274. [18] P. F. Bright, M. W. Darlington, Factors Influencing Fibre Orientation and Mechanical Properties in Fibre Reinforced Thermoplastics Injection Mouldings, Plast. Rub. Process. App. 1 (1981) 139-147. [19] N. G. McCrum, C. P. Buckley, C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., Oxford University Press, Oxford, 1997. [20] J. F. Mandell, M. G. Steckel, S. S. Chung, M. C. Kenney, Fatigue and Environmental Resistance of Polyester and Nylon Fibers, Polym. Eng. Sci. 27 (1987) 1121-1127. [21] J. A. Sauer, G. C. Richardson, Fatigue of Polymers, Int. J. Fract. 16 (1980) 499-532. [22] S. Cao, X. Wang, Z. Wu, Evaluation and Prediction of Temperature-dependent Tensile Strength of Unidirectional Carbon Fiber-reinforced Polymer Composites, J. Reinf Plast. Compos. 30 (2011) 799-807. [23] R. I. Stephens, A. A. Fatemi, R. R. Stephens, H. O. Fuchs, Metal Fatigue in Engineering, 2nd ed., Revised ed., Wiley-Interscience, John Wiley & Sons Inc., Hoboken, 2000
15
APPENDIX Table 1 The tensile strength and standard deviation of unwelded PA6GF. Test Temperature (°C)
Tensile Strength (MPa)
Standard Deviation (MPa)
24
113.3
1.5
50
94.4
1.4
70
75.3
1.8
80
62.0
1.8
120
49.2
0.6
Table 2 The tensile strength and standard deviation of welded PA6GF. Test Temperature (°C)
Tensile Strength (MPa)
Standard Deviation (MPa)
24
72.0
6.3
50
64.5
1.5
70
46.5
1.0
120
29.1
2.0
16
Figure 1. Flow direction and predominant glass fiber orientation for a plaque [11].
17
Figure 2. Pre-welding cut and welding alignment for nylon plaques used in butt welds. [11]. The arrows denote the flow direction of the polymer melt during injection molding.
18
Figure 3. Geometry of an unwelded plaque (left), welded plates (middle) and a test specimen (right) [11].
19
Figure 4 A welded PA6GF specimen clamped in the environmental chamber on the fatigue tester; the red frame marks the thermocouples used to control and measure the temperature within the chamber.
20
Figure 5. Effect of temperature on the average tensile strength of welded and unwelded PA6GF. The lines are meant to guide the eye only.
21
Normalized Tensile Strength
1.2 unwelded
1
welded
0.8 0.6 0.4 0.2 0 0
50
100 Temperature (°C)
150
Figure 6. Effect of temperature on normalized tensile strength.
22
Maximum Stress (MPa)
120 110 100 90 80 70 60 50 40 30 20 10 0 1E+0
24 °C 50 °C 70 °C 120 °C
1E+1
1E+2
1E+3 1E+4 1E+5 Cycles to Failure
1E+6
1E+7
Figure 7. Fatigue life data at various temperatures for unwelded PA6GF specimens. Each arrow denotes one (1) run-out specimen. The solid lines are meant to demonstrate the trends in the data.
23
80 24°C
Maximum Stress (MPa)
70
50°C
60
70°C
50
120°C
40 30 20 10 0 1E+0
1E+1
1E+2
1E+3 1E+4 1E+5 Cycles to Failure
1E+6
1E+7
Figure 8. Fatigue life data at various temperatures for welded PA6GF specimens. Each arrow denotes one (1) run-out specimen. The solid lines are meant to demonstrate the trends in the data.
24
(a)
(b)
Figure 9 SEM image of the fatigue failure surface after testing at 70°C for (a) an unwelded specimen tested at 40 MPa and (b) a welded specimen tested at 25 MPa; the fatigue life was 25
about
105
cycles
for
both
specimens.
26
0.6
Endurance Ratio
0.5 0.4 0.3 0.2
Unwelded
0.1
Welded
0 0
50
100 Temperature (°C)
150
Figure 10. Endurance ratios for both unwelded and welded PA6GF specimens as a function of temperature. Error bars denote the scattering of the endurance ratio caused by the error in the static tensile strength measurement.
27
1.2 1
S/Su
0.8 0.6 0.4 0.2 0 1.E+0
24 °C
50 °C
70 °C
120 °C
24 °C
50 °C
70 °C
120 °C
1.E+2
1.E+4 Cycles to Failure
1.E+6
Figure 11. Ratio fatigue strength (S) to their respective tensile strength (Su) for both unwelded (hollow symbols) and welded (filled symbols) specimens at all the tested temperatures as a function of fatigue life. The solid line represents the least squares fit to the data set. The dashed lines represent the scatter band. The difference between the fitted line and the scatter band is the ratio of the standard deviation of the fit to the local predicted value provided by Eqn (1).
28
2 1.8 1.6
Kf = Suw/Sw
1.4 1.2 1 0.8 0.6
24 °C
50 °C
0.4
70 °C
120 °C
0.2 0 1E+1
1E+2
1E+3
1E+4 1E+5 1E+6 Cycles to Failure
1E+7
1E+8
Figure 12 Fatigue notch factor (Kf) as a function of cycles to failure (N) for the welded specimens under different test temperatures.
29
Highlights • The tensile and fatigue strength of glass fiber reinforced nylon 6 were examined • The effects of temperature, from 24°C to 120°C, and vibration welding were studied • Increasing temperature causes a reduction of both tensile and fatigue strengths • For both welded and unwelded PA6GF, the endurance ratio is approximately 45% • The fatigue notch factor lies between 1.5 and 1.75
30