Read Investigations on forming of aluminum 5052 and 6061 sheet alloys at warm temperatures text version

Materials and Design 31 (2010) 2422­2434

Contents lists available at ScienceDirect

Materials and Design

journal homepage: www.elsevier.com/locate/matdes

Investigations on forming of aluminum 5052 and 6061 sheet alloys at warm temperatures

S. Mahabunphachai a,b, M. Koç a,*

a b

NSF I/UCRC Center for Precision Forming (CPF), Virginia Commonwealth University (VCU), Richmond, VA 23284, USA National Metal and Materials Technology Center, Pathumthani, Thailand

a r t i c l e

i n f o

a b s t r a c t

In an ongoing quest to realize low-mass transportation vehicles with enhanced fuel efficiency, deformation characteristics of Al5052 and Al6061 were investigated. In the first part of this study, material behavior of Al5052 and Al6061 sheet alloys were investigated under different process (temperature and strain rate) and loading (uniaxial vs. biaxial) conditions experimentally. With the biaxial, hydraulic bulge tests, flow stress curves up to 60­70% strain levels were obtained whereas it was limited to $30% strain levels in tensile tests. The microstructure analysis showed that the change of grain size due to the effects of elevated temperatures and strain rates were not significant; therefore, it was concluded that the decrease in the flow stress at high temperature levels was mainly due to the thermally activated dislocation lines. In the second part, the effect of the temperature and the pressure on the formability was further investigated in a set of closed-die warm hydroforming experiments. The test results showed that a linearly increasing pressure profile up to $20 MPa levels did not have a significant effect on the die filling ratios and thinning of the parts when a uniform temperature distribution of 300 °C was applied. Finally, in the third part of the study, finite element models were developed for the same closed-die hydroforming geometry using the material behavior models obtained from bulge and tensile tests. Flow stress curves obtained from tests were compared in terms of predicting the cavity filling ratios and thinning profiles from the experiments. Based on the comparison, it was revealed that flow stress curves obtained from e the warm hydraulic bulge tests provided accurate predictions at high strain levels (i.e., > 0:4, when part filling is above 80%) while the flow stress curves from the tensile tests did so at low strain levels (i.e., < 0:2, when cavity filling is below 80%). On the other hand, comparison of thinning values indicated e that flow stress curves from bulge tests yielded good agreement with the experimentally measured values in general. Therefore, it can be recommended that the bulge test results should be used whenever available in order to conduct accurate numerical analyses for warm sheet hydroforming where complex geometry and loading conditions exist. Ó 2009 Elsevier Ltd. All rights reserved.

Article history: Received 1 September 2009 Accepted 23 November 2009 Available online 26 November 2009 Keywords: Aluminum sheet Formability Lightweight material Hydroforming Warm forming AA5052 AA6061

1. Introduction With an increasing awareness and effects of global warming, and the scanty fossil fuel resources left when compared to the ever increasing demand of oil, car manufacturers have been seeking for alternative and sustaining solutions to the fuel efficiency problem. Many believe that the next generation cars must run on alternative and clean fuels (e.g., hydrogen via fuel cells) to prevent further increase of harmful emissions. However, this approach appears to be more of a solution that may not be practically and economically realized in short term (i.e., $10 years). On the other hand, another prominent approach that is sustaining, effective, and sooner would be the realization of low-mass vehicles. In the pursuit of latter

* Corresponding author. Tel.: +1 804 827 7029. E-mail address: [email protected] (M. Koç). 0261-3069/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.matdes.2009.11.053

solution, the car manufacturers along with various research groups have been investigating the fabrication of structural and body parts out of lightweight materials such as aluminum and magnesium alloys [1­3]. On the other hand, despite the obvious advantages of the lightweight alloys, they have a notable drawback in that their formability is significantly lower than traditional steel alloys at room temperature conditions, which is usually caused by the high alloy percentages that are required for high strength [4,5]. For example, the formability of aluminum alloys is only about twothird of a deep drawing steel grade, their Young's modulus is about one-third of the steel, which in turn causes higher susceptibility of wrinkling and springback [6], and their elongation is about half of steel's [7]. The inferior formability of aluminum alloys makes it more difficult and expensive to use them in mass production of structural and body parts, which requires high levels of elongation and ductility to be formed into complex shapes. Nevertheless, the

S. Mahabunphachai, M. Koç / Materials and Design 31 (2010) 2422­2434

2423

formability of the aluminum alloys has been shown to increase with an increasing forming temperature up to the recrystallization temperature, e.g., 300 °C for Al5xxx, and 200 °C for Al6xxx, where additional sliding planes are activated in the material [3­6,8­10]. Selective and localized heating strategies on the forming dies, causing an inhomogeneous temperature distribution on the blank, were also shown to further enhance the formability of the aluminum alloys [9­13]. In addition, high elongation could be obtained when low strain rate is used because these materials have intrinsically high strain rate sensitivity, especially at elevated temperature levels [7,14,15]. In addition to the forming at elevated temperature levels, alternative process technologies have been investigated to be used for complex and consolidated part manufacturing. The hydroforming process has been used for an increasing number of structural and body applications as it enlarges the forming limit windows of materials due to the biaxial and frictionless loading conditions, by which necking or thinning are delayed, and thus, elongation limits are extended [4,5,10,13]. The hybrid warm hydroforming process combines the advantages of both warm forming and hydroforming [12,16]. However, it is still considered as a relatively new and unknown technology waiting to be proven and validated. There are two vital aspects of this hybrid technology that demand further investigations: (1) understanding and characterization of the material behavior under warm hydroforming conditions, and (2) determination of the optimal process parameters (i.e., temperature level and distribution, pressure and blank holding profiles). In this study, our objectives were to: (1) determine proper experimental methodologies to accurately characterize the material behavior under warm hydroforming conditions (i.e., compare warm tensile and warm hydraulic bulge tests), (2) experimentally understand and quantify the effects of process parameters, such as temperature and internal pressure, on the part formability into representative die cavities with reasonably complex geometries, and (3) develop finite element models (FEM) to determine the applicability of bulge and tensile test findings to accurately predict the part formability. For this purpose, two commonly used aluminum alloys (5052 and 6061) were selected for experimentation. In the next section, experimental setup and conditions for the warm tensile and warm hydraulic bulge tests are presented. In the third section, material test results are presented and compared in terms of achievable strain and stress levels (i.e., flow stress curves). In the fourth section, experimental conditions and results of a design of experiment (DOE) study conducted using a set of closed-dies are presented and discussed to quantify the effect of process parameters on the cavity filling (i.e., formability) in warm hydroforming. Part profile, die filling ratio, and thinning on the warm hydroformed parts were reported and compared. A regression analysis was conducted to reveal the significance of pressure and temperature parameters. In the fifth section, a finite element model (FEM) of the warm hydroforming process is developed and validated by comparing the predictions with the experimental findings. The flow stress curves obtained from both the tensile and bulge tests are used in the FEA validation in order to determine which set of material test data is more accurate in predicting the part formability (i.e., cavity filling and thinning). Finally, a summary of the results and conclusions are presented in the sixth section.

CCD Cameras

ARAMIS

Temp. controller

Hydraulic Pump

Die Set Die Insert

CCD

Silicone based O-ring

Fig. 1. Warm hydraulic bulge test setup.

Laser Sensor

2. Material characterization experiments 2.1. Hydraulic bulge test setup For warm hydraulic bulge tests, a specially designed and built system composed of four major sub-systems was used as depicted in Fig. 1: (1) a pneumatic/hydraulic system: pump (Hydratron

AZ-2-180HPU-LW), pressure controller (Marsh Bellofram Type 3510), and pressure transducer (OMEGA PX605), (2) a set of bulging die: upper and lower die with a bulge diameter of 100 mm and each with a die corner radius of 6.5 mm, clamping and sealing mechanism (silicone based O-ring and copper O-ring), (3) heating system: cartridge heaters, temperature controller (OMEGA CN616tc1), and thermocouples (Type K), and (4) in-die non-contact measurement systems: laser sensor (Keyence LK-G402) and stereoscopic system (two CCD cameras with GOM ARAMIS system by Trilion). The non-contact measurement systems were used to avoid any temperature gradient at the contact location, which can influence the material behavior [4]. In order to avoid damage on the laser sensor and CCD cameras due to the splashing of the hot pressurized oil (Marlotherm SH), a thick glass was placed on the housing roof. The process parameters of interest in this test were the effect of temperature and strain rate on the flow stress behavior of the materials. The temperature of each die half was monitored and controlled independently using two separate sets of cartridge heaters and thermocouples (t/c) attached to each die half as shown in Fig. 2. With this type of control loop, the temperature variation during the test was below 5 °C. The heating cycle was made as short as possible, where the cycle time depends on the set temperature value. On the other hand, a holding time of 5­10 min was used to allow the uniform temperature distribution on the blank _ and the oil in the die cavity. The strain rate (SR or e) was also controlled using a feedback loop with a PID controller. Based on the difference between the pre-calculated dome height profile (reference value) and the instantaneous dome height value from the laser measurement, the control signals were sent to the pressure controller to regulate the air input pressure and flow rate at the pump inlet. The pressure and flow rate of the discharge fluid (oil) at the pump outlet were directly proportional to this controlled air flow at the inlet. The schematic of the control loops of the pneumatic/hydraulic system and the non-contact measurement system (laser sensor) is presented in Fig. 2.

2424

S. Mahabunphachai, M. Koç / Materials and Design 31 (2010) 2422­2434

Laser sensor CCD Cameras t/c Temp. Controller t/c Upper die Hot oil Lower die heaters

P

LabVIEW P Transducer

P controller

Pump

strain rate levels are plotted in Fig. 3. These profiles were used as a reference input signal in the feedback control loop. With this test setup, the bulging pressure and dome height could be continuously measured and recorded using the pressure transducer and the non-contact measurement systems (laser sensor and CCD cameras) during the test. Curvature of the bulging was also measured using the ARAMIS system with the CCD cameras. However, the dome height was found to be the same with the laser measurements as explained in detail in another study [18]. These pressure and dome height data were later synchronized together by the time stamp, and used for the flow curve determination. For each testing case, three specimens were tested. Overall, the variation among these three repeats was small. Hence, all the results reported in the next section are an average of these three repeats. 2.2. Tensile tests For the warm tensile tests, a 10-kN electromechanical MTS machine equipped with a furnace (max operating temperature of 315 °C) was used as illustrated in Fig. 4. A K-type thermocouple was placed in contact with the tensile specimen at the middle to measure the specimen temperature continuously. The cross-head speed, v, was calculated based on the target strain rate, SR, value (i.e., v = SR Ã l0, where l0 is the initial gauge length, which is around 50.8 mm). For each testing condition, three specimens were used. 2.3. Material preparation Two different aluminum alloys, Al5052-H32 and Al6061-T6, were tested in this study. Both had an initial thickness of 2.03 mm. The compositions of these alloys are presented in Table 1. Bulge specimens were prepared into a hexagonal shape by trimming four corners of 150 Â 150 mm square blanks, while the tensile specimens were prepared according to the ASTM standard E8-04. 3. Material testing results and discussion 3.1. Bulge test results Before using the measured and recorded pressure and dome height data for flow curve calculations, first, the accuracy of the measurement system and strain rate control were evaluated.

Fig. 2. Schematic diagram of the warm bulge test setup.

A pre-calculated dome height (hd) profile was used to obtain a constant strain rate during the tests. It was derived based on the geometrical relationships in a circular bulge testing of thin sheet blanks as follows [17]:

e ¼ ln

t0 td dc

2 2 2

ð1Þ !2 ð2Þ

td ¼ t0

dc þ 4hd

where e is the equivalent strain, t0 is the initial sheet thickness, td is the instantaneous apex thickness, dc is the bulge diameter, and hd is _ the instantaneous dome height. In addition, since strain rate ðeÞ is the rate of change in strain, one can write:

_ e¼eÁt

ð3Þ

where t is time. Combining Eq. (1)­(3), a relationship between the _ instantaneous dome height (hd) and the strain rate ðeÞ can be obtained as:

dc pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi _ hd ¼ eeÁt=2 À 1 2

ð4Þ

The relationship in Eq. (4) was used for plotting the reference hd profile as a function of time. Typical dome height profiles at various

60

MTS Machine

50

Dome height, hd (mm)

0.13 s-1

40

0.013 s-1

30 20

Specimen

0.0013 s-1

10

0 0 100 200 300 400

Furnace

time (s)

Fig. 3. Typical dome height profiles for different strain rate levels. Fig. 4. Warm tensile test setup.

S. Mahabunphachai, M. Koç / Materials and Design 31 (2010) 2422­2434

2425

40 35 Dome Height, hd (mm) 30 25 20 15 10

Al5052, t 0 = 2.03mm SR = 0.0013 s-1 Temp. = 100°C SR = 0.013 s-1 Temp. = 200°C

5 0 0

Pre-cal. Laser sensor CCD cameras

200 400 Time (sec.)

600

Fig. 5. Dome height value comparisons between the reference input (pre-cal), the laser sensor and the CCD cameras.

Measured values of dome height (hd) obtained from both the laser sensor and the CCD cameras are compared as shown in Fig. 5. The comparison indicates same measurements by both laser and CCD sensors; and hence, reliable to use in further calculations. In addition, the measured hd and the pre-calculated hd were shown to be almost identical as illustrated in Fig. 5; thus, a constant strain rate _ (SR or e) during each test could be expected. Some of the bulged samples are depicted in Fig. 6. The calculation of the flow stress was carried out based on the measured dome height (hd) and the bulging pressure (P) according to the following set of equations that were validated in another study [18]:

ða þ Rc Þ2 þ hd À 2Rc hd 2hd 2 sin a td ¼ t0 R¼

2

ð5Þ ð6Þ ð7Þ ð8Þ ð9Þ

a ¼ sinÀ1

a R

a

PR 2t d t e ¼ ln 0 td

Table 1 Typical compositions of commercial Al5052 and Al6061 alloys (www.matweb.com). wt.% AI5052 AI6061 Al 95.7­97.7 95.8­98.6 Cr 0.15­0.35 0.04­0.35 Cu Max 0.1 0.15­0.40 Fe Max 0.4 Max 0.7 Mg 2.2­2.8 0.8­1.2 Mn Max 0.1 Max 0.15 Si Max 0.25 0.4­0.8 Ti 0­0.05 Max 0.15 Zn Max 0.1 0.25 Other, each 0.05 0.05 Other, total 0.15 0.15

Room temp.

100°C

200°C

300°C

SR=0.0013 s-1 Al5052 SR=0.013 s-1

SR=0.0013 s-1 Al6061 SR=0.013 s-1 Room temp. 100°C 200°C 240°C

Fig. 6. Samples of bulged specimens.

2426

S. Mahabunphachai, M. Koç / Materials and Design 31 (2010) 2422­2434

(a)

True stress [MPa]

450

Al5052

400 350 300 250 200 150 100 50 0 0.0 0.1 0.2 0.3 0.4 0.5

Tensile-Room, 0.0013 [1/s] Tensile-100C, 0.0013 [1/s] Tensile-200C, 0.0013 [1/s] Tensile-300C, 0.0013 [1/s] Tensile-Room, 0.013 [1/s] Tensile-100C, 0.013 [1/s] Tensile-200C, 0.013 [1/s] Tensile-300C, 0.013 [1/s] Bulge-Room, 0.0013 [1/s] Bulge-100C, 0.0013 [1/s] Bulge-200C, 0.0013 [1/s] Bulge-300C, 0.0013 [1/s] Bulge-Room, 0.013 [1/s] Bulge-100C, 0.013 [1/s] Bulge-200C, 0.013 [1/s] Bulge-300C, 0.013 [1/s]

0.6

0.7

0.8

True strain

(b) 450

400 350

Al6061

True stress [MPa]

300 250 200 150 100 50 0 0.0 0.1 0.2 0.3 0.4 0.5

Tensile-Room, 0.0013 [1/s] Tensile-100C, 0.0013 [1/s] Tensile-200C, 0.0013 [1/s] Tensile-300C, 0.0013 [1/s] Tensile-Room, 0.013 [1/s] Tensile-100C, 0.013 [1/s] Tensile-200C, 0.013 [1/s] Tensile-300C, 0.013 [1/s] Bulge-Room, 0.0013 [1/s] Bulge-100C, 0.0013 [1/s] Bulge-200C, 0.0013 [1/s] Bulge-300C, 0.0013 [1/s] Bulge-Room, 0.013 [1/s] Bulge-100C, 0.013 [1/s] Bulge-200C, 0.013 [1/s] Bulge-300C, 0.013 [1/s]

0.6

0.7

0.8

True strain

Fig. 7. Comparison of flow stress curves for: (a) Al5052 and (b) Al6061 from both tensile and bulge tests.

where R is the curvature of the bulge radius, a is half bulge diameter (dc/2), Rc is the die corner radius, hd is the instantaneous height at the dome apex, td is the apex thickness, a is the angle that can be determined using Eq. (7), r is the equivalent flow stress, P is the instantaneous bulging pressure, e is the equivalent strain, and t0

is the initial thickness. The equivalent stress and strain were then combined to construct the material flow curves for different testing conditions as shown in Fig. 7. Since the assumption of the perfect spherical bulge shape, which is one of the key assumptions in deriving the Eqs. (5)­(9) for calculation of the apex thickness (td) and the dome height (hd), is far from being true at the beginning of the bulge test [19], only the measured values where hd/a > 0.2 (i.e., e ! 0:08) were used for the calculation of the flow curve in this study. Therefore, the initial flow stress value starts at around 0.08 strain as depicted in Fig. 7. Note that the flow curves shown in Fig. 7 represent the average values of the three samples tested at the same testing condition. The results in Fig. 7 reconfirm a general trend of the temperature and strain rate effects on the flow stress of aluminum alloys; in that the flow stress decreases with increasing temperature and/ or with decreasing strain rate; therefore, improving the formability. However, there is an inconsistency with this trend; that is in the case of bulging Al5052 at a low temperature level (i.e., below 100 °C), the flow stress was observed to decrease with increasing strain rate, a phenomenon that is usually caused by the solute drag and dynamic strain aging [8,15]. To elaborate deeper on the results in Fig. 7, the strain rate effect was observed to be more pronounced in the case of Al5052, especially at the elevated temperature levels, than on the Al6061. This low strain rate sensitivity in the 6xxx and 7xxx alloys has also been mentioned in the literature by Johansson et al. [20]; in their study they showed that the effect of the strain rate on the 6xxx and 7xxx alloys would not be observed before the strain rate exceeding 1000 sÀ1. In addition, with a higher percentage of Mg content in Al5052 than in Al6061, the ductility of Al5052 at elevated temperatures (i.e., 200 °C) was shown to be higher than that of Al6061, which was caused by the increasing number of the slip planes in the hexagonal structure of Mg at elevated temperatures. A similar observation on the effect of Mg content on the ductility of aluminum alloys at elevated temperatures was also reported in a tensile test study of four different aluminum alloys in [7]. Another interesting observation from the bulge test results comes from the case of bulging Al6061 at 300 °C at 0.013 sÀ1 strain

Mid-point Apex

25µm

Apex location

Base

elliptical shape

25µm

Mid-point location

25µm

elongated shape

Base location

granular shape [Images from Al5052-SR0.0013-100C#1 sample]

Fig. 8. Grain shape and distribution at different locations along the center line.

S. Mahabunphachai, M. Koç / Materials and Design 31 (2010) 2422­2434

2427

rate. All three specimens tested at this condition were fractured at the die corner region rather than at the dome apex. Thus, the flow curve in this case has a maximum strain of less than 0.2. Nonetheless, when Al6061 specimens were bulged at the lower strain rate (0.0013 sÀ1) at 300 °C, the maximum strain value is shown to be as high as 0.9. Therefore, at a low strain rate, the formability of Al6061 may continue to improve with increasing temperature even above 200 °C. 3.2. Tensile test results Due to the limitations of the electromechanical MTS system, a constant strain rate control during the tests was found to be rather difficult. Therefore, in this study a constant cross-head speed was used to provide initial strain rates of 0.0013 and 0.013 sÀ1, which were the same strain rates used in the bulge tests. The calculated true-stress­strain curves from the tensile tests are shown in Fig. 7. The flow curves are only presented here

up to the respective UTS point for each testing condition since the material data after this point is no longer meaningful for the eventual and further use in analyses. Similar effects of the temperature and strain rate on the flow stress curves were also observed; that is the flow stress decreases with increasing temperature and decreasing strain rate. All three specimens that were pulled under the same testing conditions provided almost identical flow curves, showing very reliable and repeatable results. In addition, the maximum elongation of Al5052 under the uniaxial loading condition was found to increase considerably at elevated temperature levels between 200 °C and 300 °C, however, such an increase was not observed in the case of Al6061 alloy after 200 °C. These observations agreed well with the reported results by Novotny and Geiger [5,6] and Li and Ghosh [8]. In their tensile test study, the formability of Al5xxx continuously increases with the temperature up to 300 °C, while that of the Al6xxx would increase up to 200 °C and the maximum elongation starts to decrease with further increase in temperature.

Al5052 ­ base location

Al6061 ­ base location

Room Temp.

25µm

25µm

100°C

25µm

25µm

200°C

25µm

25µm

240°C

N/A

25µm

Fig. 9. Temperature effect on grain structure at the base location.

2428

S. Mahabunphachai, M. Koç / Materials and Design 31 (2010) 2422­2434

35 30

Al5052 Al6061

23.4 21.4 21.8 23.6

Table 2 DOE of closed-die hydroforming. Runorder 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Stdorder 17 13 10 8 9 4 16 1 14 5 12 6 7 3 15 18 11 2 Temp. 1 À1 1 À1 À1 1 1 À1 À1 1 1 1 À1 À1 À1 1 1 À1 Pressure 0 À1 À1 0 1 À1 À1 À1 0 0 1 1 À1 1 1 1 0 0

Grain size [micrometer]

25 20 15 10 5 0

19.4

22.5

Room Temp.

100°C

200°C

Fig. 10. Effect of temperature on the material grain size.

As for the comparison of the hydraulic bulge and tensile test results, Fig. 7, flow stress curves from tensile test are limited to lower strain levels, particularly for Al6061 ($20%) when compared to flow stress curves from bulge test ($60%). However, with tensile tests, it was possible to obtain reliable flow stress values at low strain values (below 0.2), which was not possible in the bulge tests due to the limitations dictated by the spherical assumptions (i.e., low h/a ratio).

3.3. Microstructure analysis on bulged samples In order to better understand the effect of temperature and strain rate on the response of these Al alloys, a microstructure analysis was performed on the bulged samples. The grain structure

Al5052 ­ apex location

Al6061 ­ apex location

SR=0.0013 s-1

200°C

25µm

Room Temp.

25µm

SR=0.013 s-1

200°C

25µm

Room Temp.

25µm

Fig. 11. Strain rate effect on grain structure at the apex location.

Die insert Die insert

Fig. 12. Geometries of closed-die insert.

S. Mahabunphachai, M. Koç / Materials and Design 31 (2010) 2422­2434

2429

20

Pressure [MPa]

15

10

Al5052-300C-15MPa Al5052-300C-20MPa Al6061-200C-10MPa Al6061-200C-15MPa Reference Profile

5

0

0

20

40

60

80

100

Time [sec.]

Fig. 13. Hydroforming pressure profiles.

(i.e., size, shape, and distribution) was investigated as the grain structure was known to largely influence the overall material response. The bulged specimens were cut along the centerline and small sample strips were removed from the center region as shown in Fig. 8. These sample strips were polished and etched in Keller's reagent (2.5 ml HNO3, 1.5 ml HCl, 1 ml HF, and 95 ml water) to reveal the undeformed grain structure at the ``base location" and the deformed grain structure at the ``mid-point" and ``apex" regions as depicted in Fig. 8. Most of the grains were found to have granular structure at the base location (i.e., undeformed grains), while elongated and elliptical grain structures were observed at the midpoint and the apex locations, respectively. The grains at these regions were elongated or stretched as they underwent a large plastic deformation amount during the bugle tests.

The effect of temperature on the grain structure was investigated at the base location of the cut strips. The base location was selected because the grains in this region did not undergo any deformation (strain and strain rate independent); thus, representing the microstructure changes caused merely by the temperature effect. The microscopic images of the grain structures are illustrated in Fig. 9 for both Al5052 and Al6061. Note that the grain structures from the specimens that were bulged at 240 °C and higher could not be clearly seen, which may have been caused by the significant microstructural changes, most likely recrystallization and growth of grains or precipitates, as the bulging temperature enters into the ``warm forming" regime (i.e., temperature >0.3Tm, where Tm indicates melting point). For the samples that the grain boundary could be clearly indicated, the grain size was measured based on the ASTM Standard E112-88 (i.e., Mean Lineal Intercept or Heyn's method). The grain size measurement results are presented in Fig. 10 where slightly larger grains were observed at elevated temperature levels although the difference is statistically insignificant. According to the Hall­Petch relation [21,22], material with larger grain size was predicted to have lower strength. Despite the fact that such a case was observed in this study, it is believed that the drop in the flow stress curve is not due to the slightly and statistically indifferent grains, but mainly due to the additional slip lines activated due to the elevated temperature. The effect of the strain rate was also investigated through the grain structure analysis. Unlike in the temperature effect study, the location of interest on the cut specimen was shifted to be at the apex of the dome rather than at the base location as most deformation dependent characteristic could be seen most in this region. The microscopic images of the specimens bulged at different strain rates are shown in Fig. 11. Unfortunately, no significant difference was observed in terms of the grain structure between

(a)

200°C

Al5052

300°C

10MPa

15MPa

20MPa

Al6061

(b)

200°C

300°C

10MPa

15MPa

20MPa

Fig. 14. Closed-die hydroformed samples: (a) A5051 and (b) Al6061.

2430

S. Mahabunphachai, M. Koç / Materials and Design 31 (2010) 2422­2434

25

Al5052

20

height [mm]

15

200C-10MPa 200C-15MPa 200C-20MPa 300C-10MPa 300C-15MPa 300C-20MPa Die Profile

10

5

0

0

20

40

60

80

100

120

width [mm]

25

Al6061

20

height [mm]

Fig. 16. Response surface plots.

15

200C-10MPa

10

200C-15MPa 200C-20MPa 300C-10MPa

5

300C-20MPa Die Profile

0

0

20

40

60

80

100

120

width [mm]

Fig. 15. Profiles of hydroformed parted at different temperature and pressure levels.

the two strain rate values used in this study (0.0013 and 0.013 sÀ1). Nonetheless, the effect of the strain rates on the material response (i.e., flow curve) and the formability (i.e., maximum elongation) was obvious, especially at elevated temperature levels as discussed in the previous section. Hence, it can be concluded that it is not the change in grain size, but the thermally activated dislocation motion that causes the formability increase in warm forming.

4. Closed-die hydroforming experiments A set of closed-die warm hydroforming experiments were conducted on the same alloys using a die insert as shown in Fig. 12. These experiments were conducted under a design of experiment (DOE) plan, as tabulated in Table 2 (i.e., 18 runs for each alloy), to obtain a quantified understanding of the effect of temperature and pressure on the die cavity filling and thinning (forming limits). During the experiments, a constant blank holder force of 1000 kN was used, as in the bulge tests, to clamp the specimens at the periphery. A linearly increasing (ramp-up) pressure profile with a

slope of 0.22 MPa/s was used as a reference input (Fig. 13). The actual hydroforming pressure profiles, recorded during the tests using a pressure transducer, were shown to closely follow the reference pressure input profile (Fig. 13). After each test, the hydroformed parts were measured using the stereoscopic CCD cameras with ARAMIS software to obtain full surface profiles as illustrated in Figs. 14 and 15. For the assurance of process repeatability, three experiments were conducted for each case. An average value is reported unless otherwise is stated in this section. The effect of temperature and pressure levels on the sheet formability can clearly be seen in Figs. 14 and 15. Specifically, at 200 °C, both Al5052 and Al6061 sheet blanks showed poor formability, and a fracture occurred in the area of the die radius at the center when the pressure was increased from 10 to 15 MPa for Al5052 and from 15 to 20 MPa for Al6061. As the temperature was increased to 300 °C, the formability of Al5052 sheet blanks appeared to improve, and no premature rupture was observed up to 20 MPa. However, for Al6061, an increasing temperature only reduced the material strength (i.e., higher profile at the same forming pressure when the temperature was increased), while the elongation properties did not change. With the increasing temperature, all Al6061 specimens showed fractures at the central die radius area once a certain part height was reached. This observation agrees well with the flow curve plots of Al6061 (Fig. 7), in which the maximum strain (i.e., elongation) value did not change as temperature increased, but the flow stress values were observed to significantly decrease. Furthermore, based on the comparison of the part profiles in Fig. 15, it is clearly shown that Al5052 has a better formability when compared to Al6061 under these forming conditions (i.e., uniform temperature distribution at 200 and 300 °C and linear pressure profiles up to 15­20 MPa). Finally, it is important to point

Table 3 Percentage of die filling under different process conditions. 10 MPa (%) Al5052 Al6061 200 °C 300 °C 200 °C 300 °C 77.2 95.2 60.7 94.3 15 MPa (%) 94.2 95.6 80.8 n/a 20 MPa (%) 93.7 96.2 83.2 94.5

S. Mahabunphachai, M. Koç / Materials and Design 31 (2010) 2422­2434

2431

60

Al5052

50

% Thinning

40 30 20 10 0 0

200C-10MPa 300C-15MPa 300C-20MPa Die Profile

Table 4 Material parameters for Al5052 and Al6061 used in FEA. Material parameters Modulus of elasticity, E (MPa) Poisson's ratio, v Mass density, q (Mg/mm3) Yield strength, r0 (MPa) AI5052 70,300 0.33 2.68E À 09 89.6 AI6061 68,900 0.33 2.70E À 09 55.2

Table 5 _ ee Material constants ðr ¼ K n m Þ from tensile and bulge tests at different temperatures. Test Material AI5052 Temp. (°C) 23 200 300 23 100 200 23 100 200 300 23 100 200 300 K (MPa) 455 412 401 483 497 503 777 966 437 253 979 1058 880 474 n 0.14 0.33 0.55 0.11 0.17 0.12 0.45 0.50 0.28 0.09 0.38 0.41 0.36 0.21 m 0.010 0.075 0.135 0.013 0.020 0.075 0.000 0.027 0.051 0.151 À0.006 0.009 0.046 0.114

10

20

30

40

50

60

Bulge test

Radial distance [mm]

50

AI6061

Al6061

40

200C-10MPa 200C-15MPa Die Profile

Tensile test

AI5052

% Thinning

30

AI6061

20

10

0

0

10

20

30

40

50

60

Radial distance [mm]

Fig. 17. Thickness profiles in radial direction.

out that the die corner rupture was observed in the case of Al6061 blanks formed at 300 °C and 15 MPa. As a result, their profiles are excluded in Fig. 15 as well as in the rest of the analysis. In order to quantify the effects of the temperature and pressure on the formability of the hydroformed parts, two variables (the die

filling ratio and thinning) were measured and reported in Table 3 and in Figs. 16 and 17. Note that for thinning measurements, the hydroformed specimens were cut into two halves with an offset of 10 mm from the center line, and the thickness of the bigger half was measured using a micrometer attached with conical shape tips at several locations along the radial directions. In addition, for a meaningful comparison of the thinning, only the specimens without the fracture were used. When cavity filling comparisons in Table 3 and Fig. 16 are considered, at 200 °C, an increasing pressure leads to an increase in the cavity filling ratio for A5052 (77­93%) and Al6061 (60­83%).

Total Equivalent Plastic Strain

Fig. 18. Two-dimensional axisymmetric numerical modeling of the closed-die warm hydroforming: FE model, predicted and actual deformed part.

2432

S. Mahabunphachai, M. Koç / Materials and Design 31 (2010) 2422­2434

25

Al5052

20

15

200C-10MPa-Exp 200C-10MPa-FEA-Bulge

10

200C-10MPa-FEA-Tensile 300C-15MPa-Exp 300C-15MPa-FEA-Bulge 300C-15MPa-FEA-Tensile Die Profile

5

0 0 10 20 30 40 50 60

Width [mm]

25

Al6061

20

Height [mm].

15

(i.e., for Al5052: p = 0.238 for temperature and p = 0.246 for pressure; for Al6061: p = 0.151 for temperature and p = 0.350 for pressure). In terms of thinning comparisons as shown in Fig. 17, first of all, the highest thinning, and hence fracture in some cases, occurred around the die radius at the center region of the part. Second, the effect of temperature and pressure were as expected (i.e., increasing pressure and temperature leads to increasing thinning in general). Based on this result discussion, the use of a uniform temperature distribution (i.e., isothermal conditions) and a linearly increasing pressure profile (i.e., ramp-up pressure input) may not be the most efficient approach to increase the cavity filling and reduce thinning (i.e., part formability). Thus, process optimization investigation is needed to determine optimal process conditions (e.g., variable loading profiles: pressure and blank holder force, and temperature). This optimization study would require the use of finite element analysis (FEA) tool. In the following section, FEA models of the hydroforming process will be developed and the material properties obtained from the bulge and tensile tests will be validated for their accuracy and applicability in predicting the closed-die profiles and thinning values as measured and reported in this section. 5. Numerical modeling, validation and comparison of material behaviors In this section, Finite element models of the closed-die warm hydroforming process (Fig. 18) were developed using MSC.Marc2007r1 software. Since the problem at hand was an axisymmetric type, only a 2-D half-model analysis was developed. In the model, the sheet blank was modeled using deformable, solid, quad-4 elements. Four elements across the blank thickness were used. Both ends of the blank were fixed (no displacement) to reflect on the actual boundary condition of the process where the blank was tightly clamped between the upper and lower dies to prevent any radial flow of the material into the die cavity. Hydroforming pressure was applied from the bottom side of the blank with an increasing rate of 0.22 MPa/s until the pressure reached the preset values. According to the actual pressure profiles recorded during the experiments, the pressure pump and regulator provided close control of the pressure profile as shown in Fig. 13. Thus, a ramp pressure input

Height [mm].

10

5

200C-10MPa-Exp 200C-10MPa-FEA-Bulge 200C-10MPa-FEA-Tensile 200C-15MPa-Exp 200C-15MPa-FEA-Bulge 200C-15MPa-FEA-Tensile Die Profile

0 0 10 20 30 40 50 60

Width [mm]

Fig. 19. Comparison of hydroformed part profiles for A5052 and Al6061 alloys at different process conditions.

However, for 300 °C, the same cannot be said. A similar observation is made for an increasing temperature at low pressure value (10­15 MPa). However, in general, when a regression analysis was made for the entire ranges of temperature and pressure, their effect on the cavity filling ratio was found to be statistically insignificant

Table 6 Simulation cases. Exp. case Case 1 Case 2 Case 3 Case 4 FEA run 1 2 3 4 5 6 7 8 Material AI5052 AI5052 AI5052 AI5052 AI6061 AI6061 AI6061 AI6061 Temp. (°C) 200 200 300 300 200 200 200 200 Pressure (MPa) 10 10 15 15 10 10 15 15 Mat. flow curve Bulge Tensile Bulge Tensile Bulge Tensile Bulge Tensile

Table 7 Percentage of die filling comparisons. Material Temp. (°C) Pressure (MPa) Percentage of die filling Experiment Al5052 Al6061 200 300 200 200 10 15 10 15 77.2 95.6 60.7 80.8 FEA-bulge (%error) 86.1 (11.5) 98.2 (2.7) 67.6 (11.4) 83.8 (3.7) FEA-tensile (%error) 79.2 (2.6) 98.2 (2.7) 59.8 (1.5) 75.1 (7.1)

S. Mahabunphachai, M. Koç / Materials and Design 31 (2010) 2422­2434

2433

Fig. 20. Thickness measurement.

with a slope of 0.22 MPa/s was utilized in all validation cases. Coulomb friction model was selected with a friction coefficient of 0.1 for all contact surfaces in this study. Material input parameters: modulus of elasticity (E), poisson's ratio (v), mass density (q), and yield strength (r0) for both Al5052 and Al6061 are given in Table 4, while the material flow curves obtained from the tensile and bulge tests at different temperature levels were modeled using Field­Backofen equation (i.e., _ ee rate power law: r ¼ K n m ) as tabulated in Table 5. In order to validate the FE models, four experimental cases (Table 6) were selected which included the closed-die hydroforming results for both Al5052 and Al6061 at elevated temperatures (200­300 °C) and two different pressure levels (10 MPa and 15 MPa). The hydroformed parts in these cases were fracture-free,

which makes the thickness measurement physically possible. The other output chosen for comparison purpose was the part profiles (or percentage of die filling). In addition, with the two flow curves obtain from tensile and bulge tests at each temperature level, a total of eight simulation runs were carried out and their results are shown and compared in Fig. 19 and Table 7. According to the part profile comparisons between the experimental measurements and the FEA predictions in Fig. 19, it was found the tensile flow curves provided better profile predictions e at low pressure (i.e., low strain) levels (10 MPa, where < 0:2), while the bulge flow curves did so at the high pressure (i.e., high e strain) levels (>15 MPa, where > 0:4). The observation could be well explained by the limitations and assumptions associated with each test method and the derivations of the material flow curves based on the raw test data. Specifically, for tensile tests conducted at 200 °C, the maximum strain levels were found to be around 0.2 for Al5052, and 0.15 for Al6061. Therefore, FEA predictions for high strain levels would require extrapolation of the material flow curves. On the other hand, a higher strain levels could be achieved in the bulge tests, e.g., at 200 °C, the achievable strain was around 0.5 for Al5052 and about 0.35 for Al6061; and thus, the material data from the bulge test provided better FEA predictions at high strain or pressure levels. Furthermore, with the assumption of a non-spherical dome shape of the bulge specimen below the h/a value of 0.2 (corresponding to a strain value of 0.08), the material data below this threshold value was excluded for the flow curve determination in bulge testing case, which in turns made the FEA predictions based on the bulge flow curves less accurate at low

(a)

60 50 40 30 20 10 0 0 10 20 30 40 50 60

Al5052 200C-10MPa 300C-15MPa %Thinning FEA-Bulge FEA-Tensile FEA-Bulge FEA-Tensile Ave. 2.8 0.8 5.5 5.3 Max. 6.2 3.8 18.2 18.0

% Thinning

200C-10MPa-Exp 200C-10MPa-FEA-Bulge 200C-10MPa-FEA-Tensile 300C-15MPa-Exp 300C-15MPa-FEA-Bulge 300C-15MPa-FEA-Tensile Die Profile

Radial distance [mm]

(b)

60 50

Al6061 200C-10MPa 200C-15MPa %Thinning FEA-Bulge FEA-Tensile FEA-Bulge FEA-Tensile Ave. 0.1 2.5 1.4 6.8 Max. 6.1 4.6 4.1 12.3

% Thinning

40 30 20 10 0

200C-10MPa-Exp 200C-10MPa-FEA-Bulge 200C-10MPa-FEA-Tensile 200C-15MPa-Exp 200C-15MPa-FEA-Bulge 200C-15MPa-FEA-Tensile Die Profile

0

10

20

30

40

50

60

Radial distance [mm]

Fig. 21. Thinning comparisons in radial direction for: (a) Al5052 and (b) Al6061.

2434

S. Mahabunphachai, M. Koç / Materials and Design 31 (2010) 2422­2434

strain or pressure levels. Nonetheless, both bulge and tensile material flow curves were shown to provide reasonable predictions of the hydroformed part profiles as can be confirmed by the comparisons of the percentage die filling in Table 7 with the maximum errors of 1.5­11% based on both types of flow curves. The second comparison is based on the thinning levels and distributions. The hydroformed specimens were cut in half as shown in Fig. 20 along a line that was about 1 cm off from the centerline of the part. The measurement was then performed using a digital micrometer at several critical locations with numerous repetitions. Thinning distribution comparisons between the experimental measurements and the FEA predictions are shown in Fig. 21. The comparisons were found to be in a good agreement, except for the case of Al5052 at 300 °C and 15 MPa where the maximum thinning percentage was higher than 40% and the instability, which was not considered in the FEA, had occurred. In addition, for Al6061 specimens it was clear that the thickness values predicted based on the bulge flow curves were more accurate than those obtained from the tensile flow curves. This is due mainly to the loading condition of the bulge test (i.e., bi-axial stress) that is more relevant to the actual state of stress and strain in the hydroforming process. Therefore, both tests were shown to be appropriate for obtaining the material properties for the eventual use in FEA of warm hydroforming. However, for the parts with complex geometries or when large deformation is expected, numerical models based on bulge tests provided closer predictions. In summary, the FEA models of the hydroforming process developed in this section along with the material flow curves obtained from both bulge and tensile tests in the previous section have been shown to provide reasonable and reliable numerical predictions in terms of both part profiles and thinning distributions. 6. Conclusions In this study, mechanical characteristics of Al5052 and Al6061 sheet blanks were characterized using both tensile and bulge testing methods at temperature levels up to 300 °C, and at the strain rates of 0.0013 and 0.013 sÀ1. In addition to the expected and general trend of decreasing material flow curves with increasing temperature and/or decreasing strain rate, it was found that the flow stress curves from tensile test were limited to lower strain levels, particularly for Al6061 (<15%) when compared to the flow stress curves from the bulge tests ($30­60%). However, with tensile tests, it was possible to obtain reliable flow stress values at strain values lower than 2%, which was not possible in the bulge tests due to the limitations dictated by the spherical bulging assumptions (i.e., low h/a ratio). The microstructure analysis showed that the change of grain size was not significant at different temperature and strain rates, which leads us to conclude that the decrease in the flow stress at high temperature levels was mainly due to the thermally activated dislocation lines. The effects of the temperature and the pressure on the sheet formability were further investigated in a set of closed-die hydroforming experiments. At 200 °C, an increasing pressure leads to an increase in the cavity filling ratio for A5052 (77­93%) and Al6061 (60­83%). However, for 300 °C, the same cannot be stated. A similar observation is made for an increasing temperature at low pressure value (10 MPa), but not at high pressure level (20 MPa). The test results along with the regression analysis showed that the use of a uniform temperature distribution and a ramping pressure input do not have a significant effect on the percentage of die filling values. Therefore, process optimization investigation is needed to

determine the optimal process conditions (e.g., variable loading profiles: pressure and blank holder force, and temperature) for maximize the part formability. FE modeling findings and comparison with closed-die hydroforming experiments based on the material flow stress curves from both bulge and tensile tests at different temperature, pressure and strain rate conditions indicated that, in general, flow curves from both bulge and tensile tests are in good agreement with experimental measurements in terms of predicting the part profile, cavity filling and thinning comparisons. Overall prediction errors were less than 10­12%. However, when examined closely, it was revealed that flow curves from bulge tests resulted in better prediction accuracy, particularly at high strain (pressure) levels. Acknowledgements Authors are thankful to National Science Foundation (NSF) for the partial support on this project through NSF ENG/CMMI Grants 0703912 and NSF IIP IUCRC Grant 0638588. References

[1] Schultz RA. Aluminum for light vehicles ­ an objective look at the next ten to twenty years. In: 14th int'l aluminum conference, Montreal, Canada (Ducker Research); September 15 1999. [2] Carpenter JA. The FreedomCAR challenge and steel. American iron and steel institute, great designs in steel seminar, Livonia (MI); February 2004. p. 96­ 111. [3] Toros S, Ozturk F, Kacar I. Review of warm forming of aluminum­magnesium alloys. J Mater Process Technol 2008;207:1­12. [4] Novotny S, Celeghini M, Geiger M. Measurement of material properties of aluminium sheet alloys at elevated temperatures. In: Proceedings of the SheMet international conference, UCE, Birmingham; 2000. [5] Novotny S, Geiger M. Process design for hydroforming of lightweight metal sheets at elevated temperatures. J Mater Process Technol 2003;38:594­9. [6] Bolt PJ, Lamboo Napm, Rozier Pjcm. Feasibility of warm drawing of aluminium products. J Mater Process Technol 2001;115(1):118­21. Aug.. [7] Shehata FA. Tensile behaviour of aluminium/magnesium alloy sheets at elevated temperatures. Sheet Met Indus 1986;63(2):79­81. [8] Li D, Ghosh A. Tensile deformation behavior of aluminum alloys at warm forming temperatures. Mater Sci Eng 2003;A352:279­86. [9] Li D, Ghosh A. Biaxial warm forming behavior of aluminum sheet alloys. J Mater Process Technol 2004;145:281­93. [10] Neugebauer R, Altan T, Geiger M, Kleiner M, Stezing A. Sheet metal forming at elevated temperatures. Ann CIRP 2006;55(2):793­816. [11] Kim HS, Koç M, Ni J. Determination of proper temperature distribution for warm forming of aluminum sheet materials. J Manuf Sci Eng 2006;128(3):622­33. [12] Choi H, Koç M, Ni J. Determination of optimal loading profiles in warm hydroforming of lightweight materials. J Mater Process Technol 2007;190:230­42. [13] Choi H, Koç M, Ni J. A study on warm hydroforming of Al and Mg sheet materials: mechanism and proper temperature conditions. J Manuf Sci Eng 2008;130(4):0410071­04100714. [14] Flanigan AE, Tedsen LF, Dorn JE. Tensile properties affecting the formability of aluminum-alloy sheet at elevated temperatures. J Aeronaut Sci 1946;457­ 468(Sep.). [15] Yao X, Zajac S. The strain-rate dependence of flow stress and work-hardening rate in three Al­Mg alloys. Scand J Metall 2000;29:101­7. [16] Aginagalde A, Orus A, Esnaola JA, Torca I, Galdos L, Garcia C. Warm hydroforming of lightweight metal sheets. AIP Conf Proc 2007;908(1):1175­80 [May]. [17] Hill R. A theory of the plastic bulging of a metal diaphragm by lateral pressure. Philos Mag 1950;41(7):1113­42. [18] Billur E, Koç M. A comparative study on hydraulic bulge testing and analysis methods. In: Proceedings of the 2008 international manufacturing science and engineering conference (MSEC2008), Evanston, IL, USA; October 7­10 2008. [19] Dudderar D, Koch FB, Doerries EM. Measurement of the shapes of foil bulgetest samples. Exp Mech 1977;17(4):133­40. [20] Johansson M, Hornqvist M, Karlsson B. Influence of temperature and strain rate on the plastic deformation of two commercial high strength Al alloys. Mater Sci Forum 2006;519­521:841­6. [21] Hall EO. Deformation and ageing of mild steel. Phys Soc ­ Proc 1951;64:747­53. [22] Petch NJ. Cleavage strength of polycrystals. Iron Steel Inst 1953;174:25­8.

Information

Investigations on forming of aluminum 5052 and 6061 sheet alloys at warm temperatures

13 pages

Find more like this

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

82845


You might also be interested in

BETA
JFE()No.4-05G
Investigations on forming of aluminum 5052 and 6061 sheet alloys at warm temperatures