Power consumption and friction coefficient in the ultrasonic consolidation of aluminium alloys

Abstract

Energy input and friction behaviour are two of the key phenomena related with the welding bond of ultrasonic consolidation (UC) process. In this study, the effects of welding parameters, such as the building height, travel speed, oscillation amplitude and applied force, on the power consumption of the horn and friction coefficient of the welding interface of UC are investigated for aluminium alloys 2024 and 7075. Multiple regression equations are developed to estimate the power consumption and friction coefficient using the welding parameters. A line-contact friction test is performed to examine the friction coefficient of UC under the various applied forces and preheating temperatures. The quantitative relationships among the welding parameters, friction coefficient and power consumption are derived.

Keywords

Ultrasonic consolidation Welding Power Friction coefficient Parameters

Introduction

Ultrasonic consolidation (UC) process is a rapid additive solid-state manufacturing process, which fabricates complex 3D structure of metal and composite by employing ultrasonic welding and computer numerical control milling sub-processes. For UC working, the ultrasonic welding sub-process (Fig. 1) is conducted first. A horn carrying delivered foil gets contact with the work specimen (foil or baseplate) under applied force, and then begins to oscillate, rotate and travel. These motions cause friction, input energy, and eventually form the bond at the welding interface. Following the welding sub-process, a milling sub-process is employed to cut the needless parts. The welding and milling sub-processes are repeated alternatively to add the foils layer by layer until the final shape of the product is completed. Besides the traditional applications in similar material manufacturing, UC can also be used in the bond of dissimilar materials and forming of fibre-reinforced metal matrix composite.^1–7

Although the effects of welding parameters on the bond quality of UC have been widely studied,^8–15 some issues are still not fully understood. There are two major challenges on the bond quality control of UC. First, the bond mechanism is not clearly defined. Kong et al.⁸ concluded that the bond formation is due to atom diffusion and surface oxide layer dispersion through the mechanical and microstructure tests. Dehoff and Babu¹² proposed that the bond formation is contributed by localised plastic deformation and dynamic recrystallisation based on microstructure analysis. Yang et al.¹³ showed that the bond is related to the transmitted energy to the welding interface according to the results of linear welding density (LWD) – the percentage of bonded length to total length in the oscillation direction. Second, the suitable operating conditions are difficult to determine due to the synergetic effects of the welding parameters on the bond formation and various evaluation criterions of the bond quality. Kong et al.⁸ found that the peel load and LWD of aluminium alloy (AA)-3003 usually increase with the raise of the oscillation amplitude. However, the peel load and LWD may decrease with the raise of the oscillation amplitude under the high applied force condition. Ram et al.¹⁰ revealed that the samples made by the high applied force or oscillation amplitude have larger LWD for AA-3003. Whereas, Yang et al.¹⁴ suggested that the push-out load of the SiC fibre-reinforced aluminium matrix composite keeps the same or decreases with the increase of the applied force.

Owing to the unresolved bond mechanism and complicated relationship between the input (welding parameters) and output (bond quality) in UC, power consumption (energy input) has been introduced by some researchers as a middle variable.^13,15 Yang et al.¹³ asserted that the LWD can be evaluated by the level of energy input to the welding interface. Kelly et al.¹⁵ identified a linear relationship between the energy input and peel strength of the UC bond. Although some works have been performed to link the energy input and welding quality, experimental measured power consumption under various welding parameters is rarely reported.

During our investigation of the power consumption of UC, it is found that the friction coefficient of the welding interface is strongly influenced by the welding parameters and this is in agreement with the previous works.^16–18 Although the effects of operating conditions on the friction coefficient have been studied,^16–18 more investigation is still needed. The test conditions of the friction coefficient from the works of Naidu and Raman,¹⁶ and Zhang and Li¹⁷ are not similar to the condition of UC process. Koellhoffer et al.¹⁸ calculated the friction coefficient based on the experimental temperature results of UC process using a thermal friction model. However, the friction energy in their work was assumed to be 100% converted to heat, which might lead to overvaluation of the friction coefficient.¹⁹

In this study, a UC set experiment is designed to investigate the power consumption and friction coefficient under the various oscillation amplitudes, applied forces, travel speeds and building heights with AA-2024 and AA-7075 which are widely used in the aerospace industry. To verify the friction coefficient results obtained from the UC set experiment, a line-contact (cylindrical plane contact²⁰) friction test using SRV-4 tester under different applied forces and preheating temperatures is performed.

Materials and methods

The materials used in the present study are AA-2024, AA-7075 and Steel-350M. The related physical properties of these materials are given in Table 1.^21,22

Table 1

The related properties of the materials used. Note: E, μ and G are the Young's modulus, Poisson's ratio and shear modulus, respectively

Material	E/GPa	μ/1	G/GPa
Steel-350-M	200	0.30	77.0
AA-2024-O	73.1	0.33	28.0
AA-2024-T351	73.1	0.33	28.0
AA-7075-O	71.7	0.33	26.9
AA-7075-T651	71.7	0.33	26.9

To investigate the power consumption of the horn and friction coefficient of the welding interface of UC under the various oscillation amplitudes, applied forces, travel speeds and building heights, a UC set experiment (Fig. 1) is conducted using the Sonic Layer TM 7200 Production with a power consumption data picking frequency of 20 Hz. Baseplate with a height of 10.20 mm is mounted on an anvil via bolt fastening and heated by the heating elements. The horn is fabricated from Steel-350M with a texted surface. The power consumption of the horn in a no-load and no-rotation condition is tested to be 120 W. The single layer height, total building height and width of the foil are 0.15 mm, 3.00 mm (20 layers) and 25.40 mm, respectively. The welding length and oscillation frequency are 12 inches and 20 kHz, respectively. In the AA-2024 experiments, the baseplate and foil are made of AA-2024-T351 and AA-2024-O, respectively. The diameter, height of the horn and preheating temperature are 98.45 mm, 27.18 mm and 388.7 K, respectively. In the AA-7075 experiments, the baseplate and foil are made of AA-7075-T651 and AA7075-O, respectively. The diameter, height of the horn and preheating temperature are 97.36 mm, 27.18 mm and 377.6 K, respectively.

The power consumption under different operating conditions is initially examined in the test marked with Exp-A by conducting one test for each case. To minimise the effects of the horn contacting and leaving steps, the average power consumption is calculated from the weld mid-path. For the AA-2024 test, the ranges for oscillation amplitude, applied force and travel speed are 21–40 μm, 4000–8000 N and 33.9–63.5 mm s⁻¹, respectively. For the AA-7075 test, the ranges for oscillation amplitude, applied force and travel speed are 21–39 μm, 4000–10 000 N and 33.9–101.6 mm s⁻¹, respectively. Note that the amplitude in this work means the peak-to-peak amplitude. For the purpose of further investigating the relationships among the power consumption, friction coefficient and welding parameters, some cases in the Exp-A are reselected for multiple tests and regression analysis in the following experiment marked with Exp-B. Based on our trials for the experiment and regression (using cases as fewer as possible but ensuring accuracy of the regression equation), the cases covered different ranges of the average power consumption results from the Exp-A are selected and examined in the Exp-B as given in Table 2.

Table 2

The settings of welding parameters in the Exp-B

AA-2024	F/N	v/mm s⁻¹	A/μm	AA-7075	F/N	v/mm s⁻¹	A/μm
01	4000	63.5	21	01	4000	101.6	21
02	5500	63.5	21	02	7000	101.6	21
03	8000	63.5	21	03	10 000	101.6	21
04	4000	63.5	30	04	4000	101.6	30
05	5500	63.5	30	05	7000	101.6	30
06	5500	33.9	30	06	7000	33.9	30
07	8000	63.5	30	07	10 000	101.6	30
08	4000	33.9	40	08	4000	33.9	39
09	5500	63.5	40	09	7000	101.6	39
10	8000	33.9	40	10	10 000	33.9	39

Note: F, v and A are the applied force, travel speed and oscillation amplitude, respectively.

The schematic illustration of the power consumption and welding parameters is shown in Supplementary Figure S1. The total statistics of the power consumption and calculated friction coefficient are given in Supplementary Tables S1–S4.

By employing the following assumptions, the equations of the power consumption and friction coefficient are written in equations (1)–(3). Assumptions: (1) During welding, the welding parameters, such as the preheating temperature, building height, applied force, travel speed, oscillation amplitude and oscillation frequency are constant. (2) The power consumption consists of friction energy¹⁹ (major), and other losses (minor) which are equal to the no-load power (equation (1)). (3) Friction coefficient has many influential factors such as interface contact pressure, deformation of the contacted parts and surface condition of contacted surfaces¹⁹. The effects of the applied force, oscillation amplitude, travel speed and building height on these factors are characterised by their influences on the friction coefficient. (4) To build a semi-empirical relationship between the friction coefficient and welding parameters, the friction coefficient is assumed to be a power function of the non-dimensionalised applied force, oscillation amplitude, travel speed and building height (equation (2)). (5) The contact length of UC is calculated as a cylindrical plane contact derived from Barber's study,²⁰ and this calculation has been used in the analytic model of UC proposed by Yang et al.¹³ (equation (3)).

To develop relationships among the welding parameters, friction coefficient and power consumption, a multiple regression analysis of the power consumption results from the cases in Table 2 based on equations (1)–(3) is conducted, and the values of the friction constants in equation (2) are derived. (1) (2) (3) where E_power, E_fric, E_other and E_air are the power consumption of the horn (W), friction energy (W), other energy (W), and power consumption of the horn in no-load and no-rotation condition (W), respectively; F_fric, μ_slide and S_rel are the friction force (N), sliding friction coefficient (1) and relative sliding distance (m) of the welding interface, respectively; F_normal-UC, A_osc, f_osc, v_travel, R_sono, E_sono and μ_sono are the applied force (N), oscillation amplitude (m), oscillation frequency (Hz), travel speed (m s⁻¹), radius (m), Young's modulus (Pa) and Poisson's ratio (1) of the horn, respectively; W_foil, h_foil, H_foil, G_foil, E_foil and μ_foil are the width (m), single layer height (m), building height (m), shear modulus (Pa), Young's modulus (Pa) and Poisson's ratio (1) of the foil, respectively; L_line-UC is the line-contact length between the horn and foil in the UC set experiment; C_μ, C_A, C_F, C_v and C_H are the friction constants (1), and their values are 10, 1.92, −1.74, 0.02 and −0.05, respectively, obtained from the multiple regression analysis.

To evaluate the friction coefficient of the welding interface of UC through a directly experimental method, a line-contact (cylindrical plane contact²⁰) coefficient test (Fig. 2) under the different applied forces and preheating temperatures is performed using SRV-4 high temperature friction and wear tester. The diameter-height sets of the upper specimen and lower specimen are 15–22 mm and 24–7.88 mm, respectively. Both the test specimens are made of AA-2024-T351. The oscillation amplitude, oscillation frequency and test time are 100 μm, 10 Hz and 900 s, respectively. The average contact pressures under different applied forces in the line-contact test (equation (5)^13,20) are the same with those from the power consumption test (equation (4)^13,20). The applied forces in the line-contact friction test are 770, 1060, 1540 and 1920 N, which are corresponded to the applied forces of 4000, 5500, 8000 and 10 000 N in the power consumption test of AA-2024, respectively. Before test, all the parts are preheated to a temperature varied from 299 to 473 K. During test, the upper specimen oscillates with constant amplitude and frequency under applied force, and the friction coefficient is monitored at the same time.

The schematic illustration of the monitored friction coefficient is shown in Supplementary Figure S2. The total statistics of the monitored friction coefficient is given in Supplementary Table S5. (4) (5) where P_ave-UC and A_line-UC are the average contact pressure (Pa) and line-contact area (m²) in the UC set experiments; P_ave-fric, F_normal-fric and A_line-fric are the average contact pressure (Pa), applied force (N) and line-contact area (m²) in the line-contact friction test, respectively; R_upper, H_upper, E_upper and μ_upper are the radius (m), height (m), Young's modulus (Pa) and Poisson's ratio (1) of the upper specimen, respectively; E_lower and μ_lower are the Young's modulus (Pa) and Poisson's ratio (1) of the lower specimen, respectively.

Results and discussions

The average power consumptions of all building heights under the different oscillation amplitudes, applied forces and travel speeds are shown in Fig. 3. For both AA-2024 and AA-7075, the power consumption increases with the increase of the oscillation amplitude surprisingly, while the travel speed and applied force have minor influences on the power consumption.

The average power consumptions under the different building heights of two cases in Table 2 are shown in Fig. 4. In this relationship, the power consumption decreases linearly with the increase of the building height. In addition, by picking the building height of 0.15 mm as the baseline case, a relative level is defined as the average power consumption of a certain building height divided by the average power consumption of the baseline case. By summarising the relative levels of all the cases in Table 2 for each alloy, the average relative levels of all the ten cases under different building heights for both alloys are given in Fig. 5. It can be concluded that the average relative level drops when the building height rises. Furthermore, the average relative levels at 2.85 mm for both the studied alloys (∼87%) indicate that the loss of power consumption cannot be neglected.

Remarkably, from the experimental results of the power consumption, it is found that except for the travel speed, the influences of the applied force, oscillation amplitude and building height on the power consumption cannot be explained based on equation (1) by using a constant friction coefficient.

The multiple regression analysis (section ‘Material and methods’) reveals the relationships between the welding parameters, friction coefficient and power consumption, as given in equations (1)–(3). The obtained adj. R² value (0.97, close to 1) evidences a high reliability of the multiple regression equations. A graphical illustration of these relationships is shown in Fig. 6. Besides, by taking a constant friction coefficient of 0.3 from the Yang et al.'s analytical work of UC,¹³ the power consumption is calculated based on equation (1) and plotted in Fig. 6. The results confirm a significant scattering of the power consumption calculated based on the constant friction coefficient. On the other hand, a good correlation between the multiple regression equations calculated power consumption and monitored power consumption is obtained.

By calculating the friction coefficient (equation (1)) based on the experimental results of the power consumption from the cases in Table 2 (the data of the travel speed and building height are averaged) and gathering the data from the line-contact friction test, the friction coefficient results are shown in Fig. 7. It is evident both the power-calculated and line-contact-tested friction coefficients decrease with the increase of the average contact pressure. In addition, under the same average contact pressure, the calculated friction coefficient with high oscillation amplitude (∼40 μm) from the power consumption test is close to the tested friction coefficient from the line-contact friction test. Moreover, the negative effect of applied force (contact pressure) and positive effect of oscillation amplitude on the friction coefficient (equation (2) and Fig. 7) are consisted with the reported results.^16,18

The correlation between the friction coefficient and welding parameters is discussed next. First, according to the results from the line-contact test (Fig. 7), the effect of the preheating temperature on the friction coefficient is shown to be limited, which is different from the previous study¹⁷ under a flat plane contact condition. And this difference may be due to the different contact conditions. Moreover, it can be concluded that the effect of the travel speed on the friction coefficient (equation (2), C_v = 0.02) is quite small. Since a lower travel speed may lead to a higher temperature of the welding interface,²³ this result is consistent with the experimental results that the influence of the preheating temperature on the friction coefficient is limited under current experiment conditions. However, temperature rising could possibly diminish the effect of the applied force and enhance the deformation of the foil, and consequently affect the friction coefficient. These effects of temperature on the friction coefficient are not observed under current settings and the reason is unclear. Besides, this result also indicates that the power consumption consists of mainly the friction in the oscillation direction due to its rapid relative moving and sliding friction type (rolling friction type in the travel direction).

In contrast, the friction coefficient increases with the increase of the oscillation amplitude (equation (2), C_A = 1.92) significantly. This is because that the surface cleaning process is enhanced under a high relative moving speed, and a more thorough cleaning process has a larger friction coefficient.^18,24 Similarly, the friction coefficient is strongly influenced by the applied force. Noted that the applied force is used in both equations (2) and (3), hence the half frictional constant (equation (2), 0.5C_F = −0.87) is the value representing the ‘true effect’ of the applied force on the friction coefficient. This value suggests a diminishing effect of the applied force on the friction coefficient. One possible explanation for this phenomenon, as suggested by researchers,^11,23 is the suppressed relative motion at the welding interface under high contact pressure. To verify this explanation, the relative motions (oscillation amplitudes) near the welding interface of an aluminium–aluminium UC under different setting of the oscillation amplitudes (15, 22.5, 30 μm) and applied forces (0, 1100 N) in no-rotation condition are tested using a laser vibrometer. The monitored data indicate that the average relative motion of the three oscillation amplitudes under 1100 N applied force (97%) reduces 3% compared with that of 0 N applied force (100%).

Furthermore, for the building height, even its influence on the friction coefficient (equation (2), C_H = −0.05) is not as significant as the influence of the applied force or oscillation amplitude, this parameter still plays an important role in the UC power consumption due to its big changing range. The reason for the change of the friction coefficient can be explained by the deformation of the foil and changed relative motion at the welding interface. As shown in Fig. 8, for a cuboid made of metal (built foil), an increase of the building height leads to a decrease of the shear stiffness.²⁵ The decreased shear stiffness results in an augment of the foil displacement and a loss of the relative motion at the welding interface. Finally, the loss of relative motion causes a lower friction coefficient.^18,24

Conclusions

In this study, the power consumption of aluminium alloys 2024 and 7075 is measured under the different oscillation amplitudes, applied forces and travel speeds in a UC set experiment. The friction coefficient under different operating conditions is calculated based on the experimental power consumption results. Multiple regression equations are developed to estimate the friction coefficient and power consumption using the welding parameters based on a multiple regression analysis of the experimental power consumption results. In addition, the friction coefficient under the different applied forces and preheating temperatures is examined by a line-contact friction test using SRV-4 tester.

For the studied alloys, the power consumption increases with the increase of the oscillation amplitude surprisingly, and decreases with the increase of the building height, while the travel speed and applied force have minor influences on the power consumption. The applied force (contact pressure) and oscillation amplitude are shown to be the dominant factors to the friction coefficient. The friction coefficient increases with the increase of the oscillation amplitude and the decrease of the applied force significantly. On the contrary, the travel speed and preheating temperature are found to have insignificant influence on the friction coefficient. Furthermore, considering the big changing range of the building height, the losses of friction coefficient caused by the raise of the building height cannot be neglected.

Footnotes

Supplementary data

Supplemental data for this article can be accessed at

Acknowledgements

The authors acknowledge the financial support from Boeing Co, through Tsinghua-Boeing Joint research program. Thanks for the suggestions of Dr. S. Mironov, Dr. H.T. Fujii and Dr. H. Kokawa, from Tohoku University; and Dr. Ming Fu, from Tsinghua University

References

Yang

, Ram

G. D. J.

and Stucker

B. E.

: ‘Bond formation and fiber embedment during ultrasonic consolidation’, J. Mater. Process. Technol., 2009, 209, (10), 4915–4924. doi: 10.1016/j.jmatprotec.2009.01.014

Siddiq

and Ghassemieh

: ‘Fibre embedding in aluminium alloy 3003 using ultrasonic consolidation process-thermo-mechanical analyses’, Int. J. Adv. Manuf. Technol., 2011, 54, (9–12), 997–1009. doi: 10.1007/s00170-010-3007-6

Zhao

Y. Y.

, Li

and Zhang

Y. S.

: ‘Effect of welding energy on interface zone of Al-Cu ultrasonic welded joint’, Sci. Technol. Weld. Join., 2013, 18, (4), 354–360. doi: 10.1179/1362171813Y.0000000114

Kumar

: ‘Development of functionally graded materials by ultrasonic consolidation’, CIRP. J. Manuf. Sci. Technol., 2010, 3, (1), 85–87. doi: 10.1016/j.cirpj.2010.07.006

Robson

, Panteli

and Prangnell

P. B.

: ‘Modelling intermetallic phase formation in dissimilar metal ultrasonic welding of aluminium and magnesium alloys’, Sci. Technol. Weld. Join., 2012, 17, (6), 447–453.

Satpathy

M. P.

, Sahoo

S. K.

and Datta

: ‘Acoustic horn design and effects of process parameters on properties of dissimilar ultrasonic welding aluminum to brass’, Mater. Manuf. Process., 2016, 31, (3), 283–290. doi: 10.1080/10426914.2015.1048465

Seo

J. S.

, Jang

H. S.

and Park

D. S.

: ‘Ultrasonic welding of Ni and Cu sheets’, Mater. Manuf. Process., 2015, 30, (9), 1069–1073. doi: 10.1080/10426914.2014.880466

Kong

C. Y.

, Soar

R. C.

and Dickens

P. M.

: ‘An investigation of the control parameters for Aluminium 3003 under ultrasonic consolidation’, Proc 13th Solid Freeform Fabrication Symposium, Austin, TX, USA, August 2002, 199–211.

Kong

C. Y.

, Soar

R. C.

and Dickens

P. M.

: ‘Characterisation of aluminium alloy 6061 for the ultrasonic consolidation process’, Mater. Sci. Eng. A-Struct. Mater. Prop. Microstruct. Process., 2003, 363, (1–2), 99–106. doi: 10.1016/S0921-5093(03)00590-2

10.

Ram

G. D. J.

, Yang

and Stucker

B. E.

: ‘Effect of process parameters on bond formation during ultrasonic consolidation of aluminum alloy 3003’, J. Manuf. Syst., 2006, 25, (3), 221–238. doi: 10.1016/S0278-6125(07)80011-2

11.

Kulakov

and Rack

H. J.

: ‘Control of 3003-H18 aluminum ultrasonic consolidation’, J. Eng. Mater-T ASME, 2009, 131, (2), 021006(1)–021006(6). doi: 10.1115/1.3078303

12.

Dehoff

R. R.

and Babu

S. S.

: ‘Characterization of interfacial microstructures in 3003 aluminum alloy blocks fabricated by ultrasonic additive manufacturing’, Acta. Mater., 2010, 58, (13), 4305–4315. doi: 10.1016/j.actamat.2010.03.006

13.

Yang

Y. Z.

, Ram

G. D. J.

and Stucker

B. E.

: ‘An analytical energy model for metal foil deposition in ultrasonic consolidation’, Rapid. Prototyping. J., 2010, 16, (1), 20–28. doi: 10.1108/13552541011011668

14.

Yang

, Ram

G. D. J.

and Stucker

B. E.

: ‘An experimental determination of optimum processing parameters for Al/SiC metal matrix composites made using ultrasonic consolidation’, J. Eng. Mater-T ASME, 2007, 129, (4), 538–549. doi: 10.1115/1.2744431

15.

Kelly

G. S.

, Just

M. S.

, Advani

S. G.

and Gillespie

J. W.

: ‘Energy and bond strength development during ultrasonic consolidation’, J. Mater. Process. Technol., 2014, 214, (8), 1665–1672. doi: 10.1016/j.jmatprotec.2014.03.010

16.

Naidu

N. K. R.

and Raman:

S. G. S.

‘Effect of contact pressure on fretting fatigue behaviour of Al-Mg-Si alloy AA6061’, Int. J. Fatigue, 2005, 27, (3), 283–291. doi: 10.1016/j.ijfatigue.2004.07.001

17.

Zhang

and Li

: ‘A friction-based finite element analysis of ultrasonic consolidation’, Weld. J., 2008, 87, (7), 187s-194s.

18.

Koellhoffer

, Gillespie

J. W.

Jr , Advani

S. G.

and Bogetti

T. A.

: ‘Role of friction on the thermal development in ultrasonically consolidated aluminum foils and composites’, J. Mater. Process. Technol., 2011, 211, (11), 1864–1877. doi: 10.1016/j.jmatprotec.2011.06.011

19.

Wen

and Ping

: ‘Principles of tribology’, 2008, Beijing, Press of Tsinghua University.

20.

Barber

J. R.

: ‘Elasticity’, 2002, New York, Kluwer Academic Publishers.

21.

Davis:

J. R.

‘ASM specialty handbook – carbon and alloy steels’, 1996, Russell, ASM International.

22.

I. The Aluminum Association : ‘Aluminum standards and data’, 2000, Arlington, The Aluminum Association.

23.

Sriraman

M. R.

, Gonser

, Fujii

H. T.

, Babu

S. S.

and Bloss

: ‘Thermal transients during processing of materials by very high power ultrasonic additive manufacturing’, J. Mater. Process. Technol., 2011, 211, (10), 1650–1657. doi: 10.1016/j.jmatprotec.2011.05.003

24.

Blau

P. J.

: ‘Friction science and technology: from concepts to applications’, 2009, New York, CRC Press.

25.

Gere

J. M.

: ‘Mechanics of materials’, 2004, Belmont, Thomson Learning, Inc.