etd AT Indian Institute of Science >
Division of Mechanical Sciences >
Mechanical Engineering (mecheng) >
Please use this identifier to cite or link to this item:
|Title: ||Transfer Layer Formation And Friction In Extrusion Of Aluminum : An Experimental Study Using A High Temperature Vacuum Based Pin-On-Disc Machine|
|Authors: ||Ranganatha, S|
|Advisors: ||Kailas, Satish V|
|Submitted Date: ||Apr-2008|
|Series/Report no.: ||G22452|
|Abstract: ||Hot extrusion of aluminum is widely practiced in industries for economic production of structural components. The surface finish and tolerance of the extruded components, both from design and aesthetic requirement, are important parameters. Hot extrusion involves forcing of aluminum in the form of a billet at a predetermined temperature through a shaped opening called die. Attempts, over time, are made to evolve the die profile to produce quality components. The main geometric feature of the die enables it in imparting plastic deformation and subsequent surface generation of the extrudate. The surface of extrudate is generated on the portion of the die called bearing channel or die land. Aluminum metal which moves relative to bearing channel experiences a different state-of-stress as it passes through the bearing channel. At the entry side of the bearing channel, the stress is compressive which is large in magnitude and this magnitude of compressive stress gets diminished as metal moves towards exit side and eventually becomes zero at the exit. Temperature gradients and its distribution along the bearing channel, similar to stress gradient, are reported. Literature reports formation of the transfer layer on the bearing channel. The transfer layer is of two distinctive types, the one near entry side which virtually leads to contact between aluminum and die steel and the other nearer to exit side which isolates aluminum from die steel. The understanding of the mechanism of formation of transfer layer is most important since it is instrumental in determining the surface finish of the extruded component.
All of the previous studies were conducted either in an actual extrusion press or using an instrumented extrusion press in the laboratory. The variables during these experiments could be the temperature of billet, extrusion ratio and speed of ram. Conducting these experiments using extrusion press is expensive and time consuming. To do experiments where the condition in a bearing channel can be simulated would be useful in conducting a more comprehensive study. To simulate the condition in the bearing channel a high temperature vacuum based pin-on-disc machine is designed and built. The fact that in a bearing channel there is virtually metal to metal contact with minimal or no chance of any intervening oxide layer, necessitated studies to be conducted under vacuum. A pin-on-disc test conducted under vacuum and high temperature would almost simulate condition on a bearing channel. Using this specially designed and built experimental setup the parameters influencing the transfer layer formation, temperature, sliding speed, load on the specimen, vacuum level and surface characteristics of the die are studied. Another feature of the test rig is that the configuration of pin and disc setup is vertical, which is different from regular pin on disc tribo-system, where pin and disc are held in horizontal configuration. The advantage of holding in vertical configuration is to eliminate the possibility of trapping the debris which alters the existing friction force and conditions.
The test rig is constructed using a cylindrical shell and a matching dome shell. The test rig, to facilitate the isolation of the instrumentations used to measure the forces and wear from the heat source is partitioned into two portions. The cylindrical bottom compartment called test chamber and dome shaped top compartment called sensor chambers are separated by a plate. On the plate in the sensor chamber, a load-cell to pickup friction force, a linear variable differential transformer (LVDT) to measure linear wear and loading lever mechanisms for imparting normal load and measuring friction force are fitted. The lever mechanism, in particular the one which magnifies the normal load is designed to conserve the space in the vacuum chamber. Housing the instrumentation inside the vacuum chamber thereby reducing the number of ports required to sense mechanical signals, increases the efficiency of the pumping system. The cylindrical shell of the testing chamber is a double walled structure and water cooled in order to prevent the exposer of sensors to higher testing temperature. Rubber ‘O’ rings are used, wherever it is required to seal the vacuum.
The necessary temperature required at the contact interface in the testing chamber is obtained by an electric resistance furnace, which is configured in such a way in the chamber that the heat generated is completely directed to the area where pin and disc are positioned. The interface temperature is monitored using a chromal-alumal thermocouple which is fixed very close to sliding interface on the pin holder. The power input to the furnace is controlled using PID (proportional integral derivative).
The required sliding speed is achieved with the help of direct current (DC) servomotor. The shaft on which the disc or ring is fixed is connected to a timer-pulley. The timer-pulley in turn is connected to servomotor through a timer-belt. The diameter ratio of the driving pulley and driven -timer-pulley is selected in such a way that the rpm of driven-pulley is reduced by four times and the torque increased by four times resulting in a more stable mechanical input to the sliding pair. The necessary high vacuum level in the test chamber is created by using rotary pump and diffusion pump combination.
Following tests are carried out.
1. Compression test: The strain rate response of aluminum (6060) under compressive state of loading is studied at strain rates 10-3s-1, 10-2s-1, 10-1s-1, 1s-1, 10s-1 and 102sand temperature ranging from 573 to 823K. The compression specimen is machined out of homogenized aluminum alloy (6060) cast ingots. True stress and true strain are estimated from load-displacement data of compression test. The true stress and true strain data are made use of in predicting the friction coefficient and sliding mode during sliding of aluminum relative to die steel at various temperature and other independent variables in vacuum.
2. The tensile test: A series of tensile tests at different temperatures and 10-1s-1 strain rate are conducted. The temperatures employed are from 423K to 723K. True stress, true strain, ultimate tensile strength and total percentage of elongation are estimated using load displacement data. The estimated ultimate tensile strength and total percentage of elongation are used for qualitatively explaining the morphology of transfer layer formed in the sliding experiments under different independent variables like temperature, speed, normal load, and surface texture of steel surface.
3. Adhesion test: The interaction at different temperature between the die steel (H11) and aluminum (6060) pair under static load is studied by conducting test in vacuum. The pin is made of homogenized aluminum and disc is made of die steel whose surface is generated by polishing on diamond paste until the average surface roughness (Sa) is of the order of 0.1 microns. The test temperatures are varied from 423 to 723K. The result is used in qualitatively explaining the morphology of transfer layer formed during sliding of aluminum and die steel pair in vacuum at various temperature, speeds and, normal load.
4. Sliding experiment: Steel pin and aluminum disc Pin on disc experiments are conducted at different temperatures in vacuum of the order of 4X10-4Pa. The experiments are conducted employing factorial design. The temperature, speed and load are the experimental parameters. The pin and disc are respectively made out of die steel and aluminum. Experiments are carried out with normal loads 20N, 40N, 63N, 80N, and 100N and speeds 0.1ms-1, 0.3ms-1, 0.5ms-1, 1.0ms-1, and 1.5ms-1 and with temperature over a range from 423 to 773K. The sliding distance covered is 500 meters. The friction force during sliding is monitored and used for estimating friction coefficient. Scanning electron microscopic study is carried out on surface, subsurface, wear track. The results, specifically, the friction coefficient and morphology of transfer layer are used to evaluate the influence of independent parameters on transfer mechanisms. The data generated from subsurface study and compression tests are used for estimating friction coefficient using Rigney’s plastic deformation model at 0.1ms-1.
5. Sliding experiment: aluminum pin and Steel disc The experiments reported in the previous paragraph where aluminum disc is slid against die steel pin had developed only the stresses of the order 16MPa. The estimated magnitude of stress level on bearing surface of extrusion die by various methods including numerical analysis is found to be more than 16MPa. In order to achieve a higher magnitude of stress and preserve the transfer layer mechanism, sliding an aluminum pin over the part of the textured surface of die steel ring is carried out. The changed boundary condition resulted in a nominal stress of 28 MPa which is comparatively more at 723K. The experiments are conducted employing factorial design. The other advantage of doing these experiments is that the effect of texture on the die surface can be studied by sliding aluminum pin over various surfaces. Thus temperature and surface texture are the experimental parameters for the present test. Different textured die steel surface is generated by machining process like, milling, electro discharge machining (EDM), wire electro discharge machining, silicon carbide slurry polishing, silicon carbide wheel grinding, CNC-milling and diamond paste polishing. Thus surfaces are basically of two types 1) with a lay on the surface and 2) a random surface. The tests are conducted at ambient temperature, 423, 573, 673, and 723K with a normal load of 56N and speed of 0.129ms-1. The morphology of transfer layer on the die steel ring is studied in scanning electron microscope. The friction coefficient is estimated from monitored friction force. The average surface roughness (Sa), results of compression test, and transfer layer are made use of to identify the sliding mode. The sliding of aluminum pin on diamond polished surface showed interesting results. Hence, another series of experiments using only diamond polished surface are conducted. The surface roughness of the steel surface achieved is 0.05 micro-meters. The load is 47N and speed is 0.043ms-1. The morphology of transfer layer and pin surface is studied using a Scanning Electron Microscope.
Results of compression and tensile test:
The flow stress is found to decrease with increase in compression test temperature. The SEM micrograph indicates large amount of fragmentation of harder phase with increase in strain rate at all temperature except for 423K. The fractured surface under tensile loading shows both intergranular and transgranular failure.
Results of adhesion test:
The area covered by the material transfer is found to increase with increase in temperature. The test can be used, to study the adhesive tendency between two pair of contacting surfaces. The area covered by the material transfer is found to be maximum at 723K.
Results of sliding of die-steel pin on aluminum disc in vacuum:
1. The ANOVA (analysis of variance) results indicate the existence of transition speed of about 0.5ms-1, more than which the friction coefficient was found constant. The extrusion speeds employed in industries are in the range of 0.1ms-1 to 1.7ms-1 and the transition speed found in the present study is within this range.
2. The magnitude of friction, with a few exceptions, is found to be independent of temperature and sliding speed when the sliding speed is 0.5ms-1and above. The invariance of friction coefficient with temperature and sliding speed beyond 0.5ms-1 is beneficial in that it will not lead to any instability like stick-slip or squeal.
3. Though both ambient temperature and speed influence the morphology of transfer layer and friction coefficient, speed is found to be dominant according to statistical analysis.
4. The observed dependency of friction coefficient and morphology of transfer layer on test temperature and normal load is attributed to decrease in flow stress and increase in friction factor ‘m’, a ratio of interfacial shear stress to shear yield stress of the softer material. Though ANOVA shows the significance of speed and not that of temperature, the observed dependence of friction coefficient on temperature is attributed to the enhanced effect of adhesion at elevated temperature observed in the adhesion test.
5. The state of stress at the contacting surface is found to control the morphology of transfer layer. When the normal load exceeded 40N, it gave rise to higher magnitude of stress state at the contacting surface, resulting in formation of continuous transfer layer and hence the higher magnitude of friction coefficient.
6. Plastic deformation model based on Rigney’s approach for estimating friction coefficient can be used. The estimated friction coefficient is on the higher side. Hence, any design of equipments based on the Rigney’s plastic deformation model is a conservative design.
7. There appears to be a close relation between the morphology of wear track and quantity of wear-loss. The formation of debris and rough track, primarily for low temperature sliding indicate larger magnitude of wear-loss. The parameters, which bring about increased ductility of aluminum as observed in the case of higher temperature of sliding, results in reduced wear-loss due to large scale smearing and back transfer of material.
The results of sliding of aluminum pin on die-steel ring in vacuum:
1. The mode of sliding changes from adhesive to abrasive mode depending on depth of penetration, tan(θ) where θ is the base angle of the conical asperity and average roughness parameter Sa ,all of them in turn depend on morphology of die steel surface and test temperature.
2. The friction coefficient and morphology of transfer layer are found to depend on the mode of sliding. The sub-surface plastic deformation, which characterizes the friction coefficient and morphology of transfer layer, is dependent on temperature of sliding.
3. The sliding experiment is capable of simulating the stress state on the bearing channel of the die; elucidating evolution of transfer layer, with change in operating parameters.
4. The ANOVA has clearly indicated the significance in friction coefficient at different temperatures and surface textures. In addition, the complex comparison below and above homogenization temperature (573K) has indicated significance in friction coefficient and thereby recognizing the importance of extrusion of aluminum at a temperature where it is in a single phase. Also, ANOVA indicates the dependency of both friction coefficient and transfer layer on the texture, i.e. either a lay or random.
The results of sliding aluminum pin on diamond polished die steel ring:
1. Shearing of the cold welded junction is a probable mechanism involved in the formation of transfer layer up to 423K.
2. The transgranular and intergranular mode of fracture are identified to be the two possible modes of fracture of the asperity at temperatures greater than 573K. The large ductility of the aluminum alloy facilitated smearing resulting in a continuous transfer layer at temperatures greater than 573K.
3. The formation of a continuous transfer layer at temperatures greater than 573K is responsible for the observed high friction coefficient at these temperatures. Scanning electron microscopy observations of the fracture surfaces of the tensile test specimen revealed fracture to be a combination of both transgranular and intergranular modes.|
|Appears in Collections:||Mechanical Engineering (mecheng)|
Items in etd@IISc are protected by copyright, with all rights reserved, unless otherwise indicated.