The constant velocity joint outer race is one of the main load supporting parts in cars. Its geometry is very complicated and its required precision is high. Mohammadi and Sadeghi , mentioned that the forging process to form an outer race is a combined hot and cold forming which consists of tow performing sequence at elevated temperature, cooling and final ironing at room temperature. The process was simulated with numerical simulation tool and then physical modeling with commercial lead was performed in order to verify the reliability of numerical simulation results. Experiments with model material showed the occurrence of ductile fracture and surface defects in final ironing sequence. In order to modify tooling design, an attempt has been made to simulate failure occurrence during final ironing sequence for model material using ductile fracture failure criteria. These criteria are important tools to be used in conjunction with practical experiments and several ductile fracture criteria have been proposed and used in predicting defect occurrence in metal forming processes yet. The performance of the theoretical failure criteria depends on geometric parameters and deformation mechanics. For example Clift et al  investigated several ductile fracture criteria to model damage occurrence in simple upsetting, extrusion and strip compression and tension. Conclusions seeming to indicate that only Freudenthal's generalized plastic work per unit of volume are capable of estimating fracture initiation sites for all of the processes. Venugopal et al  evaluated various theoretical failure criteria pertaining to workability in cold forging reported in the published literature for their reliability and sensitivity in predicting the occurrence of ductile fracture in metalworking. It was claimed that no theoretical failure criterion can be described as truly geometry-independent for metalworking operations. Gouveia et al  described the utilization of ductile fracture criteria in conjunction with the finite element method for predicting surface and internal failures in cold metalworking processes. Four previously published ductile fracture criteria were selected by the author, and their relative accuracy for predicting and quantifying fracture initiation sites is investigated. The author claimed that free surface cracking and internal cracking under a tri-axial tension or tension-compression state of stress can be estimated successively using both the Cockcroft-Latham and the Oyane ductile fracture criteria. However, these criteria are unable to accumulate the damage associated with the occurrence of a material flaw limiting a stationary dead metal zone.
Get your grade
or your money back
using our Essay Writing Service!
A review of the published literature referred above reveals that the performance of the theoretical failure criteria depends on geometric parameters and deformation mechanics. Thus in order to predict damage occurrence on the parts with complicated geometry and critical application like the outer race, it is essential to evaluate damage criteria based on the coincidence between damage modeling results and practical experiments.
Considering the results of the research has been done by Venugopal et al  and Gouveia et al, and the criteria which were accessible by Deform3D software, four commonly used theoretical failure criteria were selected and integrated into the FE Software and their relative accuracy for correct prediction of the location of defects on the lead specimen (from physical modeling) were investigated.
2. Process simulation and physical modeling
Figure 1: schematically shows suggested forging sequences for a CV joint outer race.
Figure 1: Forging sequences of CV joint outer race
The forging process to form an outer race is a combined hot/cold forming which consists of tow performing sequence at elevated temperature, cooling and final ironing at room temperature. This process is investigated by Kim et all  and mohammadi and sadeghi . The first preforming sequence is a forward extrusion to make a head with an extruded rod part while the final preforming sequence is a hot closed-die forging to shape a complicated cup with a sized rod. An ironing operation is carried out as the last operation after cooling. Ironing process has the potential to enable near net shape component to be made. In this sequence that is shown in figure 1, a combination of bending and bulk deformation is carried out. Bending is done with downward movement of the punch leading to radial pressure of lower die to outer edge of the preform.
Due to symmetry of the outer race only one sixth of the billet is needed to be analyzed by numerical simulation. The process was simulated under the conditions shown in Table 1.
Always on Time
Marked to Standard
Table 1: process condition
Close die forging
Velocity of punch
To be brief, only the simulation results for ironing sequence are presented. Figure 2 shows effective stress distribution at one sixth of the specimen after deformation in final ironing sequence.
Figure 2: effective stress distribution at one sixth of the specimen at the end of deformation (forward and backward view)
To verify simulation results, physical modeling with commercial lead was preformed. The true stressââ‚¬"strain curves of the work metal can be obtained from compression tests. The stress-strain equation applied in simulation for lead billet is:
ÃÆ’ = 38.964[MPa].
The die was made from MO40 steel and no lubricant was utilized in the metal experiments and deformation was carried out with friction coefficient of 0.08. A 30 ton universal hydraulic press was used to perform experiments at room temperature. Due to use of commercial lead in physical modeling as the model material, deformation process was simulated for commercial lead to verify reliability of numerical simulation results in real process. Figure 3 shows lead part after deformation in ironing sequence from top view. Dimensional comparison between the lead billet after physical modeling and simulation results showed close accordance of simulation results and physical modeling.
Figure 3: lead specimen after deformation in ironing sequence of physical modeling
There are not any ductile fracture and surface damage and other defects in internal regions of lead specimen. As shown in figure 3 there are some wrinkle on the edge of the specimen. Figure 4 shows another view of the specimen. As seen in figure 4, there are some surface defects like lapping near the flash of the specimen (region A).
Figure 4: surface defects on the external area of the lead specimen specified by (A)
3. Evaluation of failure criteria
To start with, four commonly used failure criteria named Cockcroft-Latham (COCK), Frudenthal (FRUD), Brozzo( BROZ) and Oyane (OYAN) were chosen from the published literature.
Typical criteria for room temperature ductile fracture are usually based on combinations of stress with strain or strain rate rather than on either of these quantities separately. It was shown by Atkins and Mai  that nearly all the integrated stress-strain criteria are versions of Freudenthal's critical plastic work per unit of volume.
Where is the yield strength of the material; ÃÆ’ is the effective stress, is the incremental effective strain; C the threshold value of the criterion at the instant of fracture initiation.
In view of the importance of the largest tensile stress, Cockcroft and Latham  have suggested an alternative fracture criterion based on a critical value of the tensile energy per unit of volume. Cockcroft and Latham's criterion states that fracture will occur when the cumulative energy due to the maximum tensile stress exceeds a certain value. This criterion has provided good agreement at predicting the location of a tensile failure based on a maximum damage value. The model is based on the equation:
Explicit dependence of the level of both the largest principal stress, and the hydrostatic stress,, was proposed by Brozzo et al  by means of an empirical modification of the above mentioned criterion.
Experimental results reported by Dodd and Bai  show conclusively that ductile fracture in metalworking processes follows a void growth model. According to this model, voids initiate at inclusions or hard second-phase particles in regions of the micro-structure that are highly deformed, grow under plastic deformation caused by normal or shear stress systems and finally linkup between each other to form macroscopic cracks. Based on this hypothesis, Oyane criterion, derived under the fundamental laws of the theory of plasticity for porous materials being expressed as
Where A is a material constant to be determined experimentally.
Upsetting of cylinder is the simplest and the most widely used tool for workability test and defining threshold values () for failure criteria. In this study, threshold value of criteria is archived from published information of upsetting of lead alloy. Table2 shows critical values per criteria.
Table 2: Threshold value of the criterion
This Essay is
a Student's Work
This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.Examples of our work
Cockroft & latham
Brozzo et al
Oyane et al
3.1 DAMAGE modeling
The four theoretical failure criteria with their threshold values were incorporated into DEFORM 3D. Due to symmetry of the outer race only one sixth of the billet is needed for damage modeling. Simulation conditions were the same as mentioned for ironing sequence in table 1. The theoretical distribution of damage in work piece is achieved by DEFORM 3D for four criteria. Figure 5 shows damage modeling results for COCK. To achieve more clarity, only the upper part of the specimens on which defects will appear is shown in transparent view. The probability of defect occurrence at a region can be calculated by dividing the amount of damage value on the region by the maximum value calculated by the software. The predicted location of defect has shown by arrows.
Figure 5: damage modeling results by COCK criterion
As shown in figure 5, the damage value in the region M that is marked by arrow is 0.119 and probability of damage occurrence is 55% that can be calculated by dividing 0.119 by the maximum value of 0.214. In region N the probability of damage occurrence is 44%.
Figure 6 shows damage modeling results for FRUD criterion. As shown in figure 6, two regions are marked as the probable location of defects. The probability of defect occurrence in region D (on the flash) is 66% and in region C is 77%.
Figure 6: damage modeling results for FRUD criterion
Figure 7, 8 show damage modeling results for BRUZ and OYAN criteria. Predicted locations of damage occurrence are near the flash edge for both criterions but BRUZ shows more agreement to physical modeling results. Also the probability of damage occurrence for both of the criterions is 33% that shows similar application of these criteria for this specimen.
Figure 7: damage modeling results for BRUZ
Figure 8: damage modeling results for OYAN
3.2 EVALUATION of criteria
As shown in figure 5, COCK criterion has introduced two regions M, N as the possible location of defects. Region M completely matches the physical modeling results (region A in figure 4). But region N is not on the flash edge of the specimen and is located near the edge on the flash body.
The application of FRUD criterion is almost similar to COCK. As shown in figure 6 two regions (C, D) have been introduced as the possible location of defects. The Region C is located above the region A of physical modeling specimen. But the region D is on the edge of the flash and completely matches the physical modeling results.
Neither of BRUZ and OYAN criteria has predicted any fracture occurrence in the region A (figure 4). But both the criteria especially the former have almost truly predicted defect occurrence on the edge of flash.
As can be seen the COCK and FRUD criteria especially the former were in close accordance with experimental results in prediction of fracture occurrence in the region A. In fact, With regard to the deformation stages in final ironing sequence that is shown in figure 9, and the geometry of the preform at the region (A), it is noticeable that this defect can be due to stress concentration at mentioned region. Thus the criteria which are stress based like COCK which is based on principal stress, have better application and performance than the other.
Figure 9: deformation stages in final ironing sequence
Wrinkle occurrence on the edge of the flash that can be seen in figure 3 is due to sever plastic deformation and high values of strain at the zone. All criteria showed this defect near the flash but criteria of BRUZ and FRUD were in better agreement with experimental results respectively.
Altogether COCK and FRUD criteria showed higher performance in prediction of free surface defects on the body of the specimen and on the flash edge successively but none of the criteria could predict the location of both defects completely.
Because of the importance of the defect occurred in region A (figure 4), and with specifying the reason of this phenomena, an attempt has been made to modify the geometry of perform of final ironing sequence. In order to reduce stress concentration as the main leading cause of fracture occurrence in region A, the fillet radius on the preform was modified and physical modeling experiments was retried for modified geometry. Figure 10 shows lead specimen after modification of preform. As can be seen in figure 10, no ductile fracture and surface defects occurred on the specimen. It should be added that the flash will be trimmed so the defect on the edge of the flash will cause no problem in industrial product.
Figure 10: lead specimen without defects after modification of perform
The performance of the theoretical failure criteria depends on deformation mechanics. Hence the first step towards damage modeling is to evaluate damage criteria for the case. In the present study physical modeling experiments was preformed with commercial lead and four commonly used failure criteria were chosen and integrated into the FE Software Deform3D to model damage occurrence on the lead part. Following conclusions were obtained from comparison of experimental and calculated location of damage:
Both the physical modeling experiment and damage modeling results showed the possibility of occurrence of tow type of surface defects that the former is on the flash and is not of importance and the later is on the external surface of the specimen below the flash. It is due to stress concentration and is of great significance.
COCK criterion showed close accordance with experimental results in predicting the occurrence of rupture on the body of the specimen (region A) also BRUZ and FRUD criteria were in better agreement with experimental results in predicting wrinkle on the edge of flash but generally none of the criteria could predict the location of both defects completely and exactly.
Due to the importance of the defect occurred in region A, the results showed that the theoretical failure criterion based on the maximum principal tensile stress was more reliable and sensitive in predicting the fracture in the case.
In metal forming processes which there are the possibility of stress concentration on the deformation zone, stress based criteria like COCK will be helpful and beneficial and can be considered as the prime candidate to model damage.
With specifying the cause of defect occurrence in region A, the geometry of the preform was successfully modified. Consequently by modification of the perform geometry in simulation and physical modeling stages, a major step was taken towards reducing die try-out and lead times for actual process.
Considering the results of the present research and other researches done by Venugopal et al (2002) and Gouveia et al (2000), it is obvious that due to various states of stress in forging of complicated parts like the outer race, a single failure criterion is not able to predict all types of defects and damages. Thus developing subroutines which can apply various damage criteria in parallel with each other can help manufacturer to achieve high quality product without actual tryout.