English
中文INTRODUCTION
Roque (2000) has performed a simulation using the finite element method for cold forging of a ring of material DIN 16MnCr5. Software application used is ANSYS. The upsetting process was modelled using algorithms developed from previous research. The goal of the paper was to understand the material flow and the potential for die filling so that better parts can be produced. While upset forging is a common process, much of the die design is based on experience of the die designer and there are some inherent problems with the forging process and this results in small defects and a higher skin formation and excess allowance for machining. Chances of component rejection are higher as the material flow due to friction in the mould space depends on a number of uncontrolled variables and the actual manufacturing process is trial and error. By carrying out a simulation of the forging process, the author shows that end quality of the component can be predicted and controlled.
ANALYSIS OF THE METHOD USED BY THE AUTHOR
Roque (2000) has defined the problem statement by specifying a number of key parameters and a formed a differential equation that had to be solved using FEM simulation. A flow chart was created to depict the FEM method and the flow chart represented important nodes of the forging process. Table 1 gives the constitutive equations of plasticity and equations such as the stress strain relation yield function, flow and hardening rules and the Kuhn-Tucker conditions were fed into the applications. A number of equations were used to set the contact conditions for metal forming analysis and to calculate the stress and internal variables, the radial return mapping algorithm was used. Among other methods that were used was the penalty method to place the contact conditions between the die, tool and the billet surfaces.
The flow chart is illustrated in Figure 1 from the publication (Roque, 2000). There were three stages considered in the flow chart: input, FE analysis and Output. FE analysis was applied only for the second stage. In the first stage of Input, important elements of the metal forming process design input were taken along with the numerical model inputs. This was an important step as the inputs were made of the raw material specifications, the billet geometry, the die geometry, material properties, lubrication conditions, process history and the equipment characteristics. There formed the process parameters. ANSYS material library provides for parameters such as the Youngs modulus, yield stress, strain rate, Poisson’s ratio, plastic deformation and flow rate and other features and these are mentioned in the Nomenclature section of the paper (Roque, 2000). Identifying these parameters correctly was very important for the given raw material since flow characterises are specific to a material type at given pressure and temperature. The lubricant specified was MoS2 – bisulfide of molybdenum and specifications of the lubricant are available in the ANSYS library. The specifications of the material are given in the section ‘Ring Compression’ (Roque, 2000). The next step in the Input stage was to create a numerical model of the input. Since the ring is symmetric along the axis and radius, the simulation was performed on one fourth of the ring. This was discretized with 231 nodes and the nodal discretization gave the element geometry, the element shape functions, the integration zone and the boundary conditions. From the Ansys tool library, two elements were used to model and these are Plane42 and Vicso106. The required displacement was applied at the forging tool nodal points.
The second stage was the FE Analysis where the simulation is carried out. Material flow is a function of the friction flow and the ring test simulation used four types of friction coefficient. The friction coefficients are m =0; 0.15; 0.3; and 0.57. When the coefficient of friction is zero, then it represents the ideal condition and the component would deform in a parallel configuration and the forging process would be perfect, but this does not happen in practical applications. As the coefficient of friction rises, the material flow tends to barrel and the deformation becomes more imperfect. The simulations of the deformation at different coefficient of friction are shown in Figure 4 (Roque, 2000). As seen in the figure, the worst deformation is when the coefficient of friction is the highest at 0.57 and the component is barrelled and deformed. The author suggests that with the proper use of lubricant and optimum surface finish of the die, the coefficient of friction can be taken as 0.10. The author has also performed experimental tests and the results used to compare the simulation results with that of the actual forging process.
The third stage was the Output stage and in this stage, the deformation flow distribution was verified, stress and strain distribution was verified and mapped for the die and the part and elastic deformation of the tool was considered. The results would indicate the forming defects, areas and points of high stress and strain, instances of improper die filling, any areas of irregular flow and limitations of equipment.
To further support the findings, the author has performed a plot of the stress regions with that of the micro etchings of the experimental components. Stress patterns and work hardening of the high stress regions and the low stress region are found to match and the flow lines of the simulation closely follow the actual flow lines of the experimental sample for the die fill.
The conclusion is that the simulation is successful and it can be used for real life forging applications when punch and die design are being considered. Where possible, design modifications can be made to ensure smooth flow and optimum die fill.
CRITIQUE OF THE PAPER
The simulation exercise has not taken some very important variables that occur during the actual forging process, heat generation, determination of surface finish and punch pressure (ASM, 2005). Moreover, the focus of the paper should have been reaching the boundary level, the deformation observed and this would give the values for machining allowance and forging time, both of which have an important commercial bearing in manufacturing and this has not been done. So in effect, the simulation is for purely academic purpose and would not be very useful for manufacturing companies.
The whole simulation has been done by assuming four coefficient of friction values and while this is acceptable, what is missing is that heat generation during forging has been ignored, During forging, an enormous amount of heat is generated by the deforming part and any lubricant that is applied starts flowing away due to the heat. The loss of lubricant at the contact and flow regions creates internal stresses that have to be removed by tempering, leading to extra costs. As the number of components manufactured in a shift increases, the temperature rises between parts and lubrication becomes ineffective (ASM, 2005). There is no mention of this aspect in the paper and it is assumed that the friction will remain constant. The simulation should have had another variable and that is operating temperature and how it would effect the lubrication and the material flow. Then if the simulation was giving excessive deformation, then another lubricant with higher viscosity can be used to run the simulation.
Reaching the net shape or the end size of the forging process is not simulated. Forging is done so that a certain size of the component is achieved. Manufacturing companies are faced with the problem of extra machining allowance when an oversize billet is used. In practical forging operations, the outer diameter dimensions are set to be reached and sufficient allowance is kept on the surface and other sides and these have to be machined. The machining allowance is a necessary waste but it costs money and is done so that the part can be machined to the required size (ASM, 2005). There is no mention in the simulation of how the net shape would be achieved, amount of distortion of the planer surfaces and minimum amount of billet size required to reach the near net size. If a simulation were done for these factors then manufacturing organisations would be able to optimise their raw material and save on the production and raw material costs.
Another aspect that has not been considered is the relation of the forging pressure applied. While the Youngs Modulus of the billet material is given, there is no mention of how much pressure was applied during the pressure and if increasing the pressure would cause material to flow more faster and the effect of increased pressure on the defects and discontinuities (ASM, 2005).
FINAL COMMENTS AND CONCLUSIONS
ANSYS is a very powerful software application and all the issues mentioned above could have been answered with the proper simulation techniques. By performing the operations mentioned in the previous section, the simulation would have had more value that is practical. As such, the paper has just mentioned how simulation using FEM can be performed using the methods mentioned. The author should have at least provided section that explained these shortcomings and suggested these studies as areas for further research.
