Industrial Application of Metalforming Simulation for Automotive Forgings

This paper delivered at the CBM conference in 2000 outlined examples of simulations performed at UEF and discussed the benefits, limitations and ‘tips of the trade’ in using such tools.

1. Introduction

Metal-forming companies, such as UEF, which are involved in hot, warm and cold forging, rolling, extrusion and drawing processes have traditionally relied upon empirical methods to optimise their metal-forming processes. The skill of the die designers has developed through inherited knowledge, which can be lost if the employee retires or leaves the company, and experience through many potentially expensive plant trials and design iterations. With the rapid increase in affordable computing power, metal-forming simulation based on the finite element method (FEM) is now becoming a practical industrial tool. UEF have now been using DEFORM™-2D and 3D for twelve months and are now beginning to take advantage of this new technology ¹. Typical applications of these simulation programs so far have included:

• Progression design and improving material yield.
• Prediction of forming problems such as flow laps, fracture and under-filling of die cavities.
• Calculation of state variables such as temperatures, effective strain, damage and material flow that may affect the metallurgical structure and performance of the product.
• Prediction of forming loads and die stresses for a range of mechanical and hydraulic presses.

2. Requirements for a Simulation System

When considering the use of a simulation program within UEF, the following capabilities were
identified as requirements for successful software integration into the company design process:

• Easy-to-use graphical user interface (GUI).
• Post-processing capabilities that reflect industry needs, including – material flow tracking,
  damage contours, thermal interactions and output of press loads.
• Robust, automatic running including re-meshing and interpolation of state variables.
• Reliable interfacing with in-house CAD systems.
• Fast solution speed – not at the expense of accuracy.
• Restart facility to aid user learning, recover mistakes and facilitate multiple operation
  simulations.
• Die stress analyses capable of dealing with multiple body tool sets.

Since using the software in practice, the requirement for reliable interfacing with CAD systems has been found to be very important to UEF. For 2D problems, geometry can generally be created easily within the software pre-processor or, more typically, imported from a CAD system in DXF or IGES formats with relatively few problems. In 3D, however, the need for clean geometry with no imperfections such as overlapping surfaces or gaps is essential, otherwise the program will believe a burr or hole is present in the die. Furthermore, IGES transfer in 3D directly into DEFORM is not currently possible because of the poor quality of the surfaces imported and inconsistencies between different ‘flavours’ of IGES. Instead, STL files are used as a more reliable alternative import format.

3. General Modelling Considerations

3.1 Representing the Physics of the Process

As the material behaviour can be highly temperature-sensitive, the changing thermal conditions in hot forging often need to be modelled. This may include the effects of adiabatic heating in the work-piece due to deformation and friction, and heat transfer to the dies and environment. Surface chilling effects can be a very important influence on material flow and therefore the transient temperatures in the die are also considered. When heat transfer in the dies between forming operations is significant, such as chilling on a bottom die during tooling changeovers, these intervening stages of the manufacturing process are simulated as well. The boundary conditions are consistently changing throughout the simulation as contact between the work-piece and a die evolves. If a folding lap is anticipated, self-contact of the work-piece is activated to detected and model this flow correctly. Strategies for handling evolving contact include substepping of the time increment and projection of penetrated nodes back to the surface in 3D. The die movement control applied depends on the type of press machinery being simulated. The mechanical presses at UEF will have a distinct sinusoidal velocity-time profile dependent on the forging stroke, and infinite machine power is usually assumed. Alternatively, in cases with power-based hydraulic presses and load-based sprung secondary dies, the movement of the tools will depend on the stiffness of the work-piece material which, if hot, will in turn be affected by the forming rate. Consequently, realistic input data is required for both the material behaviour and press characteristics to obtain an accurate result.

3.2 Multiple Operation Pre-processing

A metal forming process at UEF may consist of a large number of stages. Cold forming operations, for example, can exceed half a dozen individual steps, where the flow characteristics of each step will depend on the final geometry and work hardening accumulated from the previous stages. As preform design is based on the interrelationships between these steps, a practical simulation will typically be required to consider the entire process. If intervening heat transfer stages are also considered a complete analysis may easily consist of over twenty successive simulations. The results database in DEFORM™ allows the relatively easy extraction of work-piece and die geometry and properties to manually pre-process a new forming stage. However, the multiple operation pre-processor is often preferred which enables an entire manufacturing process including tooling changeovers to be entered in a single pre-processing session using hot and cold forming pre-processing ‘templates’.

3.3 Minimising Run Times

A manufacturing process optimisation study often requires multiple metal-forming simulations to
achieve the correct die configuration or preform design. Consequently, even in 2D simulations, it is
useful to minimise run time whenever possible. For most hot and cold forming applications, the elastic behaviour of the material can be disregarded. In this case, a rigid-plastic or rigid visco-plastic model can be used which can generally afford larger time steps and converges more quickly than the elasticplastic models. Elastic deformation of the material will affect residual stresses and cause ‘springback’ phenomena following removal of the dies, and therefore the elastic modulus must be considered when these areas are of interest. Elastic or plastic die deflection during forming is usually insignificant, in which case the die components can be assumed to be rigid rather than handling deformable to deformable body contact. UEF are often interested in predicting tool stress and final work-piece geometry and perform the metal-forming simulation to obtain die loads and thermal gradients. DEFORM™ offers two strategies for calculating die stresses: fully coupled deformation analysis with deforming work-piece and elastic or elastic-plastic dies; or one step elastic die stress analysis using loads and thermal distributions previously obtained when assuming rigid dies. The fully coupled approach has been found to be computationally intensive but very useful when investigating cyclic stresses within the dies during processing. If the dies comprise a number of components, the user can model the assembly as a combination of rigid and deformable objects to reduce run time. The one-step is generally the preferred option as the die stresses can be calculated very efficiently at any step during the simulation. In addition, different designs (e.g. variation in the shrink fit interference) can be tested using the same interpolated loads without rerunning the deformation simulation.

If heat transfer is thought to be unimportant then the simulation may be performed isothermally. However, this assumption requires some care. Material flow during hot forging simulations is often significantly affected by the increased stiffness of surface material chilled by initially cold dies. Die stresses can also be greatly affected by thermal gradients. Generally, UEF run all simulations nonisothermally, although the dies are occasionally run as constant temperature heat sinks in 3D if the contact time is very quick or if only the die face surface geometry’s are available. In 3D run times are also improved by considering planes of symmetry whenever possible.

4. Training and Support

For UEF to fully exploit the capabilities of DEFORM, emphasis was placed on initial software
training. This was conducted on a one-to-one basis and with some preparation can be tailored to suit the needs of the individual forging company in terms of the equipment used, components and
materials forged, CAD data input and analysis of simulation results. Generally, a one-day intensive training course is required to introduce the software and its functions to the new user. Once the basics have been covered a course of twenty-three training laboratories can be completed at the users leisure. Each laboratory covers a different software application and become progressively more involved. By working through these exercises the new user can experience a range of metal-forming techniques and be introduced to aspects of tool design and stress analysis. Titles include – Geometry Input and Correction, Mesh Generation for Dies, Heat Transfer, Screw Press Simulation, Die Stress Analysis with Shrink Rings – all of which strengthen the users ability and forging knowledge.

Technical support from the UK agents, Wilde is available whenever required and covers all aspects of DEFORM, if necessary support can be sought from the software developers at Scientific Forming Technologies Corporation (SFTC) in the USA. Replies are prompt, usually problems are answered within a couple of days or sooner if necessary and confidentiality is maintained at all times. Past experience at UEF has proved Internet access and e-mail to be an essential communication method for efficient customer support. Keyword files containing simulation data can be sent to Wilde for assessment and then returned easily without the use of floppy disks or, as is more likely due to file size, CD-ROM’s with the resolved problem. Another advantage of Internet access is that software updates can be downloaded directly from the DEFORM web site.

Comprehensive training manuals are supplied with DEFORM™ in electronic format, which can easily be printed and bound by the user, or remain in electronic format and be accessed through Acrobat Reader. Periodically Wilde host refresher courses for users, and if necessary can organise training sessions on specific software applications, advanced simulation, die stress analysis and other intricacies of FEM techniques. In addition to this, the annual user group meeting (UGM) is a focal point where DEFORM users from around the UK can meet and discus software developments, personal experiences and applications.

5. Example Simulations

5.1 2D Multi-Stage Hot Forging Simulation

This first example demonstrates how the complete thermal cycle of the work-piece material can be
simulated using the DEFORM-2D pre-processor, where the complete production sequence can be set-up by the user in one pre-processing session. This can include – forced billet heating from the heater, temperature loss to the environment during transfer operations and heat transfer to tooling, press characteristics, friction parameters and geometry importing. The forging is produced flash-less on a Hatebur AMP50, at a rate of 55 parts per minute, from bar stock with automatic transfer grippers that move the forging through three forging stations and a final piercing station. Accurate prediction of material flow, die filling and press loading play an important role in maintaining good tool life and consequently high product quality. The heat loss during transfer from station to station and to the tooling is minimal in this simulation due to the high production rate of the machine and consequently short contact times. Typically these stages of the forging process would be excluded from a simulation thus reducing set-up and computation times. Heat loss to the tooling has the most influential effect on material flow, especially on components with long thin sections that dissipate heat easily. However, for other products manufactured on conventional mechanical presses by operators, it becomes more important to simulate the heat losses of the workpiece as transfer times can be longer than that of this example and die chilling has more of an effect on material flow characteristics.

Figure 1. Temperature gradients predicted at the end of each forming stage.

The entire process was set up in the multi-stage pre-processor, using material data from the
DEFORM™ database, and run a number of times with different initial billet temperatures and friction coefficients. The temperature gradients in the work-piece and dies predicted by one simulation are shown at the end of each forming stage in Figure 1. It can be seen that there is significant chilling at the interfaces between the punch and ejector pin and the work-piece. The final stage involves the simulation of the piercing operation, where elements are deleted from the work-piece once they exceed a user-defined level of the ‘damage’. This critical value of damage was set from experience of previous simulations using this material [2-4].

One of the objectives of the simulation was to investigate the stresses induced in the dies during
forging following experiences of tool failure around the punch nose in the final stage. Consequently this operation was simulated by DEFORM™ with elastically deforming dies to predict both the stress distributions and deflections during forming. In Figure 2, the final maximum principal stress distribution predicted in the tools is shown with the compressive and tensile regions marked in green and red respectively. These stresses are generated by a combination of forging pressure and high thermal gradients at the surface of the dies. An area of tensile stress, in red, can be seen at the punch nose radius which indicates a potential fracture mechanism.

Figure 2. Maximum principal stress distribution in the tools at the end of Station 3.

5.2 2D Flash-less Forging Development

The following example demonstrates how forging simulation can be used to modify tooling geometry and analyse material flow – enabling material yield to be maximised. Again, the simulation is a 2D axis-symmetric problem with the tooling considered as rigid and only heat transfer to the tooling being simulated. The component was previously manufactured as a hot forging in the traditional manner with flash on a 1300 ton mechanical press. By using DEFORM™-2D a flash-less version of the component has been developed, see Figure 3, which also benefits from a reduced material volume in the central wad – a resulting increase in material yield from 70% to 84%.

Figure 3. Flash-less forging development using DEFORM-2D.

Importantly, this has been achieved without having to go through the iterative process that would be necessary without the use of simulation software. Also, by making subtle changes to the mould die tooling, the consequent forging load in the finisher tooling has been reduced – helping to prolong tool life and reduce down time at the press, Figure 4.

Figure 5. Comparison of the predicted and actual final geometries of the radius arm.

5.3 3D Simulation of a Suspension Component

A temperature sensitivity study was conducted using DEFORM™-3D on a large hot forged
suspension component which is produced on a 6000 ton press suit in three stages from an initial
preform. Layout of the 6000 ton press suit is such that the billet on leaving the heater has a twenty
second transfer time through a system of conveyors before it reaches a manually operated reducer roll. The time taken for the billet to pass through the rolls varied depending upon which operator was on the machine. The resulting drop in billet temperature was measured as the preform was placed at the first forging operation.

The die and initial preform geometry was imported from the CAD system in STL format, taking
advantage of symmetry in the plane perpendicular to the forging direction to minimise computational time. After determining a suitable friction factor to closely represent the lubrication conditions, the initial work-piece temperature in the simulation model was varied to match temperature readings recorded from the actual forging operation. The load stroke graphs produced from the simulations showed a peak forging load close to the press capacity. Taking these figures into consideration and the fact that the mould and finisher operations are completed within a few seconds of each other, there was a danger of stalling the press due to the eight second flywheel energise time. As a result of this, measures were taken to ensure billet temperature remained high enough to minimise the risk of overloading the press. The outline of the final simulated geometry can be seen to agree well with the actual trimmed flash shown in Figure 5.

Figure 5. Comparison of the predicted and actual final geometries of the radius arm.

5.4 3D Simulation of a Connecting Rod

Conventional methods for calculating forging loads involved in impression forging do not always
provide accurate estimates and can be difficult to calculate as induced strains vary in differing parts of the work-piece during the press stroke. The purpose of this simulation was to provide a more accurate value of forging load produced when manufacturing a connecting rod. The connecting rod was previously hot forged with a separate cap, see Figure 6, on a 1300 ton mechanical forging press. The customer wanted to change the design to a one-piece connecting rod that could be fracture split – thus benefiting from material savings and reduced machining costs. To produce the one-piece connecting rod on the existing 1300 ton press line, UEF designers required an accurate calculation of the forging load. Previous estimates calculated by hand produced borderline load values. If the press capacity required to produce the fracture split rod was above 1300 tons then an alternative production route would be necessary. The only option available was to manufacture the new rod on an 1800 ton press but this would require a capital investment program of £300,000 to equip the press line with new reducer rolls, design and manufacture a new quick-change bolster, heater refurbishment and clip press refurbishment.

Figure 6. Hot forged connecting rod & separate cap.

6. Discussion

Simulation must be representative of what happens on the factory floor, therefore, the first tasks to be undertaken were a series of benchmark simulations. The purpose of which was to establish if the software would replicate products with known defects, allow comparisons of dimensional accuracy and grain flow to be made between predicted results and actual components. In doing this confidence was gained in the software and further investigations into the effects of process parameters such as friction and heat transfer coefficients were made. Relatively simple components, such as the synchro ring outlined in the first example, provide a good basis on which to benchmark software packages. Tool geometry is simple to generate, either with in-house CAD systems or within the DEFORM package, and results can be obtained in a short period of time. Generally, the benchmark simulations showed good correlation with the physical components. Also, the experience gained in setting-up and running the simulations served as good training and highlighted how to avoid mistakes and overcome error messages.

Both pre and post processors for DEFORM™-2D and 3D are very similar in appearance, which allows the user to easily progress from the 2D software to 3D simulation. The set-up procedure is also very similar, the main differences being that 3D geometry must be imported from a suitable source and be of a high enough quality for simulation runs to be successful – typically these are higher end CAD packages. The other main difference is in the method of generating meshes on tooling and work-piece. The Euklid CAD system currently used at UEF for generating 3D surfaces for high speed machining and EDM electrodes has been found to create poor surfaces from its models which are often noncontinuous. This can cause problems when importing 3D geometry, as the CAD files have to be manually repaired which is a tedious and time consuming process. Currently UEF are evaluating software packages that can automatically repair 3D geometry files. Also, work to overcome this problem on a long-term basis is being carried out at SFTC.

7. Conclusions

As can be seen from these examples, process simulation has found many uses within UEF. Once
successfully integrated into the company design sequence the real benefits of such packages can
clearly be identified. Process optimisation can be achieved quickly and efficiently through the use of simulation software, especially in 2D situations. Reduced product development costs and more ‘right first time’ tooling can provide significant savings. Supported by the appropriate CAD facilities 2D and increasingly more 3D process simulation can enable designers the ability to explore and evaluate more challenging components that would otherwise be declined from manufacture due to the high cost and risks associated with full scale, physical forging development trials.

Successful application of simulation software into any forging company requires commitment and a structured approach. Financial backing, time and effort from both forger and software provider are essential, but probably most important element for success is the person operating the software. The operator must have a sound knowledge of the forging process, have an understanding of FEM techniques, be proficient with a computer and preferably be capable of generating their own CAD data– all in all they must be self sufficient.

The examples outlined previously demonstrate a selection of DEFORM™ capabilities in the hot forging field. Many other metal-forming techniques can be explored using the software and clearly this technology has much to offer the metal-forming industry. With such a large portfolio of products the opportunities to exploit DEFORM™ within UEF are wide ranging, current efforts are focused on disseminating the results and benefits of simulation more widely within the UEF Group.

8. References

¹ Walters, J.; Wu, W.T.; Arvind, A.; Li, G.; Lambert, D.; Tang, J.: Recent Development of Process Simulation for Industrial Applications. Proc. 4th International Conference on Precision Forging Technology, Oct 12-14, 1998, Columbus, Ohio.
² Cockcroft, M.: G. & Latham.: D. J., Ductility and Workability of Materials, Journal of the Institute of Metals, 96, (1968), 33-39.
³ Taupin, E.; Breitling, J.; Wu, W.T.; Altan, T.: Material Fracture and Burr Formation in Blanking – Results of FEM Simulations and Comparison with Experiments, SFTC #336 (1996).
⁴ Miller, B.C.; Ward, M.J.; Davey, K.; The Numerical Simulation of Potential Forming Problems in a Railway Wheel and Tyre Manufacturing Process, ICFT’1998, IMechE, 27-28th April 1998 Birmingham, UK, 201-216.
⁵ DEFORM Users Manual 2000, Scientific Forming Technologies Corporation, Columbus, Ohio.

November 2000