Mastering Distortion: Intelligent 3D Straightening of Large Castings
1/19/2023 Technology & Processes Experts Know-how

Mastering Distortion: Intelligent 3D Straightening of Large Castings

Conventional processes like manual straightening are no longer sufficient to master distortions of large aluminium castings. Automatic, intelligent straightening systems that are adapting through self-learning algorithms are therefore more and more the solution for casters and enable the economical production of large and complex castings with high mechanical properties and tight tolerances. A guest article by experts Philipp Hettich and Martin Hartlieb.

Longitudinal straightening machine Straightening system for local and overall straightening with three degrees of freedom.

The transportation industry increasingly uses Aluminum for big and complex parts. They allow large and very complex shapes and therefore integration of different parts and functions into a single casting, avoiding complicated and costly assemblies. The rapidly increasing electrification of vehicles is now shifting the market for castings even more, away from traditional powertrain castings to applications in vehicle (sub)structures, chassis parts, cross-members, subframes, closures, and battery trays, among others. 

Precisely those new parts – which require tight tolerances and high mechanical properties to improve crashworthiness – are now bringing new challenges for casting suppliers: Ejection of such large castings as well as heat treatments and especially rapid quenching are often unavoidable to achieve the required properties, which can easily distort the castings. Even when heat treatments are completely avoided by using self-hardening alloys or a simple T5 is sufficient, large structural castings usually can’t meet tight tolerance requirements and therefore require straightening.  

Conventional processes like manual straightening are no longer sufficient to master those distortions and simply not feasible in very large castings. Manual or semi-automatic straightening processes of such large and complex castings are also very time consuming and costly; they can easily create other problems in the parts and contribute to higher scrap rates. Automatic, intelligent straightening systems that are adapting through self-learning algorithms (Artificial Intelligence - AI) are therefore more and more the solution for casters and enable the economical production of large and complex castings with high mechanical properties and tight tolerances.


Distortion in castings

In aluminum casting, distortion occurs typically when the part is cast, when it is heated up, when it is hot and when the part is (rapidly) quenched from high temperatures.

It can therefore occur during the casting process, in the die casting die or press, through 
- sticking and soldering of the part
- die surface defects,
- inadequate gating systems,
- unbalanced solidification, and
- residual stresses.

During ejection and extraction, due to
- uneven ejection forces,
- insufficient draft angles.

It also occurs in downstream processes like
- sawing,
- machining, and
- welding of assemblies.

Most often, castings distort due to heat treatment. The risk is particularly high if they are hot and therefore soft or have to be heated quickly to high temperatures. Quenching is no less problematic. The faster quenching takes place, the more the casting warps. Efficient quenching after ejection from the mould or after solution heat treatment can be very important to achieve high mechanical properties.

Rapid and efficient quenching after ejection from the die or after solution heat treatment can be very critical for achieving high mechanical properties. This is usually the largest contributing factor of distortion, especially when the quench rate is inhomogeneous. We are talking about an optimum “quench rate” that will give us the best compromise between desired properties and allowable distortion. In many cases, however, it is very difficult to find this compromise – and customers are unlikely willing to compromise on properties or tolerances, a solution has to be found to manage the distortions.

In order to control and minimize distortions technical advancements and improvements have been made with respect to simulation of the process and the part, improved heat treatment furnaces, proper support during heat treatment, new quenching techniques, as well as its temperature and agitation. As often, even all of these are still not sufficient to meet all properties and tolerance requirements, and therefore innovations in straightening of castings are now being increasingly implemented in casting and assembly plants.


Basics of straightening

Straightening means the correction of distortion with plastic deformation of a part to bring it back into its intended shape and tolerances. It means that a section of a part is pushed beyond elastic deformation (which would simply spring back) but clearly below its tensile strength. It is important to compensate for the elastic spring-back through basically “over bending” of the casting, which means that it is bent more than would be necessary to get it into the right shape and tolerance, and it will only get there once it “jumped back”. As this spring-back is not always identical (and neither is the distortion of the part), this bending (straightening) process sometimes has to be repeated several times – especially in manual or semi-automatic straightening processes. This makes it almost impossible to simply straighten a large casting with straightening systems using fixed parameters, like in the trimming press with one single pressing step. 

Stress-strain diagram showing the operating window for straightening and practical example. Figure 1 & 2: Stress-strain diagram showing the operating window for straightening and practical example.

Depending on the individual specification and type of distortion, different straightening techniques can be applied:
- Overall straightening
Local straightening
Datum/data point straightening

Overall straightening is important to correct specific general distortions of the part, like torsion or a complete bending of the entire part, or a large segment of it. In contrast, local straightening mostly refers to joining surfaces, sealing surfaces or function-critical areas where a very specific straightening stroke is required.

Datum point straightening refers to a special technique that achieves an overall effect on the part by applying local forces. In this technique, the area around the datum point (never a datum point itself!) is locally deformed with the aim of changing the datum point’s relative position to other measuring points. This mixed form is especially important when it is more effective to manipulate one datum point instead of orienting many measuring points at one datum point.

We can further distinguish straightening between
bending,
pressing and
torsional straightening.

In any case, areas of the casting that are not to be straightened should always remain below the elastic deformation limit during any straightening process. Proper support and handling of the entire casting at all times during the straightening process is therefore key, which makes it very difficult to straighten complex structural castings with one operator or even with one robot, and a more sophisticated system offers significant advantages (and is often the only possible solution).

 

Straightening of castings

In castings – especially in large ones - distortion usually does not only happen in one section and one direction, so it often has to be straightened progressively in a series of (incremental) bending, rotating and pressing operations in several shorter sections and the part needs to be well supported to avoid buckling. Each of those can now again cause a certain distortion in other sections of the casting, which makes it obvious that this is quickly becoming a complex process that requires a lot of precision and adjustments. Using the right segmentation of the casting can be vital in achieving the desired straightening.

The full range of possible distortions of the part and its behaviour needs to be known. Only then, in combination with the correct segmentation and an ideal combination of straightening techniques, can the "perfect straightening stroke" be achieved, in which all incremental individual steps are executed simultaneously to obtain a part within the tolerance specifications in minimum cycle time with one straightening stroke. It also needs to be determined, which tolerances truly need to be reached by straightening and what can later be reached through machining of certain areas of the casting.

In certain cases, a later distortion – for example in the joining process - can be anticipated and corrected through “over-straightening” beforehand, so that the joining process basically distorts the part into its intended shape and tolerance.

Another item to pay attention to for straightening of castings is the fact that casting dies change over time, which can have a significant impact on the exact shape and tolerances of the cast part. And even small changes in the process can lead to variations in residual stresses in the casting, or changes in its shape and tolerances.

 

Automatic straightening systems

Intelligent, automatic straightening systems are highly complex, require significant engineering and investment, but they can be well worth it! Especially when the caster does not need to compromise on material characteristics and at the same time can save money when it becomes obsolete to optimize every step to reduce distortion. 

These systems use synchronized forces at the part where and when it is needed to ensure local and overall results. The system will immediately counteract an unwanted deformation in an area of the casting by applying a counterforce in the right location – working with the part and not against it. These systems can learn and optimize the straightening process by themselves, achieving results otherwise impossible to obtain.

The basic procedure is like in manual straightening. The part is measured, then straightened and then again measured to determine the achieved result. This is repeated, if necessary, until the part is completely within tolerance. By evaluating past straightening effects and learning from them, the machine is then capable of minimizing the required time for straightening.

For precise measurement of the part, proper support on defined locations is vital. Instinctively one would support the casting on datum points, but this could influence the relative position of zero points and other measurement points due to gravity, which would then make all other measurements in the part incorrect and would lead to a wrong basic assumption for the straightening of the casting. Moreover, it exposes these elementary points to the risk of mechanical damage.

The key therefore is to create a force-locked state and then determine the (protected) datum points by measurement. In fractions of a second, based on this determination of the part position in space, the other measured values of the part surface are mathematically transformed to obtain a complete virtual image of the part and its distortions. 

The measurement itself is usually tactile to ensure speed and accuracy. In rare cases, laser sensors are used, for example for flatness of surfaces. Based on the input measurement and the resulting virtual model the machine compares it to the part from the CAD model and decides which straightening elements (units) will be used and to which degree.

It is understandable that the number and positioning of these straightening elements is crucial to successful straightening. To enable the machine to take the right actions, experience from castings received and measured (including all possible distortions) and experience from straightening them must be combined, the aforementioned segmentation determined, and the tool set up for the relevant straightening techniques (local, overall, datum point shift) be considered.

A 3D straightening system has three degrees of freedom for this: Y, Z and torsion around the X axis. 

 
Longitudinal member with specific straightening locations and tools for the relevant techniques  (green = local straightening, blue = overall straightening) and the respective straightening system. Figure 3 & 4: Longitudinal member with specific straightening locations and tools for the relevant techniques (green = local straightening, blue = overall straightening) and the respective straightening system. 
Straightening in each element itself is controlled by distance rather than force, as this – in combination with repeated (online) measurement of the effect - allows determining the plastic and elastic components of the applied deformation. After each straightening cycle (measurement – straightening – measurement) the software creates a record for all straightening points and parameters: combination of straightening units, straightening path, number, and effect of each stroke, etc. This enables the caster to draw additional conclusions about the manufacturing process with respect to its impact on tolerances of the part, but also regarding changes of the material characteristics like changes in properties.

Figure 4 shows the straightening system for our example longitudinal beam casting. Depending on the complexity and incoming quality of a specific casting type, the straightening cycle can range from 45 seconds (simple shock tower) to 150 seconds (highly complex battery housing or Mega/Giga-Casting).

The system can be integrated into production lines and parts loaded and unloaded automatically with robots, so that an operator is only required for setup of a new casting. Compared with manual or semi-automatic systems they can significantly reduce required floorspace and they cause very little downtime. They distinguish themselves with extreme precision, quality and repeatability. Without those systems, many of those complex castings like shock towers, longitudinal members, battery trays, etc. could not be produced competitively with aluminum casting processes.

Conclusions 

Structural aluminum castings have seen significant growth in the transportation industry in the past two decades and this is now even accelerating with the electrification of vehicles. New parts like complete front or rear body structures, or battery trays, can ideally be made with large and complex castings. In order to meet both properties and dimensional/tolerance targets, castings are increasingly requiring extensive straightening operations. As distortions can vary significantly from part to part and manual or semi-automatic straightening can be very difficult, time-consuming and costly, new automatic straightening systems have been developed and their implementation in the casting industry is increasing. These intelligent systems adjust to varying distortions and can even alert the operator if something in the die or process is changing that requires fixing. They reduce the required floor space, manpower and offer high technical availability with low cycle time. With that, casters are able to meet both property requirements and very tight tolerance requirements, which allows them to competitively produce those new parts and enter this rapidly growing market with high value-added aluminum castings.

References

Wiesner, S, Kniewallner, L, Miller, R, “Aluminum HP-DC Alloys with High Conductivity”, NADCA Transactions 2021
Schnur, M, Whealy, G, Mussler, P, Hartlieb, M, „Challenges and Innovations in Die Cast Tooling“, NADCA Transactions 2017
Amiotte, C, Desrosiers, S, Beaulieu, S, Hartlieb, M, „Heat Treatment of Structural High Integrity Die Castings”, NADCA Transactions 2015
Kalkunte, B., Sholapurwalla, A, Valente, L, Viscandi, C, “Predict and Control of Final Casting Shape through Virtual Dimensional Inspection”, Die Casting Engineer, July 2021, p 12-15 and
Gaddam, D, Jesper, T, “Integrated Modelling of Deformations and Stresses in the Die Casting and Heat Treatment Process Chain”, NADCA Transactions 2019, and
Bramann, H, Leineweber, L, Sturm, J, Gaddam, D, „Innovative Product Design and Robust Layout in Die Casting with autonomous Engineering”, NADCA Transactions 2018