Dr.-Eng. Ulf Schliephake
Area Sales Manager Scandinavia and Austria, Brechmann-Guss GmbH
Rapid Casting Part 1 How Does Rapid Casting Work?
This article series presents various possibilities along the product life cycle that innovative additive solutions offer to a tradition-conscious industry. In the first part, we describe the differences to the conventional method and introduce the process chain.
So called “Rapid Casting” is the indirect use of additive manufacturing in an (iron) foundry. It results in a drastical reduction of delivery times for cast prototypes from the usual two months to ten working days in favorable cases. The results are near-series prototypes from the series foundry in the series material in the series production process (4s).
Traditional Process Flow vs. Fast Prototyping
Pattern making involves data processing including coordination with the customer and machining of the external patterns as well as the core-forming geometries with the assembly of pattern plates and core boxes. In the usual process flow of a foundry, the process step pattern making represents the critical deadline bottleneck. This has led to many customers welding their prototypes from sheet metal or even carving them from a solid piece of metal.
Until now, this was the only way to be able to test or present real components in time for the next trade fair, the announced start of series production or the start of the field test, despite the late order from the end customer or a delayed finalization of the design process. Reducing the large amount of time required for model pattern making is a way for foundries and customers alike to quickly produce prototypes from the series material in the series production process despite process disturbances. By printing the mold components, i.e. the mold halves and cores, six weeks of model pattern making can be shortened to five days of 3D printing.
However, one restriction should be taken into account: for each casting (in sand casting) an individual lost mold is necessary in principle - this procedure can incur considerable costs due to the large volume of the molds for the outer geometry of the component. Therefore, the process of print-core and conventional construction of a model for the outer geometry is very attractive in terms of costs. Another very exciting alternative - the printing of the plastic molds in filament printing - is the subject of current studies. The time required for this process variant increases only slightly, since in conventional pattern making the production of the core boxes requires more time - and it is precisely this process step that is being substituted.
Experience from numerous projects has also shown that, in the course of the project, the customer's requests for changes generally relate only to the internal geometry - which is printed in Rapid Casting - since the mounting points and installation space (external geometry) remain unchanged. In extreme cases, the internal geometry can be varied from casting to casting by printing the different cores. The process advantage becomes significant in the case of complex core packages. In individual cases, the costs of additive production of the core package are equal to conventional production (often with gluing/joining operations) – while being faster and more precise (in iron casting, printed CT8 instead of the usual CT10 are possible).
Indirect Additive Manufacturing
Basically, the additive manufacturing processes are divided into direct and indirect processes.
Indirect processes are processes which, in the first step, create a master model/tool rapidly and with geometric accuracy. In the next process step, the master model is conventionally molded into components with defined mechanical-technological properties. Examples are sand casting and investment casting on the basis of lost generatively produced molds and cores. Direct processes, on the other hand, usually directly produce the target component from metal using the generative process1.
The process is based on 3D data sets, which are broken down into individual thin layers in the form of STL files (slicing process). In the actual printer, an extremely thin layer of sand (300 to 500 µm) is applied to the printing plate. A mobile print head with a foundry-typical binder then glues the sand grains together wherever the sand mold/sand casting core is to be created. The printing plate is lowered by one layer and the process is repeated until the sand mold/sand core is finished. This is followed by post-processing and extraction of the loose sand, which is a considerable cost and time factor in the production filigree contours.
One very important aspect is obvious: Error-prone joining processes as well as manufacturing with loose parts are eliminated and the costly reworking is reduced to a minimum. Particularly for complex components such as pump impellers or impellers produced in smaller batches, the overall process costs are therefore the same - with higher precision and less susceptibility to failure.3
Every foundryman should consider the higher amount of binder (around 10 times) contained in printed molds and cores and should consider in advance the gas surge or rather its control and the removal of the gases. However, since the dimensional stability of printed geometries is high, gas channels can be included in the data set and printed. If necessary, it is possible to simply dispense with the binder for inner areas of the (core) geometry or to print the mold or core partially hollow in these areas.
Process Chain and Project Flow
The basis for both project processes is the component data set, supplemented by a drawing if necessary, which specifies, among other things, material data, shape and support tolerances as well as critical (test) areas. Ideally, the data set is already available in a form adapted to the casting process - if not, an initial clarification is necessary in the case of a prototype. "Zero" draft and undercuts are possible when working with printed molds and cores but are not suitable for later series production using conventional core production - for example Cold Box or Hot Box. Consequently, the following distinctions are made:
- Rapid prototypes (not suitable for series production)
- Prototypes, which are manufactured identically to series in a conventional way
- Prototype components that can only be produced with printed moldings
In the context of prototype production, there is always the additional issue of testing efforts: Complete measurement and an initial sampling considering all material characteristics takes a few days.
Isn't it enough to measure a few main dimensions/some critical dimensions at random points? Experience with continuous use of the 3D data chain shows that either all important dimensions are good or almost all are bad - you will notice this already after the first five dimensions, which saves time during the project. In this respect, the measurement of 80 values brings only little or no new insights. After all, who benefits from meticulously extracting tiny specimens from a component to determine material parameters that are of little use - such as the impact force of nodular cast iron4 - if the part actually works on the test bench? If the component holds up on the test bench despite deviating from the theoretically determined material characteristics, then the specification appears to have been incorrect. There is therefore no reason to repeat the production process, but rather the utmost need to correct these errors.
But let's get back to the process chain: Once the final design has been determined, the casting technique is specified and simulated. Printing the mold - whether as a one-time mold made of sand or a reusable plastic model - and/or cores takes about five working days, plus transport from an external service provider to the caster. Casting, cooling, blasting and grinding can be assumed to take two days if the internal workflows have been defined accordingly.
In the case of a series manufacturer who produces not only prototypes but also series components in multi-shift operation, experience shows that the personnel must be sensitized (and if necessary the order marking or even the marking of the working papers must be varied in color) in order to complete the process "accident-free". A desperate project manager looking for his parts in the production department is one of the operational side effects of such unconventional, still untrained, new manufacturing processes.
Ultimately, the foundry delivers prototypes (4s) suitable for series production in the series material under series production conditions in the "series process machine molding" - there are no individual castings improvised using a variety of tricks. If the component then goes into series production, even a pilot run or a small series for field testing can still be produced using 3D print cores - albeit at higher costs. In this case, the production of the series core sets can be carried out in parallel. The "hectic pace of the process" of previous part launches is thus eliminated.
1Käfer, S.: Renault druckt Motoren aus, MM Maschinenmarkt, 31.10.2017
 Kleine, M.: Beseelt von Kunststoff, K.: Kompetenz für Konstrukteure, 04/2018, p.18 - 21
3Frank, T., Tuffentsammer, T.: Pumpenindustrie profitiert vom 3-D-Sanddruck, GIESSEREI 105, 03/2018, p. 64-65
4Oaks, M.: Letting Mr. Charpy die : Evaluating the Usefulness of Charpy Impact testing on Ductile Iron, Ductile Iron News Issue, 2, 2012, www.ductile,org