About the impact on gravity cast salt cores in high pressure die casting and rheocasting

Author links open overlay panelAndreasSchillingDanielSchmidtJakobGlückNiklasSchwenkeHusamSharabiMartinFehlbierShow moreAdd to MendeleyShareCite

https://doi.org/10.1016/j.simpat.2022.102585Get rights and content

Under a Creative Commons license

Open access


In this work, a simulations study on the impact on gravity cast salt cores was carried out for the different casting parameters in high pressure die casting as well as in rheocasting. To compare the simulation results, salt cores were created in test casts and characteristic values were determined by means of a three-point bending test. The calculated loads on the cores were determined and aligned with the bending strength. Salt cores in rheocasting showed reduced loads and tend to increase core viability compared to conventional hpdc.


Cast salt cores

Lost core

Core damage




Core testing

High pressure die casting

1. Introduction

The high pressure die casting process today accounts for 60% of aluminium castings and more than half of all non-ferrous castings in Germany. However, there are limits to the fully automatable process in terms of component design. After the casting process and solidification of the component, it is ejected perpendicular to the parting plane of the mould. Therefore, undercuts can only be created with sliders, which are removed from the casting before demoulding [1]. Rising casting complexity with internal geometries resulting from integration of several components, especially in the context of increasing e-mobility, can therefore only be realized with very complex and expensive moulds. New technologies are therefore required to meet the requirements of the geometries. The lost salt core technology offers an alternative to slide systems [2]. The cores used here are inserted into the mould tool before the casting process and reproduce the internal geometry of the casting. The solidified casting is removed from the mould together with the core and the core is removed in a further step. The advantage of salt cores over other core technologies is the water solubility, which simplifies the removal of the salt core [2]. Different types of salt cores and salt materials have been tested in high pressure die casting [[3][4][5]]. The commonly used salt mixture consists of 50 wt-% sodium chloride and 50 wt-% sodium.

Manufacturing of lost salt cores can be divided in pressing salt powders or casting processes with molten salts. Casting of salt cores offers complex geometries and the possibility of functional integration in lightweight components. Core viability is a decisive issue for the future process-safe use of salt in addition to process design through simulation of the cores and the die casting process [6]. The level of core loading in the die casting process is mainly attributed to the flow velocity of the impinging melt [3,7]. Due to the high pressures and flow speeds in the die casting process, high mechanical demands are set on the cores.

The casting of semi-solid aluminium alloys offers an alternative to conventional die casting. These rheocasting processes lower the melt temperature to induce solid fraction with simultaneous application of shear forces. The shearing of the melt causes dendritic nuclei to be broken and spread as globulites in the melt. As a result, the semi-solid melt remains flowable, since dendritic structures can inhibit the flow and the viscosity decreases under shear forces [8]. The higher viscosity compared to completely liquid melt results in lower casting speeds and a laminar melt front in the mould. The solid fraction and the reduced turbulence of the casting process lead to an improvement of the microstructure and the gas content of the components.

The aim of this study is to investigate whether the application of rheocasting technology can be applied to salt cores and simultaneously improve component quality. In the first step, a salt core will be developed which will be examined for its failure-critical material data in bending tests. These data will serve as a basis for a simulation study using Flow-3D fluid-structure-interaction (FSI) simulations to consider the impacts in conventional die casting as well as for rheocasting. After completion of these preliminary investigations, further research work will be carried out on real castings with the parameters corresponding to the simulation study in order to test the core durability and evaluate the component quality.

2. Experimental setup

In this study the core durability in high pressure die casting as well as in rheocasting is to be examined. For this purpose, a test geometry is developed, tested and the impact on the core is examined in simulations. For this study a salt core geometry and an insert were developed under the condition of using an existing die in the later casting trials. The insert provides the bearing of the salt core in transverse flow to map the maximum loads on the core, c.f. Fig. 3.

2.1. Salt core test geometry

The geometry should be simple in order to limit the final core-damaging effect as precisely as possible. The geometry of the core has a total length of 60 mm with a square cross section and a side length of 10 mm. The middle part has a side length of 6 mm and a total length of 20 mm, c.f. Fig. 1. The dimensions of the core are limited by the tool. To produce the salt cores, a die with a double cavity is used.

Fig. 1. Salt core geometry and casting die.
Fig. 1. Salt core geometry and casting die.

The salt cores for the casting tests are produced by gravity die casting. The salt mixture NaCl-Na2CO3 (50:50 wt-%) is melted in a laboratory furnace at 700 °C. The castings are made by hand with a stainless steel ladle. The steel mould is preheated to a temperature of 300 °C. The mould temperature is kept constant during casting by a heating plate. The temperature is measured with a thermocouple type K. Due to the risk of hot cracking of the salt cores, the demoulding time is set as short as possible to approx. 10 s. The cores are slowly cooled down to room temperature on a heating plate and afterwards stored in closed containers.

2.2. Core testing

The cores are tested in a modified three-point bending test, c.f. Fig. 2. The core holder simulates the bearing in the casting tool so that the core is supported in a groove at both ends, in contrast to a conventional bending test. The test geometry is a triangular fin and a rectangular test specimen which is intended to reproduce the melt impact. The salt cores are tested at 20 °C and 180 °C with a speed of 5 mm/min and a trigger force of 1 N. Three batches of 18 specimens each were tested. The tests were carried out with prism (line load) at 20 and 180 °C as well as with area load at 20 °C (6 × 8 mm rectangle).

Fig. 2. Modified three-point bending test.
Fig. 2. Modified three-point bending test.

3. Simulation model

To determine the mould filling and the forces acting on the salt core in advance of the casting tests, a simulation model is built up with Flow-3D Cast Version 5.0.1. Simulation experiments are carried out with both liquid and semi-solid metal, c.f. Fig. 3.

Fig. 3. Simulation model for the fluid structure interaction (FSI) calculation. Cells in the red marked area are filled with melt. Only cells within the three mesh blocks are included in the calculation.
Fig. 3. Simulation model for the fluid structure interaction (FSI) calculation. Cells in the red marked area are filled with melt. Only cells within the three mesh blocks are included in the calculation.

In order to reduce the model to the essential casting process, only the mould filling from the gate is calculated. For this purpose, an initial melt area is used. The boundary condition on the XZ-face of the lower mesh block defines the melt flow. This corresponds to the piston speed that is used in the real process during the mould filling phase. The speed is 2 m/s for rheocasting and 6 m/s for conventional die casting. The casting piston is not simulated and not shown, but the casting biscuit shows the counterpart of the piston in the casting.

The modelling of the rheocasting is done according to the assumptions in conventional die casting, so that the formation of the solid phase, which has to be verified within the experiments, is mapped via the temperature-dependant viscosity according to the literature. This simplification is accepted for technical purposes due to significant reduced complexity and calculation times. The convergence controls are set a max. number of 500 iterations. The time steps are controlled via stability and convergence option.

3.1. Meshing and energy analysis

The model is meshed in three separate meshing blocks. The main meshing block encloses the cavity of the step plate and has a cell size of 1 mm. Mesh block 2 encloses the biscuit and serves to apply the velocity to the melt with a cell size of 5 mm. The third mesh block encloses the salt core. The total number of cells is about 14.1 million. In order to verify the independence of the FSI results from the cell size and validate the grid sensitivity, the meshing of the salt core was separately tested with a size of 1 mm, 0.5 mm and 0.3 mm. A hexahedral mesh with uniform side lengths is used for all blocks. The FSI-mesh block contains 256,000 cells (0.3 mm), 55,296 cells (0.5 mm) or 6912 cells (1 mm). The automatic finite element usage generates a FE-mesh on this block from the flow mesh.

The main focus of the evaluation done with FlowSight is on the stress of the salt core during the casting process. The Rankine maximum-normal stress theory is used, since almost ideally brittle material behaviour and failure due to the greatest principal stress can be assumed. Furthermore, the curvature as well as the forces acting on the core are evaluated. The mesh block of the core is considered as a balance volume for the internal energy of the melt, which is considered at entry and exit. The difference determines the amount of energy that is transferred to the core and causes deformation or failure.

The aim is to compare the amount of energy with the failure energies determined from the bending tests. The specific energy difference w between incoming and outgoing flow on the inlet and outlet faces of the core balance volume is calculated. The energy conversion per time is calculated by multiplying with the mass flow recorded on the inlet face of the balance volume, c.f. Fig. 4.

Fig. 4. Detail of the core balance volume
Fig. 4. Detail of the core balance volume

3.2. Material data and boundary conditions

Table 1.

Table 1. List of calculation models used in Flow-3D.

GravityZ: −9,81 m/s
Temperature calculationEnabled, Fluid to solid heat transfer
Gas modeAdiabatic gas regions
DensityFunction of temperature
Cavitation potentialEnabled
Defect trackingFully enabled
Viscous flowThixotropic viscosity model
TurbulenceRNG model
SolidificationEnabled, no Shrinkage
AdvectionOne fluid, free surface
Flow calculationMomentum and continuity equations

The temperature-dependant material properties for the alloy A356 (AlSi7Mg0,3) are calculated in the software JMatPro 10.1 on the basis of an emission spectrometric measurement. In particular, the material data for density, solid fraction, specific heat capacity and thermal conductivity are exported to a Flow-3D Cast dataset. The temperature and shear rate dependant data for the viscosity of the semi-solid alloy are obtained from rheological investigations in [9], c.f. Table 2.

Table 2. Strain rate dependant viscosity of A356 at 600 °C obtained from [9].

Shear rate (1/s)Viscosity (Pa·s)

A constant viscosity of 0.0019 Pa·s is assumed for the liquid material, with data taken from the Flow-3D integrated data set of the alloy. Material properties of the salt core are obtained from [4] and [10].

Standard values of heat transfer coefficients are taken from the database, the HTC of the core between liquid and solidified metal is 600 and 300 W/m²K respectively, between core and mould 500 W/m²K. The movement of the casting material is applied as a constant velocity boundary condition via the lower XZ-face of mesh block 2, c.f. Fig. 3.

For the consideration of the energy change at the salt core an isothermal flow process in the core balance volume is assumed, since the balance area is relatively small and according to calculations, no temperature loss occurs. The density of the melt is considered constant. Its value was determined in simulations according to the respective temperature. For rheocasting at a temperature of 600 °C, the density is 2.49 kg/dm3 and for the conventional variant at 627 °C 2.44 kg/dm3. The velocities as well as the local pressures at the inlet and outlet limits of the core balance volume are output for each time step during the simulation. According to the casting direction, the lower interface is referred to as inlet and the upper as outlet. The direct inflow velocities are approx. 4.5 m/s for rheocasting and approx. 15 m/s for the conventional die casting.

The investigated variables, which have an effect on both the component quality and the salt core load, are the piston velocity and the melt temperature. The conventional casting parameters are represented by a piston velocity of 6 m/s and a melt temperature of 630 °C (corresponding fraction solid in the melt 0%). A lower piston velocity of 2 m/s paired with a melt temperature of 600 °C represents a rheocasting treatment (corresponding fraction solid in the melt 24%). Stirring with the SSR system always results in cooling of the melt to the solidification interval. The core temperature of 180 °C and the die temperature of 300 °C are kept constant in all simulations.

4. Results and discussion

The following section contains the results of the investigations of the numerically determined salt core loads and the three-point bending tests.

4.1. Breaking strength and deformation in the three-point bending test

The main focus of the preliminary tests is on the breaking load and the deformation at the time of breakage. The material behaviour of the core under bending load can be described as brittle and purely elastic. The variation of the test temperature results in differences in breaking load and elongation at break, c.f. Fig. 5.

Fig. 5. Breaking strength and compression resulting from three-point bending tests.
Fig. 5. Breaking strength and compression resulting from three-point bending tests.

It is shown, that a higher core temperature leads to a decreasing breaking load with triangular fin. This trend correlates with [10]. Hydration may play a role, as the hydrates only form anhydrous Na2CO3 when heated above 109 °C [11] and may damage the microstructure. As expected, the compression increases slightly at higher temperatures. The full load test shows a higher breaking load and breaking elongation with higher variance resulting from the cast surface of the cores. For this reason, testing at higher temperatures was not carried out.

It can be assumed that under bending load the critical stress on the tensile side leads to fracture of the core. The fine-grained microstructure of the approx. 0.8 - 1.0 mm thick edge zone, c.f. Fig. 6, should therefore be decisive for the bending strength. [10] also shows that the edge fibre of the cast salt cores has the highest strengths.

Fig. 6. Sectional view of the salt core (6 × 6 mm) at 40x magnification.
Fig. 6. Sectional view of the salt core (6 × 6 mm) at 40x magnification.

4.2. Calculated core load

In advance of the simulation study, three mesh sizes (1, 0.5 and 0.3 mm) of the salt core were tested and the stress distribution is shown in Fig. 7. The fine mesh variants show clearer local stress peaks, but differ significantly amongst themselves in the computation time (0.5 mm mesh about 4 h; 0.3 mm mesh about 24 h). The use of the 0.5 mm meshing of the salt core is considered reasonable in all following simulations.

Fig. 7. Mesh dependency of the stresses in the salt core after the impact of the melt.
Fig. 7. Mesh dependency of the stresses in the salt core after the impact of the melt.

During mould filling, changes in pressure and velocity occur in the balance volume, c.f. Fig. 8, resulting in a change in the internal energy of the system. The forces and the energy conversion at the salt core can be determined approximately.

Fig. 8. Balance volume of the salt core for energy consideration (purple).
Fig. 8. Balance volume of the salt core for energy consideration (purple).

In Fig. 9 the simulated core load for high pressure die casting and rheocasting are shown.

Fig. 9. Simulated impact of the melt on the salt core.
Fig. 9. Simulated impact of the melt on the salt core.

The calculated load is highest at the first impact. The specific flow velocities (15 m/s hpdc; 4.5 m/s rheocasting) result in significantly different core loads at this point. During further mould filling, the forces acting on the core remain approximately constant. Compressive forces acting on the core at the time of complete mould filling or during the holding pressure phase are not considered. The results show that the forces during rheocasting are well below the failure limit. Nevertheless, it must be noted that the deformation speed in the three-point bending test is 5 mm/min and that in the casting process the deformation occurs within about 0.001 s.

4.3. Maximum-normal stress theory to predict core damage during mould filling process

The Rankine theory is used to compare the real three-axis stress state that occurs during casting with the tensile test values given in literature. This is suitable for materials where a brittle fracture perpendicular to the greatest principal stress is expected. The maximum-normal stress, named as max principal stress, occurs after the first impact of the melt is shown in Fig. 10. With an ingate velocity of 6 m/s and a flow velocity of approximately 15 m/s, tensions up to 13 MPa occur immediately after the melt hits the core. The used salt mixture has a tensile strength of approximately 11 MPa [10]. Therefore, it can be assumed that the core will break during casting with conventional parameters. The crack initiation occurs on the upper side of the core in the thin cross-sectional area. The brittle fracture takes place after a slight deflection perpendicular to the longitudinal axis of the core.

Fig. 10. Stress condition on salt core surface immediately after the first impact of the melt. The flow velocity before impact is 15 m/s. The maximum stress on the upper core side is approximately 13 MPa and on the bottom side 4.8 MPa.
Fig. 10. Stress condition on salt core surface immediately after the first impact of the melt. The flow velocity before impact is 15 m/s. The maximum stress on the upper core side is approximately 13 MPa and on the bottom side 4.8 MPa.

The calculated displacement in melt flow direction is 0.17 mm, c.f. Fig. 11. This result is comparable to those of the of the three-point bending test. Therefore, the measured displacement in the middle of the salt core is 0.172 mm at a core temperature of 20 °C and 0.207 mm at a core temperature of 180 °C.

Fig. 11. Deformation of the salt core after first impact of the melt. The displacement in the middle of the core is in direction Z nearby 0.17 mm.
Fig. 11. Deformation of the salt core after first impact of the melt. The displacement in the middle of the core is in direction Z nearby 0.17 mm.

Another simulation run was performed with semi solid material (fs = 23%) and a piston speed of 2 m/s. The maximum stress right after the melt hits the salt core is significantly lower compared with the conventional die casting process parameters, c.f. Fig. 12. The maximum stress on the lower side of the core is approximately 2.3 MPa.

Fig. 12. Stress condition on salt core surface immediately after the first impact of the semi solid melt. The flow velocity before impact is 4.55 m/s. The maximum stress on the lower core side is approximately 2.3 MPa.
Fig. 12. Stress condition on salt core surface immediately after the first impact of the semi solid melt. The flow velocity before impact is 4.55 m/s. The maximum stress on the lower core side is approximately 2.3 MPa.

It should be noted that for this simulation, the salt core material is assumed to be homogeneous and has a tensile strength of 11 MPa in the entire sample cross section.

5. Conclusion

In this study, the boundary conditions for lost salt cores for aluminium die casting with conventional as well as rheocasting parameters were investigated. By means of a simulation model the impact during processing in conventional die casting as well as in modified die casting and the processing of semi-liquid materials was evaluated. The results of the FSI simulation show that the load from the impacting melt is higher, the faster the piston speed and the higher the flow rate of the melt. At a piston speed of 6 m/s, the tension on the upper side of the salt core is around 13 MPa. At a piston speed of 2 m/s, the maximum stress is only 2.3 MPa. The tensile strength of casted salt cores with the given salt mixture is 11 MPa. Accordingly, a core failure at high casting speeds of conventional die casting and a local increase of the tensile strength is likely. The rheocasting parameters show reduced loads on the core and tend to increase core viability. Further research is needed to verify the simulation in real experiments using die casting and rheocasting trials in the casting laboratory.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.



B. Nogowizin

Theorie Und Praxis des Druckgusses

Fachverlag Schiele&Schön, Berlin (2011)

Google Scholar[2]Jelínek, P.; Adámková, E.: Lost Cores for High-Pressure Die Casting, Archives of Foundry Engineering, 14(2) 101–104, 2014, doi:10.2478/afe-2014-0045.

Google Scholar[3]

B. Fuchs

Salzkerntechnologie Für Hohlgussbauteile im Druckguss

Cuvillier Verlag, PhD, Erlangen, Nürnberg (2014)

Google Scholar[4]

D. Pierri, C. Beck

Lost Core-Technologie – offen für alle, Chancen und Grundlagen des Verfahrens

In: Giesserei, Jg, 103 (2016)

10/2016, DVS Media GmbH, Düsseldorf

Google Scholar[5]

J. Yaokawa, D. Miura, K. Anzai

Strength of Salt Core Composed of Alkali Carbonate and Alkali Chloride Mixtures Made by Casting Technique

Mater. Trans., 48 (5) (2007), pp. 1034-1041 View PDF

CrossRefView Record in ScopusGoogle Scholar[6]Schilling, A. et al.: Approach on simulation of solidification and shrinkage of gravity cast salt cores, Simulation Modelling Practice and Theory, 107, 102231, 2021, ISSN 1569-190X, doi:10.1016/j.simpat.2020.102231.

Google Scholar[7]Kohlstädt, S.; Vynnycky, M.; Goeke, S.: On the CFD Modelling of Slamming of the Metal Melt in High-Pressure Die Casting Involving Lost Cores, Metals 2021, 11(1), 78, doi:10.3390/met11010078.

Google Scholar[8]Uggowitzer, P.J.; Gullo, G.; Wahlen, A.: Metallkundliche Aspekte bei der semi-solid Formgebung von Leichtmetallen, In: Vom Werkstoff zum Bauteil, ed. H. Kaufmann und P.J. Uggowitzer, LKR-Verlag Ranshofen, 95–107, 2000.

Google Scholar[9]

P. Das

Rheological Characterization of Semi-Solid A356 Aluminium Alloy, In: Solid State Phenomena, Vols. 192-193

Trans Tech Publications, Switzerland (2013), pp. 329-334

View Record in ScopusGoogle Scholar[10]Fickel, P.: Hohl- und Verbundguss von Druckgussbauteilen – Numerische Auslegungsmethoden und experimentelle Verifikation, PhD, Stuttgart, 2017.

Google Scholar[11]

R.K. Freier

Aqueous Solutions

Data Inorg. Org. Compd., 2 (1978)

Supplements, Gruyter, Berlin

Google Scholar