Today, high pressure casting processes account for 60% of aluminum castings and more than half of all non-ferrous metal castings in Germany. However, there are limitations to fully automated processes in terms of part design. After filling and solidification of casting, it is ejected perpendicular to parting plane of mold. Therefore, bottom can only be established by a slide that is removed from casting before demolding. Due to increasing complexity of casting internal geometries caused by combination of multiple components, especially against background of increasing popularity of electric vehicles, new technologies are needed to meet requirements of geometry. Salt core technology provides an alternative to sliding system. Here, salt core is placed in mold before casting and assumes internal geometry of casting. Solidified casting is removed from mold together with salt core, then salt core is removed in a next step. Advantage of salt cores over other core technologies is water solubility, which simplifies removal of salt core. Different types of salt cores and salt materials have been tested in high-pressure casting. Commonly used salt mixture consists of 50% by mass of sodium chloride and 50% by mass of sodium. Production of salt cores can be divided into pressed salt powder or molten salt process. Production of salt cores offers possibility of complex geometries and functional combinations in lightweight components. In addition to process design by simulating salt cores and die casting processes, feasibility of salt cores is also a decisive issue for safe use of salt in the future. Core load level in die casting process mainly comes from flow rate of impact melt. Due to high pressure and high flow rate in die casting process, high mechanical properties are required for core. Semi-solid aluminum alloy casting provides an alternative to traditional die casting. These rheocasting processes increase solid fraction while reducing melt temperature and applying shear forces. Shearing of melt causes the dendrite nuclei to break and diffuse into spherulites in melt, allowing semi-solid melt to remain flowable because dendritic structure inhibits flow. Compared with a fully liquid melt, the higher viscosity reduces casting speed, forms a laminar melt front in mold, solid part of casting process and reduced turbulence allow for improvements in microstructure and gas content of component.
Purpose of this study is to explore whether application of rheocasting technology can be applied to salt cores and improve casting quality at the same time. First, a salt core will be developed and its material mechanical properties data will be checked in bending tests. These data will serve as basis for simulation studies using FLOW-3D fluid-structure interaction (FSI) to consider effects of conventional die casting and rheocasting. After these initial investigations, further research work will be carried out on real castings using parameters corresponding to simulation studies to test salt core durability and evaluate casting quality.

In this study, durability of salt cores in high pressure casting and rheocasting will be examined. Therefore, geometries were developed and tested, effects on salt core were examined in simulation. For this study, salt core geometry and inserts were developed under conditions of existing molds used in later casting tests. Inserts provide a seat for salt core in lateral flow to correspond to maximum loads on salt core, see Figure 3.

Geometry should be simple in order to limit final core destruction effect as accurately as possible. Geometry of salt core is 60 mm in total length, with a square cross section with a side length of 10 mm, a side length of 6 mm in the middle part, and a total length of 20 mm, see Figure 1. Size of salt core is limited by mold. To produce salt core, a mold with two cavities was used. Salt cores for casting tests were produced by gravity casting. Salt mixture NaCl-Na2CO3 (50:50 wt-%) was melted in a laboratory furnace at 700 ℃. Mold material was stainless steel. Mold was preheated to 300 ℃. During casting process, mold temperature was kept constant by a heating plate. Temperature was measured with a thermocouple. Due to risk of thermal cracking of salt core, demoulding time should be as short as possible to about 10 s. After demoulding, salt core was slowly cooled to room temperature on a heating plate and then stored in a closed container.

 

Figure 1 Salt core geometry Extreme core mold

Salt cores were tested in a three-point bending test. Core head in Figure 2 simulates seat during casting, so core is supported in grooves at both ends, which is different from traditional bending test. Test geometry is a triangular tab and a rectangular test specimen. Purpose is to reproduce melt impact. Salt cores were tested at 20 ℃ and 180 ℃ at a speed of 5 mm/min and a trigger force of 1 N. Three batches of 18 specimens were tested. Tests were carried out at 20 and 180 ℃ using a prism (line load) and a surface load (6 mm×8 mm rectangle) at 20 ℃.

 

Figure 2 Three-point bending test

In order to test mold filling and forces acting on salt core during casting process, a simulation model was built using FLOW-3D CAST. Simulation experiments were carried out using liquid and semi-solid metals, see Figure 3. In order to simplify model to basic casting process, only cavity filling of ingates is calculated. For this purpose, initial melt region is used. Boundary conditions on XZ surface of lower mesh block define melt flow, which corresponds to punch speed used in actual process during mold filling phase. Speed is 2 m/s for rheocasting and 6 m/s for conventional die casting. Casting punch is not simulated and is not shown, but casting cake shows corresponding part of punch in casting.
Modeling of rheocasting is based on assumptions made in conventional die casting, so formation of solid phase must be verified experimentally, corresponding to temperature-dependent viscosity according to literature.
This simplification is accepted for technical purposes in order to significantly reduce complexity and calculation time. Convergence control is set to maximum value, number of 500 iterations, time step is controlled by stability and convergence options.

 

Figure 3 Simulation model for fluid-structure interaction (FSI) calculation, mesh in red marked area is full of melt, and only meshes of three mesh blocks are included in calculation

Model is divided in three independent mesh blocks, main block surrounds cavity of step plate with a mesh size of 1 mm, mesh block 2 contains material dry and applied to melt with a mesh size of 5 mm with velocity, and third mesh block surrounds salt core with a total of about 14.1 million meshes.
In order to verify independence of FSI results from mesh size and verify mesh sensitivity, meshing of salt core is tested with sizes of 1 mm, 0.5 mm, and 0.3 mm, respectively, and all blocks use hexahedral meshes with uniform edge lengths. FSI mesh block contains 256 000 meshes (0.3 mm), 55 296 meshes (0.5 mm) and 6 912 meshes (1 mm). Automatic generation of finite elements generates FE mesh on this block according to fluid mesh. Main focus of evaluation using post-processing is stress in salt core during casting. Rankine maximum normal stress theory is used, because it can be assumed that maximum principal stress leads to ideal brittle material behavior and failure. In addition, curvature and forces acting on salt core are evaluated. Mesh block of salt core is considered as a balance volume for energy inside melt. It is considered to be at inlet and outlet. This difference determines amount of energy transferred to core and causes deformation or failure. Purpose is to compare energy with failure energy determined by bending test. Specific energy difference (w) between inflow and outflow flow on inlet and outlet faces of salt core balance volume is calculated. Each energy conversion is calculated by multiplying mass flow recorded on inlet face of balance volume, see Figure 4.

 

Figure 4 Details of the salt core volume balance

Analytical model conditions are shown in Table 1. Temperature-dependent material properties of alloy A356 (AlSi7Mg0.3) were calculated in JMatPro10.1 software from spectroscopic measurements. Material data, mainly density, solid fraction, specific heat and thermal conductivity, were exported to FLOW-3D CAST format. Temperature and shear rate-dependent data of viscosity of semi-solid alloy were obtained from rheological studies, see Table 2. A constant viscosity of 0.0019 Pa⋅s was assumed for liquid material and data were taken from FLOW-3D database of alloy. Standard values of heat transfer coefficients were taken from software database. HTCs of salt core between liquid and solidified metal were 600 and 300 W/(m2·K), respectively, between core and mold were 500 W/(m2·K). Feed of casting material was applied as a constant velocity boundary condition through lower XZ surface of grid block 2, see Figure 3. Considering energy change of salt core, an isothermal flow process within equilibrium volume of salt core is assumed. Since equilibrium area is relatively small, according to calculation, no temperature loss occurs. Density of melt is considered constant, and its value is determined in simulation according to respective temperature.

Table 1 FLOW-3D calculation model

 

Table 2 Strain rate and viscosity of A356 at 600 ℃
For rheocasting at 600 ℃, density is 2.49 kg/m3, and for general liquid fluid at 627 ℃, density is 2.44 kg/m3. During simulation, velocity and local pressure at inlet and outlet restrictions of salt core equilibrium volume are output for each time step. According to pouring direction, lower interface is called inlet and upper interface is called outlet. Direct inflow velocity is about 4.5 m/s, and conventional die casting is 15 m/s.
Variables studied that have an impact on both casting quality and salt core load are punch speed and melt temperature. Conventional casting parameters are represented by a punch speed of 6 m/s and a melt temperature of 630 ℃ (corresponding to a solid fraction of 0% in melt). A lower punch speed of 2 m/s paired with a melt temperature of 600 ℃ represents a rheocasting treatment (corresponding to a solid fraction of 24% in melt). Stirring with an SSR system always results in melt cooling to solidification zone. Salt core temperature of 180 ℃ and mold temperature of 300 ℃ were kept constant in all simulations.

Following contains results of three-point bending tests and numerical sensitivity analysis studies of salt core load.

Main purpose of preliminary tests was fracture load and deformation at fracture. Material behavior of salt core under bending load can be described as brittle and purely elastic. Changes in test temperature lead to differences in fracture load and elongation at fracture, see Figure 5.
Results show that higher salt core temperatures lead to lower fracture loads of specimens. Hydration may play a role, since hydrates only form anhydrous Na2CO3 when heated above 109 ℃ and may destroy microstructure. As expected, compressive strength increases slightly at higher temperatures. Full load test shows higher fracture loads and elongations at break, as well as larger differences due to cast surface of salt core. Therefore, no tests at higher temperatures were performed.
It can be assumed that under bending load, critical stress on tensile side leads to core fracture. Fine-grained microstructure of edge region of 0.8-1.0 mm thickness, see Figure 6, therefore plays a decisive role in bending strength. It is also shown that edge fibers of salt core have the highest strength.

 

Figure 5 Fracture strength and compression produced by three-point bending test

 

Figure 6 Cross-section of salt core (6×6 mm) magnified 40 times

Before simulation study, three mesh sizes of salt core (1, 0.5 and 0.3 mm) were tested. Stress distribution is shown in Figure 7. Fine mesh shows clearer local stress peaks, but there is a significant difference in calculation time (about 4h for a 0.5 mm mesh; about 24h for a 0.3 mm mesh). In all following simulations, it is considered reasonable to use a 0.5 mm salt core mesh. During mold filling, changes in pressure and velocity occur in equilibrium volume, see Figure 8, resulting in changes in internal energy of system. Forces and energy conversion at salt core can be approximately determined. In Figure 9, simulated salt core loads for high pressure die casting and rheocasting are shown. Calculated loads are highest at the moment of impact. Specific flow rate (15 m/s; 4.5 m/s rheocasting) leads to significantly different salt core loads at this time. During further mold filling, forces acting on salt core remain approximately unchanged. Ignoring compressive forces acting on core during full filling or holding stage, results show that forces during rheocasting are well below failure limit. However, it must be noted that deformation speed in three-point bending test is 5 mm/min, while during casting, deformation occurs within about 0.001 s.

 

Figure 7 Grid sensitivity analysis of stress when melt impacts salt core

 

Figure 8 Balanced volume of salt core for energy consideration (purple)

 

Figure 9 Effect of melt on salt core during simulated filling process

Rankine theory is used to compare true triaxial stress state occurring during casting with tensile test values given in literature. This is applicable to materials where brittle fracture perpendicular to maximum principal stress is expected. Maximum normal stress, called maximum principal stress, occurs at the first impact of melt, see Figure 10. Gate speed is 6 m/s, flow rate is about 15 m/s, and maximum stress is 13 MPa immediately after melt hits salt core. Yield strength of salt mixture used is about 11 MPa. Therefore, it can be assumed that salt core will break during general high-pressure casting process. Crack occurs on upper side of core in thin cross-sectional area. Brittle fracture occurs after a slight deflection perpendicular to longitudinal axis of core, with a calculated displacement of 0.17 mm in direction of melt flow, Figure 11. This result is comparable to results of the three-point bending test. Therefore, at a core temperature of 20 °C, displacement measured in the middle of salt core is 0.172 mm, and at a core temperature of 180 °C it is 0.207 mm. Another simulation run was performed using semi-solid material (fs=23%) and a punch speed of 2 m/s. Compared with conventional die casting process parameters, maximum stress after melt hits salt core is significantly reduced, see Figure 12. Maximum stress on lower side of salt core is about 2.3 MPa. It should be noted that for this simulation, it is assumed that salt core material is homogeneous and has a yield strength of 11 MPa throughout sample cross section.

 

Figure 10 Stress state of salt core surface during the first impact of melt, flow rate before impact was 15 m/s, maximum stress on upper core side was about 13 MPa, and bottom was 4.8 MPa

 

Figure 11 Deformation of salt core during the first impact of melt, displacement in the middle of salt core was close to 0.17 mm in Z direction

 

Figure 12 Stress state of salt core surface after semi-solid melt impacted, flow rate before impact was 4.55 m/s, and maximum stress on lower core side was about 2.3 MPa

In this study, boundary conditions of aluminum die-casting salt core with general parameters of die casting and rheocasting parameters were studied, effects of conventional die casting and improved die casting and semi-solid material processing were evaluated through simulation models. FSI simulation results show that the higher load of impact melt, the faster punch speed, and the higher flow rate of melt. At a punch speed of 6 m/s, stress on upper side of salt core is about 13 MPa. At a punch speed of 2 m/s, maximum stress was only 2.3 MPa and yield strength of cast salt core with given salt mixture was 11 MPa. Therefore, at high casting speeds of conventional die casting, salt core fracture is likely to occur and yield strength may increase locally. Rheological parameters show a reduction in salt core loading and tend to increase salt core viability. Further studies are needed to verify comparison of real experiments performed in foundry laboratory using die casting and rheocasting tests.