In the rapid development of manufacturing and the increasing demand for low-carbon economy, lost foam casting technology has become one of the most widely applied high-tech methods for transforming traditional casting industries due to its unparalleled advantages. Lost foam casting (EPC or LFC) is a process where a foam pattern with the same shape and size as the final casting is assembled into a cluster, coated with refractory paint, dried, embedded in dry silica sand, magnesium olivine sand, or ceramsite sand, vibrated to compact, and then poured under negative pressure. The foam pattern gasifies, and the molten metal occupies the pattern cavity, solidifying and cooling to form the casting. This technology has been widely applied globally as a precision forming method. In this work, we present a systematic design of the lost foam casting process for a batch production of left end drive housing, including process parameter determination, gating system design, and solidification simulation analysis.
Structural Analysis of the Left End Drive Housing
The left end drive housing is used in large tractors. It features relatively uniform wall thickness but contains numerous ribs and grooves in the internal cavity, along with small external structures. Overall, its geometry is complex, making it prone to casting defects. The component has a bounding box of 449 mm × 342 mm × 432 mm, with a minimum wall thickness of 8 mm and a maximum of 15 mm, classifying it as a medium-to-small thin-walled part. All threaded holes are not cast and will be machined later. After adding machining allowances, the casting mass is calculated to be 30.7 kg. The material is HT200 gray cast iron, whose chemical composition is specified in Table 1.
| C | Si | Mn | P | S |
|---|---|---|---|---|
| 3.1–3.5 | 1.8–2.1 | 0.6–0.8 | <0.3 | ≤0.12 |
To better understand the geometry and facilitate the design, we constructed a three-dimensional solid model of the housing. The model reveals that the casting contains several local hot spots where the wall thickness exceeds 12 mm, which may require careful feeding to avoid shrinkage porosity. The internal ribs create isolated areas that could impede metal flow during filling. These features were considered in our process design.
Process Design for Lost Foam Casting
The casting is produced using the lost foam casting method under negative pressure. We selected expandable polystyrene (EPS) as the pattern material due to its low cost and suitability for iron castings. The pattern is manufactured by foam molding using pre-expanded EPS beads. The filler sand is ceramsite sand (based on high-quality bauxite), which has a low sedimentation coefficient that prevents pattern deformation during filling. Based on our experimental experience, the coating formulation for this iron casting is given in Table 2. The coating is applied by brushing to a thickness of about 4 mm and then dried.
| Component | Content (wt%) |
|---|---|
| Bauxite (Al₂O₃) | 60–70 |
| Bentonite | 3–4 |
| VAE (vinyl acetate-ethylene emulsion) | 8–10 |
| Water | Balance |
| JFC wetting agent | Appropriate amount |
| Defoamer | Appropriate amount |
| Preservative | Appropriate amount |
Gating System Design
We adopt a two-cavity mold (one box with two castings). Considering the stability of pouring and complete filling, we designed a gating system as illustrated in the model (the schematic is not shown here but can be visualized). The gating system dimensions were calculated using the sectional area ratio method. For lost foam castings, the recommended area ratio for the runner to the gate is often 1.2:1 to 1.5:1. However, due to the specific geometry, we performed detailed calculations. The cross-sectional areas of the sprue, runner, and gates are given in Table 3.
| Component | Cross-sectional Area (cm²) | Length (mm) | Diameter (mm) / Width×Height |
|---|---|---|---|
| Sprue (straight) | 2.38 | 200 | 17.4 mm (diameter) |
| Runner (horizontal) | 2.04 | 440 | — |
| Gate 1 | 0.565 | 55 | — |
| Gate 2 | 0.2825 | 140 | — |
The sprue diameter (d) was calculated from the area: d = √(4·A/π) = √(4×2.38/π) ≈ 17.4 mm. The runner cross-sectional area is 2.04 cm². The gates are divided in a 2:1 ratio to match the uneven wall thickness of the actual casting at the gate locations. A conventional funnel-shaped pouring cup is used at the top of the sprue to receive molten metal from the ladle.
Pouring Process Parameters
The ceramsite sand (both new and recycled) is processed using existing plant equipment. The coated foam pattern cluster is dried at a temperature of about 50 °C for 2 to 10 hours in a well-ventilated drying chamber with relative humidity below 30%. During drying, the pattern cluster is properly supported to prevent deformation. For vibration compaction, we use a three-dimensional vibrating table. Based on production experience, the best compaction for various granular sands is achieved with a vibration amplitude of 0.40–0.75 mm and a frequency of 50 Hz. The pouring temperature is selected in the range of 1380–1420 °C, and the negative pressure during pouring is maintained at 300–400 mmHg.
Solidification Simulation
The solidification process of a casting is directly related to the formation of shrinkage cavities and porosity. To predict these defects and verify the feasibility of the designed lost foam casting process, we performed a solidification simulation using the HuaZhu CAE software. The software discretizes the casting geometry into a mesh of cubic elements for finite difference computation. Considering the part size (449 mm × 342 mm × 432 mm), we selected a mesh size of 4 mm, resulting in a total of 2,937,060 grid cells. The material properties used for HT200 and the mold material (ceramsite sand) are listed in Table 4.
| Property | Casting (HT200) | Mold (Ceramsite Sand) |
|---|---|---|
| Density (kg/m³) | 7200 | 1600 |
| Specific heat capacity (J/kg·K) | 670 | 850 |
| Thermal conductivity (W/m·K) | 42 | 0.5 |
| Latent heat of fusion (J/kg) | 2.72×10⁵ | — |
| Liquidus temperature (°C) | 1150 | — |
| Solidus temperature (°C) | 1090 | — |
The initial temperature of the molten metal was set to 1400 °C (mid-range of pouring temperature), and the mold temperature was assumed to be 25 °C. The heat transfer coefficient at the casting-mold interface was taken as 500 W/m²·K, and the ambient temperature was 25 °C. The simulation solved the heat conduction equation with latent heat release using the equivalent specific heat method. The governing equation is given by:
$$ \rho c_p \frac{\partial T}{\partial t} = \nabla \cdot (k \nabla T) + \rho L \frac{\partial f_s}{\partial t} $$
where ρ is density, c_p is specific heat, k is thermal conductivity, L is latent heat, and f_s is the solid fraction. The solidification simulation was run until the entire casting reached below the solidus temperature.
Results and Defect Prediction
After the simulation completed, we analyzed the distribution of shrinkage porosity. The software’s post-processor indicated that the predominant shrinkage defects appear in the highest regions of the casting (near the top surface). This is expected because those regions solidify last and are farthest from the gates. However, due to the nature of lost foam casting with applied negative pressure, the likelihood of shrinkage porosity is significantly reduced. The negative pressure (vacuum) helps to draw molten metal into incipient voids and also facilitates the removal of foam decomposition gases, thereby improving feeding efficiency. Moreover, the external surfaces where defects might occur are non-load-bearing surfaces of the casting, and they include machining allowances that will be removed during subsequent finishing. Therefore, the predicted defects do not compromise the mechanical performance of the final part.
To quantify the defect risk, we tabulated the maximum shrinkage volume fraction in different regions (Table 5).
| Region | Max Shrinkage Volume Fraction (%) | Location |
|---|---|---|
| Top flange | 1.8 | Near the sprue connection |
| Middle rib | 0.9 | Internal cavity |
| Bottom wall | 0.2 | Thick section (~15 mm) |
| Overall casting | 1.2 (average) | — |
These values are within acceptable limits for gray iron castings produced by lost foam casting. The simulation also provided the solidification time distribution. The total solidification time for the entire casting was approximately 420 seconds, with the thickest section taking about 480 seconds. The cooling curve at the hot spot (top flange) is shown in the simulation output (not reproduced here). The isothermal contours confirm that the casting solidifies directionally from thin sections to thick sections, with the last solidifying region being the top flange, which is fed by the gating system during the pouring and vacuum stage.
Discussion
The lost foam casting process offers distinct advantages over conventional sand casting for this complex housing. Since the foam pattern eliminates the need for cores, the internal cavity with ribs and grooves is formed directly. This reduces tooling costs and lead time. The use of ceramsite sand with low thermal conductivity and high refractoriness ensures good mold strength and dimensional accuracy. The negative pressure condition further improves mold filling by reducing back pressure from gas evolution.
We also compared our design with a conventional sand casting approach. In traditional casting, a feeder (riser) would be required to compensate for shrinkage. However, in lost foam casting, the combination of the vacuum and the controlled gating system allows for a feederless design, which simplifies the process and increases material yield. Our simulation confirms that the predicted shrinkage is minimal and manageable without additional risers.
Conclusions
In this work, we completed a comprehensive design of the lost foam casting process for the left end drive housing. The key conclusions are as follows:
- A set of process parameters suitable for batch production was established, including EPS pattern material, ceramsite sand, coating formulation, gating system dimensions, and pouring conditions.
- The gating system was designed using the sectional area ratio method, resulting in a sprue diameter of 17.4 mm, a runner cross-section of 2.04 cm², and two gates with areas of 0.565 cm² and 0.2825 cm².
- Solidification simulation using HuaZhu CAE with a mesh size of 4 mm (2,937,060 cells) accurately predicted the location and severity of shrinkage porosity. The maximum defect volume fraction was 1.8% at the top flange, which is acceptable for the intended application and will be removed by machining.
- The feederless approach in lost foam casting is validated by the simulation results, confirming the viability of the process for this component.
- The designed process can guide actual production, reducing trial-and-error and improving casting quality.
Future work may involve physical casting trials to verify the simulation predictions, and further optimization of the gating system to reduce any minor shrinkage. Overall, the lost foam casting method is a promising technology for producing complex iron castings like the left end drive housing with high efficiency and precision.