In this work, I present a comprehensive sand casting process design and numerical simulation for a medium-sized aluminum alloy reducer box. The component, made of AlSi7Mg0.3, is produced in large batches and features complex multi-surface and multi-hole geometry with uneven wall thickness. After carefully analyzing the structural characteristics and manufacturability, I adopted sand casting combined with cold core box core assembly molding. The gating and riser systems were designed, and the casting process was simulated using AnyCasting software to predict and eliminate sand casting defects. The study demonstrates how proper process design and simulation can effectively reduce sand casting defects, ensuring high-quality castings.
1. Introduction
Reducer boxes are typical complex box-shaped parts with high precision and performance requirements. Due to numerous dimensional features, uneven wall thickness, and intricate internal cavities, they are prone to various sand casting defects such as shrinkage porosity, gas holes, cold shuts, and misruns. To minimize these sand casting defects, a well-designed casting process and advanced simulation tools are essential. In this project, I focused on a reducer box with overall dimensions of 590 mm × 475 mm × 810 mm and a nominal wall thickness of 12–15 mm. The material is AlSi7Mg0.3, a common aluminum alloy with good castability and mechanical properties. The production volume is large, necessitating a reliable and repeatable sand casting process.
2. CAD Modeling and Structural Analysis
I built a three-dimensional solid model of the reducer box using UG (Unigraphics) software. The model includes all geometric features: external contours, internal ribs, bosses, flanges, and mounting holes. A machining allowance was added to all surfaces requiring subsequent finishing. The internal structure consists of a thin rectangular plate (approximately 8 mm thick) that connects the walls. This thin section is particularly vulnerable to sand casting defects because it solidifies later than the thicker walls, potentially leading to insufficient feeding and shrinkage porosity. Understanding such critical regions is vital for effective defect prevention.
3. Sand Casting Process Design
3.1 Molding and Core Making
Given the large batch size and complex internal cavities, I selected sand casting using phenolic modified furan resin self-hardening sand. The mold was produced by manual molding with cold core box technology for precise core assembly. Cold core box core making offers high dimensional accuracy, flexibility in core design, low cost, and environmental benefits as it eliminates waste sand emissions. A zircon-based alcohol coating was applied to both mold and core surfaces to form a refractory barrier and fill interstices, improving surface finish and reducing sand adhesion. For ease of stripping, silver graphite powder was manually applied as a parting agent.
2.2 Parting Surface Determination
The parting surface design significantly influences mold complexity, core assembly, and potential sand casting defects. I adopted a four-part curved parting scheme, as illustrated conceptually in this work. The upper part of the box contains a circular boss and irregular side features; thus, a parting line was placed at the boss level to facilitate molding of that region. A second parting surface was located at the mid-height of the elongated box to enable easy pattern withdrawal and accurate core placement. A third parting surface was set at the bottom to form the lower geometry. This four-part arrangement ensures that all undercuts and complicated shapes are accommodated without requiring complex loose pieces, thereby reducing the risk of molding-induced sand casting defects.
3.3 Gating System Design
I employed a bottom gating system due to its advantages: smooth filling, reduced impact on cores, minimized oxidation, and efficient slag trapping. Since the interior of the box is largely hollow, the sprue was placed at the center bottom of the casting, passing through the cavity without interfering with the final shape. This central sprue location equalizes the flow distance to all parts of the mold cavity, promoting uniform filling and controlled solidification sequence. A funnel-shaped pouring cup was used for simplicity. To prevent oxide inclusions and slag from entering the mold, two ceramic foam filters were installed in the sprue at the parting plane. Additionally, four runners and four ingates were designed at the bottom, each equipped with an extra filter at the ingate–mold interface. The gating system was calculated using the restricted cross‑section method. Table 1 summarizes the calculated dimensions.
| Element | Cross‑sectional area (cm²) | Length (mm) |
|---|---|---|
| Sprue | 9 | 1115 |
| Runner (each) | 4.5 | 91 |
| Ingate (each) | 8 | 0 |
3.4 Risering and Feeding
Proper riser design is critical to eliminate shrinkage‑related sand casting defects. I used open risers placed at the top of the casting (see the distribution scheme). The riser sizes were determined by the modulus method. The feeding distance and volume were checked to ensure adequate liquid metal supply during solidification. The total pouring mass including risers was approximately 120 kg. The pouring time was calculated using the empirical formula:
$$ \tau = B \delta^m P^n $$
where \( \tau \) is the pouring time (s), \( \delta \) is the average wall thickness, \( m \) is the casting mass, and \( B, P, n \) are coefficients. With appropriate values for aluminum alloys, I obtained a pouring time of 33 seconds, yielding a metal rise velocity of 25 mm/s in the mold cavity. The pouring temperature was set between 690 and 710 °C, typical for AlSi7Mg0.3.
4. Casting Process Simulation
4.1 Simulation Setup
I used AnyCasting software to simulate the filling and solidification behavior of the designed sand casting process. The 3D CAD model, including gating and riser systems, was imported. Material properties for AlSi7Mg0.3 and the sand mold were assigned. Initial and boundary conditions were specified: pouring temperature 700 °C, mold preheat temperature 25 °C, and heat transfer coefficients. The simulation ran for both filling and solidification stages. The purpose of the simulation was to predict the location and severity of sand casting defects, particularly shrinkage porosity and misruns.
4.2 Initial Simulation Results and Defect Prediction
The solidification sequence graph obtained from the simulation indicated that the thin internal plate region solidified last, as this section had a lower modulus compared to the surrounding thicker walls. Consequently, the simulation predicted a high probability of shrinkage porosity in that area. The defect probability map confirmed the presence of significant volumetric sand casting defects exactly at the thin plate location. The defects arose because the thin plate could not be adequately fed by the risers due to its late solidification and the high volumetric shrinkage of aluminum alloys. Figure 1 (see below) illustrates the predicted defects.
4.3 Optimization by Adding Chills
To eliminate the predicted sand casting defects, I introduced internal chills (metal inserts) in the mold at the thin plate location. Chills accelerate local heat extraction, promoting earlier solidification of the critical region and ensuring that it solidifies before the surrounding risers lose their feeding capability. The chills were placed as shown in the design scheme. After adding chills, the simulation was rerun. The updated defect probability map showed that the sand casting defect at the thin plate was significantly reduced – in fact, nearly eliminated. The solidification sequence became more uniform, and the thin plate solidified earlier, allowing proper feeding from the surrounding metal and risers. Table 2 compares the defect volume before and after adding chills.
| Condition | Predicted shrinkage porosity volume (cm³) |
|---|---|
| Without chills | 12.4 |
| With chills | 0.8 |
The results confirm that the judicious use of chills effectively mitigated sand casting defects without altering the casting geometry or increasing cost significantly.
5. Discussion
The combination of bottom gating, central sprue, and open risers provided a stable filling pattern and directional solidification. However, the presence of a thin internal plate, common in box‑shaped castings, remains a vulnerable spot for sand casting defects. The simulation proved indispensable in identifying this defect risk before actual production. The addition of chills was a simple and cost‑effective solution that did not require redesigning the gating or riser system. Multiple trials in the simulation environment allowed me to optimize the chill size and placement. In practice, one might also consider modifying the casting design (e.g., increasing the local wall thickness or adding a feeder), but chills were chosen for their minimal impact on the final part and ease of implementation in sand casting.
Throughout the project, the term “sand casting defect” reappeared in every phase: from the initial structural analysis to the final optimization. Understanding the root causes of sand casting defects – such as improper solidification sequence, insufficient feeding, high shrinkage, and inadequate cooling – is fundamental to any successful sand casting process. The simulation tool enabled me to visualize these sand casting defects in a digital environment, thus saving time and material.
6. Conclusion
I have successfully designed a sand casting process for a medium‑sized aluminum alloy reducer box using sand casting with cold core box core assembly. The gating and riser systems were calculated and the casting process was simulated using AnyCasting. The simulation initially predicted severe sand casting defects (shrinkage porosity) in the thin internal plate region. By introducing chills at the critical location, the defects were effectively eliminated. The final design ensures a sound casting with minimal sand casting defects. This study demonstrates the value of integrating simulation into the sand casting process development workflow, particularly for complex geometries where sand casting defects are likely to occur. The methodology can be readily applied to similar box‑type castings to improve yield and reduce scrap caused by sand casting defects.
Future work could include experimental validation of the simulated results, investigation of chill material and geometry optimization, and exploration of advanced feeding strategies (e.g., exothermic risers) to further reduce sand casting defects in thin‑wall regions.