The reducer box under consideration is a medium-sized aluminum alloy component produced in large batches. The material is AlSi7Mg0.3, and the overall dimensions are approximately 590 mm × 475 mm × 810 mm. The main wall thickness ranges from 12 to 15 mm, which classifies it as a medium-thickness part. The geometry is essentially a hollow rectangular structure with an internal thin plate, presenting multiple faces and holes. Because the wall thickness is relatively uniform, no excessively thick sections exist, but the internal thin plate may cause insufficient feeding during solidification. This situation often leads to typical sand casting defects such as shrinkage porosity and shrinkage cavities. To ensure high quality and dimensional accuracy, a thorough process design and numerical simulation are essential. In this study, we focused on developing a reliable sand casting process and verifying it by simulation, aiming to minimize the occurrence of sand casting defects.
Three-Dimensional Modeling of the Reducer Box
We constructed a three-dimensional solid model of the reducer box using UG software. The modeling began in the sketch mode according to the part drawing and dimensions. Commands such as extrusion, rotation, and draft were used to generate the final part shape. Machining allowances were reserved on all processed surfaces. The resulting solid model served as the basis for subsequent casting process design and simulation. Accurate modeling is crucial because any geometrical simplification could distort the prediction of sand casting defects. The internal thin plate was explicitly retained in the model because it is the most critical region for feeding and solidification analysis. The model also includes the circular boss at the top and the irregular side shapes, both of which influence the parting line determination.
Selection of Casting Process
After analyzing the structure and manufacturability, we adopted sand casting as the production method. Because the batch size is large, a cold-box core assembly molding process was selected. The cold-box process provides high dimensional accuracy, flexible production, and environmental benefits due to the absence of waste sand discharge. For the mold and cores, we used a phenolic-modified furan resin self-hardening sand. Hand molding was employed for the production. An alcohol-based coating was applied to the surfaces of the mold and cores to form a refractory protective layer and to fill surface pores. For easy stripping, a silver graphite powder was manually applied as a parting agent.
The cold-box precision core assembly technique allows the production of complex internal cavities without the restrictions of a conventional parting line. This flexibility is advantageous for the reducer box since it has multiple internal features. However, the complexity also increases the risk of misalignment and consequent sand casting defects such as fins and flash. Therefore, careful core design and assembly tolerance control are necessary.
Determination of Parting Lines
The reducer box has a complex external shape with a circular boss on the top and irregular side contours. To facilitate molding and pattern withdrawal, we employed a four-box curved parting scheme. The first parting line is located at the circular boss on the top to shape the upper portion. The second parting line is set at the middle of the box to separate the long main body, which improves the accuracy of molding. The third and fourth parting lines are placed near the bottom to shape the lower features. This multi-parting approach avoids undercuts and makes the core assembly straightforward. Correct parting line selection is fundamental in avoiding sand casting defects associated with shifts, mismatches, and sand inclusion. A poorly chosen parting line can lead to excessive grinding, dimensional inaccuracy, and uneven wall thickness, all of which contribute to stress concentration and potential cracking.
Gating and Riser System Design
We designed a bottom gating system considering the overall structure and process requirements. The bottom gating system ensures smooth filling of the liquid metal, reduces the impact on sand cores, minimizes oxidation, and exhibits good slag retention capability. These characteristics are particularly important for aluminum alloys, which are prone to oxidation and dross formation. Oxides and dross can be entrapped during filling and become major sources of sand casting defects such as inclusions and cold shuts. Therefore, the gating system design plays a critical role in the final quality.
Because the internal cavity of the box is essentially empty, we chose to place the sprue at the center of the bottom. This configuration shortens the runner length and enables the metal to reach different positions at nearly the same time. As a result, a favorable sequential solidification pattern can be achieved, reducing the likelihood of sand casting defects caused by non-uniform temperature distribution. The short runner also minimizes heat loss, ensuring that the molten metal enters the cavity at a sufficiently high temperature to avoid misruns and cold shuts. A funnel-shaped pouring cup was used for easy pouring.
Aluminum alloys are highly susceptible to oxidation and slag formation during pouring. To combat this, two filters were placed in the sprue at the parting line, and an additional filter was installed at each ingate. These filters effectively trap inclusions and reduce turbulence. The gating system is an open stepped bottom design. The pouring cup was made of refractory bricks. The cross-sectional areas of the runner components were calculated using the choke area method. The resulting dimensions are summarized in Table 1.
| Component | Sprue | Runner | Ingate |
|---|---|---|---|
| Cross-sectional area (cm²) | 9 | 4.5 | 8 |
| Length (mm) | 1115 | — | 90 |
The gating ratio (sprue : runner : ingate) is approximately 1 : 0.5 : 0.89, which corresponds to an open system. In an open system, the total area increases from the sprue to the ingate, resulting in a slower flow and reduced turbulence at the ingate. This is favorable for preventing air entrainment and oxide film formation. Nevertheless, even with a well-designed gating system, sand casting defects may still occur due to improper pouring temperature or filling rate.
Pouring Time Calculation
We calculated the pouring time using the empirical formula widely adopted in sand casting design:
$$ \tau = B \delta^m P^n $$
where τ is the pouring time (s), δ is the average wall thickness (mm), m is the pouring mass (kg), and B, P, and n are coefficients determined by the alloy and mold conditions. For this aluminum alloy casting, the coefficients were selected based on handbooks. The calculated pouring time was 33 s. The corresponding rise velocity of the liquid metal in the mold cavity was 25 mm/s. This filling rate is within the recommended range for aluminum alloy sand castings to avoid both cold shuts and sand erosion.
Riser Design
The risers were designed to provide sufficient feeding during solidification. We employed open risers of the universal type. The riser dimensions were determined using the modulus method. The modulus of a casting section is defined as the ratio of its volume to its cooling surface area:
$$ M = \frac{V}{A} $$
For the feeder to be effective, its modulus must be greater than that of the section it feeds, typically by a factor of 1.1 to 1.2:
$$ M_{\text{riser}} = f M_{\text{casting}} $$
where f is the feeder safety factor. The total number and positions of risers were chosen to ensure that all hot spots are adequately fed. We placed the risers as shown in the original design, which provided a progressive solidification direction from the thinner sections toward the risers. However, the feed path to the internal thin plate was problematic due to the geometry. This is a common challenge in complex box castings, and inadequate feeding directly results in sand casting defects such as shrinkage cavities and macro-segregation in the interior.
Simulation of the Casting Process
The filling and solidification of a casting involve complex interactions between fluid flow and heat transfer. The temperature evolution and flow field are highly transient, and they directly influence the formation of sand casting defects. We used AnyCasting software to simulate the casting process. The simulation provided insights into the filling sequence, temperature distribution, solidification order, and defect probability. The process parameters used in the simulation are listed in Table 2.
| Parameter | Value |
|---|---|
| Pouring temperature (°C) | 690–710 |
| Pouring time (s) | 33 |
| Mold material | Phenolic resin sand |
| Heat transfer coefficient (mold/metal) (W/m²K) | 500 |
| Initial mold temperature (°C) | 25 |
| Alloy composition | AlSi7Mg0.3 |
The solidification sequence obtained from the simulation is depicted in Figure 1. It can be observed that the internal thin plate solidifies last. This is unexpected because the thin plate has a small volume and should cool faster than the thicker sections. However, the surrounding thicker sections act as heat sources, delaying the solidification of the thin plate. Moreover, the thin plate is surrounded by solidified metal on most sides, so the feeding path from the risers is blocked. The last-solidifying region becomes a hot spot and is highly susceptible to sand casting defects caused by volumetric contraction. The simulation’s probability defect map confirms that a major defect appears precisely at the internal thin plate, consistent with our prediction.
The defect formation mechanism can be explained by the solidification contraction of the aluminum alloy. During solidification, the alloy undergoes a volume reduction of approximately 3.5–4.5%. If the molten metal cannot flow from the risers to compensate for this contraction, porosity will form. In the case of the internal thin plate, the narrow and tortuous feeding path restricts liquid flow, resulting in a pressure drop and subsequent formation of shrinkage porosity. This is a classic example of how inadequate feeding leads to sand casting defects in complex-shaped sand castings.
Elimination of Defects by Adding Chills
To eliminate the predicted defects, we applied chills at the internal thin plate location. Chills increase the local cooling rate by absorbing heat rapidly, thereby accelerating solidification and promoting a more directional solidification pattern. The chills were placed adjacent to the thin plate, as indicated in the design. After adding the chills, the simulation was repeated. The probability defect map shows that the defects in the internal thin plate were substantially eliminated. The chills successfully changed the solidification sequence, ensuring that the thin plate solidified earlier and was no longer a hot spot. As a result, the risk of sand casting defects in that critical region was greatly reduced.
The use of chills is a common and effective method for controlling local solidification in sand casting. However, the placement and size of chills must be carefully optimized. Too many chills or excessive chill thickness can lead to premature solidification and cause misruns, while insufficient chilling may not fully eliminate the hot spot. In our design, the chills were sized based on the thermal modulus of the thin plate area. The simulation provided quantitative confirmation that the chills were adequate.
Discussion on Sand Casting Defects and Preventive Measures
Sand casting defects can be broadly classified into filling-related defects, solidification-related defects, and molding-related defects. The primary concerns for this reducer box are shrinkage-related sand casting defects such as porosity, cavities, and micro-porosity. Other possible sand casting defects include sand inclusion, blowholes, and cold shuts. Our simulation and experimental design addressed these issues through the following measures:
- Proper gating system design to ensure laminar filling and minimize air entrapment.
- Use of filters to prevent dross and sand from entering the mold cavity.
- Correct riser placement to guarantee ample feeding during solidification.
- Application of chills to modify the solidification sequence and eliminate hot spots.
- Selection of a low-moisture, high-permeability sand to avoid gas-related sand casting defects.
Even with all these measures, the occurrence of sand casting defects can never be fully eliminated in practice. Therefore, simulation plays a vital role in identifying risk areas and optimizing the process before physical tooling is made. The use of AnyCasting allowed us to visualize the defect probability and quantify the effect of chills, saving both time and cost compared to a trial-and-error approach.
Conclusion
In this work, we designed a sand casting process for a reducer box made of AlSi7Mg0.3. The process includes a four-box curved parting, a cold-box core assembly, a bottom gating system with filters, and risers sized by the modulus method. Through numerical simulation, we identified the internal thin plate as the most defect-prone region. The predicted sand casting defects were confirmed to be shrinkage-related porosity. By adding chills at that location, the solidification sequence was modified, and the defects were effectively eliminated. The simulation results demonstrate the validity and effectiveness of the proposed process. This study shows how careful process design combined with simulation can significantly reduce the risk of sand casting defects in complex aluminum alloy castings, ensuring high quality and dimensional accuracy in mass production.