In this study, we focus on the design and optimization of a sand casting process for a cross-beam type aluminum separator, commonly used in aerospace applications where high mechanical performance is critical. The goal is to achieve a defect-free casting with tensile strength exceeding 390 MPa, elongation over 8%, and hardness above 100 HBW. Sand castings are widely employed for such components due to their versatility and cost-effectiveness for small-batch production. Here, we detail our approach using numerical simulation with ViewCast software to predict and mitigate defects like shrinkage and porosity, followed by experimental validation. The process involves ZL201 aluminum alloy, a high-strength Al-Cu-Mn series known for its excellent mechanical and thermal properties. We begin by outlining the casting design, then proceed to simulation, optimization, and finally microstructural and mechanical analysis.
The ZL201 alloy composition is critical for its performance. Table 1 summarizes the chemical composition, which influences fluidity, solidification behavior, and final properties in sand castings. Proper control of these elements is essential to avoid defects and meet specifications.
| Element | Cu | Mn | Si | Al |
|---|---|---|---|---|
| Content | 4.5–5.3 | 0.6–1.0 | 0.15–0.35 | Balance |
The casting geometry, as shown in the image below, is a beam-like structure with dimensions of 571.2 mm in length, 150 mm in height, and wall thicknesses ranging from 14 mm to 20 mm. It features five bosses and a weight of approximately 3.3 kg. For sand castings of this type, the design must account for aluminum’s high reactivity, low density, and significant solidification shrinkage, which can lead to issues like oxide inclusions, cold shuts, and porosity if not properly managed.
We designed the gating system based on principles for aluminum sand castings. Aluminum alloys require rapid, tranquil filling to minimize turbulence and oxidation. We opted for a middle-injection system, which combines advantages of top and bottom pouring, suitable for medium-sized castings with uniform wall thickness. The gating system includes a sprue, runner, and four ingates. Initial dimensions were calculated using empirical relations common in sand castings. The total ingate area was set to 13 cm², with each ingate having a flat shape (height 5 mm, widths 39 mm and 41 mm) to reduce dross entrainment. The sprue diameter was determined from the casting weight relationship: for aluminum sand castings, the sprue diameter \(d\) (in mm) can be estimated from the pouring weight \(W\) (in kg) using $$ d = k \sqrt{W} $$ where \(k\) is a constant typically between 12 and 15 for aluminum. With \(W = 3.3\) kg, we chose \(d = 23\) mm. The runner was designed as a trapezoidal section with area 13 cm² (height 32 mm, widths 32 mm and 40 mm). Four top risers were placed at thick sections, with root dimensions of 39 mm × 15 mm and height 54 mm (1.3 times the root diameter), using a 5° taper. The mold material was furan resin sand, chosen for its high nitrogen content and suitability for small-batch production of non-ferrous sand castings.
To evaluate this design, we performed numerical simulation using ViewCast software. The simulation parameters are listed in Table 2. We meshed the geometry into approximately 2 million elements to ensure accuracy. The filling and solidification processes were analyzed to predict defect formation.
| Parameter | Value |
|---|---|
| Alloy | ZL201 |
| Pouring Temperature | 720°C |
| Mold Initial Temperature | 20°C |
| Mold Material | Furan Resin Sand |
| Simulation Software | ViewCast |
The filling simulation results, as shown in Figure 3 (not included here, but described), indicated stable flow without turbulence. Metal entered the sprue at 0.35 s, reached the runner at 0.68 s, began filling the cavity at 1.36 s, and completed at 6.11 s. This sequential filling from bottom to top is ideal for sand castings, as it reduces air entrapment and oxide formation. However, solidification simulation revealed critical issues. Using Chvorinov’s rule for solidification time in sand castings: $$ t = B \left( \frac{V}{A} \right)^2 $$ where \(t\) is solidification time, \(V\) is volume, \(A\) is surface area, and \(B\) is a mold constant. For ZL201 in resin sand, \(B\) is approximately 0.8–1.2 min/cm². Calculations showed that thicker sections solidified slower, leading to isolated liquid pockets. At 66.5 s, the ribs and sprue started solidifying; by 76.5 s, the front rod and sprue were fully solid, but junctions between the rod and ribs remained liquid, creating shrinkage-prone zones. By 186.5 s, the risers solidified before the thick sections, resulting in inadequate feeding. The predicted shrinkage porosity locations are summarized in Table 3, highlighting defects in thick regions and front rod junctions.
| Defect Location | Type | Severity |
|---|---|---|
| Thick Sections (Bosses) | Shrinkage Porosity | High |
| Front Rod-Rib Junctions | Microporosity | Medium |
| Riser Necks | Shrinkage Cavities | Low |
To address these, we optimized the riser design. The original risers were insufficient due to their limited feeding range. For sand castings, the feeding distance \(L\) can be estimated using: $$ L = k \cdot T $$ where \(T\) is the section thickness and \(k\) is a factor dependent on alloy and mold material (for ZL201 in resin sand, \(k \approx 4-6\)). Our calculation indicated that the risers needed enhancement. We increased the riser height by 300 mm and added insulating sleeves (25 mm thick) to slow solidification, extending the feeding range. Additionally, for the front rod defects, we introduced four new risers with rectangular sections (bottom 16 mm × 32 mm, top 16 mm × 41 mm, height 48 mm) placed strategically. The modified gating system aimed to ensure directional solidification toward the risers, a key principle in sand castings to eliminate shrinkage.
Re-simulation with ViewCast confirmed the improvement. As shown in Figure 6 (described), solidification now progressed with the risers remaining liquid longer, effectively feeding the critical zones. By 207 s, the casting was fully solid without isolated liquid regions. The defect prediction, illustrated in Figure 7, showed nearly complete elimination of shrinkage and porosity. This optimized design was then used to produce actual sand castings. After pouring and cooling, the castings underwent T5 heat treatment: solutionizing at 540°C ± 5°C for 5 hours, water quenching at 70°C, and aging at 175°C ± 5°C for 3 hours. Visual inspection revealed no obvious defects, validating the simulation-based optimization for sand castings.
We conducted microstructural analysis on samples from the castings. The etched microstructure (Figure 9) consisted of fine black precipitates within grains, identified as secondary T phase (Al12CuMn2), and eutectic of Al2Cu and α-phase along grain boundaries. This structure is typical for heat-treated ZL201 sand castings and contributes to high strength. The formation of T phase can be described by the reaction: $$ \alpha + \text{Al}_6\text{Mn} \rightarrow \text{Al}_{12}\text{CuMn}_2 $$ during aging. The volume fraction of eutectic \(f_e\) influences mechanical properties and can be estimated from composition using lever rule approximations.
Mechanical testing was performed to verify compliance with aerospace standards. Results are presented in Table 4. All values meet or exceed the requirements, demonstrating the success of our optimized sand casting process.
| Property | Value | Requirement |
|---|---|---|
| Hardness (HBW) | 117.3 | >100 HBW |
| Tensile Strength (MPa) | 413 | >390 MPa |
| Elongation (%) | 11 | >8% |
To further generalize our findings, we derived a model for riser sizing in aluminum sand castings. The riser volume \(V_r\) needed to compensate for shrinkage is given by: $$ V_r = \frac{V_c \cdot \beta}{1 – \beta} $$ where \(V_c\) is the casting volume and \(\beta\) is the solidification shrinkage rate (for ZL201, \(\beta \approx 0.06\)). For our casting, \(V_c \approx 3300 \, \text{cm}^3\), so \(V_r \approx 210 \, \text{cm}^3\). Our optimized risers collectively met this criterion. Additionally, the modulus method, where riser modulus \(M_r\) should exceed casting modulus \(M_c\) by a factor of 1.2–1.5, was applied: $$ M = \frac{V}{A} $$ We calculated \(M_c\) for thick sections as 0.5 cm and ensured \(M_r > 0.75\) cm. This approach is crucial for designing reliable sand castings.
In conclusion, we successfully designed and optimized a sand casting process for an aluminum beam separator using numerical simulation and experimental validation. The key steps included initial gating design, simulation with ViewCast to identify defects, riser optimization with insulation, and heat treatment. The final sand castings exhibited no shrinkage or porosity, with excellent mechanical properties meeting aerospace demands. This study underscores the importance of simulation tools in modern sand casting industries, enabling defect prediction and process refinement without costly trial-and-error. Future work could explore other alloy systems or larger-scale production for sand castings. Our methodology provides a template for optimizing similar components in sand castings, combining theoretical principles with practical adjustments.
Throughout this study, we emphasized the role of sand castings in producing high-integrity aluminum parts. By integrating simulation, design modifications, and thorough analysis, we achieved a robust process that ensures quality and performance. The repeated focus on sand castings in various contexts—from gating design to defect analysis—highlights its versatility and relevance in manufacturing. We hope this work contributes to advancing sand casting technologies for critical applications.