In my research, I conducted a numerical simulation of the sand casting process for aluminum alloy cooling fins used in CPU heat dissipation. The goal was to predict and eliminate sand casting defects such as shrinkage porosity and gas entrapment by optimizing process parameters. I employed vacuum negative pressure technology, calculated the gating system using the Osann equation, and designed risers based on the hot spot circle principle. Through ProCAST simulation, I analyzed mold filling and solidification, identified potential sand casting defect locations, and iteratively improved the design. The final optimized parameters—pouring time of 3 seconds, pouring temperature of 740°C, static head height of 40 mm, and nine risers of 10 mm diameter—yielded defect-free castings with smooth filling and sequential solidification.
The cooling fin geometry had a baseplate of 100 mm × 100 mm × 7 mm, fin height of 30 mm, and fin spacing of 5 mm. I began by designing a gating system using the Osann equation:
$$ F_{\text{straight}} = \frac{G}{\mu\ t\ \sqrt{H_p}} $$
Where:
| Symbol | Description | Value |
|---|---|---|
| $F_{\text{straight}}$ | Choke area (m²) | Calculated ~3 cm² |
| $G$ | Total aluminum mass in mold | 0.972 kg (1.8 × part mass) |
| $G_{\text{part}}$ | Cast part mass | 0.54 kg |
| $\mu$ | Flow coefficient | 0.6 |
| $t$ | Pouring time (s) | 3 s (using $t = S\sqrt[3]{G}$ with $S=3$) |
| $H_p$ | Average static head (minimum) | 23.4 mm (initial); later 30 mm and 40 mm used |
I set the gating ratio as $\sum F_{\text{straight}} : \sum F_{\text{runner}} : \sum F_{\text{ingate}} = 1:1.6:1.1$. For $H_p = 30$ mm and 40 mm, $F_{\text{straight}} \approx 3$ cm², giving $\sum F_{\text{runner}} = 4.8$ cm² and $\sum F_{\text{ingate}} = 3.3$ cm².
For riser design, I considered the hot spot circle at the baseplate thickness (7 mm). The casting modulus $M_f = D_r/2$, riser modulus $M_m = 1.2 M_f$, riser root diameter $D_R = 1.2 D_r$, and riser height $H_R = 1.2–1.5 D_R$. The effective feeding distance $L = 2a$, where $a$ is the casting thickness. The calculated riser dimensions are summarized below:
| Parameter | Symbol | Value (mm) |
|---|---|---|
| Hot spot circle diameter | $D_r$ | 7 |
| Riser root diameter | $D_R$ | 10 |
| Riser height | $H_R$ | 30 (initial); later 40 mm |
| Effective feeding distance | $L$ | 14 |
| Number of risers | — | 9 |
I placed nine risers on the top surface of the baseplate to feed the solidification shrinkage. The initial simulation used a static head of 30 mm with pouring time 3 s, pouring temperature 740°C, and sand mold temperature 25°C.
In the mold filling simulation with $H_p = 30$ mm, the liquid metal flowed smoothly from the sprue through the runner and ingate, rising uniformly in the cavity without turbulence or gas entrapment. However, during solidification analysis at 131 seconds after pouring, I observed that while the risers solidified completely, a portion of the baseplate remained liquid (solid fraction 53.3%) beyond the critical solid fraction of 60%, indicating a risk of sand casting defect—specifically shrinkage porosity. I examined cross-sections (e.g., section 1-1 and 2-2) and found isolated liquid regions surrounded by solidified metal, which could not be fed by the risers. This confirmed that the initial 30 mm riser height provided insufficient feeding.
To eliminate this sand casting defect, I increased the static head height to 40 mm, which effectively raised the riser height to 40 mm while keeping the same base diameter. The redesigned simulation showed a perfect sequential solidification from bottom to top. At 131 seconds, the fins began to solidify, and by 135 seconds, the entire casting was fully solidified with the risers feeding the shrinkage. The defect prediction using Niyama criterion revealed that all shrinkage porosity was confined to the risers, leaving the casting entirely sound. The optimized parameters are summarized in the table:
| Parameter | Value |
|---|---|
| Pouring time | 3 s |
| Pouring temperature | 740°C |
| Sand mold temperature | 25°C |
| Static head height | 40 mm |
| Number of risers | 9 |
| Riser diameter | 10 mm |
| Riser height | 40 mm |
| Gating system ratio | 1:1.6:1.1 |
I also validated the simulation with actual production using vacuum negative pressure casting under these optimized conditions. The resulting cooling fin castings had smooth surfaces and no internal sand casting defects, confirming the effectiveness of the numerical approach. The entire process demonstrated that careful adjustment of feeding parameters can convert potential sand casting defect locations from the casting body into the risers.
Throughout the study, I repeatedly used the Niyama criterion and solid fraction analysis to predict sand casting defect regions. By comparing the simulations for static heads of 30 mm and 40 mm, I concluded that increasing the feeding pressure head and riser volume effectively transferred shrinkage porosity away from the casting. The final optimized process not only eliminated sand casting defect but also ensured stable mold filling without gas entrapment—a critical factor for thin-walled complex castings like cooling fins.
In summary, numerical simulation is a powerful tool for predicting and preventing sand casting defect. By systematically applying the Osann equation, hot spot circle method, and iterative simulation, I achieved a robust process design that produced high-quality aluminum cooling fins. The key findings reinforce the importance of adequate riser height and proper static head in achieving sound castings. This methodology can be extended to other sand casting applications where defect-free thin sections are required.