In my research and industrial practice, I have concentrated on the optimization of investment casting processes for complex aluminum alloy components. Investment casting, often called lost-wax casting, is a near-net-shape manufacturing route that enables the production of structurally intricate parts with high dimensional precision and low surface roughness. Because of these advantages, investment casting is extensively applied in shipbuilding, aerospace, and other demanding industrial sectors. However, when an investment casting process is developed only through conventional trial-and-error experimentation, the production cycle becomes lengthy, the manufacturing cost increases, and the approach struggles to meet modern industrial requirements. With the rapid development of computer technology, numerical simulation has become a mature auxiliary method for process design in investment casting. Simulation technology allows me to effectively optimize and model complex castings, determine technical parameters for parts and products, identify potential defects, and refine the process design. ProCAST is a widely used software in casting numerical simulation. By using ProCAST, I can simulate the flow behavior of liquid metal and the solidification process, accurately predict shrinkage porosity, shrinkage cavities, and other defects in investment casting, thereby reducing the research and development cost and shortening the development cycle. In this study, I investigated a filter housing made of aluminum alloy. I used ProCAST to simulate the initial combined process for this casting, analyzed the solidification behavior and the distribution of shrinkage porosity, and then optimized the initial combined process. Finally, I obtained qualified castings through experimental trials.
1. Process Analysis of the Investment Casting Component
The filter housing that I studied is an aluminum alloy casting produced by investment casting. The material is ZL114A, and the mass of the housing is approximately 1.5 kg. The overall dimensions are 140 mm × 150 mm × 165 mm. From a structural perspective, the internal cavity of this investment casting component is complex, and the transition between thick and thin sections is very abrupt. This geometric feature makes it difficult to achieve smooth mold filling during pouring. In addition, aluminum alloy melts are prone to oxidation, which can generate oxide inclusions and slag. Consequently, this investment casting is susceptible to shrinkage porosity, gas porosity, and inclusion defects. Furthermore, because the wall thickness is nonuniform, localized overheating can occur during solidification, creating hot spots that lead to shrinkage porosity. Table 1 summarizes the key characteristics of the component and the associated investment casting challenges.
| Feature | Description | Investment Casting Challenge |
|---|---|---|
| Material | ZL114A aluminum alloy | High oxidation tendency, easy oxide inclusion formation |
| Mass | Approximately 1.5 kg | Moderate thermal mass, requires controlled cooling |
| Dimensions | 140 mm × 150 mm × 165 mm | Complex geometry, difficult mold filling |
| Wall thickness | Nonuniform, abrupt transitions | Hot spots, isolated liquid regions, shrinkage porosity |
| Internal cavity | Complex shape | Difficult to feed, possible misruns and cold shuts |
To quantify the solidification behavior of this investment casting, I considered the fundamental heat transfer equation that governs the process. The transient temperature field in the casting and shell mold can be described by:
$$\rho C_p \frac{\partial T}{\partial t} = \nabla \cdot (k \nabla T) + Q$$
where \(\rho\) is density, \(C_p\) is specific heat capacity, \(T\) is temperature, \(t\) is time, \(k\) is thermal conductivity, and \(Q\) represents the latent heat released during solidification. For investment casting of aluminum alloys, the latent heat term is crucial because it governs the local solidification time and the formation of shrinkage defects. I also used the solid fraction, \(f_s\), to track the progression of solidification:
$$f_s = \frac{T_l – T}{T_l – T_s}$$
where \(T_l\) is the liquidus temperature and \(T_s\) is the solidus temperature. When \(f_s = 1\), the alloy is fully solid. When isolated liquid regions remain after the feeding channels have solidified, shrinkage porosity is likely to form. In my investment casting simulation, I monitored the solid fraction field to identify such isolated liquid regions.
2. Initial Process Design and Simulation Results
Based on the structural characteristics of the casting and my experience with investment casting process design, I placed risers at the thick sections of the component to prevent metallurgical defects caused by insufficient feeding. The initial riser arrangement is shown in my design scheme. Because the metal is an aluminum alloy, I adopted a bottom-gating system to prevent oxidation and slag entrapment during pouring. This bottom-gating design ensures smooth mold filling and reduces turbulence in the investment casting. I modeled the entire gating system using a 3D CAD software and then imported the model into the Visual-Mesh module of ProCAST for mesh generation. The mesh size for the casting and gating system was set to 5 mm, and the shell mold thickness was 6.5 mm. After checking the mesh quality, I generated the volume mesh. Then I set the casting parameters in the Cast module of ProCAST. Table 2 lists the numerical simulation parameters that I used for the initial investment casting process.
| Parameter | Value |
|---|---|
| Metal material | ZL114A |
| Shell mold material | Mullite |
| Shell mold thickness | 6.5 mm |
| Interface heat transfer coefficient | 200 W·m-2·K-1 |
| Pouring time | 15 s |
| Pouring temperature | 710 °C |
| Shell mold temperature | 350 °C |
| Cooling method | Air cooling |
2.1 Solidification Process Analysis
I analyzed the solidification process of the investment casting using the solid fraction module in ProCAST. The solid fraction indicates the degree of solidification; a higher solid fraction means a larger proportion of solid phase. When the solid fraction reaches 100%, the entire casting has completely solidified. If an isolated liquid region appears during solidification—meaning the feeding channel has already solidified while a local region remains liquid or semi-solid—then that location is prone to shrinkage porosity. From my simulation, at t = 90 s, the casting began to solidify from the thinnest regions and gradually progressed toward the thicker regions. At t = 250 s, most of the casting had solidified, and no isolated liquid region had yet appeared. At t = 305 s, however, an isolated liquid region formed inside the casting because the feeding channel closed too early. This region was highly likely to produce shrinkage porosity. Table 3 presents the solidification sequence and the corresponding observations from my investment casting simulation.
| Time (s) | Solidification State | Observation |
|---|---|---|
| 90 | Solidification starts at thin sections | Progressive freezing from thin to thick regions |
| 250 | Most of casting solidified | No isolated liquid region detected |
| 305 | Isolated liquid region forms | Feeding channel closed prematurely; high risk of shrinkage porosity |
2.2 Shrinkage Porosity Analysis
I used the porosity module in ProCAST to analyze the shrinkage porosity in the investment casting. Porosity represents the volume fraction of shrinkage voids in the casting and is a critical criterion for evaluating whether shrinkage defects will occur. If shrinkage porosity exists even when the porosity threshold is greater than 1%, then the investment casting process is likely to produce shrinkage defects in actual production. In my simulation, with a porosity threshold of 2%, the casting still exhibited shrinkage porosity. Moreover, the defect location coincided exactly with the isolated liquid region. Therefore, I concluded that this initial investment casting process would generate shrinkage porosity. Table 4 summarizes the porosity prediction results.
| Porosity Threshold (%) | Shrinkage Volume Present? | Defect Location | Conclusion |
|---|---|---|---|
| 1 | Yes | Thick internal region | Defect likely |
| 2 | Yes | Same as isolated liquid region | Shrinkage porosity expected |
To further quantify the feeding behavior, I applied the Niyama criterion, which is often used in investment casting simulation to predict shrinkage porosity:
$$N = \frac{G}{\sqrt{R}}$$
where \(G\) is the temperature gradient and \(R\) is the cooling rate. A lower Niyama value indicates a higher probability of shrinkage porosity. In my initial investment casting simulation, the Niyama value in the thick internal region fell below the critical threshold, confirming the risk of porosity. I also evaluated the cooling rate using:
$$R = \frac{\partial T}{\partial t}$$
and the temperature gradient using:
$$G = \|\nabla T\|$$
These quantities helped me understand why the feeding channel solidified before the thick region could be adequately fed. The thin wall connecting the riser to the thick section acted as a bottleneck. Because the thin wall solidified rapidly, the liquid metal in the thick region could not receive additional feed metal, resulting in an isolated liquid pocket and subsequent shrinkage porosity.
3. Process Optimization and Simulation Results
Based on the analysis of the initial investment casting simulation, I determined that the process would likely produce shrinkage porosity. The defect was located in the thick internal region of the casting. The adjacent thin walls caused the thick region to solidify later than the thin sections, creating an isolated liquid zone. From a structural standpoint, the defect was internal, and simply adding more risers would not provide effective feeding because the feeding path was too thin and would freeze before the thick region solidified. Therefore, I adopted a method of adding a padding block to increase the thickness of the feeding channel. This approach was intended to improve the solidification sequence and eliminate shrinkage porosity. I placed the padding block at the critical location, redesigned the riser arrangement, and modified the gating system accordingly. I then simulated the improved investment casting process using the same simulation parameters as the initial process.
Table 5 compares the initial and optimized investment casting designs.
| Design Aspect | Initial Process | Optimized Process |
|---|---|---|
| Riser placement | Thick sections only | Thick sections with padding block |
| Feeding channel thickness | Thin, rapid solidification | Increased by padding block |
| Gating system | Bottom gating | Modified bottom gating with improved feeding |
| Solidification sequence | Isolated liquid region formed | No isolated liquid region |
| Predicted shrinkage porosity | Present at 2% threshold | Absent at 2% threshold |
3.1 Solidification Process After Optimization
I analyzed the solidification process after the investment casting optimization. From my simulation, at t = 95 s the casting began to solidify, and at t = 330 s the solidification was essentially complete. Throughout the entire solidification process, no isolated liquid region appeared. This indicates that after the process improvement, the liquid feeding in the originally defective region was effectively enhanced, and the solidification sequence became favorable. Consequently, the shrinkage porosity was eliminated. Table 6 presents the solidification timeline for the optimized investment casting process.
| Time (s) | Solidification State | Observation |
|---|---|---|
| 95 | Solidification begins | Smooth progressive freezing |
| 250 | Majority solidified | No isolated liquid region |
| 330 | Solidification complete | No isolated liquid region throughout |
3.2 Shrinkage Porosity After Optimization
I also analyzed the shrinkage porosity after the investment casting optimization. With a porosity threshold of 2%, the casting exhibited no shrinkage volume. This means that after the process improvement, the defect region received effective feeding, and the original shrinkage porosity was resolved. Table 7 summarizes the porosity results for the optimized investment casting.
| Porosity Threshold (%) | Shrinkage Volume Present? | Defect Location | Conclusion |
|---|---|---|---|
| 1 | No | None | No shrinkage porosity |
| 2 | No | None | Defect eliminated |
To further validate the feeding improvement, I calculated the feeding efficiency using the following relationship:
$$\eta_f = \frac{V_{\text{feed}}}{V_{\text{shrinkage}}} \times 100\%$$
where \(V_{\text{feed}}\) is the volume of liquid metal available for feeding and \(V_{\text{shrinkage}}\) is the volume shrinkage of the solidifying alloy. In the initial investment casting process, \(\eta_f\) was less than 100% in the thick region because the feeding channel solidified too early. In the optimized investment casting process, \(\eta_f\) exceeded 100%, ensuring complete feeding and eliminating shrinkage porosity.
4. Experimental Validation
To verify the accuracy of my simulation results and the feasibility of the optimized investment casting process, I conducted trial productions using both the initial and optimized combined processes. I assembled the wax patterns according to the two process designs, then performed shell making, pouring, and other investment casting steps to obtain the castings. I then subjected the castings to X-ray nondestructive testing. The results showed that the casting produced by the initial investment casting process exhibited shrinkage porosity, and the defect location matched the simulation prediction. In contrast, the casting produced by the optimized investment casting process showed no defects and met the product quality requirements. Table 8 compares the experimental validation results.
| Process Version | Nondestructive Testing Result | Defect Location | Compliance |
|---|---|---|---|
| Initial investment casting | Shrinkage porosity detected | Matches simulated location | Non-compliant |
| Optimized investment casting | No defects detected | None | Compliant |
The experimental results demonstrate that numerical simulation can effectively prevent shrinkage porosity in investment casting. By combining simulation with experimental validation, I was able to confirm that the optimized investment casting process is robust and reliable. Table 9 presents a summary of the key performance indicators before and after optimization.
| Indicator | Initial Investment Casting | Optimized Investment Casting |
|---|---|---|
| Isolated liquid region | Present | Absent |
| Shrinkage porosity at 2% threshold | Present | Absent |
| Niyama criterion value | Below critical threshold | Above critical threshold |
| Feeding efficiency | < 100% | > 100% |
| Experimental defect | Detected | Not detected |
| Product quality | Non-compliant | Compliant |
5. Discussion
My study confirms that investment casting combined with numerical simulation is a powerful approach for producing complex aluminum alloy components. The initial investment casting design, which used only risers at thick sections, failed to provide adequate feeding because the thin connecting walls solidified before the thick region. This created an isolated liquid region and led to shrinkage porosity. By adding a padding block to increase the thickness of the feeding channel, I improved the solidification sequence and ensured that liquid metal could feed the thick region until it fully solidified. The optimized investment casting process eliminated the isolated liquid region and produced sound castings.
The thermal history of the investment casting can be further analyzed using the cooling curve equation:
$$T(t) = T_0 + (T_p – T_0) \exp\left(-\frac{h A}{\rho V C_p} t\right)$$
where \(T_0\) is the initial mold temperature, \(T_p\) is the pouring temperature, \(h\) is the heat transfer coefficient, \(A\) is the surface area, \(V\) is the volume, and \(C_p\) is the specific heat capacity. This equation shows that the cooling rate depends on the surface-to-volume ratio. Thin sections have a high surface-to-volume ratio and therefore cool faster, while thick sections cool more slowly. In investment casting, this differential cooling is the primary cause of isolated liquid regions and shrinkage porosity. My padding block increased the effective volume of the feeding channel, reduced the surface-to-volume ratio, and delayed solidification of the feeding path. As a result, the feeding channel remained open longer, allowing liquid metal to compensate for shrinkage in the thick region.
I also considered the effect of the shell mold on the cooling rate. The shell mold thickness was 6.5 mm, and the interface heat transfer coefficient was 200 W·m-2·K-1. The mold preheat temperature was 350 °C, and the pouring temperature was 710 °C. These parameters were kept constant between the initial and optimized simulations to ensure a fair comparison. The only major change was the addition of the padding block and the corresponding modification of the gating system. This controlled approach allowed me to isolate the effect of the padding block on the solidification behavior and shrinkage porosity.
In addition to the Niyama criterion, I evaluated the shrinkage porosity using the following expression for volumetric solidification shrinkage:
$$\beta = \frac{\rho_s – \rho_l}{\rho_s} \times 100\%$$
where \(\rho_s\) is the solid density and \(\rho_l\) is the liquid density. For ZL114A aluminum alloy, \(\beta\) is typically between 3% and 6%. If the feeding path cannot supply enough liquid metal to compensate for this volumetric shrinkage, porosity forms. My optimized investment casting process ensured that the feeding path remained open until the thick region reached a sufficient solid fraction, thereby compensating for the shrinkage.
Table 10 lists the thermophysical properties of ZL114A that I used in the simulation.
| Property | Value |
|---|---|
| Liquidus temperature | 610 °C |
| Solidus temperature | 555 °C |
| Density (solid) | 2680 kg·m-3 |
| Density (liquid) | 2480 kg·m-3 |
| Specific heat capacity | 900 J·kg-1·K-1 |
| Thermal conductivity | 150 W·m-1·K-1 |
| Latent heat of fusion | 389 kJ·kg-1 |
These properties were used in the heat transfer and solidification equations to predict the temperature field and solid fraction evolution. The simulation results showed good agreement with the experimental observations, validating the accuracy of my numerical model.
6. Conclusions
Based on my numerical simulation and experimental validation of the investment casting process for the filter housing, I draw the following conclusions:
(1) I used ProCAST to simulate the initial investment casting process. By analyzing the solidification field, I found that an isolated liquid region formed inside the casting. The simulation accurately predicted the location of shrinkage porosity.
(2) By improving the initial investment casting process, I adopted a combined scheme of risers and a padding block. This increased the thickness of the feeding channel, improved the solidification sequence, and effectively solved the shrinkage porosity problem.
(3) I conducted trial production using both the initial and optimized investment casting processes and performed nondestructive testing on the castings. The results showed that the casting produced by the initial process had shrinkage porosity at the simulated location, while the casting produced by the optimized process had no defects and met the product quality requirements. These results verified the accuracy of the simulation and the feasibility of the optimized investment casting process.
Overall, my work demonstrates that numerical simulation is an indispensable tool for modern investment casting. By integrating ProCAST simulation with experimental validation, I can significantly reduce trial-and-error cycles, lower production costs, and ensure high-quality investment casting components. The methodology that I developed for this filter housing can be applied to other complex aluminum alloy investment castings, especially those with nonuniform wall thicknesses and intricate internal cavities. Future work will focus on extending this approach to other alloy systems and optimizing the thermal management of the shell mold to further improve the investment casting process.
To summarize the quantitative outcomes, Table 11 provides the final comparison of the initial and optimized investment casting processes.
| Metric | Initial Investment Casting | Optimized Investment Casting |
|---|---|---|
| Isolated liquid region formation time | 305 s | None |
| Solidification completion time | 305 s (with defect) | 330 s (sound) |
| Shrinkage porosity at 2% threshold | Present | Absent |
| Niyama value in critical region | Below critical | Above critical |
| Feeding efficiency \(\eta_f\) | < 100% | > 100% |
| Experimental X-ray result | Defect detected | No defect |
| Product quality | Non-compliant | Compliant |
I believe that this systematic approach to investment casting optimization will contribute to the broader adoption of simulation-driven process design in the foundry industry. The combination of ProCAST simulation, solidification analysis, defect prediction, and experimental validation forms a robust framework for developing reliable investment casting processes for high-performance components.