The pursuit of manufacturing high-integrity, geometrically intricate components for demanding applications, such as turbo-machinery in aerospace and power generation sectors, consistently drives advancements in foundry technologies. Among these, the investment casting process stands out for its exceptional capability to produce net-shape parts with superior surface finish, dimensional accuracy, and the ability to cast complex internal features that are challenging or impossible for other methods. This research delves into the optimization of this very investment casting process for a critical component—an aluminum alloy impeller—by integrating modern additive manufacturing (AM) techniques and computational numerical simulation, thereby addressing traditional challenges associated with mold fabrication.
Conventionally, the investment casting process is a multi-step procedure. It begins with the creation of a precise expendable pattern, typically from wax, which is an exact replica of the desired final part. This pattern is then assembled into a cluster with a gating system, repeatedly dipped into ceramic slurries, and stuccoed with refractory sands to build a robust, multi-layered shell mold. The wax is subsequently melted out in a dewaxing autoclave or furnace, leaving behind a hollow ceramic mold cavity. After high-temperature firing to develop strength and remove residues, the mold is filled with molten metal. Upon solidification, the ceramic shell is broken away, and the castings are separated from the gating system and finished. While effective, the traditional method of producing the wax pattern via injection molding requires costly and time-consuming metal tooling (dies), making it economically viable only for medium to high-volume production and posing a significant barrier for prototyping and low-volume runs.
The integration of Additive Manufacturing, specifically Fused Deposition Modeling (FDM), presents a paradigm shift. By 3D printing the pattern directly from a CAD model using a sacrificial polymer like Polylactic Acid (PLA), the need for hard tooling is completely eliminated. This synergy between AM and the investment casting process—often termed “rapid investment casting”—drastically compresses lead times, reduces upfront costs for prototype and small-batch production, and enhances design flexibility. However, simply replacing a wax pattern with a printed one does not guarantee a defect-free casting. The design of the gating and feeding system remains paramount to control metal flow, heat transfer, and solidification, which directly influence the final part’s soundness.
This study focuses on a specific centrifugal impeller with complex, curved thin-walled blades. The key dimensions of the impeller are summarized below. The primary challenge lies in ensuring complete, turbulence-free filling of the thin, contoured blade passages and promoting directional solidification to prevent shrinkage porosity, a common defect in such geometries. To systematically address this, the research workflow encompassed: (1) Geometric analysis and pattern fabrication via FDM, (2) Design and numerical simulation of multiple gating systems to predict and eliminate defects, and (3) Experimental validation through a complete investment casting process using the optimized design.
| Feature | Dimension (mm) |
|---|---|
| Base Diameter | 104 |
| Top Diameter | 25 |
| Overall Width | 49 |
| Blade Wall Thickness | 2.5 |
The first step involved translating the 3D CAD model into a physical pattern. An FDM 3D printer was employed using PLA filament. The printing parameters (layer height, infill density, print speed) were optimized to create a pattern with a smooth surface finish and sufficient strength to withstand subsequent handling during shell building. This printed PLA pattern directly substitutes the traditional wax pattern in the investment casting process. The success of this substitution hinges on the polymer’s ability to be completely removed from the ceramic shell without residue during the burn-out stage.
The core of the process optimization lies in the design of the gating and risering system. Three distinct schemes were conceived based on fundamental principles of fluid flow and heat transfer in casting:
- Top-Gating System: Metal enters the mold cavity from the top, near the impeller’s hub.
- Side-Gating System: Metal is introduced from a lateral position, aiming for a more balanced fill.
- Bottom-Gating System: Metal enters from the lowest point of the mold cavity, typically promoting a calm, progressive fill from the bottom up.
To virtually analyze and compare these designs without costly trial-and-error experiments, numerical simulation using AnyCasting software was performed. The simulation models the entire investment casting process, including mold filling, solidification, and defect prediction. The material was defined as ZL104 aluminum alloy, and the mold was set as a preheated ceramic shell. The critical process parameters input into the simulation are listed below.
| Parameter | Value |
|---|---|
| Casting Alloy | ZL104 |
| Pouring Temperature | 750 °C |
| Mold (Shell) Preheat Temperature | 650 °C |
| Pouring Velocity | 25 cm/s |
| Mold Material | Ceramic Shell |
The simulation results for the top-gating system revealed significant drawbacks. The molten metal fell freely from the sprue, impacting the bottom of the mold with high velocity, creating severe turbulence. This turbulent flow increases the risk of oxide entrapment and gas porosity. Furthermore, the thermal analysis showed an unfavorable solidification pattern. The top (entry point) and thin blades solidified first, potentially isolating liquid metal in the thicker sections of the base, leading to macro-shrinkage. The defect prediction module, often based on criteria like the Niyama criterion (a function of thermal gradient G and cooling rate R), highlighted a high probability of shrinkage porosity in the lower regions of the impeller. The Niyama criterion Ny is given by:
$$Ny = \frac{G}{\sqrt{\dot{T}}}$$
where $\dot{T}$ is the cooling rate. Regions with a low Niyama value are prone to shrinkage porosity. The top-gating simulation consistently yielded low Ny values in the impeller base.
The side-gating system showed improved metal flow, reducing turbulence compared to the top-gate. The filling was more sequential. However, the solidification analysis indicated that while the overall sequence was better, thermal hot spots formed at the junctions where the gates met the impeller body and at the center of the blade span. These localized areas, which were the last to solidify and feed, were flagged by the software as potential sites for micro-porosity and shrinkage. This outcome underscored that while filling was improved, the feeding logic was not yet optimal for the specific geometry.
The bottom-gating system demonstrated the most favorable simulation outcomes. The metal entered calmly at the base and rose steadily to fill the impeller blades, minimizing turbulence and air entrapment. Most importantly, the solidification progressed sequentially from the thin blade tips (farthest from the gate) back towards the thicker hub and finally into the gating system itself, which acted as an effective feeder (riser). This directional solidification is ideal for sound casting production. The defect prediction plot showed that the critical areas of the impeller—the blades and the structural hub—were virtually free of predicted shrinkage. Any potential defects were isolated to the gating system, which is subsequently removed and discarded. This confirmed the principle of establishing a strong thermal gradient from the casting to the riser. The solidification time ts for a section can be approximated by Chvorinov’s rule:
$$t_s = B \left( \frac{V}{A} \right)^n$$
where $V$ is volume, $A$ is surface area, $B$ is a mold constant, and $n$ is an exponent (typically ~2). The bottom-gate design ensured that sections of the casting had a higher $V/A$ ratio (like the hub) solidified later than those with a lower $V/A$ ratio (like the blades), facilitating effective feeding.
| Gating Scheme | Filling Behavior | Solidification Sequence | Predicted Defect Location | Suitability |
|---|---|---|---|---|
| Top-Gating | Highly turbulent; free-fall impact. | Reverse (top & blades first). | Substantial shrinkage in impeller base. | Poor |
| Side-Gating | Moderate turbulence; improved fill. | Partially directional with hot spots. | Micro-porosity at gate-blade junctions. | Moderate |
| Bottom-Gating | Calm, progressive bottom-up fill. | Excellent directional solidification. | Defects confined to gating system. | Excellent |
Based on the compelling simulation results, the bottom-gating system was selected as the optimal design for the physical investment casting process. The complete assembly, including the impeller pattern and the optimized gating/feeding channels, was 3D printed as a single PLA piece.
The experimental investment casting process proceeded with this printed pattern. A high-strength, fine-grained石膏 (gypsum-based) investment material was chosen for the shell due to its excellent replication fidelity, low thermal conductivity (promoting better feeding), and suitability for aluminum alloys. The slurry was prepared under vacuum to eliminate air bubbles. The printed pattern cluster was repeatedly dipped, stuccoed, and dried to build a shell of adequate thickness.
The crucial burn-out stage was carefully controlled. The mold was placed in a furnace and the temperature was gradually ramped to over 750°C. This profile ensured the PLA pattern completely pyrolyzed and volatilized without causing shell cracking from rapid gas expansion, leaving a perfectly clean cavity. The mold was then cooled to the preheat temperature of approximately 650°C to minimize thermal shock during pouring and to slow solidification, aiding feeding.
ZL104 aluminum alloy was melted in a crucible furnace and degassed. At a superheat of 750°C, the metal was poured into the preheated ceramic mold in a smooth, continuous motion. After solidification and cooling, the ceramic shell was mechanically broken away. The casting cluster was retrieved, and the individual impeller castings were separated from the gating system using a cutting wheel.
The final as-cast impeller was visually inspected and then cleaned by light grit blasting and machining. The result validated the simulation predictions. The impeller blades were fully formed with excellent surface detail replication from the 3D printed pattern. No visible defects such as mistruns, cold shuts, or gross shrinkage cavities were present on the critical functional surfaces of the blades and hub. A small amount of oxide film was observed, which is typical and removable by standard finishing processes. The success of this trial confirms the fidelity of the numerical simulation in optimizing the investment casting process parameters and the flawless integration of the FDM-printed pattern into the traditional shell-building and metal-pouring workflow.
In conclusion, this research successfully demonstrates a robust framework for producing complex thin-walled castings. The integration of FDM-based additive manufacturing for pattern fabrication eliminates tooling costs and lead times. More critically, the use of numerical simulation software like AnyCasting provides a powerful, pre-emptive tool to design and optimize the gating and feeding system—the heart of a successful investment casting process. By simulating filling and solidification, potential defects can be identified and rectified in the virtual domain before any metal is poured. For the specific impeller, the bottom-gating system was analytically and empirically proven to establish the necessary directional solidification for soundness. This hybrid approach, combining the flexibility of AM with the predictive power of simulation and the proven capabilities of the investment casting process, presents a highly efficient and reliable pathway for prototyping and low-volume manufacturing of high-performance components across aerospace, automotive, and energy industries. It effectively decouples the economic feasibility from volume, allowing for the cost-effective production of complex geometries even in single-digit quantities.