The transition from a stamped and welded fabrication method to investment casting for a complex turbine component presented a significant manufacturing challenge. The fundamental difficulty stemmed from the part’s extreme geometry: a thick-walled shell housing 21 intricate, thin-walled curved vanes, each connected to an internal semicircular ring. This disparity in wall thicknesses, coupled with the vane’s slender profile (1 mm) and their integration with the ring, rendered traditional fabrication methods inefficient and inadequate for meeting stringent performance parameters. This account details the first-hand, meticulous process refinement undertaken to successfully produce this component via conventional (non-vacuum) investment casting.
The core challenge in investment casting this part lay in replicating its internal geometry. The vanes and the internal ring formed an inaccessible cavity that was impossible to mold in a single, solid piece of tooling. Therefore, the strategy involved splitting the wax pattern into two major components: one for the main shell and another for the complex internal core comprising the vanes and ring. These would be fabricated separately and then joined with precision.
1. Foundational Precision: Wax Pattern Engineering
The success of any investment casting process is predicated on the quality of the initial wax pattern. For this turbine, pattern engineering was paramount.
1.1 Pattern Assembly via Adhesive Bonding
Instead of using traditional hot irons for welding wax components—a method prone to causing local distortion or wax bead formation—a specialized adhesive wax was employed. The key to a seamless, strong joint was controlling the wax’s viscosity and application timing. The adhesive wax was heated to approximately 70°C, significantly higher than the typical 60°C, to achieve optimal fluidity. The ring pattern was dipped into this wax for no more than two seconds. Crucially, it was not joined immediately. The wax was brushed to ensure even distribution, and a pause of 5-7 seconds allowed slight cooling and thickening, preventing “slump” or pooling at the joint interface. The ring was then carefully positioned and bonded into the vane assembly cavity. This method produced a clean, dimensionally stable joint critical for the final casting integrity.
1.2 Rigorous Wax Reclamation and Preparation
To ensure the wax pattern material was free from contaminants that could cause surface defects, a three-stage filtration system was implemented for reclaimed wax:
- First Filtration: As wax is melted out from the ceramic shell (using hot water), it passes through a filter before entering the reclamation tank.
- Second Filtration: After chemical treatment (e.g., acid washing to remove impurities), the wax is filtered again before flowing into a settling tank.
- Third Filtration: Finally, before being cast into large ingots for reuse, the wax undergoes a final filtration.
Furthermore, these large wax ingots were not simply melted down. They were placed on a flaking machine and shaved into thin, uniform flakes. This flake form allows for much quicker, more homogeneous, and gentler mixing into a new paste for injection, effectively eliminating unmelted granules or inhomogeneities that could translate into pattern surface imperfections.
1.3 Controlled Pattern Cooling and Handling
Pattern dimensional stability was enforced through strict environmental control. The pattern injection room was maintained at or below 25°C. Upon extraction from the die, the delicate vane and ring patterns were not quenched in water—a common but potentially distorting practice. Instead, they were laid flat on a table and allowed to cool naturally to ambient temperature for a minimum of two hours before any handling or gating. The assembly via adhesive wax was completed within three hours of demolding to minimize stress from differential aging or warpage.
2. Strategic Gating and Feeding System Design
Given the part’s thin sections and the need for complete filling, centrifugal casting was identified as essential. Three distinct gating system designs were developed and trialed to optimize feeding and yield.
| System Design | Configuration | Primary Objective | Observations & Outcome |
|---|---|---|---|
| Design 1: Spherical Riser with Triple Feeder | A large spherical feeder head located above the part, connected via three substantial feeder gates. | Maximize feeding distance according to riser effect zone principles, aid dewaxing, venting, and provide shell rigidity. | Provided good feeding but resulted in significant shell material usage and challenging cleanup. Moderate yield. |
| Design 2: Spherical Riser with Quad-Feeder | Spherical feeder head with four radial, branching gates leading to the part’s top circumference. | Improve metal distribution and temperature gradient compared to Design 1. | Better fill and thermal distribution than Design 1, but cleanup of the multiple gate junctions remained complex. |
| Design 3: Integrated Top Runner Ring | A continuous, ring-shaped runner attached to the entire top perimeter of the turbine shell, fed by several vertical sprues from a central pouring cup. | Provide uniform, rapid filling from all points; create a massive thermal reservoir to feed thin vanes; simplify post-casting cutoff. | Optimal. Achieved complete, turbulence-free filling. The ring acted as an excellent thermal feeder, preventing shrinkage in vanes. Cleanup was simpler despite larger runner mass. Demonstrated superior performance for this geometry. |
The superiority of Design 3 can be analyzed through the concept of the feeding range. For a cylindrical feeding channel, the effective feeding distance (L) can be approximated by considering the section thickness (T). A common rule in steel casting is:
$$ L \approx 4.5 \times T $$
For a thin section like the 1 mm vane, the isolated feeding distance would be very limited (~4.5 mm). The integrated top ring in Design 3 effectively acts as a massive, hot section directly attached to each vane root, vastly extending the effective thermal feeding distance and ensuring soundness throughout the vane network.
3. Ceramic Shell Building: A Multi-Stage Approach
The shell-building process was tailored to create a robust mold capable of withstanding centrifugal forces while allowing for the removal of ceramic from the intricate internal passages post-casting.
3.1 Primary Shell Build Sequence
The initial shell-building process was designed for maximum strength and refractoriness:
| Layer | Slurry (Mullite) | Stucco (Mullite Grit) | Drying Conditions | Special Notes |
|---|---|---|---|---|
| 1st (Prime) | Dip: 35 s | 80-100 mesh | 10 h @ 23°C, 65% RH | Critical for surface finish replication. |
| 2nd | Dip: 22 s | 60-80 mesh | 12 h @ 23°C, 65% RH | Builds thickness. |
| 3rd | Dip: 15 s | 60-80 mesh | 12 h @ 23°C, 50% RH (forced air) | Wire reinforcement applied after drying. |
| 4th | Dip: 14 s | 30-60 mesh | 12 h @ 23°C, 50% RH (forced air) | Intermediate strength layer. |
| 5th & 6th | Dip: 14 s | 16-30 mesh | 12 h each @ 23°C, 50% RH (forced air) | Provides bulk and mechanical strength. |
| Seal Coat | Dip: 14 s | None | 16 h @ 23°C, 50% RH (forced air) | Smooths outer surface and seals stucco. |
3.2 Process Optimization for Cleanability
A major post-casting issue was the difficulty in removing the ceramic core from between the thin vanes. The process was modified after the 2nd coat. Before applying the 3rd slurry layer, the internal vane cavity was manually filled (“invested”) with loose 60-80 mesh mullite sand. The openings were then temporarily sealed with a thick paste made of slurry and mullite flour. This created a permeable, friable core within the cavity. After the remaining shell layers were applied and fired, this internal sand core remained loosely bonded, allowing it to be easily broken up and removed after casting, significantly improving cleanability without compromising the mold’s integrity during pouring.
3.3 Dewaxing and Shell Preparation
Steam autoclave dewaxing was employed for its speed and ability to minimize wax expansion stress on the shell. Immediately after dewaxing, while the shell was still hot, it was immersed in and vigorously flushed with near-boiling water. This critical step served to dissolve and flush out any residual wax that had migrated into the ceramic matrix, which if left behind, could burn out during firing and leave carbonaceous defects on the casting surface.
4. Thermal Preparation: Two-Stage Shell Firing
To ensure complete removal of volatiles and achieve high-temperature strength, a two-stage firing process was adopted.
- Pre-firing: Shells were fired to approximately 950°C. This burns out any remaining organic material, completes the ceramic sintering, and develops initial strength. After cooling, the shells underwent another internal washing to remove any ash or loose material from the pre-fired internal sand core.
- Packing and Final Firing: The pre-fired shells were then placed into cylindrical steel drums. The space around the shells was filled with a coarse granular refractory, and the top surface of this packing material was brushed with a thin coating of silica sol binder to prevent loose sand from contaminating the molten metal during pouring. The entire drum was then fired to the final pouring temperature of 1100-1150°C and held at this temperature for a minimum of 30 minutes to ensure thermal uniformity. This packing method supports the shell against centrifugal force and minimizes heat loss during transfer to the casting station.
5. Melting, Pouring, and Solidification Dynamics
The final phase of the investment casting process required precise synchronization of metal treatment, mold temperature, and pouring dynamics.
5.1 Centrifugal Casting Parameters
A dedicated centrifugal casting machine was utilized. The rotational speed (ω) was calculated based on the desired gravitational multiplier (G-factor), which is the ratio of centrifugal acceleration to gravitational acceleration. The G-factor is given by:
$$ G = \frac{\omega^2 r}{g} $$
where:
- \(\omega\) is the angular velocity in rad/s,
- \(r\) is the effective radius of rotation (from axis to part) in meters,
- \(g\) is the acceleration due to gravity (9.81 m/s²).
For effective filling of thin sections, a G-factor between 40 and 80 is often targeted. For this turbine with an effective radius of approximately 0.15 m, a speed of 293 RPM was determined empirically. Converting to rad/s:
$$ \omega = \frac{293 \times 2\pi}{60} \approx 30.7 \text{ rad/s} $$
Thus, the calculated G-factor was:
$$ G = \frac{(30.7)^2 \times 0.15}{9.81} \approx 14.4 $$
While this G-factor appears lower than typical literature values, it proved sufficient when combined with the optimal gating of Design 3 and precise temperature control, highlighting the importance of system-specific tuning in investment casting.
5.2 Non-Vacuum Melting and Pouring Practice
The alloy was ZG310-570 (a cast carbon steel). Melting was conducted in a medium-frequency induction furnace under a protective slag cover.
- Deoxidation Practice: After complete melting and superheating to 1560-1570°C, preheated manganese iron (0.20% addition) and silicon iron (0.10% addition) were added for preliminary deoxidation. Slag was thoroughly removed.
- Final Deoxidation: Following slag removal, a final, strong deoxidizer, aluminum (0.03% addition), was plunged into the bath to ensure a fully killed steel, minimizing the risk of gas porosity during solidification.
- Temperatures: The target tap temperature was 1610-1620°C. The high mold temperature (1100-1150°C) significantly reduced the thermal shock and allowed the metal to remain fluid long enough to fill the thin vanes under centrifugal force.
- Pouring Protocol: Small, preheated “teapot” style ladles (10 kg capacity) were used. This ensured one full ladle per shell mold, preventing temperature drop from leftover metal in a larger ladle. Any excess metal was returned to the furnace. Pouring was completed rapidly within a 5-8 second window. Rotation was initiated just before metal entered the mold and was maintained until the feed ring was full, at which point rotation was stopped.
5.3 Controlled Cooling and Finishing
To prevent thermal cracking from overly rapid cooling or uneven stresses, the packed drums containing the hot castings were not broken open immediately. They were allowed to cool slowly within the insulating packing sand for at least three hours at room temperature (longer in colder ambient conditions). Water quenching was strictly prohibited. After shakeout, the gating system (the integrated top ring) was removed via abrasive cutting or sawing. Shell removal was followed by gentle blasting with fine steel shot (≤0.3 mm diameter) to clean the surface without peening over or damaging the delicate vane edges.
6. Conclusion
The successful production of this complex turbine component through conventional investment casting was not the result of a single revolutionary change, but rather the cumulative effect of rigorous, detailed control across every stage of the process. Key takeaways include:
- Pattern Integrity is Foundational: The shift to adhesive wax bonding and stringent wax management directly contributed to dimensionally accurate, defect-free starting patterns.
- System Design is Critical: The evolution from multiple discrete feeders to an integrated top-feeding ring was the decisive factor in achieving complete filling and sound solidification for the thin-walled vane structure.
- Process Adaptability is Essential: Modifications like the internal sand core for cleanability and the two-stage firing with packing demonstrate how standard investment casting processes must be tailored to specific part geometries.
- Synchronization is Key: The interlocking requirements of high mold temperature, precise metal temperature and chemistry, controlled centrifugal rotation, and planned cooling demand seamless coordination between furnace, mold preparation, and casting teams.
This project conclusively demonstrates that even for highly challenging geometries with extreme section variations, a meticulously controlled, non-vacuum investment casting process is capable of producing high-integrity castings that meet demanding performance specifications, offering a viable and efficient alternative to fabricated assemblies.