In the development of advanced turbine components for gas turbine engines, the investment casting process plays a pivotal role due to its ability to produce complex, high-integrity parts with excellent surface finish and dimensional accuracy. As an engineer specializing in casting technology, I have focused on optimizing the investment casting process for critical components like turbine nozzles, which operate under extreme thermal conditions. The turbine nozzle is a core component that directs high-temperature gas flow onto turbine blades, enabling energy conversion. Its service environment demands exceptional resistance to thermal stress and creep, making the quality of the investment casting process paramount. Traditional trial-and-error methods for process development are time-consuming and costly, often failing to reveal underlying physical phenomena during casting. Therefore, in this study, I employed numerical simulation techniques to analyze and optimize the investment casting process, aiming to enhance metallurgical quality, reduce defects, and improve economic efficiency through increased metal yield. The integration of simulation tools like ProCAST allows for a deep understanding of filling and solidification dynamics, facilitating rapid iteration and robust process design. This approach aligns with modern manufacturing paradigms where simulation-driven development is essential for competitiveness.
The investment casting process involves creating a ceramic mold around a wax pattern, which is melted out to form a cavity for molten metal. For the turbine nozzle, which is typically made from high-temperature alloys such as K4169, the process must ensure complete filling of thin sections, controlled solidification to avoid defects, and efficient use of material. My initial investigation centered on a specific turbine nozzle design characterized by an intricate geometry with an inner ring, outer ring, flange, and 51 airfoil-shaped blades. The blades exhibit a crescent-like profile with thickness variations, where the thinnest section is only 0.5 mm, while the flange region reaches 24 mm in thickness. This disparity in section thickness poses significant challenges in the investment casting process, as it can lead to issues like mist runs, cold shuts, shrinkage porosity, and hot tearing if not properly managed. To address these, I first developed an initial gating system design based on conventional wisdom, utilizing a top-and-bottom filling approach with multiple risers. However, through numerical simulation, I identified several shortcomings that necessitated optimization.
To quantitatively assess the turbine nozzle geometry, I summarized key dimensions in Table 1. These parameters are critical for setting up the numerical model and understanding the thermal gradients during the investment casting process.
| Component | Dimension | Value (mm) |
|---|---|---|
| Inner Ring | Diameter | 236 |
| Inner Ring | Thickness | 8 |
| Outer Ring | Diameter | 366 |
| Outer Ring | Minimum Thickness | 2 |
| Flange | Thickness | 24 |
| Blades | Number | 51 |
| Blades | Minimum Thickness | 0.5 |
For the numerical simulation of the investment casting process, I used ProCAST software, which employs finite element methods to solve governing equations for fluid flow, heat transfer, and solidification. The key equations include the Navier-Stokes equations for incompressible flow during filling:
$$ \rho \left( \frac{\partial \mathbf{v}}{\partial t} + \mathbf{v} \cdot \nabla \mathbf{v} \right) = -\nabla p + \mu \nabla^2 \mathbf{v} + \rho \mathbf{g} $$
where $\rho$ is density, $\mathbf{v}$ is velocity, $p$ is pressure, $\mu$ is dynamic viscosity, and $\mathbf{g}$ is gravity. The energy equation for heat transfer is:
$$ \rho c_p \frac{\partial T}{\partial t} = \nabla \cdot (k \nabla T) + Q $$
with $c_p$ as specific heat, $T$ as temperature, $k$ as thermal conductivity, and $Q$ as latent heat release during solidification. The solidification kinetics can be described using a fraction solid model, such as:
$$ f_s = 1 – \exp(-k (T_{liq} – T)^n) $$
where $f_s$ is solid fraction, $k$ and $n$ are material constants, and $T_{liq}$ is liquidus temperature. These equations are solved iteratively to predict flow patterns, temperature distribution, and defect formation in the investment casting process.
My initial investment casting process design featured a gating system with a top-and-bottom filling arrangement, including a pouring cup, horizontal runners, and six large risers attached to the flange. The mold shell was made of mullite with a thickness of 10 mm, and the simulation parameters were set as summarized in Table 2. The alloy used was K4169, a nickel-based superalloy commonly employed in high-temperature applications due to its excellent creep resistance and oxidation stability.
| Parameter | Setting |
|---|---|
| Alloy Material | K4169 |
| Mold Material | Mullite |
| Mold Thickness | 10 mm |
| Pouring Temperature | 1500°C |
| Mold Preheat Temperature | 1050°C |
| Cooling Environment | Vacuum Cooling |
| Pouring Time | 4 s |
| Interface Heat Transfer Coefficient | 500 W/m²K |
| Simulation Convergence Tolerance | 1×10⁻⁵ |
The simulation of the initial investment casting process revealed several issues during filling. At 40% filling, metal entered the inner ring and blades smoothly, but as filling progressed to 45%, turbulent flow occurred in the outer ring and flange region due to the confluence of metal from top runners and bottom gates. This turbulence can be analyzed using the Reynolds number:
$$ Re = \frac{\rho v D_h}{\mu} $$
where $D_h$ is hydraulic diameter. High Re values indicate turbulent flow, which increases the risk of gas entrapment and slag inclusion. In this case, Re exceeded critical thresholds locally, leading to potential defects. By 60% filling, the risers were filled unevenly, further exacerbating flow instability. The filling sequence deviated from ideal laminar flow, compromising the integrity of the investment casting process.
During solidification, the initial process showed non-sequential solidification, with thin sections like blades solidifying first but isolated liquid pockets forming in thicker regions. The solidification time can be estimated using Chvorinov’s rule:
$$ t_s = C \left( \frac{V}{A} \right)^2 $$
where $t_s$ is solidification time, $C$ is a constant dependent on mold material and alloy, $V$ is volume, and $A$ is surface area. For the flange with high $V/A$ ratio, solidification was delayed, creating hot spots. As a result, shrinkage porosity and microporosity defects were predicted in the blades, blade-root junctions, and between risers on the flange. The defect formation is often linked to inadequate feeding, which can be quantified by the Niyama criterion:
$$ Ny = \frac{G}{\sqrt{\dot{T}}} $$
where $G$ is temperature gradient and $\dot{T}$ is cooling rate. Low Ny values indicate a high risk of shrinkage porosity. Simulation results showed Ny values below critical levels in defect-prone areas. Additionally, the metal utilization rate, defined as the ratio of casting weight to total poured weight, was only 12.13% for the initial investment casting process, indicating excessive gating and riser material that added to cost without benefit.
To optimize the investment casting process, I implemented several modifications based on the simulation insights. First, I changed the filling method from top-and-bottom to a bottom-gating system with a “funnel + cylinder” pouring cup to promote steady, upward filling and reduce turbulence. Second, I increased the number of risers from 6 to 8 while reducing their length from 110 mm to 40 mm, distributing them evenly around the flange to enhance feeding and eliminate thermal hot spots. Third, I incorporated iron sand filling between the blades to accelerate cooling and enforce directional solidification from thin to thick sections. These adjustments aimed to achieve a controlled investment casting process with sequential solidification and minimized defects.
The optimized investment casting process was modeled and simulated with the same software and material properties. Key changes are summarized in Table 3, along with a comparison of metal utilization rates. The optimization significantly reduced the gating system weight while improving feeding efficiency.
| Parameter | Initial Process | Optimized Process |
|---|---|---|
| Filling Method | Top-and-Bottom | Bottom Gating |
| Pouring Cup Shape | Gyroscopic | Funnel + Cylinder |
| Number of Risers | 6 | 8 |
| Riser Length | 110 mm | 40 mm |
| Additional Cooling | None | Iron Sand Between Blades |
| Gating System Weight | 59.77 kg | 16.79 kg |
| Casting Weight | 7.25 kg | 7.25 kg |
| Metal Utilization Rate | 12.13% | 43.18% |
The simulation of the optimized investment casting process demonstrated marked improvements. During filling, metal flowed smoothly from the pouring cup through the runner into the bottom gate, gradually filling the inner ring, blades, outer ring, and flange without turbulent effects. The flow velocity remained within optimal ranges, ensuring a stable investment casting process. By 60% filling, the blades were completely filled, and the outer ring showed uniform filling progression. At 90% filling, only minor sections of risers remained unfilled, indicating efficient use of metal.
Solidification analysis revealed sequential solidification, starting from the blades and moving toward the risers and bottom gate. The temperature gradient $G$ and cooling rate $\dot{T}$ were calculated using:
$$ G = \frac{\Delta T}{\Delta x}, \quad \dot{T} = \frac{dT}{dt} $$
where $\Delta T$ is temperature difference over distance $\Delta x$. In the optimized investment casting process, $G$ was higher in critical regions, and $\dot{T}$ was controlled to promote directional solidification. The solid fraction evolution followed a predictable pattern, with no isolated liquid zones. Defect prediction models, including the Niyama criterion, indicated no shrinkage porosity or hot tears in the casting. The metal utilization rate increased to 43.18%, representing a 3.6-fold improvement over the initial investment casting process, which translates to significant cost savings in material and energy consumption.
To validate the optimized investment casting process, I oversaw the production of trial castings using the modified gating design. Ten turbine nozzles were manufactured under controlled conditions, employing the same K4169 alloy and mold materials as in the simulation. After casting, each component underwent non-destructive testing using X-ray radiography in accordance with ASTM E1742 standards. The inspection focused on detecting internal defects such as cracks, porosity, inclusions, and shrinkage cavities. All castings met the stringent requirements of specification EMS52301/2, with no discernible defects in critical areas. This successful verification confirms the accuracy of the numerical simulation and the robustness of the optimized investment casting process. Subsequently, a batch of 50 castings was produced using this process, consistently achieving high quality and demonstrating scalability for industrial applications.
In conclusion, numerical simulation is an indispensable tool for optimizing the investment casting process, particularly for complex components like turbine nozzles. Through detailed analysis of filling and solidification dynamics, I identified inefficiencies and defect mechanisms in the initial process. By implementing targeted modifications—such as switching to bottom gating, resizing risers, and adding cooling media—I developed an optimized investment casting process that ensures stable filling, sequential solidification, and high metal yield. The simulation-driven approach reduced development time and cost while enhancing product reliability. This study underscores the value of integrating advanced simulation techniques into the investment casting process to achieve superior metallurgical quality and economic efficiency. Future work could explore further refinements, such as adaptive mesh refinement for higher accuracy or multi-objective optimization to balance multiple performance metrics in the investment casting process.