My research focuses on the investment casting process of a K4222 superalloy thin-walled casting, specifically an aeroengine pre-swirl nozzle. This component is characterized by a maximum outer ring diameter of 365.73 mm, a minimum central wall thickness of only 1.53 mm, and a total weight of approximately 7.21 kg. The geometry includes numerous curved surfaces, abrupt wall thickness transitions, and hollow thin-walled struts, making it a challenging casting to produce without defects. The primary objective of my study was to develop a reliable casting process that minimizes casting defects while meeting stringent dimensional and mechanical property requirements.
1. Introduction
K4222 superalloy is a nickel-based precipitation-hardened cast superalloy, an equivalent of GTD222, designed for long-term service at temperatures below 1000°C. It offers a balanced combination of high-temperature strength, excellent elongation, fatigue resistance, and outstanding oxidation and corrosion resistance. These properties make it suitable for static components such as guide vanes and pre-swirl nozzles in aeroengines. However, its wide melting temperature range (approximately 1292°C–1344°C) and tendency to form shrinkage porosity during solidification present significant challenges for thin-walled precision castings.
Investment casting, also known as lost-wax casting, is the preferred manufacturing process for complex near-net-shape components. It provides excellent surface finish and dimensional accuracy, which are indispensable for aeroengine parts. Nevertheless, the process involves multiple interrelated parameters, including pouring temperature, mold preheating temperature, pouring time, and the design of the gating and riser system. Inappropriate combinations of these parameters frequently lead to casting defects such as misruns, cold shuts, shrinkage cavities, and porosity.
The aim of my work was to establish a robust investment casting route for the pre-swirl nozzle by combining thermodynamic calculations, numerical simulation, and experimental validation. I systematically studied the effects of pouring temperature, shell temperature, and pouring speed on the formation of casting defects using thin-walled benchmark samples. Based on these findings, I optimized the gating system and process parameters for the actual nozzle. Finally, I evaluated the quality of the castings through non-destructive testing, dimensional inspection, and mechanical property tests.
2. Materials and Methods
2.1 Material Composition and Properties
The chemical composition of K4222 superalloy used in this study is listed in Table 1. The main strengthening phase is γ′-Ni₃(Al,Ti), which precipitates coherently with the γ matrix. The alloy also contains MC-type carbides formed by niobium, tantalum, and titanium. These microstructural constituents influence the solidification behavior and, consequently, the susceptibility to casting defects.
Table 1: Chemical composition of K4222 superalloy (wt.%)
| Ni | C | Cr | Co | W | Al | Ti | Nb | B | Ta | Zr | O |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Bal. | 0.1 | 22.73 | 18.95 | 1.9 | 1.19 | 2.35 | 0.86 | 0.0053 | 1.06 | 0.012 | 0.0013 |
For numerical simulation, accurate thermophysical properties are essential. I calculated the thermal conductivity, density, specific heat, and solid fraction of K4222 using two independent approaches: JMatPro software and the ProCAST material database. The predicted values were compared with data from the high-temperature alloy handbook. A best-fit set of thermophysical parameters was then selected for simulation, as shown in Figure 1. These parameters were used as input for all subsequent simulation work.
Figure 1: Optimum thermophysical properties of K4222 alloy used for simulation.
2.2 Numerical Simulation Setup
The casting process was simulated using ProCAST, a finite-element based software. The mathematical model includes the continuity equation, momentum conservation, and energy conservation. For incompressible Newtonian fluids, the governing equations are:
Continuity:
$$ \frac{\partial u}{\partial x} + \frac{\partial v}{\partial y} + \frac{\partial w}{\partial z} = 0 \tag{1} $$
Momentum (Navier-Stokes):
$$ \rho \left( \frac{\partial \mathbf{v}}{\partial t} + \mathbf{v} \cdot \nabla \mathbf{v} \right) = -\nabla p + \rho \mathbf{g} + \mu \nabla^2 \mathbf{v} \tag{2} $$
Energy:
$$ \rho c_p \left( \frac{\partial T}{\partial t} + \mathbf{v} \cdot \nabla T \right) = \nabla \cdot (k \nabla T) + S \tag{3} $$
where \( \rho \) is density, \( \mathbf{v} \) is velocity, \( p \) is pressure, \( \mu \) is dynamic viscosity, \( c_p \) is specific heat, \( k \) is thermal conductivity, and \( S \) is the heat source term accounting for latent heat release during solidification.
Prediction of casting defects, particularly shrinkage porosity and cavities, was performed using the Niyama criterion and temperature gradient methods. The Niyama parameter is defined as:
$$ N = \frac{G}{\sqrt{R}} \tag{4} $$
where \( G \) is the temperature gradient and \( R \) is the cooling rate. A low Niyama value indicates insufficient feeding and high risk of microporosity.
2.3 Experimental Procedures
I designed a thin-walled benchmark sample (100 mm × 100 mm × 1 mm) to systematically investigate the influence of process parameters on casting defects. The benchmark was top-gated with a large ingate to avoid premature freezing. I varied the pouring temperature (1400°C, 1450°C, 1500°C, 1550°C), the shell preheating temperature (1050°C, 1100°C, 1150°C, 1200°C), and the pouring time (3.5 s, 4.0 s, 4.5 s, 5.0 s). The simulations provided the optimum process window for the nozzle.
After determining the optimal parameters, I produced the pre-swirl nozzle castings using a vacuum investment casting furnace with a vacuum level below 3 Pa. The gating system was optimized iteratively through simulation and experimental trials. The final castings were subjected to non-destructive testing, including fluorescent penetrant inspection, X-ray radiography, and blue-light scanning for dimensional verification. Additionally, heat-treated test bars were machined for tensile and stress-rupture testing at room temperature and elevated temperatures.
3. Initial Gating System Design
3.1 Structural Analysis of the Pre-Swirl Nozzle
The pre-swirl nozzle is a cylindrical ring-like component with a conical outer wall and a hollow inner ring. It contains twelve hollow thin-walled struts, three of which include flanges. The wall thickness varies significantly: the upper and lower flanges are relatively thick, while the conical wall and struts are as thin as 1.53 mm. More than 90% of the surface area is classified as thin-walled (below 4 mm). This geometry poses risks of misrun, cold shut, and shrinkage-related casting defects.
To overcome these challenges, I adopted a top-gated gravity pouring system. The alloy is poured from the top, and the riser system is designed to provide adequate feeding. The pouring temperature was initially set in the range of 1400°C–1550°C, based on the alloy’s liquidus temperature. The shell preheating temperature was set between 1050°C and 1200°C to reduce the cooling rate of the melt and improve fillability.
3.2 Pouring Time and Liquid Rise Velocity
For thin-walled castings, the pouring time \( t \) can be estimated by the empirical formula:
$$ t = S \sqrt[3]{G} \tag{5} $$
where \( G \) is the total pouring mass in kg and \( S \) is an experience coefficient. For wall thicknesses of 1.53 to 4 mm, I selected \( S = 3 \) to 4.3. With a total pouring mass of about 7.21 kg, the pouring time ranged from 3.5 s to 5 s.
The liquid rise velocity \( v \) is given by:
$$ v = \frac{H}{t} \tag{6} $$
where \( H \) is the total height of the gating system (approximately 480 mm). Substituting the values, \( v \) ranges from 96 mm/s to 137 mm/s, which is appropriate for thin-walled castings below 4 mm thickness (typically 30–100 mm/s, but higher speeds are acceptable for complex geometries).
3.3 Riser Design and Verification
I designed the risers based on the modulus concept. The modulus \( M \) is defined as the ratio of volume to surface area:
$$ M = \frac{V}{A_{\text{cooling}}} \tag{7} $$
For directional solidification, the modulus of the riser must be greater than that of the casting section it feeds. I also verified the feeding capacity using the volume shrinkage criterion:
$$ \eta V_{\text{riser}} \geq \varepsilon (V_{\text{casting}} + V_{\text{riser}}) \tag{8} $$
where \( \eta \) is the riser efficiency and \( \varepsilon \) is the volume shrinkage of the alloy. With \( \varepsilon \) ≈ 5% and \( \eta \) ≈ 0.15, the designed risers provided a sufficient feed metal volume.
4. Numerical Simulation of Casting Defects
4.1 Effect of Pouring Temperature on Casting Defects
Simulations on the benchmark sample revealed a strong influence of pouring temperature on the formation of casting defects. Figure 2 shows the temperature field at 35% filling for different pouring temperatures. At 1400°C, the melt rapidly loses temperature and the front becomes irregular, leading to potential cold shuts. At 1500°C, the melt spreads uniformly with a smooth fan-shaped front, indicating good fluidity.
Figure 3 displays the shrinkage porosity prediction for each pouring temperature. The porosity fraction was quantified as a function of pouring temperature, as listed in Table 2. At 1550°C, the cooling time becomes so long that grain coarsening and increased microporosity occur. Therefore, 1500°C was identified as the optimal pouring temperature.
Table 2: Effect of pouring temperature on porosity fraction.
| Pouring Temperature (°C) | Relative Porosity (%) | Observation |
|---|---|---|
| 1400 | 0.42 | Misrun risk, high porosity near surfaces |
| 1450 | 0.31 | Improved filling, still casting defects |
| 1500 | 0.22 | Minimum porosity, smooth filling |
| 1550 | 0.29 | Increased porosity due to slow cooling |
4.2 Effect of Shell Preheating Temperature
The shell preheating temperature affects the cooling rate and, consequently, the intensity of casting defects. As shown in Figure 4, increasing the shell temperature from 1050°C to 1200°C reduces the temperature gradient and prolongs solidification. While this promotes filling, it also increases the tendency for centerline shrinkage. The porosity–shell temperature relationship is given in Table 3.
Table 3: Effect of shell preheating temperature on porosity fraction.
| Shell Temperature (°C) | Relative Porosity (%) | Filling Behavior |
|---|---|---|
| 1050 | 0.19 | Risk of misruns, incomplete filling |
| 1100 | 0.22 | Balanced filling and feeding |
| 1150 | 0.27 | Better filling but more shrinkage casting defects |
| 1200 | 0.35 | Wide mushy zone, severe shrinkage |
The optimum shell temperature was chosen as 1100°C, which provided sufficient fluidity without excessive porosity.
4.3 Effect of Pouring Time
Pouring time significantly influences the temperature distribution at the end of filling. In simulations, a shorter pouring time (3.5 s) leads to a more uniform temperature field because the melt remains hotter. However, extremely short times increase the filling velocity and erosion of the shell, potentially generating inclusions and cold shots. Table 4 summarizes the effect of pouring time on the final solidification profile.
Table 4: Effect of pouring time on casting quality.
| Pouring Time (s) | Filling Velocity (mm/s) | Casting Defect Tendency |
|---|---|---|
| 3.5 | 137 | Less shrinkage, but risk of mold erosion |
| 4.0 | 120 | Good compromise |
| 4.5 | 107 | Optimal for uniform filling |
| 5.0 | 96 | Temperature loss, cold shuts |
Based on these systematic studies, I selected 1500°C pouring temperature, 1100°C shell preheating temperature, and 4.5 s pouring time as the optimized parameters for the pre-swirl nozzle.
4.4 Initial Simulation of the Nozzle Casting
Before optimization, I simulated the pre-swirl nozzle with an initial gating system. The shrinkage porosity prediction in Figure 5 showed significant casting defects in the thick flanges and near the ingate contacts. The defects were primarily located in regions where the modulus was highest and where the riser feeding was insufficient. The initial gating system employed cold risers, which solidified before the casting sections, leading to inadequate compensation for volumetric contraction.
5. Optimization of the Gating System
5.1 Improved Design
To mitigate the casting defects, I modified the gating system by adding nine horizontal runners that connected the central sprue directly to the risers on the lower flange. This design allowed the risers to be fed with hot liquid metal from the sprue during the late stages of solidification, converting the cold risers into hot risers. The improved gating system is shown in Figure 6. The key change was to provide a secondary feeding path from the sprue to the bottom cruciform risers, ensuring that the riser melt remained molten longer than the casting sections.
5.2 Filling Simulation of the Optimized System
I simulated the filling process with the optimized gating system. The liquid metal descended smoothly from the pouring cup through the sprue, split into the inner and outer runners, and filled the thin-walled regions with a nearly horizontal melt front. No severe turbulence or splashing was observed. The total filling time was 4.5 s as designed. At t = 1.25 s, the melt began filling the bottom flanges and struts; at t = 3.85 s, most of the casting was filled; and at t = 4.5 s, the casting was completely filled with all risers topped up.
5.3 Solidification Simulation and Casting Defect Prediction
The solidification sequence of the optimized casting is depicted in Figure 7. Solidification initiated at the thin-walled sections, progressed toward the thicker flanges, and the last liquid remained in the risers and gating system. The temperature gradient promoted directional solidification from bottom to top, which is essential for preventing shrinkage casting defects. The final solidification time was approximately 2.5 h.
Shrinkage porosity was predicted with a threshold of 2%. As shown in Figure 8, most porosity was located in the risers and in the top flange regions, which would be removed in subsequent machining. The defect volume in critical areas, such as the conical wall and the hollow struts, was dramatically reduced. A quantitative comparison of defect volume between the initial and optimized gating systems is given in Table 5.
Table 5: Defect volume comparison between initial and optimized gating systems.
| Location | Initial System Defect Volume (mm³) | Optimized System Defect Volume (mm³) |
|---|---|---|
| Top flange | 152 | 25 |
| Conical wall | 187 | 18 |
| Bottom flange | 342 | 37 |
| Hollow struts | 96 | 8 |
| Total | 777 | 88 |
The optimized gating system reduced the predicted casting defects by more than 85%, indicating that the new design successfully eliminated the adverse effects of cold risers.
6. Experimental Verification
6.1 Visual and Fluorescent Penetrant Inspection
The actual castings produced with the optimized parameters and gating system were first visually inspected. The cast surfaces were smooth, free from flash, cracks, and cold shuts. The contours of the thin-walled features were clearly defined, demonstrating excellent mold filling.
Fluorescent penetrant inspection (FPI) was carried out in accordance with industrial standards. The casting surface was cleaned, coated with a fluorescent penetrant, and after a suitable dwell time, the excess penetrant was removed and a developer was applied. Under ultraviolet light, any surface-connected casting defects would appear as bright indications. The inspection revealed no relevant indications larger than 0.3 mm in length. The surface quality met the acceptance criteria for aeroengine components.
6.2 X-Ray Radiography
Internal casting defects were evaluated using X-ray radiography. The castings were placed in an industrial radiography system and exposed with parameters listed in Table 6. Radiographic films were evaluated against reference radiographs per ASTM E 192.
Table 6: X-ray inspection parameters.
| Machine | Film Type | Exposure Time | Focal Distance | Tube Current | Tube Voltage |
|---|---|---|---|---|---|
| XYD3010 | N-III | 3 min | 1000 mm | 5 mA | 80 kV |
The X-ray images showed only minor diffuse microporosity in the thin-walled sections, which was within the acceptable limits. No significant shrinkage cavity or macro-porosity exceeding 0.3 mm was detected. The positions of the few porosity indications were consistent with the simulation predictions, confirming the accuracy of the numerical model.
6.3 Dimensional Accuracy Using Blue-Light Scanning
I employed a blue-light three-dimensional scanner (StereoScan neo R8) to obtain a point cloud of the cast nozzle. The scanning parameters are listed in Table 7. The point cloud was aligned with the original CAD model using Imageware software, and the deviations were color-mapped.
Table 7: Blue-light scanner parameters.
| Parameter | Value |
|---|---|
| Camera resolution | 8 megapixels |
| Minimum point spacing | 17 μm |
| Minimum measurement range | 75 mm |
| Maximum measurement range | 1100 mm |
| Measurement speed | ≥1 s |
The dimensional deviation analysis showed that more than 98.12% of the scanned data points laid within a tolerance band of ±0.75 mm. The maximum deviation was 0.755 mm. This level of dimensional accuracy satisfies the requirements for the pre-swirl nozzle, which demands minimal machining allowance.
7. Heat Treatment and Mechanical Properties
7.1 Heat Treatment Regime
After casting, the pre-swirl nozzle and the accompanying test bars were heat-treated in an argon-protected vacuum furnace. The heat treatment schedule was: solution treatment at 1150°C for 4 hours, air cooling to room temperature, followed by aging at 800°C for 8 hours and air cooling to room temperature. This standard treatment for K4222 alloy promotes the precipitation of γ′ and optimizes the mechanical properties.
7.2 Room-Temperature Tensile Properties
Test bars machined from the heat-treated coupons were subjected to room-temperature tensile tests using a Zwick Z250 universal testing machine. The results are summarized in Table 8.
Table 8: Room-temperature tensile properties of K4222 alloy test bars.
| Temperature (°C) | Yield Strength (MPa) | Tensile Strength (MPa) | Elongation (%) | Reduction in Area (%) |
|---|---|---|---|---|
| Requirement | ≥724 | ≥896 | ≥4 | ≥6 |
| Room temperature (actual) | 755 | 980 | 11 | 14 |
The measured strength and ductility significantly exceeded the specification, indicating a sound casting without harmful casting defects.
7.3 High-Temperature Tensile Properties
Tensile tests at 760°C were performed using an SHIMADZU AG-25TA machine. The results are listed in Table 9.
Table 9: High-temperature (760°C) tensile properties.
| Temperature (°C) | Yield Strength (MPa) | Tensile Strength (MPa) | Elongation (%) | Reduction in Area (%) |
|---|---|---|---|---|
| Requirement | ≥552 | ≥703 | ≥6 | ≥8 |
| Actual | 661.5 | 840.0 | 10.0 | 17.5 |
Again, all values were well above the minimum required, demonstrating the effectiveness of the optimized casting process in eliminating detrimental casting defects that could compromise high-temperature mechanical properties.
7.4 Stress-Rupture Properties
Stress-rupture tests were conducted at 900°C with an initial stress of 172 MPa. The results are shown in Table 10.
Table 10: Stress-rupture properties at 900°C/172 MPa.
| Temperature (°C) | Initial Stress (MPa) | Test Duration (h) | Elongation (%) | Reduction in Area (%) | Remarks |
|---|---|---|---|---|---|
| Requirement | 172 | ≥30 | ≥7 | ≥12 | — |
| Actual | 172 | 47.4 | 28.5 | 29.0 | Fractured |
The stress-rupture life exceeded 47 hours, with high elongation and reduction of area, indicating excellent resistance to intergranular failure and the absence of significant internal casting defects.
8. Conclusion
In this research, I successfully developed an investment casting process for K4222 superalloy thin-walled pre-swirl nozzle castings. The following conclusions can be drawn:
- The pre-swirl nozzle, with its large thin-walled areas and abrupt thickness changes, requires careful control of pouring temperature, shell preheating temperature, and pouring time to avoid casting defects. The optimal process parameters were determined as 1500°C pouring temperature, 1100°C shell preheating temperature, and 4.5 s pouring time.
- Numerical simulation using ProCAST, combined with JMatPro thermophysical calculations, accurately predicted the tendency for casting defects. The optimized gating system, featuring additional horizontal runners connecting the sprue to the lower risers, reduced the predicted porosity volume by more than 85% compared to the initial design.
- Fluorescent penetrant inspection and X-ray radiography confirmed that the castings were free from surface and internal casting defects larger than 0.3 mm. The dimensional deviation, measured by blue-light scanning, was within ±0.75 mm for over 98% of the surface, satisfying the technical requirements.
- Heat-treated test bars exhibited excellent mechanical properties: room-temperature yield strength of 755 MPa, tensile strength of 980 MPa, elongation of 11%; at 760°C, yield strength of 661.5 MPa, tensile strength of 840.0 MPa, elongation of 10%; and at 900°C/172 MPa, stress-rupture life of 47.4 h with elongation of 28.5%. All values exceeded the acceptance criteria.
The combination of simulation-driven design and experimental validation proved to be an effective approach to eliminate casting defects and manufacture high-quality thin-walled superalloy components. Further work may focus on reducing the residual microporosity in the thick flanges through hot isostatic pressing and optimizing the machining allowance to enhance productivity.