The reliability and structural integrity of critical components like the engine brake are paramount in aerospace applications. As a vital part of an aero-engine, its performance is directly dictated by the quality of the manufacturing process. The casting of such components, often featuring complex geometries with large diameters and extensive thin-walled sections, presents significant challenges. In the realm of aerospace casting, minimizing defects is non-negotiable due to the extreme operational environments. The specific geometry of a slanted support plate engine brake makes it particularly susceptible to shrinkage porosity and cavities during solidification, defects intrinsically linked to the chosen pouring process and the subsequent cooling sequence. Consequently, the optimization of the casting process through advanced simulation techniques holds substantial practical significance for enhancing the final component’s performance and reliability.
1. Physical Model of the Brake Casting Assembly
The simulation of this aerospace casting process requires a comprehensive three-dimensional model that captures all significant thermal interactions. While the brake casting itself exhibits rotational symmetry, the heating environment inside the vacuum furnace—with its non-uniform radiation fields—breaks this symmetry, necessitating a full-scale model. The assembly comprises the ceramic shell (mold), the brake casting (with its integrated gating system), a sand flask enclosing the shell, and insulating blankets. The blankets fill the gap between the shell and the flask and also cover the top of the shell assembly.
The process initiates by preheating the entire shell and flask assembly to 830°C before placing it inside a large, water-cooled vacuum furnace with a diameter of 3 meters. The superalloy melt is then poured at 1500°C over a period of 9-13 seconds, followed by cooling under vacuum. The thermal interfaces within this assembly are complex and dynamic, primarily consisting of:
- Heat transfer interface between the sand flask and the insulating blanket.
- Heat transfer interface between the insulating blanket and the ceramic shell.
- Heat transfer interface between the ceramic shell and the metal casting.
The thermal contact resistance at each of these interfaces varies with temperature. Furthermore, as the cooling occurs in a vacuum, heat loss from the outer surfaces of the sand flask and the top blanket occurs predominantly through gray-body radiation to the water-cooled furnace walls, which are maintained at a constant temperature. The emissivity of the blanket surface itself is also temperature-dependent. Given the wide temperature range involved—from 1500°C down to ambient—the thermophysical properties of all materials must be defined as functions of temperature rather than constants.
The brake casting material is the nickel-based superalloy K4169, a staple in aerospace casting for high-temperature components. The ceramic shell is made of alumina-based material. Their key temperature-dependent properties are summarized in the tables below.
| Temperature (°C) | Thermal Conductivity (W/m·K) | Specific Heat Capacity (J/kg·K) |
|---|---|---|
| 300 | 15.02 | 479 |
| 400 | 16.73 | 504 |
| 500 | 18.33 | 518 |
| 600 | 20.01 | 552 |
| 700 | 21.29 | 569 |
| 800 | 22.33 | 588 |
| Temperature (°C) | Thermal Conductivity (W/m·K) | Specific Heat Capacity (J/kg·K) |
|---|---|---|
| 26 | 6.93 | 547 |
| 593 | 4.72 | 1103 |
| 790 | 3.98 | 1137 |
| 1180 | 3.88 | 1198 |
| 1350 | 3.76 | 1229 |
The alloy has a density of 8193 kg/m³, with a liquidus temperature of 1362°C and a solidus temperature of 1198°C. The shell density is 3096 kg/m³.
2. Mathematical Framework and Numerical Method
To accurately model the formation of defects in this aerospace casting, a transient coupled analysis of fluid flow and heat transfer is essential. The filling stage must account for fluid flow to track temperature changes in the liquid stream and the cooling effect of the mold. This is seamlessly coupled with a full thermal analysis for the solidification phase. The governing equations are solved for an incompressible Newtonian fluid (the molten metal), considering the release of latent heat during phase change.
2.1 Governing Equations
The core of the model is a set of coupled partial differential equations:
Mass Conservation (Continuity):
$$ \nabla \cdot \vec{u} = 0 $$
where $\vec{u}$ is the velocity vector.
Momentum Conservation (Navier-Stokes with Darcy term):
$$ \rho \left( \frac{\partial \vec{u}}{\partial t} + \vec{u} \cdot \nabla \vec{u} \right) = -\nabla p + \mu \nabla^2 \vec{u} + \rho \vec{g} + S_{Darcy} $$
Here, $\rho$ is density, $p$ is pressure, $\mu$ is dynamic viscosity, $\vec{g}$ is gravity. The Darcy source term $S_{Darcy}$ is crucial for modeling flow in the mushy zone during solidification and is given by:
$$ S_{Darcy} = – \frac{\mu}{K} \vec{u} $$
where the permeability $K$ is a function of the liquid fraction $g_l$, often modeled using the Kozeny-Carman relation: $K = K_0 \frac{g_l^3}{(1 – g_l)^2}$.
Energy Conservation:
$$ \rho c_{eff} \frac{\partial T}{\partial t} + \rho c \vec{u} \cdot \nabla T = \nabla \cdot (k \nabla T) $$
The effective specific heat $c_{eff}$ incorporates the latent heat $L$ of fusion:
$$ c_{eff} = c + L \frac{\partial f_s}{\partial T} $$
where $f_s$ is the solid fraction, and $T$ is temperature.
Volume of Fluid (VOF) Equation (for tracking the melt front during filling):
$$ \frac{\partial F}{\partial t} + \nabla \cdot (F \vec{u}) = 0 $$
where $F$ is the volume fraction of fluid in a cell ($F=1$: liquid, $F=0$: gas/vacuum).
2.2 Defect Prediction Methodology
The solid fraction $f_s$ at any point and time is determined from the local temperature using the alloy’s solidification range. A common approach for binary/alloy systems is the lever rule or Gulliver-Scheil approximation within the mushy zone. A simplified linear relationship in the mushy zone can be expressed as:
$$ f_s(T) = \begin{cases}
0 & T \geq T_{liquidus} \\
\frac{T_{liquidus} – T}{T_{liquidus} – T_{solidus}} & T_{solidus} < T < T_{liquidus} \\
1 & T \leq T_{solidus}
\end{cases} $$
The formation of shrinkage porosity is predicted using a criterion based on mass conservation. As the metal solidifies and cools, its density increases, leading to volumetric shrinkage. If this shrinkage cannot be compensated by feeding from liquid-rich regions (risers), microscopic or macroscopic voids form. The local volumetric shrinkage deficit, $\Delta V$, can be related to the local solidification conditions. The Niyama criterion is a widely used predictive index for shrinkage porosity in castings, defined as:
$$ N_y = \frac{G}{\sqrt{\dot{T}}} $$
where $G$ is the temperature gradient at the solidus front and $\dot{T}$ is the local cooling rate. Regions where $N_y$ falls below a critical threshold (specific to the alloy) are predicted to be susceptible to shrinkage porosity. For this aerospace casting simulation, a more direct method of calculating the fractional porosity volume based on the cumulative shrinkage that cannot be fed is employed, derived from the solved thermal and fluid fields.
2.3 Numerical Discretization and Mesh
The Finite Volume Method (FVM) is employed for the numerical solution of the coupled equations due to its inherent strength in conserving mass, momentum, and energy. The computational domain for the coupled fluid-thermal analysis is discretized into a high-resolution mesh. The details of the mesh for the original process model are as follows:
| Component | Number of Elements (×10⁴) | Maximum Element Size (mm) | Minimum Element Size (mm) |
|---|---|---|---|
| Casting (with gates) | 209.3 | 12 × 12 | 4 × 1 |
| Ceramic Shell | 280.2 | 12 × 12 | 4 × 2 |
| Insulating Blanket | 109.8 | 36 × 36 | 12 × 12 |
| Sand Flask | 4.7 | 36 × 6 | 36 × 6 |
3. Simulation Analysis and Process Optimization
3.1 Analysis of the Original Casting Process
The simulation of the original process, with a preheat temperature of 830°C, provided critical insights. During the brief filling stage (~10-13 seconds), the temperature of the molten metal remained relatively uniform, ranging from 1470°C to 1500°C, with the most significant cooling occurring at narrow sections and sharp corners of the gating system.
The solidification analysis revealed the root cause of potential defects. Due to the relatively low mold preheat, the thin-walled sections of the brake casting cooled very rapidly. The thermal analysis showed that the solidification sequence was chaotic and poorly controlled. The risers (feeders), intended to supply liquid metal to compensate for shrinkage in thick sections, were ineffective because they solidified prematurely or concurrently with the casting body. This disordered solidification led to isolated hot spots and regions where shrinkage compensation was impossible, resulting in a high predicted propensity for shrinkage cavities and porosity, particularly at junctions between thick and thin sections and in isolated thick regions.
The validity of the simulation was confirmed by comparing its predictions with real-world X-ray inspection results of a cast component. All major defect locations identified in the physical X-ray inspection were accurately predicted by the numerical model. While the model predicted some additional minor defect-prone areas not visible in the specific X-ray sample, the strong correlation validated the simulation’s effectiveness as a tool for guiding aerospace casting process design. This step is crucial in high-integrity aerospace casting to preemptively identify failure points.
3.2 Proposed Process Modifications and Optimized Simulation
Based on the analysis, two key modifications to the original aerospace casting process were proposed to enforce a more favorable, directional solidification pattern:
- Enhanced Insulation at the Base: Replace the insulating blanket at the bottom of the flask with quartz sand. The sand is mounded slightly higher in the center to ensure direct contact with the lower surface of the shell at the brake’s inner and outer flanges. This modification aims to retard cooling from the casting’s bottom, encouraging solidification to progress vertically from the thin walls towards the top risers.
- Increased Mold Preheat Temperature: Raise the initial preheat temperature of the shell and flask assembly from 830°C to 1090°C. This significantly reduces the initial thermal shock, slows down the overall cooling rate, and provides a larger thermal reservoir to promote better feeding.
All other parameters, including pouring temperature and time, were kept constant. The modified process was then simulated using the same coupled-field model.
3.3 Results of the Optimized Process Simulation
The simulation results for the modified process demonstrated a dramatic improvement. The temperature field and solid fraction evolution showed a clear and orderly solidification sequence. The thin-walled sections at the periphery and the central slanted support plate now cooled significantly earlier than the heavy flange sections. Crucially, within the thick sections, solidification progressed directionally from areas farthest from the risers towards the risers themselves, which remained liquid for the longest time. This created a consistent temperature gradient favorable for feeding.
The defect prediction model output for the optimized process was remarkably clean. The vast majority of the casting volume was predicted to be free from shrinkage porosity. Only a very small, isolated region in the middle of the outer thin-walled ring showed a slight susceptibility to micro-shrinkage, a drastic reduction compared to the original process. The risers now functioned effectively as thermal reservoirs, continuously feeding liquid metal to compensate for solidification shrinkage throughout the critical period. This outcome underscores the power of physics-based simulation in aerospace casting optimization, allowing for virtual testing of “what-if” scenarios without costly and time-consuming physical trials.
| Parameter / Outcome | Original Process | Optimized Process |
|---|---|---|
| Mold Preheat Temperature | 830 °C | 1090 °C |
| Base Insulation | Flat Insulating Blanket | Mounded Quartz Sand |
| Solidification Sequence | Disordered, rapid thin-wall cooling | Directional, controlled from thin to thick sections |
| Riser (Feeder) Effectiveness | Poor, early solidification | High, prolonged liquid life |
| Predicted Shrinkage Defects | Extensive in thick sections & junctions | Negligible; minor in one thin-wall area |
4. Conclusion
The numerical investigation employing a coupled temperature-fluid flow model has proven highly effective in diagnosing and solving quality issues in a complex aerospace casting. The simulation of the original process correctly identified the disordered solidification and poor feeding as the root causes of shrinkage defects, a prediction validated by experimental X-ray results. The proposed modifications—significantly increasing the mold preheat temperature and strategically altering the bottom insulation to promote directional solidification—were virtually tested and shown to transform the solidification pattern. The optimized process simulation predicts a casting largely free of shrinkage porosity, demonstrating a successful virtual optimization cycle.
This study highlights the indispensable role of advanced numerical simulation in the development and refinement of high-integrity aerospace casting processes. It moves the industry beyond trial-and-error methods, enabling a scientific, physics-driven approach to design robust processes that ensure the reliability of mission-critical components. The coupled-field modeling framework presented provides a powerful template for addressing similar challenges in the aerospace casting of other intricate superalloy components.