The investment casting process is a critical manufacturing technique widely employed in aerospace, automotive, and medical industries due to its ability to produce components with high dimensional accuracy, excellent surface finish, and complex geometries. In particular, titanium alloys, known for their low density, high strength-to-weight ratio, and superior corrosion resistance, are extensively used in aerospace applications, with over 90% of titanium alloy castings fabricated via the investment casting process. However, the high chemical reactivity of titanium at elevated temperatures poses significant challenges, as it tends to react with ceramic molds, leading to defects such as shell cracking and casting deformation. Therefore, understanding and optimizing the sintering stage of ceramic shell fabrication is paramount to ensuring shell integrity and final casting quality. This study focuses on the numerical simulation of temperature distribution and deformation behavior during the sintering process of ceramic shells used in titanium alloy investment casting. By developing advanced thermo-mechanical models and validating them through experiments, we aim to provide insights that can enhance the investment casting process, reduce defects, and improve yield rates.
The ceramic shell in the investment casting process is typically constructed through multiple steps: pattern assembly, slurry coating, stuccoing, drying, dewaxing, and sintering. The sintering stage, where the shell is heated to high temperatures, induces profound physical and chemical changes, including sintering shrinkage, phase transformations, and stress evolution. These phenomena directly influence the shell’s mechanical properties, dimensional stability, and interaction with molten titanium. Traditional approaches to optimizing the investment casting process have relied heavily on trial-and-error experiments, which are time-consuming, costly, and often limited in scope. Numerical simulation offers a powerful alternative, enabling detailed analysis of temperature fields, stress distributions, and deformation patterns under various process conditions. In this work, we employ an improved Monte Carlo method to model radiative heat transfer and establish a coupled thermo-mechanical-damage constitutive model that accounts for thermal damage effects. Through secondary development based on ABAQUS, we create specialized simulation software to investigate the sintering process. This article presents our methodology, experimental validation, and comprehensive simulation results for ceramic shells, emphasizing the role of temperature gradients and material behavior in the investment casting process.
To accurately simulate the investment casting process, we first developed a radiative heat transfer model using an enhanced Monte Carlo method. This approach is particularly suitable for modeling heat exchange in furnace environments where radiation dominates. The model considers the emission, absorption, and scattering of thermal radiation between surfaces, such as the heating elements and the ceramic shell. The radiative heat flux between surfaces can be expressed as:
$$ q_{rad} = \sigma \epsilon (T^4 – T_{\infty}^4) $$
where \( \sigma \) is the Stefan-Boltzmann constant, \( \epsilon \) is the emissivity, \( T \) is the surface temperature, and \( T_{\infty} \) is the ambient temperature. In our simulation, we set the emissivity of the heating surfaces and shell surfaces to 0.9, and other furnace surfaces to 0.75, with an initial temperature of 22°C. The Monte Carlo method tracks a large number of energy bundles (set to 1000 rays) to compute view factors and radiative exchange, ensuring accurate temperature predictions in complex geometries.
For the mechanical response, we formulated a thermo-mechanical-damage constitutive model that incorporates the effects of thermal damage on the ceramic material. The total strain \( \epsilon_{total} \) is decomposed into elastic, thermal, and inelastic components:
$$ \epsilon_{total} = \epsilon_{el} + \epsilon_{th} + \epsilon_{in} $$
The elastic strain follows Hooke’s law, \( \sigma = E \epsilon_{el} \), where \( E \) is the temperature-dependent elastic modulus. The thermal strain is given by \( \epsilon_{th} = \alpha (T – T_0) \), with \( \alpha \) as the coefficient of thermal expansion. The inelastic strain accounts for creep and sintering shrinkage, which are significant during the investment casting process. To model damage, we introduce a damage variable \( D \) ranging from 0 (no damage) to 1 (complete failure), which evolves based on thermal stress and temperature history. The effective stress is then \( \sigma_{eff} = \sigma / (1 – D) \). The evolution law for damage is expressed as:
$$ \dot{D} = A \left( \frac{\sigma_{eff}}{\sigma_0} \right)^n \exp\left(-\frac{Q}{RT}\right) $$
where \( A \) and \( n \) are material constants, \( \sigma_0 \) is a reference stress, \( Q \) is activation energy, \( R \) is the gas constant, and \( T \) is absolute temperature. This model captures the degradation of mechanical properties due to microcracking and phase changes during sintering, which is crucial for predicting shell deformation and potential failure in the investment casting process.
Key thermo-physical parameters of the ceramic shell, essential for the simulation, were experimentally measured. The shell consisted of an inner layer of Y₂O₃ and outer layers of Al₂O₃ and SiO₂, fabricated via repeated slurry coating and stuccoing. We tested thermal diffusivity (\( \alpha \)), specific heat (\( c \)), and elastic modulus (\( E \)) over a temperature range of 25°C to 1200°C using specialized equipment. The thermal conductivity (\( \lambda \)) was derived from the relation \( \lambda = \alpha \rho c \), where \( \rho \) is density. The results are summarized in Table 1, showing the temperature dependence of these properties, which significantly influence heat transfer and stress development during the investment casting process.
| Temperature (°C) | Thermal Diffusivity, \( \alpha \) (m²/s) | Specific Heat, \( c \) (J/g·K) | Thermal Conductivity, \( \lambda \) (W/m·K) | Elastic Modulus, \( E \) (GPa) |
|---|---|---|---|---|
| 25 | 0.8 × 10⁻⁶ | 0.8 | 2.8 | 250 |
| 200 | 1.2 × 10⁻⁶ | 1.0 | 3.2 | 200 |
| 400 | 1.6 × 10⁻⁶ | 1.2 | 3.0 | 150 |
| 600 | 2.0 × 10⁻⁶ | 1.4 | 2.8 | 100 |
| 800 | 2.4 × 10⁻⁶ | 1.6 | 2.4 | 50 |
| 1000 | 2.8 × 10⁻⁶ | 1.8 | 2.0 | 25 |
| 1200 | 3.2 × 10⁻⁶ | 2.0 | 1.6 | 10 |
To validate our models, we conducted sintering experiments on a flat-plate ceramic specimen with dimensions 180 mm × 109 mm × 25 mm. The sintering schedule involved three stages: heating to 500°C in 1 hour, holding for 2 hours; heating to 700°C in 1 hour, holding for 2 hours; and heating to 1050°C in 1 hour, holding for 4 hours. Temperature was monitored at four points on the specimen, and deformation was measured after sintering. The simulation results for temperature evolution showed good agreement with experimental data, as depicted in Figure 1 (simulated temperature fields at different times). The maximum deformation predicted was 0.77 mm, compared to an experimental measurement of 0.65 mm, confirming the model’s accuracy for the investment casting process.
Building on this validation, we performed numerical simulations for an annular-stepped ceramic shell, a geometry representative of complex investment casting components. The shell, with a total height of 410 mm and an average thickness of 10 mm, was placed in a furnace model measuring 3400 mm × 2200 mm × 2200 mm. We defined three sintering schemes to study the effect of final holding temperature, as outlined in Table 2. Each scheme followed the same heating and holding times for the first two stages, but varied the third-stage holding temperature (950°C, 1000°C, or 1050°C) for 2 hours. Temperature monitoring points were set at the sprue bottom, mid-section of the annular part, and the pouring cup to capture spatial variations during the investment casting process.
| Scheme | Stage 1: Holding at 500°C | Stage 2: Holding at 700°C | Stage 3: Holding Temperature |
|---|---|---|---|
| 1 | 1 h heat, 2 h hold | 1 h heat, 2 h hold | 950°C for 2 h |
| 2 | 1 h heat, 2 h hold | 1 h heat, 2 h hold | 1000°C for 2 h |
| 3 | 1 h heat, 2 h hold | 1 h heat, 2 h hold | 1050°C for 2 h |
The simulation results revealed pronounced non-uniform temperature distributions during the sintering process, a key factor in the investment casting process. In the initial heating phase, the shell edges and protruding structures like the pouring cup heated faster due to direct radiation from heating elements, while interior regions lagged due to the low thermal conductivity of ceramics. This temperature gradient, quantified as ΔT, increased with higher sintering temperatures. For Scheme 1, ΔT was about 11°C; for Scheme 2, 30°C; and for Scheme 3, over 30°C. The temperature evolution at monitoring points followed the set schedules, but with delays in interior regions, highlighting the importance of heat transfer dynamics in the investment casting process.
The temperature non-uniformity directly influenced thermal stress and deformation. The mechanical model, incorporating the damage variable, allowed us to compute stress distributions and displacements. The von Mises stress \( \sigma_{vM} \) was used to assess stress intensity:
$$ \sigma_{vM} = \sqrt{\frac{1}{2}\left[(\sigma_1 – \sigma_2)^2 + (\sigma_2 – \sigma_3)^2 + (\sigma_3 – \sigma_1)^2\right]} $$
where \( \sigma_1, \sigma_2, \sigma_3 \) are principal stresses. High stress concentrations were observed at structural protrusions, such as the connection between the pouring cup and sprue, where temperature gradients were steepest. The deformation, represented by displacement magnitude \( u \), showed that these areas experienced the largest distortions. For Scheme 1, maximum deformation was around 0.7 mm; for Scheme 2, 1.5 mm; and for Scheme 3, 2.65 mm. This trend correlates with the reduction in viscosity of the glassy phase formed at higher temperatures, which exacerbates stress relaxation and creep deformation. The relationship between glass phase viscosity \( \eta \) and temperature can be described by the Arrhenius equation:
$$ \eta = \eta_0 \exp\left(\frac{E_\eta}{RT}\right) $$
where \( \eta_0 \) is a pre-exponential factor and \( E_\eta \) is activation energy for viscous flow. As temperature rises in the investment casting process, \( \eta \) decreases, facilitating grain boundary sliding and enhancing inelastic strain, thereby contributing to deformation.
To further analyze the impact of sintering parameters, we examined the phase transformations occurring in the ceramic shell. During the investment casting process, the shell undergoes transitions such as the formation of mullite and glass phase from Al₂O₃ and SiO₂. The volume change associated with these transformations induces additional stresses. The fraction of phase transformation \( f \) can be modeled using the Johnson-Mehl-Avrami-Kolmogorov equation:
$$ f = 1 – \exp(-kt^n) $$
where \( k \) is a rate constant dependent on temperature, and \( n \) is the Avrami exponent. In our simulation, we considered the effect of phase change on thermal expansion coefficient and elastic modulus, which were updated based on temperature history. This refinement improved the accuracy of stress predictions, particularly during holding stages where phase evolution is pronounced.
The investment casting process also involves complex interactions between the shell and the furnace environment. Our radiation model accounted for multiple reflections and shadowing effects, which are critical in large furnaces. The net radiative heat transfer rate \( Q_{net} \) between surfaces i and j is given by:
$$ Q_{net} = \sum_j A_i F_{ij} \sigma (\epsilon_i T_i^4 – \epsilon_j T_j^4) $$
where \( A_i \) is area, \( F_{ij} \) is view factor, and \( \epsilon_i, \epsilon_j \) are emissivities. The Monte Carlo method computed \( F_{ij} \) accurately for the complex shell geometry, ensuring realistic temperature fields. This approach is essential for simulating the investment casting process, where uniform heating is often challenging due to geometric constraints.
In addition to temperature and stress, we evaluated the damage evolution in the ceramic shell. The damage variable \( D \) increased with accumulated inelastic strain and temperature, particularly above 700°C where glass phase formation weakens the material. The critical damage \( D_{cr} \) for crack initiation was set at 0.8 based on experimental observations. In Scheme 3, localized regions near the pouring cup approached \( D_{cr} \), indicating potential cracking risks. This insight underscores the need for controlled heating rates and holding times in the investment casting process to mitigate damage.
To summarize the simulation findings, Table 3 compares key outcomes for the three sintering schemes. It highlights how higher temperatures in the investment casting process lead to greater temperature gradients, stress, and deformation, emphasizing the trade-offs between sintering quality and process parameters.
| Scheme | Max Temperature Gradient ΔT (°C) | Max von Mises Stress (MPa) | Max Deformation (mm) | Damage Indicator (Max D) |
|---|---|---|---|---|
| 1 (950°C) | 11 | 48 | 0.70 | 0.2 |
| 2 (1000°C) | 30 | 140 | 1.50 | 0.5 |
| 3 (1050°C) | 30+ | 325 | 2.65 | 0.75 |
The investment casting process benefits greatly from such numerical analyses, as they enable optimization without extensive physical trials. For instance, based on our results, we recommend using intermediate sintering temperatures (e.g., 1000°C) with prolonged holding times to reduce temperature gradients while achieving sufficient densification. Additionally, modifying shell geometry to minimize sharp protrusions can alleviate stress concentrations. These adjustments can enhance shell performance and casting quality in the investment casting process.
In conclusion, our study demonstrates the effectiveness of numerical simulation in analyzing the sintering stage of ceramic shells for titanium alloy investment casting. By integrating radiation heat transfer, thermo-mechanical-damage constitutive models, and experimental data, we accurately predicted temperature distributions and deformation behaviors. The investment casting process is highly sensitive to sintering parameters, with higher temperatures leading to increased deformation due to temperature gradients and glass phase viscosity reduction. This work provides a foundation for optimizing sintering schedules and shell designs, ultimately improving the reliability and efficiency of the investment casting process. Future research could extend these models to include chemical interactions with molten titanium, further advancing the investment casting process for high-performance applications.