In the field of turbomachinery, impellers play a critical role in converting mechanical energy into fluid kinetic and pressure energy, enabling efficient power output. The complex, streamlined profiles of impeller blades are challenging to achieve through machining but can be directly fabricated using casting methods, particularly the lost wax investment casting process. This technique allows for near-net-shape production of intricate components while avoiding stress concentrations induced by machining, thereby enhancing mechanical properties. However, defects such as shrinkage porosity and gas holes remain significant quality concerns in the production of small to medium-sized impellers. This article details the optimization of the lost wax investment casting process for a duplex stainless steel closed impeller, utilizing numerical simulation to address these issues and improve production efficiency.
The lost wax investment casting process involves creating a wax pattern, coating it with a ceramic shell, melting out the wax, and pouring molten metal into the cavity. For the duplex stainless steel closed impeller studied here, the initial casting design led to shrinkage defects in critical areas like the hub and caliber ring. Through iterative simulations with ProCAST software, modifications to the gating and feeding systems, as well as the pouring method, were implemented to eliminate these defects. The optimized process was validated through actual production, resulting in higher yield and reduced scrap rates.
Introduction to Lost Wax Investment Casting and Impeller Design
Lost wax investment casting is renowned for its ability to produce complex geometries with high dimensional accuracy and surface finish. In this process, a wax pattern is injected into a mold, assembled into a tree, and coated with multiple layers of ceramic slurry to form a shell. After dewaxing, the shell is fired, and molten metal is poured into the cavity. For impellers, this method is ideal due to the intricate blade shapes and the need for internal soundness. The duplex stainless steel material, such as A890 3A, offers excellent corrosion resistance and strength, but its solidification characteristics require careful control to prevent defects.
The closed impeller in this study features a single-suction, six-channel design with significant variations in wall thickness, ranging from 8.5 mm to 175 mm. Key dimensions are summarized in Table 1, highlighting the challenges in achieving uniform solidification. The hub and caliber ring, being thicker sections, are prone to shrinkage porosity due to thermal gradients and inadequate feeding.
| Parameter | Value (mm) |
|---|---|
| Hub Diameter | 110 |
| Flow Channel Width | 110 |
| Caliber Ring Diameter | 495 |
| Impeller Diameter | 611 |
| Blade Thickness Range | 8.5–10 |
The solidification behavior in lost wax investment casting can be modeled using heat transfer equations. The general heat conduction equation is given by:
$$ \frac{\partial T}{\partial t} = \alpha \nabla^2 T $$
where \( T \) is temperature, \( t \) is time, and \( \alpha \) is the thermal diffusivity. For metallic alloys, the solidification process involves latent heat release, which can be incorporated into the energy equation as:
$$ \rho C_p \frac{\partial T}{\partial t} = \nabla \cdot (k \nabla T) + \rho L \frac{\partial f_s}{\partial t} $$
Here, \( \rho \) is density, \( C_p \) is specific heat, \( k \) is thermal conductivity, \( L \) is latent heat, and \( f_s \) is the solid fraction. In the lost wax investment casting process, these parameters are critical for predicting shrinkage defects.
Initial Casting Process Design and Numerical Simulation
The initial lost wax investment casting process was designed with the impeller oriented such that the caliber ring faced upward. A top-pouring system was employed, with a main ingate located at the spherical hub and three side ingates with feeders attached to the caliber ring. These side ingates were intended to provide feeding, venting, and wax removal during dewaxing. The gating system used a cup-shaped feeder to promote directional solidification, a key principle in steel casting to minimize shrinkage.
Numerical simulation parameters were set based on typical lost wax investment casting conditions. The mesh generation involved a unit step size of 10 mm for the casting and 20 mm for the shell, resulting in 75,192 two-dimensional elements and 706,712 three-dimensional elements. Key process parameters are listed in Table 2.
| Parameter | Value |
|---|---|
| Pouring Temperature | 1,620 °C |
| Shell Preheat Temperature | 650 °C |
| Pouring Rate | 7 kg/s |
| Metal-Shell Heat Transfer Coefficient | 500 W/(m²·K) |
| Feeder Top-Air Heat Transfer Coefficient | 100 W/(m²·K) |
| Shell-Air Heat Transfer Coefficient | 45 W/(m²·K) |
The simulation results revealed shrinkage defects in the hub and caliber ring regions, with a critical shrinkage porosity criterion set at 2.3%, based on the material’s contraction behavior. The solid fraction and temperature fields indicated that the three side ingates provided insufficient feeding coverage for the caliber ring hotspots. Additionally, the feeder root solidified prematurely due to higher heat dissipation to the environment, reducing its modulus and effectiveness in feeding the hub. The solidification time \( t_s \) for a section can be estimated using Chvorinov’s rule:
$$ t_s = C \left( \frac{V}{A} \right)^2 $$
where \( C \) is a constant dependent on the material and mold properties, \( V \) is volume, and \( A \) is surface area. For the hub, the low \( V/A \) ratio contributed to slower solidification, exacerbating shrinkage.
Optimization of Gating and Feeding Systems
To address the deficiencies in the initial lost wax investment casting process, the gating and feeding systems were optimized. First, the number of side ingates was increased from three to six, ensuring that each hotspot at the caliber ring-blade junction received adequate feeding. This modification also helped delay solidification at the feeder roots by distributing heat more evenly. Second, the feeder size was enlarged to a bottom diameter of 170 mm, with a hexagonal top design for easier wax tree assembly. The modulus \( M \) of a feeder is given by:
$$ M = \frac{V}{A} $$
where a higher modulus indicates slower solidification, improving feeding capacity. The optimized feeder design aimed to achieve \( M_{\text{feeder}} > M_{\text{casting}} \) at critical sections.
Simulation of the optimized lost wax investment casting scheme showed a significant reduction in shrinkage defects. The caliber ring areas were fully fed by the additional side ingates, as evidenced by the solid fraction plots. However, minor shrinkage persisted in the hub due to unfavorable temperature gradients. The solidification sequence indicated a “V”-type feeding pattern, which posed quality risks by creating isolated hot spots.
The effectiveness of feeding in lost wax investment casting can be quantified using the feeding distance \( L_f \), which for steel castings is approximately:
$$ L_f = 4.5 \sqrt{T} $$
where \( T \) is the section thickness. By increasing the number of ingates, the effective feeding distance covered the entire caliber ring, satisfying \( L_f \geq \text{ring circumference} / \text{number of ingates} \).
Pouring Process Optimization and Numerical Analysis
The initial pouring method involved central introduction of molten metal from the pouring cup, which led to excessive heat accumulation at the hub bottom. This caused delayed solidification and increased shrinkage susceptibility. To mitigate this, the pouring process was altered to a side-pouring approach, where metal enters along the cup wall. This reduced direct impingement on the hub, promoting more uniform cooling and minimizing thermal gradients.
Numerical simulations of the optimized pouring scheme in the lost wax investment casting process demonstrated improved temperature distribution. The side-pouring method reduced the hub bottom temperature by approximately 50–100 °C during initial stages, as shown by the thermal fields. The revised pouring velocity profile can be described by:
$$ v(t) = v_0 e^{-\beta t} $$
where \( v_0 \) is the initial velocity and \( \beta \) is a damping coefficient accounting for flow resistance. Side pouring lowered \( v_0 \) at impact zones, decreasing erosion risk and thermal shock.
The final simulation results confirmed the elimination of shrinkage defects in both the hub and caliber ring. The solid fraction evolution showed progressive solidification from the blades toward the feeders, ensuring continuous feeding paths. The Niyama criterion, often used to predict shrinkage porosity, is expressed as:
$$ G / \sqrt{\dot{T}} \geq K $$
where \( G \) is the temperature gradient, \( \dot{T} \) is the cooling rate, and \( K \) is a material constant. After optimization, the \( G / \sqrt{\dot{T}} \) values exceeded \( K \) in all critical regions, indicating soundness.
Production Validation and Results
The optimized lost wax investment casting process was implemented in actual production, involving steps such as tree assembly, shell building, pouring, and finishing. The final impeller castings were inspected mechanically and through non-destructive testing, revealing no defects in the hub or caliber ring. The yield rate improved by over 20%, and production efficiency increased due to reduced rework. This success underscores the value of numerical simulation in refining lost wax investment casting parameters for complex components like closed impellers.
Key performance metrics from production are summarized in Table 3, highlighting the benefits of optimization in the lost wax investment casting process.
| Metric | Initial Process | Optimized Process |
|---|---|---|
| Defect Rate in Hub (%) | 15 | 0 |
| Defect Rate in Caliber Ring (%) | 20 | 0 |
| Yield Rate (%) | 65 | 85 |
| Production Cycle Time (hours) | 48 | 40 |
Conclusion
In summary, the lost wax investment casting process for a duplex stainless steel closed impeller was successfully optimized through numerical simulation and practical adjustments. By increasing the number of side ingates and enlarging the feeder modulus, shrinkage defects in the caliber ring were eliminated. Furthermore, changing the pouring method to side introduction resolved hub-related issues by improving thermal management. The validated approach demonstrates that lost wax investment casting, when combined with advanced simulation tools, can achieve high-quality impeller production with enhanced efficiency. This methodology provides a reliable reference for similar applications in turbomachinery component manufacturing.
The continuous improvement in lost wax investment casting techniques relies on integrating thermodynamic principles with real-world process controls. Future work could explore dynamic parameter adjustments during pouring or advanced alloy formulations to further reduce defects. Overall, the lost wax investment casting process remains a cornerstone for producing high-integrity cast components in demanding industries.