This paper presents a systematic study on the casting performance of a large steel axle housing used in heavy engineering trucks. The work integrates numerical simulation, process optimization, and experimental validation to address severe casting defects such as shrinkage porosity, shrinkage cavity, and hot tearing. By employing the commercial software MAGMA, I simulated the mold filling, solidification, and thermal stress fields of the original gating system and proposed an improved bottom gating system combined with cracking strips and cold irons. The simulation results were validated through production trials, showing a significant reduction in casting defect rates. The methodology and findings provide practical guidance for the design and manufacturing of large steel castings.
1. Introduction
Large engineering truck axle housings are critical structural components that support the main reducer, half shafts, and differential. Due to their harsh service conditions and high load requirements, the quality of these castings is of utmost importance. During the initial trial production, the axle housing exhibited severe casting defects, particularly in the abdominal region where shrinkage porosity and shrinkage cavity occurred at a rate exceeding 80%, while hot tearing defects appeared at a rate close to 100%. These casting defects not only deteriorated product quality but also extended production cycles and increased manufacturing costs.
To solve these problems, I adopted a combined approach of numerical simulation and experimental verification. The casting simulation software MAGMA was used to analyze the flow field, temperature field, and stress field during the filling and solidification processes. Based on the simulation results, I proposed an optimized casting process that replaced the original middle gating system with a bottom gating system, and additionally employed cracking strips and cold irons to mitigate hot tearing. The effectiveness of these modifications was confirmed through production trials.
The main objectives of this research are:
- To analyze the root causes of shrinkage porosity, shrinkage cavity, and hot tearing in the axle housing.
- To establish accurate numerical models for flow, temperature, and stress fields.
- To propose and validate process improvements that reduce casting defect rates.
- To provide a reliable engineering methodology for similar large steel castings.
2. Theoretical Basis of Casting Process Simulation
2.1 Governing Equations for Mold Filling
The mold filling process of liquid alloy is governed by the conservation laws of mass, momentum, and energy. For an incompressible Newtonian fluid, the continuity equation is expressed as:
$$ \nabla \cdot \mathbf{v} = 0 $$
where \(\mathbf{v}\) is the velocity vector. The momentum equation (Navier-Stokes) in the \(x\), \(y\), and \(z\) directions can be written as:
$$ \rho \frac{\partial \mathbf{v}}{\partial t} + \rho (\mathbf{v} \cdot \nabla) \mathbf{v} = -\nabla p + \mu \nabla^2 \mathbf{v} + \rho \mathbf{g} $$
where \(\rho\) is the fluid density, \(p\) is the pressure, \(\mu\) is the dynamic viscosity, and \(\mathbf{g}\) is the gravitational acceleration. The energy equation, which accounts for heat transfer and phase change, is given by:
$$ \rho c_p \frac{\partial T}{\partial t} + \rho c_p \mathbf{v} \cdot \nabla T = \nabla \cdot (k \nabla T) + Q $$
where \(c_p\) is the specific heat at constant pressure, \(T\) is the temperature, \(k\) is the thermal conductivity, and \(Q\) represents the internal heat source including latent heat release during solidification.
2.2 Treatment of Latent Heat
During solidification, the release of latent heat must be accurately modeled. I employed the enthalpy method, which defines the enthalpy \(H\) as:
$$ H(T) = \int_{T_0}^{T} c_p(\tau) d\tau + L (1 – f_s(T)) $$
where \(L\) is the latent heat of fusion and \(f_s\) is the solid fraction. The energy equation can then be reformulated in terms of enthalpy:
$$ \rho \frac{\partial H}{\partial t} + \rho \mathbf{v} \cdot \nabla H = \nabla \cdot (k \nabla T) $$
This approach allows a unified treatment of both pure metals and alloys with a mushy zone.
2.3 Criteria for Predicting Shrinkage Defects
Shrinkage porosity and shrinkage cavity are common casting defects caused by insufficient feeding during solidification. I used the Niyama criterion, which evaluates the local thermal gradient \(G\) and cooling rate \(R\):
$$ \text{Niyama} = \frac{G}{\sqrt{R}} $$
When the Niyama value falls below a critical threshold, the risk of shrinkage porosity increases. Additionally, the temperature gradient method and the critical solid fraction method were adopted to assess the feeding efficiency.
2.4 Hot Tearing Prediction
Hot tearing occurs in the semi-solid state when the accumulated thermal strain exceeds the material’s ductility. The hot tearing susceptibility can be evaluated using the strain-based criterion:
$$ \text{Hot Tearing Index} = \frac{\varepsilon_{loc}}{\varepsilon_{crit}} $$
where \(\varepsilon_{loc}\) is the local strain in the mushy zone and \(\varepsilon_{crit}\) is the critical strain at fracture. A value greater than unity indicates a potential hot tear. In MAGMA, the hot tearing tendency is visualized using the maximum principal strain rate and the temperature distribution in the semi-solid range.
3. Numerical Simulation of the Original Casting Process
3.1 Description of the Axle Housing
The axle housing is a large steel casting with overall dimensions of approximately 2000 mm × 460 mm × 450 mm, weighing about 310 kg. The material is C-grade steel with the following mechanical properties: ultimate tensile strength \( \sigma_b \geq 450\) MPa, yield strength \( \sigma_s \geq 240\) MPa, elongation \( \geq 13\%\), reduction of area \( \geq 18\%\), and Brinell hardness HB 130–170. The cross-section shows a complex internal structure with varying wall thicknesses: the main body wall is 17 mm thick, while the maximum wall thickness reaches 80 mm. The casting is produced using furan resin self-hardening sand for both mold and core due to its high strength and good collapsibility.
3.2 Original Gating System
The original casting process employed a middle gating system, as shown in the process drawing. The cross-sectional area ratio of sprue: runner: ingate was 1:1.45:1.35. Two ingates were placed on the parting line far apart from each other. Multiple insulating risers were symmetrically arranged on the upper part of the casting to provide feeding during solidification. The pouring temperature was 1560±10 °C, the pouring time was 30 seconds, and the mold was cooled in air. The casting defects observed in practice included extensive shrinkage porosity and cavity in the abdominal region and at the bottom of the two flanges, as well as hot tears at the bowl edge, neck, bottom platform, and inside the flanges.
3.3 Simulation Setup
I built a three-dimensional solid model of the casting process using SolidWorks, including the casting, gating system, risers, and sand mold. The model was exported in STL format and imported into MAGMA for meshing and simulation. The material of the casting was defined as GS20Mn5, whose composition is close to the C-grade steel used in practice. Table 1 lists the thermophysical properties used for the simulation.
Table 1: Thermophysical properties of GS20Mn5
| Temperature (°C) | Specific heat (kJ/kg·K) | Thermal conductivity (W/m·K) |
|---|---|---|
| 0 | 0.47 | 51.8 |
| 200 | 0.52 | 48.6 |
| 400 | 0.59 | 42.6 |
| 600 | 0.75 | 35.6 |
| 800 | 0.95 | 25.9 |
| 1000 | 0.64 | 27.2 |
| 1200 | 0.66 | 29.7 |
| 1495 | 0.71 | 32.28 |
| 1525 | 0.73 | 26.89 |
Table 2 shows the high-temperature mechanical properties employed for the stress analysis.
Table 2: Mechanical properties of GS20Mn5 at elevated temperatures
| Temperature (°C) | Elastic modulus (GPa) | Plastic hardening modulus (GPa) | Yield strength (MPa) | Poisson’s ratio | Thermal expansion (×10⁻⁵/°C) |
|---|---|---|---|---|---|
| 50 | 201.7 | 20.17 | 415 | 0.278 | 1.4 |
| 400 | 190.5 | 19.05 | 311 | 0.310 | 1.5 |
| 800 | 175.0 | 17.5 | 283 | 0.344 | 1.7 |
| 1464 | 80.0 | 8.0 | 46 | 0.390 | 2.2 |
| 1475 | 4.03 | 0.403 | 6.9 | — | — |
| 1519 | 0.403 | 0.00403 | 2.07 | — | — |
Interfacial heat transfer coefficients were selected from the MAGMA material database for analogous material pairs (Steel–Sand, Steel–Mold, etc.). The pouring parameters are summarized in Table 3.
Table 3: Simulation parameters
| Parameter | Value |
|---|---|
| Pouring temperature | 1560 °C |
| Initial mold temperature | 20 °C |
| Liquidus temperature | 1508 °C |
| Solidus temperature | 1435 °C |
| Pouring time | 30 s |
| Riser efficiency | 25% |
| Cooling condition | Air cooling |
3.4 Simulation Results and Analysis
3.4.1 Flow Field
The flow field simulation of the original middle gating system revealed severe turbulence and splashing during filling. At the initial stage, the liquid metal stream directly impacted the mold wall, causing splashing and potential oxidation. When the two filling fronts met at the central abdominal region, a strong opposing flow occurred, leading to vortex formation and air entrapment. This unstable filling behavior contributes to casting defects such as gas porosity, inclusions, and shrinkage porosity. Figure 1 shows the filling states at different fractions.
3.4.2 Temperature Field
The solidification sequence was analyzed using temperature field snapshots at various times. At 120 s after filling, most of the casting was above the solidus, except regions near internal chillers. At 379 s, isolated liquid pools appeared in the abdominal area, indicating potential shrinkage defects. At 1349 s, the temperature distribution was non-uniform with steep gradients, particularly at the bottom platform and neck, which can lead to hot tearing. The two flange regions remained hot for a long time, delaying solidification and causing insufficient feeding at the bottom.
3.4.3 Prediction of Shrinkage Defects
Using the Niyama criterion and shrinkage porosity module, I obtained the distribution of shrinkage porosity and cavity, as shown in Figures 2 and 3. Large dark regions were predicted in the abdominal area, matching the actual defect locations observed in the trial casting. The main cause is the early closure of feeding channels due to the thin wall and rapid cooling, resulting in isolated liquid pockets that cannot be compensated by the risers.
3.4.4 Hot Tearing Prediction
The thermal stress simulation coupled with the temperature field produced the hot tearing tendency distribution in the semi-solid range. Figures 4–7 compare the predicted hot tearing zones with actual cracks. The bowl edge showed two symmetrical high-risk areas at the inner boss edge, which matched the observed external cracks. The neck region exhibited high maximum principal strain rates at four symmetric locations, again correlating with actual cracks. The bottom platform edge showed a distinct crack source, and the flange interior displayed a hot tearing region associated with shrinkage porosity. The good agreement between simulation and practice confirmed the reliability of the simulation model.
4. Optimization of the Gating System
4.1 Design Methodology
To improve the casting quality, I redesigned the gating system according to the principles of smooth and rapid filling, minimizing oxidation and turbulence. The main change was to replace the middle gating system with a bottom gating system, as shown in Figure 8. The ingates were relocated to the bottom of the casting, directly beneath the risers in the abdominal area. This configuration promotes a more uniform upward filling and reduces the impact on mold walls. The cross-sectional area ratio remained unchanged.
4.2 Simulation of the Optimized Gating System
4.2.1 Flow Field
The bottom gating system produced a calm and steady upward flow. The liquid metal entered from the bottom and advanced horizontally toward both ends, avoiding the opposing flow seen previously. No significant turbulence or vortex was observed. This stable filling minimizes the entrapment of air and oxides, thereby reducing casting defects such as gas porosity and inclusions.
4.2.2 Temperature Field
The solidification sequence was improved. The abdominal region still showed some thermal gradients, but the isolated liquid pools were eliminated. The bottom platform and neck still exhibited relatively dense isotherms, indicating potential hot tearing risk. The flange regions solidified more uniformly, promoting better feeding from the risers.
4.2.3 Shrinkage Defect Prediction
After the gating system change, the predicted shrinkage porosity and cavity in the abdomen were significantly reduced. The area of potential casting defects decreased substantially, as shown in Figures 9 and 10. The bottom gating system allowed a more favorable temperature gradient, enabling directional solidification toward the risers.
4.2.4 Hot Tearing Assessment
The hot tearing tendency was still present at the bowl edge, neck, bottom platform, and flange interior, though the intensity was somewhat reduced. This indicates that simply changing the gating system is insufficient to completely eliminate hot tearing. Therefore, additional measures such as cracking strips and cold irons were necessary.
4.3 Production Validation of the Bottom Gating System
A production trial was conducted using the improved bottom gating system. The results showed that the shrinkage porosity and cavity in the abdominal region were effectively controlled, with the defect rate dropping from approximately 80% to about 15%. The size of the remaining defects was significantly smaller, making welding repair easier. However, hot tearing at the bowl edge and bottom platform remained prominent, confirming the simulation prediction that further optimization was required.
5. Control of Hot Tearing in the Bottom Gating System
5.1 General Prevention Measures for Hot Tearing
Hot tearing is one of the most serious casting defects in steel castings. Common prevention measures include improving mold collapsibility, using chillers and padding, optimizing pouring temperature and speed, employing cracking strips, refining the gating system, and reducing impurity elements such as phosphorus and sulfur. Based on the specific defect analysis, I selected the following measures for the axle housing:
- Adding cracking strips at locations prone to tensile stress concentration.
- Placing an external chill at the flange bottom to accelerate cooling and improve feeding.
- Adjusting pouring temperature and pouring time within safe ranges.
5.2 Influence of Pouring Temperature and Pouring Time
I performed simulations with pouring temperatures of 1550 °C, 1560 °C, and 1570 °C, while keeping other parameters constant. The hot tearing tendency distributions for the bowl edge and bottom platform showed only negligible differences, although 1570 °C appeared slightly better. Similarly, I simulated pouring times of 25 s (fast) and 35 s (medium). The hot tearing tendency was slightly lower for a pouring time of 35 s. Therefore, I adopted a pouring temperature of 1560 °C and a pouring time of 35 s as a compromise to minimize thermal stress while ensuring complete filling.
5.3 Design and Placement of Cracking Strips and Cold Irons
Based on the hot tearing mechanisms at each location, I designed the following countermeasures:
(a) Bowl edge: The cracks at the inner boss edge were caused by the last solidification at the hot spot beneath the riser, combined with abrupt wall thickness changes and constrained contraction. I increased the fillet radius at the boss edge and added several cracking strips perpendicular to the bowl face, as shown in Figure 11(a). The strips solidify earlier than the casting surface, providing reinforcement and reducing stress concentration.
(b) Flange bottom: The internal cracks were associated with shrinkage porosity. The wavy solidus profile closed off the feeding path from the riser. I placed a chill at the flange bottom to accelerate local cooling and create a V-shaped solidus profile that maintains feeding until the end. This cold iron reduced the shrinkage porosity and consequently prevented internal hot tearing.
(c) Neck region: The external cracks were due to restrained contraction of the mould. I added cracking strips around the neck to reinforce the surface and share the tensile stress, as shown in Figure 11(c).
(d) Bottom platform: Similar to the neck region, external cracks arose from stress concentration due to mould restraint. Cracking strips were placed along the platform edge to strengthen the area.
5.4 Simulation Results after Optimization
After adding the cracking strips and cold iron, I repeated the stress simulation. The hot tearing tendency distributions are compared in Figures 12–15.
For the bowl edge, the dark region indicating high hot tearing probability almost disappeared. For the neck region, the maximum principal strain rates were drastically reduced. For the bottom platform, the crack source at the boss edge vanished. For the flange interior, the shrinkage porosity area was notably smaller, indicating a denser microstructure. These simulation results confirmed the effectiveness of the proposed measures.
5.5 Production Validation after Full Optimization
A full production trial was conducted with the optimized process including the bottom gating system, cracking strips, and cold iron. The results are summarized in Table 4.
Table 4: Comparison of casting defect rates before and after optimization
| Defect type | Original process rate | Optimized process rate |
|---|---|---|
| Shrinkage porosity/cavity in abdomen | ~80% | ~15% |
| Hot tears at bowl edge | ~100% | ~0% |
| Hot tears at neck | ~100% | ~20% |
| Hot tears at bottom platform | ~100% | ~20% |
| Hot tears inside flange | ~100% | ~0% |
The results showed that the bowl edge microcracks were completely eliminated, and the flange internal cracks were no longer detected by radiographic inspection and dissection. The neck and bottom platform cracks occurred less frequently and were narrower, facilitating welding repair. The overall quality of the axle housing was significantly improved, with lower scrap rates and higher production efficiency.
6. Discussion
The numerical simulation methodology used in this study proved to be a powerful tool for diagnosing and solving casting defects. The key findings can be summarized as follows:
First, the flow field analysis revealed that the original middle gating system caused turbulent filling and oxide entrapment, which contributed to shrinkage porosity and gas defects. The bottom gating system provided laminar filling and a uniform temperature distribution, reducing these defects.
Second, the hot tearing prediction based on stress-strain coupled analysis was able to locate high-risk zones accurately. Hot tears occur in the semi-solid range when the local strain exceeds the material’s ductility. The use of cracking strips effectively reinforced these areas and reduced the strain concentration.
Third, internal hot tears often accompany shrinkage porosity. By placing a chill to alter the solidification sequence, I achieved better feeding and eliminated the porosity, thereby preventing the associated internal cracks.
The simulation results were consistent with production trials, demonstrating the reliability of the modeling assumptions and the effectiveness of the optimization measures. The combination of bottom gating, cracking strips, and cold iron significantly reduced the casting defect rates from near 100% to around 20% or less.
7. Conclusion
In this research, I performed a comprehensive numerical simulation and experimental investigation of casting defects in a large engineering truck axle housing. The main conclusions are:
- The original middle gating system caused turbulent filling and non-uniform solidification, leading to severe shrinkage porosity and cavity in the abdominal region.
- Replacing the middle gating with a bottom gating system improved the flow pattern and promoted directional solidification, reducing the shrinkage defect rate from 80% to 15%.
- Hot tearing was accurately predicted using the coupled stress-strain-temperature analysis. The main causes were stress concentration at geometric discontinuities and constrained contraction.
- Adding cracking strips at the high-stress locations significantly reduced hot tearing at the bowl edge, neck, and bottom platform. Placing a cold iron at the flange bottom eliminated internal hot tearing by improving feeding and reducing shrinkage porosity.
- After full optimization, the overall casting defect rate dropped to approximately 20% for the remaining minor cracks, which were easily repairable, while the critical defects were completely eliminated.
The methodology presented here provides a practical and effective approach for optimizing casting processes of complex large steel castings, helping to reduce production costs and improve product reliability.
Future Work
Although the current study achieved significant improvements, there are still opportunities for further research. The dimensions of cracking strips and cold irons could be optimized using parametric studies to minimize material usage while maintaining their effectiveness. Additionally, a quantitative relationship between the hot tearing tendency index and the actual crack probability could be established through more extensive experiments. Moreover, the influence of microalloying elements and heat treatment on hot tearing resistance could be investigated for further enhancement of casting quality.