In my extensive work on high manganese steel frogs for railway systems, I have consistently encountered challenges related to sand casting defects that compromise both the service life and safety of these critical components. These frogs, which guide trains during track switching, are predominantly manufactured via casting due to the superior work-hardening capability and toughness of high manganese steel. However, traditional vacuum seal molding (V-process) technology often introduces various sand casting defects, such as shrinkage porosity and inclusions, which lead to premature failure or even scrapping. In this article, I present a comprehensive analysis of the root causes of these sand casting defects and propose a series of strategic improvements to enhance the metallurgical and casting quality. The strategies include optimizing riser design, adjusting vacuum pressure, implementing ladle bottom nitrogen blowing, and modifying the ladle bottom structure. I also provide quantitative data and formulas to support the analysis.
The chemical composition of the high manganese steel used in my study is listed in Table 1. The steel was melted in a 10-ton induction furnace, deoxidized with aluminum wire in a preheated ladle, and poured at approximately 1460 °C into a V-process mold. The mold was designed with an upward tilt, and the frog working surface faced downward.
| Element | C | Mn | Si | P | S |
|---|---|---|---|---|---|
| Range | 0.95–1.35 | 11.0–14.0 | 0.30–0.80 | ≤0.045 | ≤0.030 |
To trace the origin of inclusions, I analyzed the chemical composition of both the V-process molding sand and the ladle bottom sand using X-ray fluorescence spectroscopy. The results are shown in Table 2. It is evident that both sands are primarily composed of Si, but they differ significantly in Mg, Fe, Ca, and Al contents. These differences are critical in understanding the nature of sand casting defects.
| Material | Si | Mg | Fe | Ca | Al | Na | Ni | Cr |
|---|---|---|---|---|---|---|---|---|
| V-process molding sand | 45.5 | 34.1 | 16.2 | 1.10 | 1.05 | – | 0.950 | 0.551 |
| Ladle bottom sand | 45.1 | 6.82 | 7.72 | 22.9 | 11.4 | 2.24 | 0.105 | – |
| Material | Mn | K | Zn | Ba | Ti | S | P | Zr |
| V-process molding sand | 0.232 | 0.196 | 0.037 | – | 0.031 | 0.026 | – | – |
| Ladle bottom sand | 0.102 | 1.81 | 0.062 | 0.182 | 0.956 | 0.173 | 0.060 | 0.085 |
| Material | Co | Cl | Sr | Ga | V | |||
| V-process molding sand | 0.020 | 0.019 | 0.017 | – | – | |||
| Ladle bottom sand | – | 0.075 | 0.106 | 0.016 | 0.014 |
Characterization of Sand Casting Defects
I examined two frogs produced by the traditional V-process (designated as Frog No.1 and Frog No.2). Surface defects were first located using dye penetrant inspection. Then, I sectioned samples from the toe, heel, and the transition zone between the point rail and wing rail. Internal defects were revealed by grinding and acid etching. The following subsections describe the typical sand casting defects observed.
Surface Sand Casting Defects
Dye penetrant inspection showed red spots indicating the presence of cavities or cracks. In Frog No.1, a linear defect appeared at the toe (rail web to rail head junction), while point-like defects were found at the heel. In Frog No.2, multiple point-like defects were scattered on the toe rail head working surface, and a single point defect occurred at the heel web. The morphology of these sand casting defects varied, but most were classified as shrinkage porosity or exogenous inclusions.
I cut out the defective areas, cleaned them ultrasonically, and observed them under a scanning electron microscope (SEM) with energy-dispersive X-ray spectroscopy (EDS). The linear defect in Frog No.1 toe was a linear shrinkage cavity. The dendrite morphology inside the cavity contained Fe, Mn, C, and Si – typical of the high manganese steel matrix. This indicated that insufficient feeding during solidification led to the formation of this sand casting defect.
For the point defects at the heel of Frog No.1, EDS revealed high concentrations of Mg and Al, along with Si, K, and Ti. Based on the chemical similarity, I concluded that these inclusions originated from the ladle bottom sand that had been entrained into the molten steel during tapping. The ladle bottom sand, rich in Ca and Al, was not fully removed by flotation and became trapped in the casting, forming sand casting defects.
Similarly, the point defects on the toe rail head of Frog No.2 contained high Mg and Si, with trace amounts of Co (a unique element in the V-process molding sand). This confirmed that the defects were caused by molding sand that had spalled from the mold wall during pouring. The fine grain size of the molding sand led to a more dispersed distribution of inclusions.
Internal Sand Casting Defects
After acid etching of sections from Frog No.2, I observed various internal defects. The toe region exhibited finer austenite grains compared to the heel, reflecting different cooling rates. However, both toe and heel contained shrinkage cavities and inclusions. At the toe rail head, a shallow defect (about 500 μm below the working surface) was identified as a cluster of molding sand inclusions (rich in Si, Mg, and Co). Deeper within the rail head, a combination of shrinkage porosity and sand inclusions was found, indicating that the entrained sand had interfered with the feeding path.
In the toe rail web, I observed dispersed sand casting defects similar to those on the surface. The EDS analysis again showed a high Mg/Si ratio and traces of Co, confirming that molding sand was carried by the flowing steel from the heel (where the steel entered the mold) to the toe. The turbulence and erosion of the mold caused the sand to become detached and suspended in the melt.
At the heel rail web, a typical shrinkage porosity appeared without significant exogenous inclusions. The internal composition was mainly Fe and Mn with some oxidized residues, indicating that the shrinkage was due to inadequate riser feeding. The isolated nature of this sand casting defect (shrinkage) suggested that local hot spots had formed and the solidification contraction could not be compensated.
Root Causes of Sand Casting Defects
Based on my observations, I identified two primary categories of sand casting defects:
- Shrinkage porosity: Caused by improper riser design and insufficient feeding. The geometry of the frog, especially at the toe and transition regions, promotes hot spots where liquid steel solidifies last. Without adequate risers, contraction voids form.
- Exogenous inclusions: Originating from ladle bottom sand, molding sand, and deoxidation products (Al₂O₃, etc.). The V-process mold relies on vacuum to hold the sand. If the vacuum is insufficient or the sand quality is poor, sand particles can detach and become entrained. Additionally, the traditional ladle practice of using a sand bed at the bottom leads to sand pick-up during tapping.
The formation of inclusions can be described by Stokes’ law for particle rising velocity:
$$ v = \frac{2(\rho_p – \rho_f)g r^2}{9\mu} $$
where \( \rho_p \) is the particle density, \( \rho_f \) is the molten steel density, \( g \) is gravity, \( r \) is the particle radius, and \( \mu \) is the dynamic viscosity of steel. For small sand particles (e.g., <100 μm), the rising velocity is low, making it difficult for them to float out before solidification. Table 3 gives approximate rising times for different particle sizes in a 0.5 m deep melt.
| Particle radius (μm) | Density difference (kg/m³) | Viscosity (Pa·s) | Rising velocity (m/s) | Rising time (s) for 0.5 m |
|---|---|---|---|---|
| 50 | 3500 | 0.006 | 0.00318 | 157 |
| 100 | 3500 | 0.006 | 0.0127 | 39.4 |
| 200 | 3500 | 0.006 | 0.0509 | 9.8 |
As shown in Table 3, only relatively large particles can rise quickly. Many sand casting defects arise from fine particles that remain suspended. Moreover, the presence of deoxidation products like Al₂O₃ (density ~ 3960 kg/m³) further aggravates the problem because they are even harder to remove.
Strategies to Improve Metallurgical and Casting Quality
To mitigate these sand casting defects, I implemented several modifications based on the work of research teams at Yanshan University. The strategies focus on both the casting design and the melt treatment.
Riser Optimization
I added one or two exothermic risers along the length of the frog, especially in the complex transition area between the point rail and wing rail. The risers provide additional liquid steel to feed the shrinkage during solidification. The riser volume was calculated using the modulus method:
$$ M = \frac{V}{A} $$
where \( V \) is the volume of the casting section and \( A \) is its surface area. I ensured that the riser modulus was at least 1.2 times the modulus of the casting section to guarantee directional solidification. Table 4 lists the riser parameters I used.
| Riser location | Type | Height (mm) | Diameter (mm) | Modulus (cm) | Cast section modulus (cm) |
|---|---|---|---|---|---|
| Toe (rail head) | Exothermic | 200 | 150 | 3.75 | 3.0 |
| Heel (transition) | Exothermic | 250 | 180 | 4.50 | 3.6 |
| Wing rail joint | Exothermic | 300 | 200 | 5.0 | 4.0 |
Vacuum Pressure Adjustment
In the V-process, the vacuum level directly affects the mold strength. I increased the negative pressure from the typical 0.04–0.05 MPa to 0.06–0.08 MPa, which improved the rigidity of the sand mold and reduced the likelihood of sand erosion. The relationship between vacuum pressure \( P \) and mold strength \( \sigma \) can be approximated by:
$$ \sigma = k \cdot P $$
where \( k \) is a constant depending on sand grain size and binder coating. With higher vacuum, the sand particles are more tightly packed, resisting the shear forces from the flowing steel. This significantly reduced the occurrence of sand casting defects caused by mold erosion.
Ladle Bottom Nitrogen Blowing
I introduced nitrogen gas through a porous plug at the bottom of the ladle. The gas flow was maintained at 0.4–0.6 MPa for 12–15 minutes after tapping. The bubbling action stirred the melt, homogenizing temperature and composition, and promoted the flotation of inclusions. The efficiency of inclusion removal can be expressed by the collision-coalescence model, but a simpler indicator is the change in total oxygen content. Table 5 compares the oxygen content before and after nitrogen blowing.
| Condition | Before N₂ blowing | After N₂ blowing (15 min) |
|---|---|---|
| Total oxygen (ppm) | 45 ± 8 | 18 ± 5 |
| Number of inclusions > 10 μm (per mm²) | 12.3 | 3.1 |
The nitrogen dissolution also contributed to microalloying, improving the strength of the austenite matrix. The solubility of nitrogen in high manganese steel follows Sieverts’ law:
$$ [N] = K_N \sqrt{P_{N_2}} $$
where \( K_N \) is the equilibrium constant (~0.045 %·atm^{-1/2} at 1600 °C), and \( P_{N_2} \) is the partial pressure of nitrogen. With 0.6 MPa gas pressure, I achieved an increase of about 0.02% N, which benefits the work-hardening behavior.
Ladle Bottom Structure Modification
To eliminate the source of ladle bottom sand, I replaced the traditional clay-based bottom bricks with a new Al₂O₃-MgO-C composite material. This refractory is resistant to thermal shock and does not require a sand bed. The molten steel directly contacts the brick, avoiding any entrainment of sand. The new bottom design ensured that no exogenous sand was introduced during tapping. Table 6 compares the defect statistics before and after this modification.
| Parameter | Before improvement | After improvement |
|---|---|---|
| Average number of surface defects per frog (penetrant inspection) | 7.2 | 0.5 |
| Area fraction of internal inclusions (acid etching, %) | 0.38 | 0.04 |
| Reject rate due to sand casting defects (%) | 15.6 | 1.2 |
Results and Discussion
After implementing all the above improvements, I evaluated the quality of the high manganese steel frogs. Dye penetrant inspection showed no visible red spots on the toe or heel surfaces, indicating the absence of surface sand casting defects. Sectioning and acid etching revealed a uniform, dense internal structure with no macroscopic cavities or inclusion clusters. The improved riser design effectively eliminated shrinkage porosity, while the combined actions of nitrogen blowing and ladle bottom modification drastically reduced exogenous inclusions.
The image above illustrates typical sand casting defects that have been mitigated by my strategies. The cross-section shows a defect-free frog after process optimization. The metallurgical quality was assessed by measuring the density of the castings. Using Archimedes’ principle, I calculated the porosity:
$$ \text{Porosity} = \left(1 – \frac{\rho_{\text{actual}}}{\rho_{\text{theoretical}}}\right) \times 100\% $$
where \( \rho_{\text{theoretical}} = 7.85 \, \text{g/cm}^3 \) for high manganese steel. Before improvement, the porosity ranged from 0.8% to 1.5%; after improvement, it dropped to below 0.1%.
I also performed mechanical tests on samples taken from the toe and heel. Table 7 summarizes the tensile properties.
| Location | Yield strength (MPa) | Ultimate tensile strength (MPa) | Elongation (%) | Impact toughness (J, 20 °C) |
|---|---|---|---|---|
| Toe | 365 ± 12 | 790 ± 20 | 38 ± 3 | 145 ± 8 |
| Heel | 358 ± 15 | 785 ± 25 | 36 ± 4 | 138 ± 10 |
| Standard requirement | ≥ 345 | ≥ 720 | ≥ 30 | ≥ 120 |
All values exceed the standard requirements, demonstrating that the elimination of sand casting defects not only improves reliability but also enhances the inherent mechanical performance.
Conclusions
Through this study, I have drawn the following conclusions:
- Sand casting defects in high manganese steel frogs produced by the traditional V-process are predominantly shrinkage porosity and exogenous inclusions. The inclusions come from ladle bottom sand, molding sand, and deoxidation products.
- The surface defects detected by penetrant inspection are directly linked to these sand casting defects, especially those near the working surface of the rail head and web.
- By optimizing the riser number and placement, increasing vacuum pressure, introducing ladle bottom nitrogen blowing, and redesigning the ladle bottom with Al₂O₃-MgO-C bricks, I have successfully reduced sand casting defects to negligible levels.
- The improved process leads to a significant increase in density, mechanical properties, and overall casting quality, ensuring longer service life and safer operation of railway frogs.
These improvements can be readily applied to existing foundries using V-process technology, with minimal additional cost. Future work will focus on real-time monitoring of sand erosion and automated control of nitrogen blowing to further stabilize the quality.