In our production of ductile iron railway switch components, we encountered significant challenges with large flat surfaces, particularly the upper surface of a casting. These surfaces are prone to various sand casting defect types, including sand inclusions, sand scabs, mold adhesion, and gas porosity. Our experience with a small batch of approximately 300 pieces of a ductile iron base plate (dimensions roughly 400 mm × 300 mm × 80 mm) forced us to develop effective countermeasures. Initially, we employed conventional green sand molding on a Z145 jolt-squeeze molding machine, with the pattern parting along the horizontal plane. The green sand mixture contained high moisture content, resulting in low green strength (only 10%–20% of dry strength). This made the casting highly susceptible to sand casting defect issues. After several trials, we systematically improved the process through material modifications, mold design changes, and pouring techniques. Below I describe our analysis and the comprehensive measures we implemented.
1. Initial Observations and Defect Analysis
Our first castings showed severe surface defects on the top flat area. The bottom surface was acceptable (surface roughness Ra ~ 12.5 μm), but the top exhibited widespread sand inclusions, scabs, mold-metal adhesion, and gas holes. We identified the following root causes:
- Gas evolution and porosity: In green sand molds, when molten ductile iron (high temperature) contacts the moist mold surface, water vaporizes violently and migrates inward, increasing moisture content in the deeper layers and reducing strength. If the mold permeability is inadequate and venting is poor, gas can become trapped, causing invasive or reaction-type gas holes.
- Sand expansion and scabbing: The mold surface rapidly heats and expands, while the underlying cooler sand resists expansion. This creates compressive thermal stress in the surface layer, leading to buckling and scabbing. The problem is more pronounced over large flat areas because of the greater thermal load.
- Other factors: Sand inclusions arose from loose sand falling into the cavity; mold-metal adhesion occurred due to insufficient refractory coating; and poor venting contributed to gas defects.
2. Initial Material and Process Parameters
We used a quartz-feldspar sand with grain size 0.15–0.212 mm (100–70 mesh). Green sand composition is shown in Table 1. Despite adding coal dust and heavy oil to improve anti-scab and anti-adhesion properties, defects persisted.
| Component | Face sand (wt%) | Backing sand (wt%) |
|---|---|---|
| New sand | 20–30 | 0–10 |
| Return sand | 70–80 | 90–100 |
| Bentonite | 5–8 | 1–2 |
| Coal dust | 3–5 | 0 |
| Soda ash | 0.5–1.0 | 0 |
| Heavy oil | 0.2–0.5 | 0 |
| Moisture | 4.5–5.5 | 4.5–5.5 |
| Permeability (AFS) | ≥ 100 | ≥ 40 |
| Green compression strength (MPa) | 0.06–0.08 | 0.06–0.08 |
Molding was done on a Z145 machine with compaction degree ~50%. After closing, we placed a weight and poured manually. The top surface consistently showed defects, while the bottom remained acceptable. This indicated that the thermal load on the upper surface was critical.
3. Theoretical Basis for Defect Formation
To understand the sand casting defect mechanisms, we considered the thermal expansion behavior and gas evolution. The volumetric expansion of a sand layer heated from initial temperature \(T_0\) to final temperature \(T_f\) is:
$$ \frac{\Delta V}{V} = 3 \alpha (T_f – T_0) $$
where \(\alpha\) is the linear thermal expansion coefficient of sand (approximately \(1.5 \times 10^{-5}\,\text{K}^{-1}\) for silica). For a temperature rise of 1200–1300 K, the volumetric expansion can reach 5–6%. The surface layer tries to expand, but the underlying cold sand constrains it, generating compressive stress \(\sigma\):
$$ \sigma = E \alpha (T_f – T_0) $$
where \(E\) is the Young’s modulus of the compacted sand (typically 1–5 MPa). This stress can exceed the sand’s hot tensile strength, leading to layer separation and scabbing.
Simultaneously, the moisture content in the sand leads to steam evolution. The mass of water vapor formed per unit volume of sand is:
$$ m_v = \rho_s \cdot w $$
with \(\rho_s\) the sand density and \(w\) the moisture fraction. This vapor must escape through the permeable sand; otherwise gas pressure builds and causes blowholes. The maximum gas pressure \(P_{\text{max}}\) in a closed pore can be estimated from the ideal gas law:
$$ P_{\text{max}} = \frac{m_v R T}{V_{\text{pore}}} $$
where \(V_{\text{pore}}\) is the pore volume. High pressure can cause the surface layer to lift, creating a scab or a gas hole.
4. Process Improvements
Because the batch size was small (only ~300 pieces), we switched from green sand to a surface-dried sand mold technique. We also redesigned the pouring layout. The key changes are summarized below.
4.1 Mold Material & Preparation
- Sand: We used silica sand of grain size 0.125–0.180 mm (80–120 mesh) to improve permeability.
- Binder: Bentonite (5–7%) plus 3–5% sawdust to reduce thermal expansion and increase hot strength.
- Moisture control: Reduced to 3–4% to minimize gas evolution.
- Surface drying: Before pouring, we dried the mold cavity surface using a kerosene blowtorch to remove residual moisture and create a strong, low-moisture skin.
Table 2 shows the optimized sand composition for the surface-dried molds.
| Component | Weight percentage |
|---|---|
| Silica sand (80–120 mesh) | 100 |
| Bentonite | 6–8 |
| Sawdust | 3–5 |
| Coal dust | 2–3 |
| Moisture | 3–4 |
| Permeability (AFS) | ≥ 150 |
| Green compression strength (MPa) | 0.08–0.10 |
4.2 Mold Design & Tilting
Instead of horizontal pouring, we placed the mold on a specially made support plate and tilted it at an angle of about 10°–15° (see figure below). The pouring basin was located at the low end. This arrangement promoted directional solidification and better gas escape. The inclined orientation reduced the direct thermal load on the large flat surface because the molten metal rose gradually, and gases could exit through the top vents more easily.
4.3 Pouring System & Parameters
- Gating system: We used a single downsprue with a well-rounded junction to ensure smooth flow and reduced erosion.
- Pouring temperature: Lowered to 1350–1380°C (from ~1420°C) to reduce thermal shock.
- Pouring speed: Fast but controlled, filling the mold in 8–10 seconds to minimize heat exposure time.
- Venting: Multiple small vent holes (1.5–2 mm diameter) were drilled in the upper part of the mold, especially near the large flat area, to allow trapped gases to escape.
4.4 Additional Measures for Each Defect Type
Gas Porosity
- Use rounded sand grains for higher permeability.
- Ensure smooth flow in the gating system to avoid turbulence and entrained gas.
- Maintain uniform mold compaction – not too dense – to keep high permeability.
- Reduce binder content (bentonite) and add permeable additives (sawdust).
- Drill extra vents in the cope directly above the large flat area.
- Tilt the mold to help gas rise to the vents.
Sand Inclusions
- Place the ingate at the low end of the tilted mold to reduce direct erosion of the cavity roof.
- Minimize the time between surface drying and pouring to avoid moisture reabsorption and sand fall-off.
- Design core prints with appropriate clearances to prevent crushing during closing.
- Clean the cavity thoroughly before closing, and cover the pouring basin immediately.
Mold-Metal Adhesion
- Increase mold compaction to reduce intergranular voids.
- Add 2–3% coal dust to the sand mixture to create a reducing atmosphere and form a lustrous carbon layer that prevents metal penetration.
Sand Scabs
- Improve hot wet tensile strength: control clay quality and quantity, activate bentonite properly, and increase its content slightly.
- Insert steel nails or chaplets in the areas prone to scabbing to mechanically reinforce the sand.
- Reduce thermal expansion by adding sawdust, coal dust, or heavy oil – these additives burn out and create voids that absorb expansion.
- Improve venting to reduce gas pressure that can lift the sand layer.
- Adopt low-temperature, fast-pouring practice with multiple, distributed ingates to avoid local overheating.
5. Results and Discussion
After implementing all the above measures, we produced a batch of 300 castings with a drastically reduced rejection rate. The top surface defects fell from over 50% to less than 5%. The surface roughness met the required Ra ≤ 12.5 μm. No gas holes were visible, and sand inclusions were virtually eliminated. The slight increase in production time (for surface drying) and minor cost (for sawdust and extra bentonite) were well justified by the quality improvement.
We also derived an empirical formula to estimate the required vent area \(A_v\) for a given mold surface area \(A_m\) and pouring rate \(Q\):
$$ A_v = k \cdot \frac{Q \cdot \rho \cdot w}{v_g \cdot P_{\text{max}}} $$
where \(k\) is a safety factor (≈ 1.5), \(\rho\) is the metal density, \(w\) is the moisture fraction in the sand, \(v_g\) is the gas velocity through the vents, and \(P_{\text{max}}\) is the allowable back pressure (typically 0.1–0.2 atm). For our case, \(A_v ≈ 2.5\) cm² per 1000 cm² of mold surface.
The tilt angle \(\theta\) was optimized using the condition that the hydrostatic pressure at the top of the inclined plane should not exceed the sand’s hydrostatic resistance:
$$ \rho g h \sin\theta \le 0.5 \sigma_{\text{hot}} $$
where \(h\) is the vertical height of the cavity, \(\sigma_{\text{hot}}\) is the hot tensile strength of the dried sand (about 0.5–1.0 MPa). For our geometry, \(\theta = 12^\circ\) was found safe.
6. Conclusions
Large flat surfaces of ductile iron castings are especially vulnerable to sand casting defect phenomena, including gas porosity, sand scabs, inclusions, and adhesion. By systematically analyzing the root causes – excessive moisture, thermal expansion mismatch, poor venting, and inadequate sand properties – and by shifting from high-moisture green sand to surface-dried molds, tilting the mold, optimizing the gating system, and adding scavenging materials like sawdust and coal dust, we successfully eliminated these defects. Our approach demonstrates that even for small batches, careful control of material and process parameters can yield high-quality castings. The experience has strengthened our capability to handle similar sand casting defect challenges in the future.