In my work with lost foam casting of ductile iron casting, the reducer housing is one of the most demanding parts I have handled. The component weighs approximately 112 kg, has a maximum wall thickness near 54 mm, a minimum wall thickness near 14 mm, and contains several concentrated geometric hot spots. The specified material is QT450-10, which requires high strength, good toughness, wear resistance, and vibration damping. Because lost foam casting offers excellent surface quality, high dimensional precision, and high process yield, I selected this route for the ductile iron casting. However, the original production process produced two persistent defects: wrinkling on the end faces and shrinkage porosity or shrinkage cavities in the geometric hot spots. The defects appeared not only in trial production but also at a significant rate during small-batch manufacturing. I therefore studied the filling behavior, pyrolysis behavior, feeding behavior, and local thermal modulus of the ductile iron casting, redesigned the pouring system, and developed a heat-dissipation process using sacrificial foam fins. This article summarizes the analysis, calculations, tables, and validation results from that work.
1. Component and Material Basis
The reducer housing is a rotational body with a thick upper region and a thinner lower wall. The abrupt change in section size creates a strong thermal gradient during solidification of the ductile iron casting. The chemical composition range I used is shown in Table 1. The carbon and silicon levels place the alloy in a near-eutectic ductile iron casting range, which promotes graphitization, nodularity, and feeding behavior. The magnesium and rare-earth contents are controlled to achieve nodularity while avoiding excessive slag, dross, or gas defects.
| Element | Content, mass fraction % | Function in ductile iron casting |
|---|---|---|
| C | 3.5–4.0 | Promotes graphite formation; excessive carbon increases carbon defect risk |
| Si | 2.0–3.0 | Promotes graphitization and ferrite formation; affects toughness |
| Mn | ≤ 0.45 | Controls matrix strength; too high may reduce ductility |
| P | ≤ 0.05 | Low phosphorus is required for toughness |
| S | ≤ 0.025 | Low sulfur is required for nodularization |
| Mg | 0.02–0.06 | Nodularizing element |
| RE | 0.015–0.04 | Auxiliary nodularizing and desulfurizing element |
The carbon equivalent can be estimated as follows:
$$CE = C + \frac{Si}{3} + \frac{P}{3}$$
Using the midpoint of the composition range, the carbon equivalent of the ductile iron casting lies near the eutectic region, which is favorable for castability but also increases the risk of carbon-related surface defects in lost foam casting. I confirmed that Y-block mechanical properties met the QT450-10 requirement and that nodularity was in the 2–3 range. Therefore, the defect problem was not caused by a basic material specification failure. The problem was process-related, especially gating, filling stability, and local thermal modulus.
| Item | Target or observed value | Comment |
|---|---|---|
| Material designation | QT450-10 | Ductile iron casting with high toughness |
| Cast mass | 112 kg | Medium heavy section |
| Maximum wall thickness | 54 mm | Hot spot risk |
| Minimum wall thickness | 14 mm | Cold-end risk |
| Geometry | Rotational housing | Concentrated thermal nodes |
| Nodularity | 2–3 | Acceptable for the ductile iron casting |
| Y-block properties | Met QT450-10 | Material base was acceptable |
2. Original Lost Foam Casting Process
In the original process, I used a top pouring system. The end face had a machining allowance of 4 mm. The tapping temperature was 1,580–1,600 °C, the pouring temperature was 1,370–1,440 °C, the vacuum level was −0.06 to −0.04 MPa, and the vacuum holding time was 900 s. The chemical composition was controlled according to Table 1. Although the material met the required mechanical properties, the ductile iron casting still showed wrinkling, shrinkage porosity, and shrinkage cavities. The defects were concentrated on the upper end face and in the thick machined boss regions.
| Parameter | Value |
|---|---|
| Pouring system | Top pouring |
| Machining allowance on end face | 4 mm |
| Tapping temperature | 1,580–1,600 °C |
| Pouring temperature | 1,370–1,440 °C |
| Vacuum during pouring | −0.06 to −0.04 MPa |
| Vacuum holding time | 900 s |
| Observed defects | Wrinkling, shrinkage porosity, shrinkage cavities |
The original top pouring arrangement behaved like a combined middle and bottom gating system because the cavity itself acted as a runner. High-temperature liquid iron entered the top face, penetrated the thick upper wall, spread laterally, and then moved downward through the thin wall in a fan-shaped flow. After reaching the bottom, the flow reversed upward. This created turbulence, cold-end zones, and large dead corners. In a ductile iron casting produced by lost foam casting, such flow behavior strongly affects pyrolysis product removal and carbon defect formation.
3. Root Cause Analysis of Wrinkling
Wrinkling in lost foam casting of ductile iron casting is closely related to carbon defects, incomplete pyrolysis, and unstable filling. The foam pattern decomposes when it contacts liquid iron. The decomposition products include gases, liquids, and solid carbonaceous residues. If these products cannot escape or be consumed quickly, they accumulate near the last-filled regions or the cold ends. The result is a rough, folded, orange-peel-like surface commonly called wrinkling.
The local carbon concentration in the cavity can be expressed conceptually as:
$$C_{\text{local}} = \frac{m_{\text{carbon}}}{V_{\text{cavity}}}$$
When the local carbon concentration is high and the local temperature is low, solid carbon and heavy hydrocarbon residues are more likely to remain on the surface. For the ductile iron casting, the carbon content is already between 3.5% and 4.0%, so the alloy has a limited capacity to absorb additional carbon without forming carbon defects. I selected a copolymer foam material to balance gas evolution and solid carbon generation. This reduced, but did not eliminate, the risk of carbon-related defects because the gating system still produced severe turbulence and cold ends.
| Stage | Temperature range | Products | Effect on ductile iron casting |
|---|---|---|---|
| Softening | Below 200 °C | Pattern deformation | Dimensional variation if support is poor |
| Liquefaction | 200–400 °C | Liquid monomers and oligomers | Sticky residue, coating penetration |
| Gas evolution | 400–700 °C | Hydrocarbons, CO, H2, soot precursors | Gas porosity and back pressure |
| Carbonization | 700–1,000 °C | Solid carbon, soot, heavy tar | Carbon defects and wrinkling |
| High-temperature reaction | Above 1,000 °C | CO, CO2, H2, carbon dissolution | Surface carbon pickup and gas defects |
The flow regime also influenced wrinkling. I estimated the Reynolds number to compare stable and unstable filling:
$$Re = \frac{\rho v D}{\mu}$$
When the liquid iron velocity is high and the channel dimension is large, the Reynolds number rises, the flow becomes turbulent, and the pyrolysis products are mixed into the bulk liquid rather than being pushed smoothly to the upper machining allowance. In the original top pouring system, the metal stream directly impinged on the top surface, broke the foam front, and generated strong eddies. The last-filled regions became cold ends, where the pyrolysis products remained. These regions later appeared as wrinkled surfaces on the ductile iron casting.
| Filling condition | Flow character | Pyrolysis product removal | Wrinkling risk |
|---|---|---|---|
| Top pouring with cavity runner | Turbulent, reversing flow | Poor; products trapped at cold ends | High |
| Middle pouring | Mixed flow, eddies | Moderate to poor | Moderate to high |
| Bottom pouring | Smooth, progressive upward flow | Good; products pushed to top allowance | Low |
The wrinkles were not caused by low pouring temperature alone, because the casting showed no cold shut and no sand burn-on. The surface was metallurgically sound in many areas, but the last-filled upper and side regions had carbon-rich residues. I therefore concluded that the main cause was the gating system and the resulting unstable filling pattern, not the base chemistry of the ductile iron casting.
4. Root Cause Analysis of Shrinkage
Shrinkage porosity and shrinkage cavities in the ductile iron casting originated in the geometric hot spots. In a ductile iron casting, liquid contraction, solidification contraction, and graphite expansion interact. If a hot spot cannot be fed by liquid metal, a cavity forms. The local thermal modulus controls how long the hot spot remains liquid:
$$M = \frac{V}{A}$$
where M is the modulus, V is the local volume, and A is the local heat-dissipating surface area. A larger modulus means a longer solidification time:
$$t_s = B M^2$$
where B is a constant depending on the alloy, mold, and thermal conditions. The thick sections around the bolt holes had a high modulus, so they solidified later than the surrounding thin walls. Because the gating system did not provide a positive feeding gradient, the last liquid in those regions could not be replenished. The shrinkage volume can be estimated as:
$$V_s = V \left[ \alpha_v (T_p – T_s) + \beta \right]$$
where alpha-v is the volumetric liquid contraction coefficient, T-p is the pouring temperature, T-s is the solidification temperature, and beta is the solidification contraction. For the ductile iron casting, graphite expansion can offset part of the shrinkage, but it cannot compensate for an isolated hot spot with no feeding path.
| Region | Relative thickness | Modulus tendency | Feeding condition | Defect tendency |
|---|---|---|---|---|
| Upper end face | Thick | High | Poor in original top gating | Wrinkling and shrinkage |
| Bolt-hole bosses | Very thick | Very high | Isolated by thin walls | Shrinkage cavities |
| Central hub | Medium to thick | Medium to high | Depends on gating | Shrinkage porosity |
| Thin side wall | Thin | Low | Solidifies early | Cold-end wrinkling |
I also considered the Niyama criterion as a useful indicator of shrinkage tendency:
$$N_y = \frac{G}{\sqrt{R}}$$
where G is the local temperature gradient and R is the cooling rate. A low Niyama value indicates a high risk of shrinkage porosity. In the original process, the temperature gradient around the bolt-hole bosses was low because the surrounding thin walls cooled quickly and isolated the hot spot. The hot spot therefore had a low Niyama value, which matched the observed shrinkage cavities. Traditional solutions include adding risers or using chills. However, for this ductile iron casting produced by lost foam casting, both solutions had serious drawbacks.
| Method | Principle | Advantage | Limitation in this ductile iron casting |
|---|---|---|---|
| Riser | Provides liquid feeding | Directly compensates shrinkage | Reduces process yield; increases cleaning and cutting cost |
| External chill | Accelerates local cooling | Creates directional solidification | Difficult to place in lost foam; may fall off; may distort pattern |
| Internal chill | Acts as heat sink inside casting | Strong local cooling | Must match chemistry; risk of incomplete fusion |
| Heat sink fin | Increases local surface area and air cooling | Simple, low cost, high yield | Requires careful placement and foam bonding |
5. Gating System Optimization
To eliminate wrinkling, I changed the pouring system from top pouring to a center closed bottom gating system. The new design removed the upper slag trap and introduced four ingates at the bottom. The filling pattern became progressive and upward. High-temperature liquid iron entered from the bottom, pushed the foam pyrolysis products ahead of the front, and finally deposited low-temperature iron and residues in the top machining allowance. This allowed the sound metal to form the functional surfaces of the ductile iron casting.
I calculated the pouring time using a foundry relationship:
$$t = S \sqrt{G}$$
where t is the pouring time, S is a process coefficient, and G is the casting mass. The calculated pouring time was approximately 24 s. The average static pressure head for bottom gating was calculated as:
$$H_P = H_0 – \frac{C}{2}$$
Using the original height terms, the average static pressure head was approximately 34 cm. The minimum ingate area was estimated as:
$$A_g = \frac{G}{0.31\, t \sqrt{H_P}}$$
The theoretical minimum ingate area was about 3.46 cm² according to this foundry calculation. Because lost foam casting requires slightly larger ingates than green sand casting, and because the ductile iron casting has a complex rotational shape, I selected a center sprue with four ingates. Each ingate was approximately 70 mm by 4 mm, giving a total ingate area of about 11.2–12.8 cm². The sprue length was set to 480 mm with a pressure head near 200 mm.
| Parameter | Symbol | Value |
|---|---|---|
| Casting mass | G | 112 kg |
| Pouring time | t | 24 s |
| Average static pressure head | H_P | 34 cm |
| Theoretical minimum ingate area | A_g | 3.46 cm² |
| Selected ingate count | n | 4 |
| Selected ingate size | — | 70 mm × 4 mm each |
| Total selected ingate area | A_total | 11.2–12.8 cm² |
| Sprue length | L_s | 480 mm |
| Pressure head for design | H_0 | 200 mm |
The new gating system also reduced the Reynolds number at the ingate because the metal entered through multiple smaller ports instead of a single high-velocity stream. The flow front became more uniform. The Bernoulli relationship helps explain the pressure and velocity behavior:
$$P + \frac{1}{2}\rho v^2 + \rho g h = \text{constant}$$
By using bottom gating, I lowered the effective impact velocity on the foam pattern and avoided the reversing flow that had created cold ends. The ductile iron casting filled from the bottom upward, and the pyrolysis products were continuously swept toward the top allowance. This is a key reason why bottom gating is so effective for reducing wrinkling in lost foam casting of ductile iron casting.
| Feature | Original top gating | Improved bottom gating |
|---|---|---|
| Filling direction | Top to cavity, then reversing | Bottom to top, progressive |
| Flow character | Turbulent | Relatively smooth |
| Cold-end location | Top and side thick corners | Top machining allowance |
| Pyrolysis product removal | Poor | Good |
| Wrinkling result | Frequent | Eliminated in batch production |
| Process yield | Acceptable but defects reduced yield | High |
After the redesign, I produced 2,000 pieces in a batch. The surface quality of the ductile iron casting was acceptable, and no batch wrinkling occurred. The end faces and side walls were clean after shot blasting and machining. This confirmed that the wrinkling mechanism had been correctly identified and that bottom gating was an effective remedy.
6. Heat Sink Process for Shrinkage
For the shrinkage cavities in the geometric hot spots, I developed a heat sink process rather than using conventional risers or chills. The method uses small foam fins bonded to the pattern at the hot spot locations. During coating, drying, flasking, pouring, and vacuum cooling, the fins become part of the mold cavity. After casting, they create additional surface area and promote local heat extraction. In effect, they act as an external chill without the handling problems of metal chills.
For the ductile iron casting, I placed 12 foam fins at the bolt-hole boss regions. Each fin measured 50 mm × 30 mm × 7 mm. The fins were bonded to the pattern before coating. The rest of the process remained unchanged. During vacuum pouring and solidification, the vacuum pump continued to extract air. Cool air entered from the top of the flask, passed through the sand and near the fins, and removed heat by convection. Because the fins had a high surface-area-to-volume ratio, the local modulus dropped and the hot spot solidified earlier.
| Parameter | Value |
|---|---|
| Fin material | Foam compatible with pattern |
| Fin length | 50 mm |
| Fin width | 30 mm |
| Fin thickness | 7 mm |
| Number of fins | 12 |
| Placement | Bolt-hole hot spots |
| Bonding stage | Pattern assembly |
| Process change | None after coating and flasking |
The heat transfer benefit can be described by Newton’s law of cooling:
$$Q = h A \Delta T$$
where Q is the heat transfer rate, h is the convective heat transfer coefficient, A is the surface area, and delta-T is the temperature difference between the casting and the cooling medium. By adding fins, I increased A. Because the vacuum system maintained air movement through the sand, h was also improved near the fins. The combined effect increased Q and reduced the local solidification time.
The fin efficiency can be estimated using standard fin equations:
$$m = \sqrt{\frac{h P}{k A_c}}$$
$$\eta_f = \frac{\tanh(mL)}{mL}$$
where P is the fin perimeter, k is the thermal conductivity, A-c is the fin cross-sectional area, L is the fin length, and eta-f is the fin efficiency. Even if the foam fin is consumed during pouring, the resulting cavity and sand flow path create a local cooling channel. The effective modulus of the hot spot becomes:
$$M_{\text{eff}} = \frac{V}{A + \eta_f A_f}$$
Because the denominator increases, the effective modulus decreases. A lower modulus means a shorter local solidification time and a lower shrinkage tendency. The local cooling rate also increases:
$$\dot{T} = \frac{h A (T – T_\infty)}{\rho V c_p}$$
where rho is density, c-p is specific heat, T is the local casting temperature, and T-infinity is the cooling air temperature. This equation shows that increasing A or h raises the cooling rate. The Biot number helps indicate whether the cooling is controlled by surface convection or internal conduction:
$$Bi = \frac{h L_c}{k}$$
For the thin hot-spot regions, the added fins increased the surface heat transfer enough to shift the local solidification pattern toward directional solidification. The Niyama value increased because the temperature gradient rose and the cooling rate also rose. The result was a reduction in shrinkage porosity and shrinkage cavities in the ductile iron casting.
| Quantity | Without fins | With fins | Effect |
|---|---|---|---|
| Local surface area | Low | Increased by fin area | More heat extraction |
| Local modulus | High | Lower | Earlier solidification |
| Temperature gradient | Low | Higher | Better feeding direction |
| Cooling rate | Lower | Higher | Reduced shrinkage |
| Niyama value | Low | Higher | Lower porosity risk |
| Process yield | Acceptable | High | No large riser needed |
The heat sink process is especially suitable for lost foam casting of ductile iron casting because the foam pattern can be modified easily before coating. Unlike a metal chill, the foam fin does not fall off during flasking, does not create a mismatch with the pattern, and does not require special surface treatment. It also does not reduce process yield because it is not a riser that must be cut and remelted. The fins are simple to attach, and their position can be adjusted according to the hot spot map of the casting.
7. Experimental Validation
I validated the improved process through batch production. The bottom gating system was used for the ductile iron casting, and the heat sink fins were applied at the bolt-hole hot spots. A total of 2,000 pieces were produced. The castings were inspected after cleaning and machining. The results are summarized in Table 12.
| Inspection item | Original process | Improved process | Result |
|---|---|---|---|
| Batch size | Small batch and trial | 2,000 pieces | Production validation |
| Surface wrinkling | Frequent | No batch wrinkling | Defect eliminated |
| Shrinkage in hot spots | Frequent | No batch shrinkage after machining | Defect eliminated |
| Bolt-hole machining | Shrinkage found | Normal | Accepted |
| Mechanical properties | Met QT450-10 | Met QT450-10 | Unchanged |
| Nodularity | 2–3 | 2–3 | Unchanged |
| Process yield | Reduced by defects | High | Improved |
| Post-processing | Rework and scrap | Simple cleaning | Improved |
After machining, the bolt-hole regions showed no shrinkage defects. The surface of the ductile iron casting was smooth, and the top machining allowance successfully collected the cold-end metal and pyrolysis residues. Because the pouring system was changed to bottom gating, the filling process became stable, and the carbon-rich residues did not remain on the functional surfaces. The heat sink fins effectively reduced the hot spot modulus and eliminated the shrinkage cavities.
8. Discussion
The results show that wrinkling and shrinkage in lost foam casting of ductile iron casting often have different root causes and require different countermeasures. Wrinkling is primarily a filling and pyrolysis product management problem. It is strongly influenced by gating direction, flow velocity, turbulence, cold-end location, foam material, coating permeability, and vacuum level. Shrinkage is primarily a thermal and feeding problem. It is influenced by hot spot modulus, temperature gradient, pouring temperature, graphite expansion, and local cooling rate.
The bottom gating system solved wrinkling because it changed the order in which the mold cavity filled. Instead of allowing the metal to enter from the top and reverse direction, the new system introduced metal from the bottom and pushed the front upward. This is consistent with the basic principle that the last-filled region should be placed in a machining allowance or a non-functional area. The top allowance acted as a trap for low-temperature iron and pyrolysis products. As a result, the functional surfaces of the ductile iron casting were formed by cleaner, hotter metal.
The heat sink process solved shrinkage because it changed the local thermal balance. A hot spot with a high modulus tends to remain liquid after the surrounding thin walls have solidified. By adding fins, I increased the surface area and allowed the vacuum-driven air flow to remove heat from the hot spot. The local modulus decreased, the cooling rate increased, and the hot spot solidified earlier. This created a more favorable temperature gradient and improved feeding. The fins acted in a manner similar to chills but without the handling and alignment problems of metal chills in lost foam casting.
| Defect | Primary mechanism | Key variables | Remedy used | Result |
|---|---|---|---|---|
| Wrinkling | Pyrolysis residue and turbulent filling | Gating, flow velocity, cold end, foam chemistry | Bottom gating | Eliminated |
| Shrinkage cavity | Isolated hot spot with poor feeding | Modulus, temperature gradient, cooling rate | Foam heat sink fins | Eliminated |
| Shrinkage porosity | Low Niyama value and slow local cooling | Hot spot geometry, vacuum, sand cooling | Heat sink fins and vacuum flow | Reduced |
| Carbon defect | Solid carbon from foam decomposition | Copolymer foam, filling pattern, coating | Bottom gating and top allowance | Reduced |
I also found that the heat sink process has practical advantages for ductile iron casting. It does not require a large riser, so the process yield remains high. It does not introduce a metallic insert that could cause a cold shut or a mismatch in composition. It is easy to apply during pattern assembly, and it can be placed exactly where the hot spot analysis indicates. The foam fins can be made in standard sizes and bonded with the same adhesive used for pattern assembly. After casting, the fins leave a small surface feature that can be removed during normal cleaning or machining.
The process also has limits. The fin size, number, and placement must be determined from the local modulus and the hot spot map. If the fins are too small, they will not provide enough cooling. If they are too large or placed incorrectly, they may cause local cold shuts or incomplete filling. The coating must remain intact over the fins, and the vacuum level must be high enough to maintain air flow through the sand. The pouring temperature must still be controlled within the proper range for the ductile iron casting. These variables should be monitored in production.
9. Practical Guidelines
Based on the analysis and batch validation, I recommend the following guidelines for lost foam casting of ductile iron casting with wrinkling and shrinkage risk.
- Use bottom gating or a carefully designed step gating system to promote smooth upward filling.
- Place the last-filled region in a machining allowance or a non-functional area.
- Select a foam material that balances gas evolution and solid carbon generation.
- Control pouring temperature and vacuum level to avoid both cold-end defects and excessive turbulence.
- Identify geometric hot spots by modulus analysis before pattern assembly.
- Apply heat sink fins at hot spots to increase surface area and local cooling rate.
- Keep fin dimensions large enough to reduce the local modulus but small enough to avoid cold shuts.
- Verify coating coverage over fins and maintain vacuum during pouring and solidification.
- Use batch machining inspection to confirm that shrinkage has been eliminated in critical holes and bosses.
- Record process data such as pouring time, vacuum, temperature, and fin placement for traceability.
| Variable | Recommended range or practice | Reason |
|---|---|---|
| Gating type | Center closed bottom gating | Reduces turbulence and cold ends |
| Ingate area | About 11.2–12.8 cm² total | Balances filling speed and foam decomposition |
| Pouring time | Near 24 s | Provides stable filling |
| Pouring temperature | 1,370–1,440 °C | Maintains fluidity without excessive gas |
| Vacuum | −0.06 to −0.04 MPa | Removes gases and supports sand rigidity |
| Vacuum holding | 900 s | Ensures sufficient solidification time |
| Heat sink fins | 50 mm × 30 mm × 7 mm | Provides local cooling without excessive chilling |
| Fin count | 12 at bolt-hole hot spots | Matches the observed shrinkage locations |
| Top allowance | 4 mm or more | Collects cold-end metal and pyrolysis residue |
10. Conclusion
Through systematic analysis and production validation, I found that wrinkling in the lost foam casting of ductile iron casting was caused by unstable filling and pyrolysis product accumulation at cold ends. The original top gating system behaved like a combined middle and bottom gating system, producing turbulent flow and trapping carbonaceous residues on the upper and side surfaces. By changing to a center closed bottom gating system, I made the filling process smooth and progressive. The pyrolysis products and low-temperature iron were pushed into the top machining allowance, and the wrinkling defect was eliminated in a 2,000-piece batch.
The shrinkage cavities were caused by isolated geometric hot spots in the bolt-hole boss regions. The local modulus was high, the temperature gradient was low, and feeding was insufficient. Traditional risers and chills were not suitable for this ductile iron casting because they reduced process yield or were difficult to apply in lost foam casting. I developed a heat sink process using foam fins bonded to the pattern at the hot spots. The fins increased the local surface area, promoted vacuum-driven air cooling, reduced the effective modulus, and increased the local cooling rate. After machining, the bolt-hole regions showed no shrinkage defects, and the process remained simple with high process yield.
The combined approach of bottom gating and heat sink fins provides a practical solution for ductile iron casting produced by lost foam casting. It controls both filling-related surface defects and solidification-related internal defects without adding complex tooling or reducing productivity. The method is based on measurable principles such as pouring time, static pressure head, ingate area, modulus, solidification time, heat transfer rate, fin efficiency, and Niyama criterion. These relationships can be used to design and optimize other ductile iron casting parts with similar geometry and defect risks. In my experience, the most important lesson is to separate the filling problem from the feeding problem, then apply a targeted process change for each. When the gating system is stable and the hot spots are properly cooled, the lost foam casting of ductile iron casting can achieve both excellent surface quality and internal soundness.