I carried out this work to solve two persistent defects in a lost foam casting process for a reducer housing made of ductile iron casting. The component was produced in a foundry environment using an expandable pattern, dry sand, and vacuum-assisted pouring. The material was QT450-10, a ferritic-pearlitic ductile iron casting grade requiring good strength, toughness, wear resistance, and vibration damping. The reducer housing weighed 112 kg. Its maximum wall thickness was 54 mm, its minimum wall thickness was 14 mm, and its geometry created several concentrated hot spots. I selected lost foam casting because it offers good surface quality, high dimensional accuracy, and high process yield. However, during initial trials and small-batch production, the ductile iron casting exhibited wrinkles on the surface and shrinkage cavities in the geometric hot spots. The wrinkles were especially visible on the upper face and vertical dead zones. The shrinkage cavities appeared near the bolt-hole bosses and other thick sections. My objective was to identify the formation mechanisms and develop a robust production method that would eliminate both defects without sacrificing process yield.
Baseline process and material conditions. I began by recording the original process parameters and material specification. The original gating design used top ingates. Although it appeared to be a top-pouring arrangement, the actual filling pattern behaved like a middle-to-bottom pour because the metal used the mould cavity as a runner. The foam pattern was made from a copolymer material that balanced gas evolution and solid carbon generation compared with pure EPS or EPMMA. The pouring temperature was 1370–1440 °C, the tapping temperature was 1580–1600 °C, the vacuum level was −0.06 to −0.04 MPa, and the vacuum holding time was 900 s. The chemical composition was within the required range for QT450-10. A Y-block test bar met the mechanical property requirements of QT450-10, and the nodularity was 2–3. These results told me that the base metallurgy was not the primary cause of the defects. The problem was therefore related to filling behaviour, pattern decomposition, local thermal modulus, and solidification feeding.
| Parameter | Value or condition |
|---|---|
| Material | QT450-10 ductile iron casting |
| Casting mass | 112 kg |
| Maximum wall thickness | 54 mm |
| Minimum wall thickness | 14 mm |
| Pattern material | Copolymer foam |
| Pouring temperature | 1370–1440 °C |
| Tapping temperature | 1580–1600 °C |
| Vacuum level | −0.06 to −0.04 MPa |
| Vacuum holding time | 900 s |
| Original gating | Top ingates, cavity acting as runner |
| Observed defects | Surface wrinkles, shrinkage cavities |
Chemical composition and metallurgical baseline. I verified the chemical composition of the ductile iron casting before changing the process. The carbon and silicon levels were typical for ductile iron, but the high carbon content also increased the risk of carbon-related defects in lost foam casting. The magnesium and rare earth contents were sufficient for nodularisation. The phosphorus and sulphur contents were kept low to avoid embrittlement and hot tearing. The carbon equivalent was calculated using a standard relationship for cast iron. The result was consistent with a ductile iron casting that can achieve QT450-10 properties when the nodularity and matrix are controlled.
| Element | Range, wt.% | Role in ductile iron casting |
|---|---|---|
| C | 3.5–4.0 | Promotes graphite formation, affects fluidity and shrinkage |
| Si | 2.0–3.0 | Promotes ferrite, affects nodularity and carbon equivalent |
| Mn | ≤0.45 | Controls matrix strength, but excess increases segregation |
| P | ≤0.05 | Should be low to avoid embrittlement |
| S | ≤0.025 | Must be low for spheroidal graphite |
| Mg | 0.02–0.06 | Nodularising element |
| RE | 0.015–0.04 | Supports nodularisation and neutralises impurities |
The carbon equivalent was estimated as follows:
$$ CE = C + \frac{Si}{3} + \frac{P}{3} $$
For the specified ranges, the carbon equivalent remained within the family of hypereutectic and near-eutectic ductile iron casting compositions. This supported good fluidity but also increased the importance of controlled filling and feeding. I concluded that the defects were not caused by an incorrect grade specification. They were caused by the interaction between the lost foam pattern, the gating system, the vacuum field, and the local solidification modulus.
Defect appearance and classification. I classified the defects into two groups. The first group was surface wrinkling, also called orange-peel skin or carbon-related surface roughness. The wrinkles appeared on the upper surface, vertical side walls, and dead corners. They were rough, folded, and sometimes accompanied by carbon deposits. The second group was shrinkage, which appeared as irregular cavities with rough internal walls in the thick bolt-hole bosses and geometric hot spots. These cavities were not gas holes. They were feeding-related cavities caused by volume contraction during liquid cooling and solidification. The two defects had different causes, so I treated them as two separate process problems and then combined the solutions.
| Defect | Location | Morphology | Likely mechanism |
|---|---|---|---|
| Wrinkle | Top face, side walls, dead corners | Folded skin, orange-peel texture, carbon-rich deposits | Turbulent filling and incomplete decomposition of foam |
| Shrinkage cavity | Bolt-hole bosses, thick hot spots | Rough irregular internal cavity | Insufficient feeding of the last solidifying region |
| Cold shut | Not dominant | Not observed as a batch problem | Filling was not globally cold, but locally unstable |
| Gas hole | Not dominant | Not the primary defect | Vacuum and copolymer pattern reduced gas risk |
Root cause of wrinkles in the ductile iron casting. I found that the original gating system was the main cause of wrinkling. The ingates were placed on the top face, but the actual metal flow did not fill the pattern from top to bottom in a stable sequence. Instead, the high-temperature metal entered the top thick section and used the cavity as a runner. It then reached the thin wall, spread downward in a fan-like pattern, and finally filled upward from the bottom. The actual flow was therefore a mixed middle-to-bottom pour with strong turbulence. This produced a cold end at the top and in side dead zones. In lost foam casting, the pattern decomposes into gas, liquid, and solid products. The solid carbon and liquid residues can be trapped at the cold end. This creates carbon defects, folds, and wrinkles on the surface of the ductile iron casting.
I described the pattern decomposition as a multi-phase process:
$$ Pattern \rightarrow Gas + Liquid + Solid\ Carbon $$
The gas pressure at the filling front can be represented as a balance between atmospheric pressure, decomposition gas pressure, and vacuum pressure:
$$ P_g = P_{atm} + P_{decomp} – P_{vac} $$
When the filling front is unstable, the decomposition products cannot be removed uniformly. The local Reynolds number increases, and the flow becomes turbulent. I used the Reynolds number to explain the change in flow regime:
$$ Re = \frac{\rho v D}{\mu} $$
Here, \(\rho\) is the metal density, \(v\) is the local flow velocity, \(D\) is a characteristic channel dimension, and \(\mu\) is the dynamic viscosity. In the original top-gating arrangement, the metal velocity and the effective channel dimension both increased in the thick upper section. As a result, \(Re\) rose, the filling front broke down, and the decomposition products were pushed into the top and side dead zones. Those zones became the cold ends where wrinkles formed. The wrinkle tendency can be summarised by a carbon defect index:
$$ CDI = \frac{m_C}{m_m} \frac{1}{t_e} $$
In this expression, \(m_C\) is the mass of solid carbon residue, \(m_m\) is the local metal mass, and \(t_e\) is the evacuation time for the decomposition products. To reduce \(CDI\), I needed to reduce the amount of carbon residue retained at the cold end and increase the evacuation time or improve the evacuation path. A stable bottom pour would accomplish both goals.
| Factor | Original top-gating behaviour | Effect on wrinkle formation |
|---|---|---|
| Filling direction | Top entry, cavity used as runner | Created mixed middle-to-bottom flow |
| Flow regime | Turbulent | Trapped decomposition products |
| Cold end location | Top face and side dead corners | Wrinkles appeared in these regions |
| Pattern decomposition | Gas, liquid, solid carbon | Carbon-rich skin and folds |
| Vacuum removal | Uneven at dead zones | Residue remained on surface |
Root cause of shrinkage in the ductile iron casting. I ruled out incorrect chemistry, incorrect nodularity, low pouring temperature, and inadequate vacuum as the main causes of shrinkage. The Y-block properties were satisfactory, the nodularity was acceptable, and the surface did not show cold shuts or burn-on. The shrinkage cavities were located in the thick bolt-hole bosses. These regions had a high geometric modulus and solidified later than the surrounding thin walls. In ductile iron casting, shrinkage occurs when the last solidifying region cannot receive enough liquid metal to compensate for volume contraction. The shrinkage volume can be approximated as:
$$ V_{sh} = V \left[ \alpha_v (T_p – T_s) + \beta \right] $$
Here, \(V\) is the affected volume, \(\alpha_v\) is the volumetric contraction coefficient in the liquid state, \(T_p\) is the pouring temperature, \(T_s\) is the solidification temperature, and \(\beta\) represents the contraction during solidification. The local modulus is the controlling factor for solidification time. I used the standard modulus definition:
$$ M = \frac{V}{A} $$
The solidification time was estimated using a Chvorinov-type relationship:
$$ t_s = B M^2 $$
In this equation, \(B\) is a mould constant that depends on the mould material, vacuum condition, and interface heat transfer. The hot spots had a larger \(M\), so they remained liquid longer. Because the surrounding thin walls froze earlier, the hot spots could not be fed. This produced the shrinkage cavities that I observed after machining. The shrinkage problem was therefore a geometric hot-spot problem, not a global pouring problem.
| Candidate cause | Observation | Conclusion |
|---|---|---|
| Incorrect chemistry | Composition within specification | Not the main cause |
| Poor nodularity | Nodularity 2–3 | Acceptable for QT450-10 |
| Low pouring temperature | No cold shuts or misruns | Not the main cause |
| Insufficient vacuum | No general gas defects | Not the main cause |
| Geometric hot spot | Shrinkage at bolt-hole bosses | Primary cause |
First solution: bottom gating for wrinkle elimination. I redesigned the gating system from a top-entry arrangement to a central closed bottom-gating arrangement. The goal was to make the filling sequence stable from the bottom upward. In a proper bottom pour, the metal enters the cavity at the lowest level and rises smoothly. The decomposition products are pushed upward ahead of the metal front and are finally collected in the upper machining allowance. The low-temperature metal and the incompletely decomposed foam products remain in the top allowance, which is later removed. This protects the functional surfaces of the ductile iron casting. I also removed the top-face slag pocket because it was no longer needed and because it disturbed the filling pattern.
I calculated the pouring time and gating area using practical foundry relationships. The target pouring time was set to 24 s. The average static head for bottom gating was 34 cm. The minimum ingate area was calculated with a discharge coefficient to account for friction and flow losses:
$$ t_p = 24 \text{ s} $$
$$ H_p = H_0 – \frac{P}{2} = 34 \text{ cm} $$
$$ A_g = \frac{G}{\rho t_p C_d \sqrt{2gH_p}} $$
Using \(G = 112 \text{ kg}\), \(\rho = 7000 \text{ kg/m}^3\), \(t_p = 24 \text{ s}\), \(C_d = 0.75\), and \(H_p = 0.34 \text{ m}\), I obtained:
$$ A_g = \frac{112}{7000 \times 24 \times 0.75 \times \sqrt{2 \times 9.81 \times 0.34}} \approx 3.44 \times 10^{-4} \text{ m}^2 = 3.44 \text{ cm}^2 $$
The theoretical minimum ingate area was therefore about 3.46 cm² when rounded by the original calculation method. Because lost foam casting usually requires a slightly larger ingate area than sand casting, I selected a practical range based on experience. For this ductile iron casting, the final ingate area was set between 11.2 and 12.8 cm². I used four central ingates. Each ingate had a cross-section of approximately \(7 \times 40 \text{ mm}\) or \(8 \times 40 \text{ mm}\), depending on the final pattern layout. The sprue length was 480 mm, and the available head was 200 mm. The gating system was a closed central bottom-gating system. The filling front became stable, the turbulence was greatly reduced, and the top machining allowance collected the cold metal and decomposition residue.
| Gating parameter | Original design | New design |
|---|---|---|
| Gating type | Top ingates, cavity as runner | Central closed bottom gating |
| Filling pattern | Mixed middle-to-bottom, turbulent | Bottom-to-top, stable |
| Pouring time | Not optimised | 24 s |
| Average static head | Low and unstable | 34 cm |
| Theoretical ingate area | Not controlled | 3.46 cm² minimum |
| Selected ingate area | Top ingates | 11.2–12.8 cm² |
| Number of ingates | Top entry | 4 |
| Sprue length | Not specified | 480 mm |
| Top slag pocket | Present | Removed |
After the bottom-gating redesign, I produced a trial batch and then scaled to 2000 pieces. The batch results showed that the surface wrinkles were eliminated. The upper machining allowance collected the cold metal and foam residue, and the functional surfaces were sound. The process yield remained high because the gating system did not require large risers or complex side feeders. This was a major improvement for the ductile iron casting process.
Second solution: heat-dissipation fins for shrinkage elimination. The shrinkage problem required a different approach. The conventional methods are risers and chills. A riser provides liquid metal to compensate for contraction, but it reduces process yield and increases cleaning work. A chill creates a local cold zone and promotes directional solidification, but in lost foam casting, external chills are difficult to place. They can fall off during pattern handling, they can distort the pattern, and they can increase process variation. For this ductile iron casting, I needed a method that would reduce the local modulus without adding a large riser and without using loose chills.
I developed a heat-dissipation fin process. The idea was to attach thin foam plates to the pattern at the geometric hot spots. After coating, drying, sand filling, and pouring, these foam plates became metal fins on the casting. The fins increased the local surface area. During solidification, the vacuum pump continued to draw air through the sand. The air entered from the top of the flask, passed through the sand, and exchanged heat with the casting and the fins. The fins created a large surface area and a kind of micro-channel flow. The flowing air removed heat and created a local chilling effect. The local modulus decreased, and the solidification mode became closer to directional solidification. The shrinkage cavities were eliminated.
The heat transfer can be described by the following relationships:
$$ Q = h A_s \Delta T $$
$$ \dot m = C_d A \sqrt{2 \rho \Delta P} $$
$$ \dot Q = \dot m c_p (T_{out} – T_{in}) $$
Here, \(Q\) is the heat removed, \(h\) is the convection coefficient, \(A_s\) is the effective surface area, \(\Delta T\) is the temperature difference, \(\dot m\) is the air mass flow, \(C_d\) is the discharge coefficient, \(A\) is the flow area, \(\rho\) is the air density, \(\Delta P\) is the pressure difference, \(c_p\) is the specific heat of air, and \(T_{out} – T_{in}\) is the air temperature rise. The effective modulus with fins becomes:
$$ M_{eff} = \frac{V + V_f}{A + A_f} $$
The relative modulus reduction is:
$$ \Delta M = \frac{M – M_{eff}}{M} $$
I attached twelve foam fins at the hot-spot regions. Each fin had dimensions of \(50 \times 30 \times 7 \text{ mm}\). The fins were bonded during the pattern assembly stage. The rest of the process remained unchanged. The fins were easy to attach, did not disturb the coating, and did not change the pattern dimensions in critical areas. After pouring, the fins became part of the casting and were removed during normal finishing. Because the fins were small, the process yield was almost unaffected. The finishing work remained simple.
| Fin parameter | Value |
|---|---|
| Fin material | Copolymer foam |
| Fin dimensions | 50 × 30 × 7 mm |
| Number of fins | 12 |
| Attachment method | Bonded to hot-spot regions |
| Location | Bolt-hole bosses and geometric hot spots |
| Added surface area per fin | Approximately 26.2 cm² net |
| Total added surface area | Approximately 314.4 cm² net |
| Process impact | Minimal, simple, low cost |
| Post-processing | Removed with normal finishing |
The modulus reduction can be estimated with a representative hot-spot volume. Suppose a local hot spot has \(V = 100 \text{ cm}^3\) and \(A = 150 \text{ cm}^2\). The original modulus is \(M = 0.667 \text{ cm}\). The twelve fins add about \(V_f = 126 \text{ cm}^3\) and \(A_f = 314.4 \text{ cm}^2\). The effective modulus becomes \(M_{eff} = (100 + 126)/(150 + 314.4) = 0.487 \text{ cm}\). The modulus reduction is about 27%. For a larger hot spot with \(V = 200 \text{ cm}^3\) and \(A = 200 \text{ cm}^2\), the original modulus is 1.0 cm, and the effective modulus becomes \((200 + 126)/(200 + 314.4) = 0.634 \text{ cm}\), a reduction of about 36.6%. These estimates show why the fins improve feeding and reduce shrinkage. The fins do not need to be large; they only need to change the local heat balance enough to make the hot spot solidify earlier or in a more directional manner.
| Case | Original volume, cm³ | Original area, cm² | Original modulus, cm | Effective modulus with fins, cm | Reduction, % |
|---|---|---|---|---|---|
| Small hot spot | 100 | 150 | 0.667 | 0.487 | 27.0 |
| Medium hot spot | 150 | 175 | 0.857 | 0.548 | 36.0 |
| Large hot spot | 200 | 200 | 1.000 | 0.634 | 36.6 |
Mechanism of the heat-dissipation fin process. I examined the mechanism in detail. In lost foam casting, the vacuum system continuously removes decomposition gases and draws air through the sand. When a metal fin is created at the hot spot, the fin increases the contact area between the solidifying metal and the flowing air. The fin also increases the surface area exposed to the sand. Because the fin is thin, it cools quickly. This local cooling reduces the temperature of the adjacent hot spot. The surrounding sand forms a micro-channel region around the fin. Air flows through this region and carries heat away. The fin therefore behaves like a chill, but it is formed from the same metal and is attached to the casting. It cannot fall off, and it does not introduce a separate material interface. This is a major advantage for lost foam casting of ductile iron casting components.
I also observed that the fin process did not increase gas defects. The fins were thin, and the copolymer pattern already balanced gas evolution. The vacuum level was sufficient to remove the additional decomposition products. The fins were placed only at hot spots, so they did not disturb the overall filling pattern. The bottom gating system provided a stable fill, and the fins provided local cooling. The two solutions worked together. The bottom gating solved the wrinkle problem, and the fins solved the shrinkage problem.
| Mechanism | Effect on solidification | Benefit for ductile iron casting |
|---|---|---|
| Increased surface area | Lowers local modulus | Reduces shrinkage tendency |
| Air flow through vacuum | Removes heat from fins | Creates local chilling |
| Micro-channel around fins | Improves heat exchange | Promotes directional solidification |
| Thin metal fins | Solidify quickly | Act like attached chills |
| No loose chills | Avoids displacement | Improves process stability |
Comparison of conventional methods and the new process. I compared risers, chills, and heat-dissipation fins. Risers can solve shrinkage, but they reduce process yield and require more cleaning. Chills can create local cooling, but they are difficult to fix in lost foam patterns. They can fall off, shift, or create pattern distortion. The heat-dissipation fin process is simple, stable, and cost-effective. It uses the same foam material as the pattern, so it is compatible with the lost foam process. It does not require a separate chill material or a large riser. It also does not significantly reduce process yield. For this ductile iron casting, the fin process was the best solution.
| Method | Feeding or cooling effect | Process difficulty | Effect on yield | Suitability for lost foam ductile iron casting |
|---|---|---|---|---|
| Riser | Provides liquid feeding | Moderate to high | Reduces yield | Limited |
| External chill | Creates local cold zone | High | Small effect | Poor due to placement and stability |
| Heat-dissipation fin | Increases local cooling | Low | Almost no effect | Excellent |
Validation and production results. I validated the combined process in two stages. First, I tested the bottom-gating design on trial castings and then on 2000 production pieces. The surface wrinkles were eliminated. The top machining allowance collected the cold metal and foam residue. The functional surfaces were clean and sound. Second, I tested the heat-dissipation fin process at the bolt-hole hot spots. After machining, the bolt holes were inspected. No batch shrinkage cavities were found. I then combined both solutions and produced another 2000 pieces. The combined process produced sound ductile iron casting components with stable quality. The process yield remained high, and the cleaning work did not increase significantly.
| Production stage | Quantity | Wrinkle result | Shrinkage result | Overall quality |
|---|---|---|---|---|
| Original process | Small batch | Large proportion of wrinkles | Shrinkage at hot spots | Not acceptable |
| Bottom gating only | 2000 pieces | Batch wrinkles eliminated | Still present | Improved but incomplete |
| Fins only | Trial castings | Not addressed | Hot-spot shrinkage reduced | Partial improvement |
| Combined process | 2000 pieces | No batch wrinkles | No batch shrinkage | Acceptable |
I also checked the mechanical properties and nodularity after the process changes. The Y-block properties continued to meet QT450-10. The nodularity remained 2–3. The chemical composition remained within the specification. The process changes affected filling and solidification, not the metallurgical grade. This was important because the reducer housing required both strength and toughness. The ductile iron casting also required wear resistance and vibration damping. The combined process maintained these properties while eliminating the surface and internal defects.
| Quality item | Requirement | Result after optimisation |
|---|---|---|
| Material grade | QT450-10 | Met |
| Tensile strength | QT450-10 requirement | Met by Y-block test |
| Elongation | QT450-10 requirement | Met by Y-block test |
| Nodularity | 2–3 | 2–3 |
| Surface wrinkles | None on functional surfaces | Eliminated |
| Shrinkage cavities | None in machined holes | Eliminated |
| Process yield | High | Maintained |
Discussion of the filling stability. The success of the bottom gating system can be explained by the reduction of the filling front velocity and the Reynolds number. In the original process, the metal entered the top thick section and accelerated. The cavity acted as a runner, so the flow was not controlled. In the new process, the metal entered through four central bottom ingates. The flow split into several streams and then rose smoothly. The filling front was more uniform, and the decomposition products were pushed upward. The top machining allowance acted as a collector for the cold metal and residue. The surface of the ductile iron casting was therefore free from wrinkles. I found that the bottom gating system also reduced the risk of sand collapse and pattern deformation because the filling force was more balanced.
The filling front velocity can be estimated as:
$$ v_f = \frac{Q}{A_c} $$
Here, \(Q\) is the volumetric flow rate and \(A_c\) is the local cross-sectional area. In the original top-gating design, \(A_c\) changed abruptly, so \(v_f\) fluctuated. In the new bottom-gating design, the ingates were sized to maintain a more stable \(v_f\). The flow rate itself was controlled by the pouring time and the static head:
$$ Q = \frac{G}{\rho t_p} $$
For \(G = 112 \text{ kg}\), \(\rho = 7000 \text{ kg/m}^3\), and \(t_p = 24 \text{ s}\), the average volumetric flow rate was about \(6.67 \times 10^{-4} \text{ m}^3/\text{s}\). This flow rate was distributed through the four ingates. The local velocity in each ingate was therefore moderate. The lower velocity reduced turbulence and improved the removal of decomposition products. This is a key reason why the bottom-gating design solved the wrinkle defect in the ductile iron casting.
| Flow parameter | Original process | New bottom-gating process |
|---|---|---|
| Flow path | Top entry, cavity as runner | Four central bottom ingates |
| Flow stability | Low | High |
| Filling front | Fan-like and turbulent | Uniform and upward |
| Decomposition product removal | Poor at dead zones | Collected in top allowance |
| Wrinkle risk | High | Low |
Discussion of local modulus and feeding. The heat-dissipation fins changed the local modulus and the solidification sequence. In a ductile iron casting, graphite expansion can partially compensate for shrinkage, but it cannot compensate for an isolated hot spot that is surrounded by thin walls. The hot spot needs either liquid feeding or local cooling. The fins provided local cooling without a large riser. The effective modulus decreased, so the hot spot solidified earlier. The surrounding thin walls no longer froze before the hot spot. The solidification front moved in a more directional manner. The feeding distance increased, and the shrinkage cavities were eliminated.
The feeding condition can be expressed qualitatively as:
$$ M_{hot} \leq M_{feeding} $$
When the hot-spot modulus is greater than the feeding modulus, shrinkage occurs. The fins reduced \(M_{hot}\) by increasing the surface area. The vacuum air flow increased the heat transfer coefficient. The combination made \(M_{hot}\) smaller than the critical value. This is why the fins worked. The process did not require a large riser, so the process yield remained high. The fins also acted as local cooling paths during the early solidification stage, which helped to reduce the temperature gradient between the hot spot and the thin wall.
| Solidification factor | Without fins | With fins |
|---|---|---|
| Local surface area | Baseline | Increased |
| Local modulus | High | Reduced |
| Solidification time | Long | Shortened |
| Feeding distance | Limited | Improved |
| Shrinkage risk | High | Low |
Process control and reproducibility. I paid attention to several process control points during scale-up. The foam pattern density had to be uniform because density affects gas evolution and filling. The coating thickness had to be controlled because coating permeability affects vacuum removal. The fins had to be bonded securely so they would not detach during coating or sand filling. The vacuum level had to be maintained at −0.06 to −0.04 MPa during pouring and holding. The pouring temperature had to be kept within 1370–1440 °C. The pouring time had to be close to 24 s. These parameters were recorded for each batch. The production data showed that the combined process was reproducible. The 2000-piece validation runs confirmed that the wrinkles and shrinkage cavities could be eliminated in routine production.
| Control point | Target | Reason |
|---|---|---|
| Pattern density | Uniform | Controls gas evolution and filling stability |
| Coating thickness | Controlled range | Affects vacuum removal and surface quality |
| Fin bonding | Secure | Prevents detachment and local modulus loss |
| Vacuum level | −0.06 to −0.04 MPa | Controls decomposition product removal and cooling |
| Pouring temperature | 1370–1440 °C | Ensures fluidity without excessive gas |
| Pouring time | 24 s | Controls filling front velocity |
| Vacuum holding time | 900 s | Ensures sufficient solidification and gas removal |
Additional observations on carbon defects. The wrinkle defect was closely related to carbon residue. In lost foam casting, the foam decomposes into hydrocarbons and carbon. If the decomposition products cannot escape, they form a carbon-rich layer on the surface. This layer creates folds and wrinkles. The bottom gating design helped because it pushed the decomposition products to the top allowance. The copolymer pattern also helped because it balanced gas and solid carbon. I found that the combination of copolymer pattern, bottom gating, and proper vacuum was essential. If any one of these factors was not controlled, the wrinkle risk increased. For example, a lower vacuum level would reduce the removal rate of decomposition products. A higher pouring temperature would increase gas evolution. A longer pouring time would allow more carbon to deposit. Therefore, the process window had to be maintained.
The carbon defect index can be used as a qualitative tool:
$$ CDI = \frac{m_C}{m_m} \frac{1}{t_e} $$
To reduce \(CDI\), I reduced \(m_C\) by using a copolymer pattern, increased \(t_e\) by using bottom gating and a top allowance, and maintained a sufficient vacuum level. The result was a clean surface on the ductile iron casting. The wrinkles no longer appeared in batch production.
| Carbon defect factor | Action taken | Effect |
|---|---|---|
| Pattern material | Used copolymer foam | Balanced gas and solid carbon |
| Filling pattern | Changed to bottom gating | Pushed residue to top allowance |
| Top allowance | Used as residue collector | Protected functional surfaces |
| Vacuum | Maintained −0.06 to −0.04 MPa | Improved gas removal |
| Pouring time | Set to 24 s | Reduced turbulence and carbon deposition |
Additional observations on shrinkage and graphite expansion. Ductile iron casting has a unique solidification behaviour because graphite expansion can partially offset liquid and solidification shrinkage. However, this expansion is only effective if the graphite nodules form in a connected or uniformly distributed manner and if the mould has enough rigidity. In lost foam casting with vacuum, the mould is relatively rigid, so graphite expansion can be used to reduce feeding requirements. However, if the hot spot is isolated and the surrounding thin walls freeze early, the expansion cannot compensate for the local volume deficit. The fins changed the local cooling and allowed the graphite expansion to occur more uniformly. This reduced the need for a large riser. The process therefore combined the natural expansion of ductile iron with local cooling. This is a key reason why the fin process worked so well for this ductile iron casting.
The shrinkage condition can be summarised as:
$$ V_{sh} = V \left[ \alpha_v (T_p – T_s) + \beta \right] – V_{graphite} $$
Here, \(V_{graphite}\) is the volume change due to graphite expansion. When \(V_{graphite}\) is large and distributed, shrinkage is reduced. When the hot spot is isolated, \(V_{graphite}\) is not effective, and shrinkage remains. The fins improved the distribution of cooling and allowed \(V_{graphite}\) to act more effectively. This is consistent with the observed elimination of shrinkage cavities after the fin process was introduced.
| Shrinkage compensation factor | Without fins | With fins |
|---|---|---|
| Liquid contraction | Higher local deficit | Reduced by earlier solidification |
| Solidification contraction | Concentrated at hot spot | More distributed |
| Graphite expansion | Not fully effective | Better utilised |
| Feeding requirement | High | Lower |
| Shrinkage risk | High | Low |
Engineering summary of the combined process. The final process for the ductile iron casting included the following steps. I designed a central closed bottom-gating system with four ingates. I set the pouring time to 24 s and the average static head to 34 cm. I used a theoretical minimum ingate area of about 3.46 cm² and a selected practical ingate area of 11.2–12.8 cm². I removed the top-face slag pocket and used the top machining allowance as a collector for cold metal and decomposition residue. I attached twelve foam fins of \(50 \times 30 \times 7 \text{ mm}\) at the bolt-hole hot spots. I maintained the vacuum level at −0.06 to −0.04 MPa and the holding time at 900 s. I kept the pouring temperature at 1370–1440 °C. The rest of the process remained unchanged. This combined process eliminated wrinkles and shrinkage cavities in the ductile iron casting and maintained high process yield.
| Process step | Final condition |
|---|---|
| Gating system | Central closed bottom gating |
| Number of ingates | 4 |
| Ingate area | 11.2–12.8 cm² |
| Pouring time | 24 s |
| Average static head | 34 cm |
| Sprue length | 480 mm |
| Top slag pocket | Removed |
| Hot-spot treatment | 12 foam fins, 50 × 30 × 7 mm |
| Vacuum level | −0.06 to −0.04 MPa |
| Vacuum holding time | 900 s |
| Pouring temperature | 1370–1440 °C |
Why the process is suitable for ductile iron casting. The process is suitable for ductile iron casting because it addresses the specific characteristics of ductile iron. Ductile iron has high carbon content, a wide solidification range, and a tendency to form carbon-related surface defects in lost foam casting. It also has geometric hot spots that can produce shrinkage cavities. The bottom gating system controls the filling front and reduces turbulence. The copolymer pattern reduces solid carbon. The vacuum system removes decomposition gases. The heat-dissipation fins reduce the local modulus and improve feeding. The combination of these measures is robust and practical. It does not rely on expensive equipment or complex tooling. It can be applied to other lost foam ductile iron casting components with similar geometry.
| Ductile iron casting challenge | Process response | Result |
|---|---|---|
| High carbon and carbon defects | Copolymer pattern and bottom gating | Wrinkles eliminated |
| Turbulent filling | Central bottom gating | Stable filling front |
| Geometric hot spots | Heat-dissipation fins | Shrinkage eliminated |
| Need for high yield | No large risers | Yield maintained |
| Need for simple finishing | Small fins | Easy removal |
Final verification and conclusions. I verified the combined process through repeated production. The bottom-gating design eliminated surface wrinkles in 2000 pieces. The heat-dissipation fin process eliminated shrinkage cavities in 2000 pieces. The combined process produced a sound ductile iron casting with stable dimensions, acceptable surface quality, and reliable internal quality. The mechanical properties and nodularity remained within the QT450-10 requirements. The process yield was high, and the finishing work was simple. I concluded that the wrinkle defect was caused by unstable filling and trapped decomposition products, and that the shrinkage defect was caused by geometric hot spots with insufficient local cooling. The bottom-gating system solved the filling problem, and the heat-dissipation fins solved the solidification problem. The new process is a practical and effective method for lost foam casting of ductile iron casting components with thin walls and concentrated hot spots.
| Conclusion item | Finding |
|---|---|
| Wrinkle cause | Top-entry turbulence and trapped foam decomposition products |
| Shrinkage cause | Geometric hot spots with insufficient feeding |
| Wrinkle solution | Central closed bottom gating with top allowance collection |
| Shrinkage solution | Heat-dissipation fins at hot spots |
| Validation | 2000-piece batch for bottom gating and 2000-piece batch for fins |
| Quality result | No batch wrinkles and no batch shrinkage |
| Process benefit | Simple, high yield, low cost, easy post-processing |
The work showed me that lost foam casting of ductile iron casting components can be controlled effectively when filling stability and local solidification are treated as coupled process variables. The bottom gating system created a stable fill and removed the cold-end carbon defects. The heat-dissipation fins created local cooling and removed the geometric hot-spot shrinkage. I recommend this combined approach for similar ductile iron casting parts with thin walls, high carbon equivalent, and concentrated hot spots. The method is especially useful when risers or external chills would reduce process yield or increase process difficulty. The heat-dissipation fin process is a new and practical way to solve shrinkage in lost foam ductile iron casting without the disadvantages of traditional feeding and chilling methods.