I was assigned to develop and qualify a lost foam casting route for a heavy-section ductile iron hub with the grade QT500-7. The part had previously been produced by sand casting, but the overall scrap rate was high, the productivity was low, and the manufacturing cost was difficult to justify. My objective was to convert the component to lost foam casting while keeping the specified hardness, elongation, spheroidization level, and internal soundness. The most critical quality requirement was that the hub must not contain shrinkage porosity or shrinkage cavities. Because the average wall thickness was approximately 12 mm but local sections were much heavier, the solidification behavior created several hot spots that demanded careful attention in the ductile iron casting process. The finished hub had a mass of about 120 kg and an overall envelope of approximately 556 mm in diameter and 414.5 mm in height. After normalizing and tempering, the hardness was required to be in the range of 170 to 230 HB. I therefore treated the project as a combined problem of gating design, riser design, lost foam pattern handling, vacuum control, pouring practice, and chemistry control for ductile iron casting.
I began with a structured review of the hub geometry. The part is essentially a flanged wheel-like body with a central bore region, an outer flange, an upper rim, and several internal features. The heavy flange and hub sections feed the thinner walls, so the thermal gradient during cooling is not uniform. In ductile iron casting, the graphite expansion can partly compensate for liquid and solidification shrinkage, but only when the riser and gating system allow enough liquid metal to remain available while the mold cavity is still connected to the feeding path. If the connection solidifies too early, the expansion cannot be transmitted to the hot spot, and shrinkage porosity appears. I therefore analyzed the section modulus, the feeding distance, and the probable last-solidifying regions before selecting any gating arrangement.
For a first-order thermal assessment, I used the modulus concept:
$$M = \frac{V}{A}$$
where \(M\) is the cooling modulus, \(V\) is the volume of the section, and \(A\) is the heat-dissipating surface area. For the heavy flange and hub regions, the modulus was significantly larger than that of the surrounding thin walls. This indicated that the last-solidifying zones would be located in the central mass and in the upper rim, which matched the defect locations observed in the earlier sand-cast trials. I also used the approximate solidification time relationship:
$$t_s = k M^n$$
where \(t_s\) is the local solidification time, \(k\) is a constant depending on the alloy and mold, and \(n\) is typically between 1.5 and 2 for cast iron. Because the lost foam process uses a gasifiable pattern and a dry sand mold under vacuum, the effective heat transfer is different from that of a green sand or resin sand mold. The pattern decomposition absorbs heat, the vacuum level changes sand compaction, and the coating permeability influences gas escape. All of these factors affect the local cooling rate and therefore the final shrinkage behavior in ductile iron casting.
I summarized the main hub characteristics in the following table to keep the design decisions traceable.
| Parameter | Value or requirement |
|---|---|
| Material | QT500-7 ductile iron |
| Heat treatment | Normalizing and tempering |
| Hardness requirement | 170–230 HB |
| Average wall thickness | Approximately 12 mm |
| Mass | Approximately 120 kg |
| Overall size | Approximately 556 mm diameter × 414.5 mm height |
| Internal quality | No shrinkage porosity or shrinkage cavity |
| Previous process | Sand casting with high scrap rate, low efficiency, and high cost |
| New process target | Lost foam casting with stable production and acceptable performance |
Because lost foam casting requires the unbonded sand to fill the pattern cavity and to support the pattern during pouring, the orientation of the pattern is not only a metallurgical decision but also a filling and compaction decision. I considered the need for sand to enter all recesses without bridging or shadowing. The large circular opening had to face upward so that the sand could flow freely into the internal passages and around the pattern. At the same time, the pouring position had to promote a stable liquid metal front, avoid severe pattern decomposition back-pressure, and place the last-solidifying regions where risers could feed them or where machining could remove them. I compared five possible pouring positions: top pouring, outer round inlet at the middle flange, inner round inlet at the middle, bottom pouring, and an internal-side structure arrangement. Each had a different influence on gas evolution, flow stability, cold shut risk, and shrinkage location.
For top pouring, the pattern mass was about 500 g because the foam was a copolymer. When the liquid iron contacted the top surface, the foam decomposed rapidly and released a large volume of gas. That gas tried to escape upward, which was opposite to the downward metal flow in top pouring. The result was severe back-pressure, spitting, and unstable filling. In addition, the top circular opening was about 443 mm in diameter, so the metal had to travel a long distance before the separate streams could meet. The meeting fronts could be too cold, producing cold shuts and an inconsistent filling pattern. I rejected top pouring as a primary production route.
For the outer round inlet at the middle flange, the external diameter was about 556 mm. The feeding points were far apart, and the metal had to travel a long distance around the circumference. In lost foam casting, the foam degradation products must be swept away continuously by the advancing metal front. If the front velocity is too low at the far side, the decomposition gases can become trapped, and the local temperature can drop below the safe filling limit. I predicted cold shuts, shrinkage porosity, and shrinkage cavities on the side opposite the inlet and near the upper part of the casting. The first trial later confirmed this prediction. For the inner round inlet at the middle, the inlet circle was only about 202 mm in diameter. The distance between inlet points was short, the flow was stable, and the middle and lower parts of the casting could be fed reliably. However, the uppermost region still tended to form shrinkage porosity because it was the last region to solidify and was far from the feeding path. This was a more favorable starting point because the top region could be fed with risers or removed by machining.
For bottom pouring, the filling itself was stable, but the upper part again became the last-solidifying region. In addition, the inner side of the bottom small circle contained a recess with a wall thickness of about 6 mm. That thin section reduced the local metal velocity and delayed the upward rise of the metal front, increasing the risk of shrinkage porosity and misrun in the body. I therefore did not select bottom pouring. For the internal-side structure arrangement, the geometry created sand-filling difficulties and feeding restrictions that were not compatible with a robust lost foam process. After this screening, I selected two candidate designs for trial: the outer round inlet at the middle flange and the inner round inlet at the middle.
I list the qualitative comparison of the candidate pouring positions in the table below.
| Pouring position | Main advantage | Main risk | Decision |
|---|---|---|---|
| Top pouring | Simple pattern orientation | Severe back-pressure, spitting, cold shut, unstable filling | Rejected |
| Outer round inlet at middle flange | Good access to heavy flange | Long flow distance, cold shut, shrinkage on opposite side and top | Trial candidate |
| Inner round inlet at middle | Short inlet distance, stable flow, good feeding of middle and bottom | Top region may shrink | Trial candidate |
| Bottom pouring | Stable bottom filling | Top shrinkage, thin recess reduces velocity | Rejected |
| Internal-side structure | Possible local feeding | Sand filling and feeding path problems | Rejected |
For the outer round inlet at the middle flange, I designed the sprue as a circular channel with a diameter of 50 mm. Its cross-sectional area was:
$$A_{\text{sprue}} = \frac{\pi d_{\text{sprue}}^2}{4} = \frac{\pi (50)^2}{4} \approx 1960 \text{ mm}^2$$
I considered the filter and the lower flow velocity required for ductile iron. The flow velocity was reduced to about 60% of the nominal value because of the filter and the viscous behavior of ductile iron. Using the modified choking relationship and the area ratio for a larger casting, I set the ratio of ingate area to sprue area to obtain a total ingate area of about 2450 mm². In other words:
$$A_{\text{ingate,total}} = 2450 \text{ mm}^2$$
The number of ingates had to be large enough to distribute the metal and avoid a single hot spot, but not so large that multiple streams would meet turbulently and create cold shuts or oxide films. Considering the circular geometry and the feeding distance, I chose four ingates arranged in a half-circle pattern. The area of each ingate was therefore:
$$A_{\text{ingate,each}} = \frac{2450}{4} \approx 613 \text{ mm}^2$$
However, in the actual design I used a practical value of about 700 mm² per ingate for the half-circle arrangement. I also placed ten risers on the top surface. Each riser had a rectangular section of 20 mm × 100 mm × 100 mm. These risers were intended to compensate for liquid shrinkage and to maintain a positive pressure gradient toward the last-solidifying regions. The complete gating arrangement for the outer round inlet is summarized in the table below.
| Element | Design value |
|---|---|
| Sprue diameter | 50 mm |
| Sprue area | Approximately 1960 mm² |
| Total ingate area | Approximately 2450 mm² |
| Number of ingates | 4 |
| Ingate arrangement | Half-circle distribution |
| Individual ingate area | Approximately 700 mm² |
| Top risers | 10 risers, each 20 mm × 100 mm × 100 mm |
For the inner round inlet at the middle, I used a total inlet area of about 2820 mm². I divided this into four equally spaced points on the inner circle. Each point therefore had an area of about:
$$A_{\text{ingate,each}} = \frac{2820}{4} = 705 \text{ mm}^2$$
Because the four-point inner-circle arrangement produced a short and stable flow path, the middle and lower regions were not expected to form shrinkage defects. The main risk remained at the top of the casting, where the metal streams converged. I therefore placed four risers at the top, equally spaced. The upper part of each riser had a diameter of 80 mm and a height of 120 mm. The contact between the riser and the pattern was kept small to facilitate cleaning, but the contact still had to provide sufficient feeding. I designed the contact face as a circular section with a diameter of 40 mm. The riser volume and contact area were approximately:
$$V_{\text{riser}} = \pi \left(\frac{80}{2}\right)^2 \times 120 \approx 6.03 \times 10^5 \text{ mm}^3$$
$$A_{\text{contact}} = \pi \left(\frac{40}{2}\right)^2 \approx 1.26 \times 10^3 \text{ mm}^2$$
The inner round inlet design is summarized below.
| Element | Design value |
|---|---|
| Total inlet area | Approximately 2820 mm² |
| Number of inlets | 4 |
| Arrangement | Equally spaced on inner circle |
| Individual inlet area | Approximately 705 mm² |
| Top risers | 4, equally spaced |
| Riser upper size | 80 mm diameter × 120 mm height |
| Riser contact diameter | 40 mm |
After the gating and riser designs were established, I moved to the pouring process parameters. For ductile iron casting, the pouring temperature must be high enough to avoid cold shuts and misruns but not so high that excessive shrinkage, gas defects, or burn-on occur. I set the initial pouring temperature window at 1450 to 1490 °C. The vacuum level in lost foam casting is another strong variable. A higher vacuum increases sand compaction and dimensional stability, but it also increases the risk of metal penetration and surface burn-on. A lower vacuum reduces burn-on but may allow the mold to move or the pattern to collapse. I therefore selected a vacuum range of -0.06 to -0.08 MPa. I also required film coverage and a holding time after pouring. The preliminary holding time was 8 to 10 min. The shortest holding time that prevents mold movement is preferred because it improves the production rhythm.
The nominal pouring parameters are listed in the table below.
| Parameter | Range or value |
|---|---|
| Pouring temperature | 1450–1490 °C |
| Vacuum level | -0.06 to -0.08 MPa |
| Film coverage | Required |
| Holding time after pouring | 8–10 min |
| Pouring sequence | To be optimized in trials |
I then conducted the first trial. I produced three castings with the outer round inlet and three castings with the inner round inlet. The pouring temperature was set between 1460 and 1500 °C, and the vacuum level was kept between -0.06 and -0.08 MPa. After film coverage, the vacuum was about -0.06 MPa. The pouring temperature was about 1500 °C. No mold expansion or box lifting was observed during pouring. After pouring, I held the vacuum for 8, 9, and 10 min, marking the castings as 1, 2, and 3, respectively. No obvious box lifting occurred after holding. After cleaning, the surfaces showed slight burn-on. Casting 1 showed slight lifting at the top, while castings 2 and 3 were normal. Both outer round inlet castings exhibited wrinkling on the ring opposite the inlet. The inner round inlet castings cleaned normally.
After sectioning, I found shrinkage porosity and shrinkage cavities in the outer round inlet castings at the inlet point and at the opposite half-circle region. The inner round four-point castings were normal. Based on these results, I eliminated the outer round inlet at the middle flange. I selected the inner round inlet at the middle as the production design. I set the holding time to 8 min and kept the vacuum at -0.06 to -0.08 MPa. The first trial results are summarized in the table below.
| Trial item | Outer round inlet | Inner round inlet |
|---|---|---|
| Number of castings | 3 | 3 |
| Pouring temperature | 1460–1500 °C | 1460–1500 °C |
| Vacuum | -0.06 to -0.08 MPa | -0.06 to -0.08 MPa |
| Holding time | 8, 9, 10 min | 8, 9, 10 min |
| Box lifting | Slight on casting 1 | None |
| Surface | Slight burn-on; wrinkling opposite inlet | Normal after cleaning |
| Sectioned defects | Shrinkage porosity and cavity at inlet and opposite half-circle | No shrinkage defects |
| Decision | Rejected | Selected |
The wrinkling and shrinkage in the outer round inlet trial were consistent with the thermal and flow analysis. The long circumferential flow path allowed the leading metal front to cool before it met the incoming stream on the opposite side. The foam decomposition products were not fully swept out, and the local gas pressure disturbed the metal front. The result was a cold fold or wrinkle on the surface and a lack of feeding below it. In ductile iron casting, such a cold front also reduces the ability of graphite expansion to compensate for shrinkage because the expansion cannot be transmitted through a partially solidified network. The inner round inlet avoided this problem because the four streams met over a short distance and the metal front remained hot and continuous.
I then conducted a second trial to solve the burn-on problem and to confirm the inner round design. I poured nine castings using the middle inner-round four-point inlet. The pouring temperature was 1460 to 1480 °C, the vacuum was -0.06 to -0.08 MPa, and the holding time was 8 min. The pouring time was 46 to 50 s. I used a slow-fast-slow pouring sequence. No box lifting occurred. After cleaning, there was no burn-on. After sectioning, no shrinkage porosity, shrinkage cavity, or misrun was found. I then moved to small-batch production. Cleaning, machining, and leak testing were all normal. The hardness was 170 to 200 HB, the elongation was 7% to 10%, and the spheroidization grade was 1 to 2. These results met the customer requirements.
The second trial and small-batch results are shown in the table below.
| Parameter or result | Value |
|---|---|
| Number of trial castings | 9 |
| Gating design | Middle inner-round four-point inlet |
| Pouring temperature | 1460–1480 °C |
| Vacuum | -0.06 to -0.08 MPa |
| Pouring time | 46–50 s |
| Pouring sequence | Slow-fast-slow |
| Holding time | 8 min |
| Box lifting | None |
| Burn-on | None after cleaning |
| Sectioned defects | No shrinkage porosity, cavity, or misrun |
| Small-batch machining | Normal |
| Leak testing | Normal |
| Hardness | 170–200 HB |
| Elongation | 7%–10% |
| Spheroidization grade | 1–2 |
The slow-fast-slow pouring sequence was important in the lost foam casting process. The initial slow phase allowed the first metal to contact the pattern gently and to establish a stable decomposition front without explosive gas evolution. The fast middle phase maintained a high metal velocity so that the foam decomposition products were swept toward the coating and the vacuum system instead of accumulating at the front. The final slow phase reduced turbulence at the top and allowed the risers to fill without excessive splashing. For ductile iron casting, this sequence also helps to preserve magnesium treatment effectiveness because excessive turbulence and contact with atmospheric oxygen can cause magnesium fade and dross formation.
I also controlled the chemistry to ensure the required hardness, elongation, and spheroidization grade. The base iron was treated with a spheroidizing agent, and the final composition was maintained within narrow windows. I used a graphitizing carburizer with low impurity content. The final composition control range is given in the table below.
| Element | Control range, wt.% |
|---|---|
| C | 3.4–3.8 |
| Si | 2.4–2.8 |
| Mn | 0.35–0.5 |
| P | <0.08 |
| S | <0.03 |
| Cu | 0.30–0.35 |
| Fe | Balance |
I used the carbon equivalent to guide the balance between fluidity, graphite expansion, and mechanical properties. For ductile iron, a common expression is:
$$CE = C + \frac{1}{3}Si + \frac{1}{4}P$$
With a carbon content of about 3.6% and a silicon content of about 2.6%, the carbon equivalent is approximately:
$$CE \approx 3.6 + \frac{2.6}{3} + \frac{0.05}{4} \approx 4.48$$
This value is close to the eutectic range, which promotes good fluidity and graphite expansion. The expansion can offset part of the liquid and solidification shrinkage, but it can also create expansion defects if the mold is too rigid or if the risers are not properly designed. In lost foam casting, the dry sand mold under vacuum has a moderate rigidity that can accommodate some expansion, but the coating and vacuum must be controlled to avoid mold wall movement. I therefore kept the vacuum in the range of -0.06 to -0.08 MPa and used a holding time of 8 min after pouring.
The hardness and elongation requirements were satisfied by the combination of chemistry, spheroidization, and heat treatment. The hardness is related to the matrix structure, the pearlite content, and the nodule count. A simplified expression for hardness is:
$$HB = \frac{F}{A_{\text{indentation}}}$$
where \(F\) is the applied load and \(A_{\text{indentation}}\) is the projected area of the indentation. In practice, I controlled the copper content to about 0.30 to 0.35% to stabilize pearlite and to help achieve the required hardness without excessive alloying. The manganese content was kept between 0.35 and 0.5% to avoid excessive segregation, while the phosphorus and sulfur were kept low to maintain ductility. The elongation was measured as:
$$\delta = \frac{L_f – L_0}{L_0} \times 100\%$$
where \(L_0\) is the original gauge length and \(L_f\) is the final gauge length after fracture. The results of 7% to 10% elongation met the QT500-7 requirement. The spheroidization grade was evaluated by the nodularity and the number of graphite nodules per unit area:
$$N = \frac{n}{A}$$
where \(n\) is the number of nodules and \(A\) is the observed area. The grade of 1 to 2 indicated a high degree of nodularity and a uniform distribution.
I also considered the thermal and filling conditions in more detail. In lost foam casting, the metal front must displace the decomposed pattern products. The gas pressure at the front can be expressed qualitatively as:
$$P_{\text{gas}} = P_{\text{decomposition}} + P_{\text{back-pressure}} – P_{\text{vacuum}}$$
If \(P_{\text{gas}}\) becomes too high, the metal front becomes unstable, and gas defects or cold shuts can form. The vacuum reduces the effective back-pressure and helps to draw the decomposition products through the coating. However, too much vacuum increases sand compaction and can cause metal penetration. The balance I selected was -0.06 to -0.08 MPa, which was low enough to reduce burn-on but high enough to keep the mold stable.
The filling velocity is also important. For a given pouring time, the average velocity in the sprue can be estimated from the volume flow rate:
$$v = \frac{Q}{A} = \frac{G}{\rho A t}$$
where \(Q\) is the volume flow rate, \(G\) is the mass of metal poured, \(\rho\) is the density, \(A\) is the cross-sectional area, and \(t\) is the pouring time. For a 120 kg casting and a pouring time of about 48 s, the average mass flow rate is:
$$\dot{m} = \frac{120}{48} = 2.5 \text{ kg/s}$$
Using a density of about \(7.1 \times 10^3 \text{ kg/m}^3\), the volume flow rate is:
$$Q = \frac{\dot{m}}{\rho} \approx \frac{2.5}{7.1 \times 10^3} \approx 3.52 \times 10^{-4} \text{ m}^3/\text{s}$$
This flow rate must be distributed among the four ingates. The average velocity in each ingate depends on its cross-sectional area. For an individual ingate area of about 705 mm², the average velocity is:
$$v_{\text{ingate}} = \frac{Q}{4 A_{\text{ingate}}} \approx \frac{3.52 \times 10^{-4}}{4 \times 705 \times 10^{-6}} \approx 0.125 \text{ m/s}$$
This value is only an average. During the fast middle phase, the local velocity is higher, which helps to sweep the foam decomposition products. During the initial and final slow phases, the velocity is lower, which reduces turbulence and gas entrapment. This is why the slow-fast-slow sequence worked well.
I also checked the riser contact area. The riser must remain liquid long enough to feed the casting. The feeding distance depends on the modulus of the riser and the modulus of the casting section. A common requirement is that the riser modulus be larger than the casting modulus by a factor that depends on the alloy and the feeding efficiency. For ductile iron, the graphite expansion reduces the required riser size, but the riser still must maintain a connection to the hot spot. The contact diameter of 40 mm provided a sufficient channel without creating a large heat sink that would be difficult to clean. The riser volume of approximately \(6.03 \times 10^5 \text{ mm}^3\) was enough to supply liquid metal during the final stages of solidification.
The gating ratio also influenced the filling behavior. In the selected inner round design, the total ingate area was about 2820 mm². The sprue area was about 1960 mm², so the ingate-to-sprue area ratio was:
$$\frac{A_{\text{ingate,total}}}{A_{\text{sprue}}} = \frac{2820}{1960} \approx 1.44$$
This ratio is slightly higher than the ratio used in the outer round trial, which helps to reduce the velocity in the gating system and to promote a smoother fill. A lower velocity reduces the risk of mold erosion and turbulence, while the large total ingate area still allows a short pouring time. The pouring time of 46 to 50 s was consistent with a controlled fill for a 120 kg casting in a lost foam mold.
I found that the inner round four-point inlet also improved the temperature distribution. The four streams met near the center and then moved outward and upward together. This created a more uniform thermal field in the middle and lower sections. The top risers were placed at the four convergence zones, so any last-solidifying liquid could be fed directly. In contrast, the outer round inlet created a long circumferential flow path that cooled the metal before it reached the far side. The surface wrinkling was a visible sign of a cold front, and the subsurface shrinkage was a consequence of poor feeding. By eliminating that design, I removed the main source of scrap.
I also monitored the pattern mass and the decomposition behavior. The pattern mass was about 500 g for the top-pouring trial. Although the selected inner round design used a different pattern configuration, the total pattern mass was similar. The amount of gas generated during decomposition is roughly proportional to the pattern mass. For a pattern mass \(m_p\), the gas volume at the decomposition temperature can be estimated from the ideal gas law:
$$V_g = \frac{m_p R T}{M_w P}$$
where \(R\) is the gas constant, \(T\) is the absolute temperature, \(M_w\) is the average molecular weight of the decomposition products, and \(P\) is the pressure. The vacuum reduces \(P\), which increases the volume that the gas occupies. However, the vacuum also removes the gas through the coating and sand. The balance between gas generation and gas removal determines whether the metal front remains stable. The inner round design produced a shorter and more direct path to the coating, so the gas could escape more easily. This was another reason for its success.
After the process was stabilized, I prepared a standard operating window for production. The key parameters are listed in the table below.
| Production parameter | Standard window |
|---|---|
| Gating design | Middle inner-round four-point inlet |
| Top risers | 4, equally spaced, 80 mm × 120 mm upper size, 40 mm contact |
| Pouring temperature | 1480–1500 °C |
| Vacuum level | -0.06 to -0.08 MPa |
| Film coverage | Required before pouring |
| Pouring time | 46–50 s |
| Pouring sequence | Slow-fast-slow |
| Holding time after pouring | 8 min |
| Spheroidizing time per ladle | 9–10 min |
| Carburizer type | Graphitizing carburizer with low impurities |
The final performance results are shown below.
| Property | Result | Requirement |
|---|---|---|
| Hardness | 170–200 HB | 170–230 HB |
| Elongation | 7%–10% | QT500-7 minimum |
| Spheroidization grade | 1–2 | Acceptable nodularity |
| Shrinkage porosity | Not found after sectioning | Not permitted |
| Shrinkage cavity | Not found after sectioning | Not permitted |
| Burn-on | None after cleaning | Acceptable surface |
| Machining | Normal | Normal |
| Leak testing | Normal | Normal |
I also compared the original sand casting route with the new lost foam route. The lost foam process reduced the need for cores, improved dimensional consistency, and allowed the pattern to be assembled before molding. However, it introduced new variables such as pattern decomposition, coating permeability, vacuum level, and sand compaction. By controlling these variables and by selecting the correct pouring position, I achieved a stable ductile iron casting process. The scrap rate decreased, the efficiency increased, and the cost per acceptable casting was reduced.
The main lessons from this project can be summarized as follows. First, the pouring position in lost foam casting must satisfy both sand-filling requirements and metallurgical feeding requirements. Second, for a heavy-section ductile iron casting, the last-solidifying regions must be either fed by risers or placed in a location that can be removed by machining. Third, the gating system should provide a short, stable flow path with a sufficient ingate area to avoid cold shuts and to sweep away decomposition products. Fourth, the pouring sequence and vacuum level must be balanced to avoid burn-on while maintaining mold stability. Fifth, chemistry control is essential for achieving the required hardness, elongation, and spheroidization grade.
In conclusion, I successfully developed a lost foam casting process for the QT500-7 ductile iron hub. The selected design used a middle inner-round four-point inlet and four top risers. The pouring temperature was 1480 to 1500 °C, the vacuum was -0.06 to -0.08 MPa, the pouring time was 46 to 50 s, and the holding time was 8 min. The spheroidizing time per ladle was 9 to 10 min. The resulting hardness was 170 to 200 HB, the elongation was 7% to 10%, and the spheroidization grade was 1 to 2. Sectioning and small-batch production confirmed that the casting was free from shrinkage porosity and shrinkage cavities. The process met the customer requirements and provided a reliable production route for this heavy-section ductile iron casting.
For further optimization, I would continue to monitor the coating thickness, the sand compaction, and the vacuum decay curve. These variables influence the local cooling rate and the gas removal rate. A small change in coating permeability can shift the balance between burn-on and gas entrapment. I would also use thermal simulation to refine the riser size and to verify the feeding distance for future variants of the hub. The same design logic can be applied to other heavy-section ductile iron castings produced by lost foam, especially those with large flanges, internal passages, or thick-to-thin transitions. By combining thermal analysis, gating design, riser design, and chemistry control, it is possible to produce sound ductile iron castings with consistent mechanical properties.