In our foundry, the production of ductile iron castings by the lost foam casting process has become an important route for components that require both dimensional accuracy and reliable internal soundness. Among these components, the A16 wheel hub is a representative part used in the transmission systems of agricultural and industrial machinery, especially in tracked combined wheels. The service conditions of this part demand high internal integrity and a clean external surface. The casting weighs approximately 64 kg and is made from QT450-10 ductile iron. Its maximum outer profile ranges from about 322 mm to 285 mm. The geometry is divided into an outer running section and an inner connecting section, which are joined by relatively thin ribs. This mixed wall-thickness design creates a difficult solidification environment: the ribs and connecting webs are thin, while the top and side sections are comparatively thick. In lost foam casting of ductile iron castings, such geometry often magnifies the conflict between feeding, cooling rate, and graphite morphology control.
During a six-month statistical period, we produced 2,100 wheel hubs. We inspected them from three main perspectives: dimensional accuracy, surface quality, and metallographic structure. The dimensional results were acceptable. However, the rejection rate caused by shrinkage cavities and shrinkage porosity reached 7.95%. In addition, the rejection rate caused by unacceptable metallographic structure reached 5.38%. The metallographic failures were mainly associated with poor nodularization caused by graphite flotation. Because the total rejection rate was too high for stable production, I focused our technical team on these two defect categories: shrinkage and graphite flotation in lost foam ductile iron castings.
1. Product and Production Baseline
The A16 wheel hub is produced with a top gating system in the lost foam process. The molten iron is melted in a medium-frequency induction furnace. The baseline process control parameters were as follows: tapping temperature 1,520–1,540 °C, pouring temperature 1,440–1,480 °C, pouring vacuum −0.065 to −0.075 MPa, and holding time after pouring 4–5 min. The target chemical composition of QT450-10 is summarized in Table 1. The alloying and treatment windows are relatively narrow because ductile iron castings require both adequate nodularization and controlled solidification contraction.
| Parameter | Target or Baseline Value | Remarks |
|---|---|---|
| Casting mass | 64 kg | A16 wheel hub |
| Material grade | QT450-10 | Ductile iron |
| Maximum profile | ϕ322–ϕ285 mm | Outer running section |
| Wall thickness distribution | Thin ribs and webs; thick top and side sections | Mixed wall thickness |
| Melting furnace | Medium-frequency induction furnace | Batch melting |
| Gating system | Top gating | Lost foam casting |
| Tapping temperature | 1,520–1,540 °C | Baseline |
| Pouring temperature | 1,440–1,480 °C | Baseline |
| Pouring vacuum | −0.065 to −0.075 MPa | Baseline |
| Holding time | 4–5 min | After pouring |
| Statistical production | 2,100 castings | Six-month period |
| Dimensional result | Acceptable | No major dimensional rejection |
| Shrinkage rejection | 7.95% | Shrinkage cavity and porosity |
| Metallographic rejection | 5.38% | Mainly graphite flotation |
Table 2 lists the chemical composition specification that we used for the A16 wheel hub. The carbon and silicon levels are particularly important because they determine the carbon equivalent, the graphite expansion behavior, and the tendency for graphite flotation in heavy sections. The magnesium and rare earth contents are also critical because they control nodularization, but excessive residual magnesium can increase shrinkage tendency in ductile iron castings.
| Element | Specified Range, wt.% | Function in Ductile Iron Castings |
|---|---|---|
| C | 3.65–3.85 | Graphite formation, carbon equivalent |
| Si | 2.65–2.85 | Graphitization, matrix control |
| Mn | 0.2–0.6 | Hardenability, matrix strength |
| P | ≤ 0.07 | Impurity, embrittlement risk |
| S | ≤ 0.02 | Nodulizer consumption |
| Mg | 0.03–0.065 | Nodularization, residual magnesium |
| RE | 0.02–0.04 | Nodularization, impurity control |
2. Defect Statistics and Evaluation Method
I established a simple and transparent method for evaluating the quality of the ductile iron castings. The rejection rate for each defect type was calculated as the number of rejected castings divided by the total number of castings inspected. This allowed us to compare the effects of process changes over time. The general formula is shown below:
$$ R_d = \frac{N_d}{N_t} \times 100\% $$
where \(R_d\) is the rejection rate for defect type \(d\), \(N_d\) is the number of castings rejected because of that defect, and \(N_t\) is the total number of castings inspected. For the A16 wheel hub, the two main defect categories were shrinkage and metallographic failure due to graphite flotation. The initial monthly data showed that both defects were persistent rather than random. This indicated that the causes were related to the process design and chemistry control rather than isolated operational mistakes.
Table 3 gives the initial monthly rejection rates before the systematic improvement. The numbers show that shrinkage was already above 7.5% in every month, and metallographic rejection related to poor nodularization and graphite flotation was between 4% and 6.5%. Such a level of rejection is not acceptable for stable production of ductile iron castings, especially for a safety-related transmission component.
| Month | Shrinkage Rejection, % | Poor Nodularization Rejection, % | Main Observation |
|---|---|---|---|
| Month 1 | 8.27 | 4.00 | Shrinkage dominant |
| Month 2 | 8.00 | 5.64 | Graphite flotation increasing |
| Month 3 | 7.67 | 5.86 | Mixed defects |
| Month 4 | 8.93 | 6.48 | Both defects high |
3. Root Cause Analysis of Shrinkage in Ductile Iron Castings
In the trial production stage, we used the plunging method for nodularization treatment. In this method, the nodulizer is added into the molten iron and the recovery rate can fluctuate. To avoid the risk of poor nodularization, our operators intentionally increased the nodulizer addition. As a result, the residual magnesium content in the liquid iron was often controlled at a relatively high level of 0.045%–0.065%. This high residual magnesium level was the main reason for the high shrinkage rejection rate. The typical shrinkage defect appeared as internal cavities or dispersed porosity after machining and sectioning.
To confirm the relationship, we produced two heats of A16 wheel hubs under otherwise similar conditions. Each heat contained four ladles, and each ladle contained five castings. Thus, each heat represented 20 castings. All 40 castings were surface machined, and selected castings were sectioned. The results are summarized in Table 4. In the first heat, the nodulizer addition was 1.76%, and the residual magnesium was approximately 0.056%–0.060%. The shrinkage and porosity rejection rate reached 8%. In the second heat, the nodulizer addition was reduced to 1.18%, and the residual magnesium was approximately 0.039%–0.041%. The shrinkage rejection rate decreased to 3%. This comparison provided direct evidence that excessive nodulizer addition and high residual magnesium increase the shrinkage tendency of ductile iron castings.
| Heat | C, wt.% | Si, wt.% | Mn, wt.% | P, wt.% | S, wt.% | Mg, wt.% | RE, wt.% | Nodulizer Addition, % | Shrinkage Rejection, % |
|---|---|---|---|---|---|---|---|---|---|
| Heat 1, Ladle 1 | 3.74 | 2.85 | 0.45 | 0.035 | 0.018 | 0.060 | 0.030 | 1.76 | 8 |
| Heat 1, Ladle 2 | 3.76 | 2.80 | 0.46 | 0.036 | 0.016 | 0.058 | 0.027 | 1.76 | 8 |
| Heat 1, Ladle 3 | 3.76 | 2.79 | 0.48 | 0.040 | 0.016 | 0.058 | 0.024 | 1.76 | 8 |
| Heat 1, Ladle 4 | 3.70 | 2.79 | 0.48 | 0.040 | 0.016 | 0.056 | 0.025 | 1.76 | 8 |
| Heat 2, Ladle 1 | 3.69 | 2.66 | 0.45 | 0.032 | 0.015 | 0.041 | 0.021 | 1.18 | 3 |
| Heat 2, Ladle 2 | 3.72 | 2.66 | 0.46 | 0.035 | 0.018 | 0.041 | 0.019 | 1.18 | 3 |
| Heat 2, Ladle 3 | 3.67 | 2.63 | 0.46 | 0.036 | 0.017 | 0.039 | 0.021 | 1.18 | 3 |
| Heat 2, Ladle 4 | 3.66 | 2.63 | 0.46 | 0.033 | 0.018 | 0.039 | 0.020 | 1.18 | 3 |
The mechanism can be explained by the effect of magnesium on graphite nucleation and eutectic solidification. After nodularization treatment, elements that can act as graphite nucleation substrates, such as sulfur, oxygen, hydrogen, and nitrogen, combine with nodularizing elements and lose their ability to promote graphite nucleation. This reduces the graphitization tendency and increases the undercooling of the eutectic reaction. Carbide eutectic is more likely to form. Because carbide eutectic has a larger contraction than graphite eutectic, increasing the amount of carbide eutectic raises the solidification contraction of ductile iron castings.
In addition, when the nodulizer addition increases, the residual magnesium content rises. The resulting carbide eutectic makes the shrinkage cavity problem more pronounced. High residual magnesium also promotes the precipitation of primary austenite at a higher temperature. The primary austenite then has a larger temperature range in which to develop into coarse dendrites. These coarse dendrites obstruct the flow of external liquid iron and reduce the ability to feed volumetric contraction. Magnesium-containing inclusions can also accumulate and create favorable conditions for shrinkage cavity and porosity formation. Therefore, the shrinkage problem in our lost foam ductile iron castings was not only a feeding issue but also a chemistry and nucleation issue.
For a more quantitative view, we can express the carbon equivalent as follows:
$$ CE = C + \frac{Si + P}{3} $$
where \(C\), \(Si\), and \(P\) are the mass fractions in percent. For ductile iron castings, the eutectic carbon equivalent is often estimated by an expression such as:
$$ CE_e = 4.26 – 0.30 Si – 0.03 Mn $$
These formulas are simplified, but they help explain why small changes in carbon and silicon can shift the solidification path. When the carbon equivalent is too high, graphite flotation becomes more likely. When the residual magnesium is too high, shrinkage becomes more likely. The challenge is to balance these two opposing risks in the same casting.
We can also express the shrinkage tendency qualitatively as a function:
$$ S_t = f(Mg_{res}, CE, CR, T_p, G_f) $$
where \(S_t\) is shrinkage tendency, \(Mg_{res}\) is residual magnesium, \(CE\) is carbon equivalent, \(CR\) is cooling rate, \(T_p\) is pouring temperature, and \(G_f\) is graphite flotation tendency. In our process, increasing \(Mg_{res}\) and decreasing \(CR\) both raised \(S_t\). Increasing \(CE\) raised \(G_f\). Thus, the two defects were linked through the same set of process variables.
4. Root Cause Analysis of Graphite Flotation in Ductile Iron Castings
Graphite flotation is a defect specific to ductile iron castings. It usually occurs in thick-walled castings and is essentially a density segregation phenomenon. The upper surface of the casting accumulates a large amount of graphite. Graphite explosion, or exploded graphite, is often observed in slow-cooling regions of thick sections and in graphite flotation zones. Under microscopic observation, exploded graphite has an outer contour close to a circle, as if the graphite had burst from the center. On the casting surface or on a machined surface, dark gray graphite-rich zones may be visible to the naked eye. The mechanical properties of castings with such dark gray spots are significantly reduced.
The flotation of graphite can be understood by comparing densities. The crystal density of graphite is about 2.2 g/cm³, while the density of liquid iron is about 7 g/cm³. When the carbon equivalent is too high, the density difference causes spherical graphite in the liquid iron to float upward. This leads to an uneven distribution of graphite in the solidified body and forms graphite flotation. Silicon is a graphitizing element. Silicon-containing compounds exist in the core of spheroidal graphite and form part of the nucleation substrate. They promote graphite nucleation and growth, thereby intensifying graphite flotation.
A simplified Stokes-type expression for the upward velocity of a graphite particle can be written as:
$$ v_g = \frac{2 r_g^2 (\rho_l – \rho_g) g}{9 \eta} $$
where \(v_g\) is the flotation velocity, \(r_g\) is the effective graphite particle radius, \(\rho_l\) is the density of liquid iron, \(\rho_g\) is the density of graphite, \(g\) is gravitational acceleration, and \(\eta\) is the apparent viscosity of the liquid. This equation shows that larger graphite particles and lower liquid viscosity increase flotation velocity. Higher pouring temperature lowers viscosity and extends the time during which graphite can float. Thick sections also extend the local solidification time, which increases the total flotation distance.
For the A16 wheel hub, we identified several specific factors that contributed to graphite flotation:
- Carbon equivalent: In the early production stage, the carbon equivalent was controlled at 4.65%–4.80%. This high value increased the amount of graphite and raised the flotation tendency in ductile iron castings.
- Wall thickness: The top and side sections of the wheel hub are thick. After pouring, the liquid iron in these regions remains molten for a long time. Under such slow cooling, graphite has more time to float.
- Charge composition and charging method: Different charge ratios have a strong influence on graphite flotation. A high proportion of pig iron in the charge tends to promote flotation. The timing and method of carburization, inoculation, and nodularization also affect graphite flotation.
- Pouring temperature: An excessively high pouring temperature extends the time that the iron remains liquid and reduces the resistance to graphite movement. This significantly increases the flotation tendency.
- Lost foam process characteristics: The foam pattern is made of STMMA copolymer, which contains about 69.6% carbon. When the pattern contacts the liquid iron, it decomposes and produces a large amount of free carbon. This carbon is absorbed by the liquid iron, further increasing the local carbon concentration and aggravating graphite flotation. This effect is stronger than in conventional sand casting.
- Other operational factors: Sand temperature, vacuum level, holding time, and the waiting time from tapping to pouring all influence cooling rate, nodularization fading, and inoculation fading. If these factors are not stable, graphite flotation and shrinkage can both become worse.
In our early practice, we used the plunging method for nodularization. The ladle bottom was flat, and the nodulizer and inoculant were lost quickly, resulting in low absorption. The A16 wheel hub has thick top and side sections, so the liquid iron cools slowly after filling. To improve fluidity and reduce shrinkage, we initially controlled the carbon equivalent at 4.65%–4.80%. This high carbon equivalent, combined with the thick sections and the carbon pickup from the lost foam pattern, made graphite flotation a serious problem in the ductile iron castings.
Table 5 summarizes the main factors and their effects on graphite flotation and shrinkage. The table shows that many variables act in opposite directions. For example, increasing carbon equivalent can improve fluidity and reduce shrinkage risk, but it increases graphite flotation. Increasing residual magnesium can improve nodularization, but it increases shrinkage. Therefore, optimization must be based on a balanced window rather than a single-variable adjustment.
| Factor | Effect on Graphite Flotation | Effect on Shrinkage | Required Direction |
|---|---|---|---|
| Carbon equivalent | Higher CE increases flotation | Higher CE may improve feeding | Control to a moderate window |
| Silicon content | Higher Si promotes graphitization and flotation | Higher Si may affect matrix and contraction | Limit final Si |
| Residual magnesium | Indirect effect through nodularization and carbides | Higher Mg increases shrinkage | Reduce to an adequate minimum |
| Pouring temperature | Higher temperature increases flotation | Higher temperature may improve fluidity | Lower within a safe range |
| Wall thickness | Thicker sections increase flotation | Thicker sections require more feeding | Use process control and feeding design |
| Lost foam pattern carbon | Free carbon pickup increases flotation | May alter local carbon and contraction | Control pattern density and decomposition |
| Carburizer addition | Late addition increases flotation | Poor absorption increases variability | Add in batches before melting |
| Vacuum and sand temperature | Affect cooling rate and flotation time | Affect solidification and feeding | Stabilize and monitor |
5. Systematic Corrective Actions
Based on the root cause analysis, I organized a systematic improvement plan. The plan addressed chemistry, pouring temperature, charging practice, process control, and lost foam specific factors. The goal was to reduce residual magnesium and carbon equivalent while maintaining nodularization and castability in the ductile iron castings.
5.1 Chemistry Adjustment
We reduced the nodulizer addition while ensuring that the nodularization grade remained acceptable. The residual magnesium content was controlled to 0.035%–0.060%. At the same time, the carbon equivalent was controlled to 4.5%–4.7%. The base iron silicon content was controlled to 1.5%–1.7%. The final carbon content was kept at \(\omega(C) \le 3.8\%\), and the final silicon content was kept at \(\omega(Si) \le 2.7\%\). These changes reduced the amount of excess graphite and lowered the flotation tendency without creating unacceptable carbide formation or poor nodularization.
| Parameter | Before Improvement | After Improvement | Reason |
|---|---|---|---|
| Residual Mg, % | 0.045–0.065 | 0.035–0.060 | Reduce shrinkage tendency |
| Carbon equivalent, % | 4.65–4.80 | 4.50–4.70 | Reduce graphite flotation |
| Base Si, % | Higher and variable | 1.5–1.7 | Control graphitization |
| Final C, % | Up to 3.85 | ≤ 3.8 | Lower excess carbon |
| Final Si, % | Up to 2.85 | ≤ 2.7 | Reduce flotation and matrix variation |
| Nodulizer addition, % | 1.76 in trial | About 1.18 or optimized | Lower residual Mg |
5.2 Pouring Temperature and Ladle Practice
We preheated the ladle and ladle cover by baking and by passing molten iron through them before the treatment. The pouring temperature was lowered to 1,440–1,460 °C. This reduced the burning loss of the nodulizer and inoculant and shortened the time during which graphite could float. Before tapping, we ensured that the liquid iron was held at 1,520–1,540 °C for about 10 minutes. The furnace temperature was then lowered before tapping. This practice improved the stability of the treatment and reduced the superheat at pouring.
The effect of pouring temperature on solidification time can be approximated by a cooling-time relationship. A simplified form is:
$$ t_s = B \left(\frac{V}{A}\right)^n $$
where \(t_s\) is solidification time, \(V\) is volume, \(A\) is surface area, \(B\) is a mold constant, and \(n\) is an exponent usually close to 2. This relationship shows that thick sections with a high \(V/A\) ratio solidify slowly. Lowering the pouring temperature reduces the initial superheat and shortens the time available for graphite flotation. For the A16 wheel hub, this was especially important in the thick top and side sections.
5.3 Charge Quality and Charging Sequence
We changed the carburizer addition practice. Instead of adding carburizer directly into the molten iron in the furnace, we added it in batches before the charge was fully melted. This improved the melting and absorption efficiency of the carburizer. Late addition of carburizer was found to increase the risk of graphite flotation because it created local high-carbon regions that were not fully homogenized. By adding carburizer early and in batches, we achieved a more uniform carbon distribution and reduced the flotation tendency in the ductile iron castings.
5.4 Process Control and Operator Training
We improved process control through theoretical and practical training. The operators were trained to shorten the time between nodularization and pouring. Reducing this waiting time decreased nodularization fading and inoculation fading. We also standardized the weighing of the foam pattern every day to ensure pattern density stability. This reduced gas evolution during pouring and helped control the carbon pickup from the foam pattern. We measured temperature regularly and cleaned the vacuum pipeline. Sand temperature, vacuum level, and holding time were strictly controlled to reduce uncontrolled variation in solidification.
| Control Item | Before Improvement | After Improvement | Expected Effect |
|---|---|---|---|
| Ladle preheating | Inconsistent | Baked and preheated with iron | Lower treatment loss |
| Pouring temperature | 1,440–1,480 °C | 1,440–1,460 °C | Less flotation time |
| Holding before tapping | Variable | 10 min at 1,520–1,540 °C | Stable dissolution |
| Carburizer addition | Sometimes late | Batched before melting | Better absorption |
| Nodularization-to-pouring time | Long and variable | Shortened and standardized | Less fading |
| Foam pattern weight | Not checked daily | Checked daily | Stable density and gas |
| Sand temperature | Fluctuating | Controlled | Stable cooling |
| Vacuum pipeline | Cleaned irregularly | Cleaned regularly | Stable vacuum |
| Holding time | 4–5 min, variable | Controlled within window | Consistent solidification |
6. Quantitative Process Windows and Formulas
To make the improvement reproducible, I summarized the key process windows with formulas and tables. The carbon equivalent was used as a central control parameter. For the A16 wheel hub, the target window was:
$$ CE = C + \frac{Si + P}{3} = 4.50\% \sim 4.70\% $$
The residual magnesium window was:
$$ Mg_{res} = 0.035\% \sim 0.060\% $$
The final carbon and silicon limits were:
$$ \omega(C) \le 3.8\% $$
$$ \omega(Si) \le 2.7\% $$
The pouring temperature window was:
$$ T_p = 1,440 \sim 1,460 \, ^\circ C $$
The vacuum window was:
$$ P_v = -0.065 \sim -0.075 \, MPa $$
The holding time after pouring was:
$$ t_h = 4 \sim 5 \, min $$
These windows were not independent. For example, if the carbon equivalent was close to the upper limit, the pouring temperature had to be kept lower to reduce flotation. If the residual magnesium was close to the lower limit, the nodulizer addition had to be controlled carefully to avoid poor nodularization. Table 6 presents the optimized process window as a control matrix.
| Control Variable | Symbol | Optimized Window | Primary Defect Controlled |
|---|---|---|---|
| Carbon equivalent | \(CE\) | 4.50–4.70% | Graphite flotation |
| Residual magnesium | \(Mg_{res}\) | 0.035–0.060% | Shrinkage |
| Final carbon | \(\omega(C)\) | ≤ 3.8% | Graphite flotation |
| Final silicon | \(\omega(Si)\) | ≤ 2.7% | Graphite flotation |
| Base silicon | \(Si_0\) | 1.5–1.7% | Graphitization control |
| Pouring temperature | \(T_p\) | 1,440–1,460 °C | Flotation and shrinkage |
| Tapping temperature | \(T_t\) | 1,520–1,540 °C | Treatment stability |
| Pouring vacuum | \(P_v\) | −0.065 to −0.075 MPa | Mold filling and cooling |
| Holding time | \(t_h\) | 4–5 min | Solidification control |
7. Results After Improvement
After implementing the improvements, we tracked the rejection rates from August to December. The results showed a clear and sustained reduction in both poor nodularization and shrinkage defects. Table 7 summarizes the monthly rejection rates before and after the improvement. The poor nodularization rejection rate decreased from 6.48% in the worst pre-improvement month to 2.00% in August, then to 1.25% in September, 0.35% in October, and zero in November and December. The shrinkage rejection rate decreased from a high of 8.93% to 4.46% in August, then to 3.29% in September, 2.88% in October, 2.30% in November, and 2.48% in December.
| Month | Poor Nodularization Rejection, % | Shrinkage Rejection, % | Status |
|---|---|---|---|
| March | 4.00 | 8.27 | Before improvement |
| April | 5.64 | 8.00 | Before improvement |
| May | 5.86 | 7.67 | Before improvement |
| June | 6.48 | 8.93 | Before improvement |
| August | 2.00 | 4.46 | Improvement implemented |
| September | 1.25 | 3.29 | Improvement continuing |
| October | 0.35 | 2.88 | Stable improvement |
| November | 0 | 2.30 | Best result |
| December | 0 | 2.48 | Stable production |
The trend in Table 7 can also be expressed as a relative reduction. The average poor nodularization rejection before improvement was approximately 5.50%, and after improvement it fell below 1.00%. The average shrinkage rejection before improvement was approximately 8.22%, and after improvement it fell to approximately 3.08%. This demonstrated that the systematic changes were effective for both defect types in the ductile iron castings.
We can calculate the relative reduction as follows:
$$ \Delta R = \frac{R_{before} – R_{after}}{R_{before}} \times 100\% $$
For poor nodularization, using the initial and final stable values:
$$ \Delta R_{nod} = \frac{5.50 – 0.70}{5.50} \times 100\% \approx 87.3\% $$
For shrinkage, using the initial and final stable values:
$$ \Delta R_{shr} = \frac{8.22 – 3.08}{8.22} \times 100\% \approx 62.5\% $$
These calculations show that the improvement was not marginal. It changed the process from a high-rejection state to a controlled state. The remaining shrinkage rejection of about 2.3%–2.5% was still significant, but it was much more stable and manageable. The complete elimination of poor nodularization rejection in the last two months was particularly encouraging because it confirmed that the nodulizer addition and residual magnesium window were properly balanced.
8. Discussion
The results of this work show that shrinkage and graphite flotation in lost foam ductile iron castings are strongly coupled. In our process, the attempt to avoid poor nodularization by adding more nodulizer led to high residual magnesium. The high residual magnesium increased the solidification contraction and promoted coarse dendritic growth, which in turn increased shrinkage. At the same time, the high carbon equivalent used to improve fluidity and reduce shrinkage increased the graphite flotation tendency. The lost foam process made the situation more complex because the foam pattern decomposed and released free carbon into the liquid iron. The thick top and side sections of the A16 wheel hub provided a long solidification time, which allowed graphite to float.
The corrective actions worked because they addressed the root causes rather than the symptoms. Lowering the nodulizer addition reduced residual magnesium and shrinkage. Lowering the carbon equivalent reduced graphite flotation. Lowering the pouring temperature reduced the time available for flotation and decreased superheat. Improving carburizer addition ensured a more uniform carbon distribution. Strengthening process control reduced variation in sand temperature, vacuum, holding time, and treatment time. These changes were mutually reinforcing.
It is important to note that the optimized window for ductile iron castings must be adjusted for each product. A thin-walled casting may require a higher carbon equivalent or a higher pouring temperature to avoid cold shuts and misruns. A thick-walled casting may require a lower carbon equivalent and a lower pouring temperature to avoid graphite flotation. The A16 wheel hub is a mixed-wall-thickness casting, so the window had to be selected to satisfy both the thin ribs and the thick top and side sections. In our case, the final window was \(CE = 4.50\%–4.70\%\), \(Mg_{res} = 0.035\%–0.060\%\), and \(T_p = 1,440–1,460 °C\).
Another important lesson is that the lost foam process introduces additional carbon. The STMMA copolymer pattern contains a high carbon content, and its decomposition products can be absorbed by the liquid iron. This means that the carbon equivalent measured before pouring may not fully represent the local carbon content at the solidification front. For ductile iron castings produced by lost foam, the carbon equivalent should be controlled slightly lower than in conventional sand casting when thick sections are present. In our production, this adjustment was essential for suppressing graphite flotation.
Process control also played a major role. The best chemistry window will not produce consistent results if the treatment time, pouring time, sand temperature, and vacuum level vary from ladle to ladle. By training operators and standardizing these variables, we reduced the scatter in the results. The monthly rejection data show that the improvement was sustained over several months, which indicates that the changes were embedded in the process rather than depending on a single trial.
9. Practical Rules for Lost Foam Ductile Iron Castings
From this work, I extracted several practical rules that can be applied to similar ductile iron castings produced by lost foam:
- Control residual magnesium to the lowest level that still provides acceptable nodularization. Excess magnesium increases shrinkage.
- Control carbon equivalent to a moderate window. High carbon equivalent increases graphite flotation, especially in thick sections.
- Lower the pouring temperature within the range required for mold filling. Lower superheat reduces flotation time.
- Add carburizer in batches before melting. Avoid late addition into liquid iron.
- Preheat the ladle and cover. Reduce nodulizer and inoculant loss.
- Shorten the time between nodularization and pouring. Reduce fading.
- Monitor foam pattern density and weight. Reduce gas evolution and carbon pickup.
- Control sand temperature, vacuum, and holding time. Keep solidification conditions stable.
- Use sectioning and machining checks to verify internal soundness, not only surface appearance.
- Track rejection rate by defect type every month. Use the data to confirm that process changes are effective.
Table 8 presents a summary of the defect mechanisms, corrective actions, and observed results. This table can be used as a quick reference for process engineers working with ductile iron castings in lost foam casting.
| Defect | Main Mechanism | Key Variables | Corrective Action | Observed Result |
|---|---|---|---|---|
| Shrinkage cavity and porosity | High residual Mg increases carbide eutectic and coarse dendrites, blocking feeding | Nodulizer addition, \(Mg_{res}\), cooling rate, dendrite growth | Reduce nodulizer, control \(Mg_{res}\) to 0.035–0.060%, shorten treatment time | Shrinkage rejection decreased from about 8% to about 2.3–2.5% |
| Graphite flotation | Density segregation of graphite in high-CE, slow-cooling thick sections | \(CE\), Si, pouring temperature, wall thickness, foam carbon pickup | Lower \(CE\) to 4.50–4.70%, lower pouring temperature to 1,440–1,460 °C, control carburizer | Poor nodularization rejection decreased from about 5–6% to 0% in the final months |
| Poor nodularization | Insufficient Mg or fading after treatment | Nodulizer addition, treatment time, pouring time | Maintain adequate \(Mg_{res}\), shorten nodularization-to-pouring time | Stable nodularization with lower nodulizer addition |
| Process variation | Uncontrolled sand temperature, vacuum, holding time, pattern density | Sand temperature, vacuum, holding time, pattern weight | Daily checks, regular cleaning, standardized holding time | More consistent monthly results |
10. Conclusion
In our production of A16 wheel hubs by lost foam casting, the main quality problems were shrinkage cavities and graphite flotation in ductile iron castings. The shrinkage problem was primarily caused by excessive nodulizer addition, which led to high residual magnesium. The high residual magnesium increased the solidification contraction tendency, promoted carbide eutectic, and encouraged coarse dendritic growth that hindered feeding. The graphite flotation problem was caused by a combination of high carbon equivalent, slow cooling in thick sections, and carbon pickup from the lost foam pattern. The two defects were linked through the chemistry and cooling conditions of the process.
By systematically optimizing the process, we achieved a significant reduction in both defects. The key measures were reducing residual magnesium to 0.035%–0.060%, controlling carbon equivalent to 4.50%–4.70%, limiting final carbon to \(\omega(C) \le 3.8\%\) and final silicon to \(\omega(Si) \le 2.7\%\), lowering the pouring temperature to 1,440–1,460 °C, improving carburizer addition, preheating the ladle, shortening the treatment-to-pouring time, and strengthening routine process control. After these changes, the poor nodularization rejection rate fell to zero in the last two months of the study, and the shrinkage rejection rate fell from about 8% to about 2.3%–2.5%.
The experience from this work confirms that quality control in lost foam ductile iron castings must be based on a balanced process window. It is not enough to adjust one variable in isolation. Chemistry, pouring temperature, charging practice, pattern quality, vacuum, sand temperature, and holding time all interact. When these variables are controlled together, the result is a stable process with fewer shrinkage defects, less graphite flotation, and more reliable mechanical properties. This approach can be applied to other thick-section ductile iron castings produced by lost foam casting, especially those with mixed wall thickness and high internal quality requirements.
In future production, I will continue to monitor the residual magnesium, carbon equivalent, pouring temperature, and rejection rate by defect type. The monthly data will be used to refine the process window and to ensure that the improvements remain stable over longer production periods. For ductile iron castings, the most important principle is to maintain a narrow and well-understood control window rather than relying on excessive additions of nodulizer or carbon equivalent. That principle has proven effective in our foundry and has directly improved the quality of the A16 wheel hub.