In the production of heavy-duty truck axle components, the main reducing shell cover is a critical structural part that must be assembled together with the main reducing shell. Over the years, I have been directly involved in optimizing the manufacturing route for this component using ductile iron casting. The casting we produce has an overall envelope of 554 mm × 350 mm × 138 mm, with a nominal wall thickness of 7 mm and a locally thickened section of 37 mm. The component weight is 20 kg, and the material specification is QT400-15, which corresponds to a ferritic ductile iron casting with a minimum tensile strength of 400 MPa and an elongation of 15%.
Our production environment consists of a horizontal molding line with green sand molds, hot-box cores for internal cavities, and medium-frequency induction furnaces for melting. The customer requirement is extremely strict: no shrinkage cavity, no shrinkage porosity, and no other volumetric defects are allowed on machined surfaces or in critical internal areas. During the initial phases of process development, we encountered frequent shrinkage defects, especially in a thick isolated section around a 125 mm diameter circular boss. This article summarizes the investigation, the root-cause analysis, and the countermeasures that eventually allowed us to produce sound ductile iron casting components on a mass-production basis.
1. Component Characteristics and Process Requirements
Before designing the gating and feeding system, we carefully analyzed the geometrical features of the main reducing shell cover. The casting is essentially a shallow lid with a large planar flange, a central large-diameter circular area, and several bosses and ribs. The ratio of the largest dimension (554 mm) to the nominal wall thickness (7 mm) is almost 80:1, which makes the casting relatively rigid but also prone to thermal-center formation in the local thick sections.
Table 1 summarizes the principal casting parameters that influenced our ductile iron casting process design.
| Parameter | Value | Remarks |
|---|---|---|
| Overall size | 554 mm × 350 mm × 138 mm | Flange to upper boss |
| Nominal wall thickness | 7 mm | Main body |
| Maximum local thickness | 37 mm | Boss and pad areas |
| Casting weight | 20 kg | Finished part weight |
| Material grade | QT400-15 | Ferritic ductile iron |
| Molding process | Horizontal green sand | Flaskless line |
| Core process | Hot-box | For central recess and bolt holes |
| Melting | Medium-frequency induction furnace | Batch melting |
| Quality requirement | No shrinkage defects | Radiographic and sectioning checks |
From the perspective of solidification, ductile iron casting is more prone to shrinkage-related defects than gray iron because of the mushy solidification mode and the larger volumetric contraction during eutectic solidification. The carbon equivalent, inoculation practice, mold rigidity, and feeding efficiency all play a decisive role in producing a sound casting.
2. Initial Process Design
2.1 Parting Line and Mold Filling Direction
On our horizontal molding line, the parting line is generally selected at the middle plane of the large flange. This choice offers good pattern withdrawal and also gives flexibility for adjusting dimensional errors caused by mold shift. After considering the two alternative pouring positions — flange face up or flange face down — we compared the advantages and disadvantages for this particular ductile iron casting.
If the flange face is placed up, the majority of the casting lies in the lower mold. The gating system can be arranged as a top sprue with ingates directly on the upper surfaces. This produces a high static metal head, which is beneficial for gravity feeding of lower hot spots. The shrinkage tendency is therefore relatively small. However, the falling metal stream during pouring can cause sand erosion and inclusions, which often results in surface defects and dirt.
If the flange face is placed down, the gating system is necessarily a bottom fill system. The metal rises smoothly inside the mold cavity, which greatly improves surface quality. The downside is that the upper part of the mold must contain a higher volume of casting, and the hot spots located in the upper region become difficult to feed because the feeding head is reduced and the metal temperature drops before reaching those sections. In our ductile iron casting experiments, we initially saw a higher shrinkage incidence with the flange-down position, but the improvement in surface quality was so significant that we decided to keep this orientation and solve the feeding problem by means of better riser design.
Table 2 summarizes the trade-off we observed.
| Pouring position | Advantages | Disadvantages | Selected? |
|---|---|---|---|
| Flange face up | High pressure head; better feeding; lower shrink tendency | Sand erosion; inclusions; poor surface | No |
| Flange face down | Smooth filling; good surface; fewer inclusions | Upper hot spots hard to feed; higher shrink risk | Yes |
After several trial runs, we confirmed that the flange-down position was the only way to meet the aesthetic and internal quality requirements of the final ductile iron casting. The increased shrink risk had to be addressed with carefully engineered risers and controlled pouring conditions.
2.2 Hot Spot Identification and Solidification Modulus
In ductile iron casting, the solidification time of a local region is proportional to the square of its modulus, where the modulus \(M\) is defined as the ratio of the cooling volume to the cooling surface area:
$$M = \frac{V}{A}$$
Using Chvorinov’s rule, the local solidification time \(t_s\) can be written as:
$$t_s = B M^2$$
where \(B\) is a mold-material constant. A larger modulus means a longer solidification time and therefore a greater demand on feeding. For the main reducing shell cover, we identified three structural “natural” hot spots:
- Location 1: a relatively large boss on the lower flange surface;
- Location 2: the 125 mm diameter circular pad on the upper part of the casting;
- Location 3: the intersection of the flange with lateral ribs, though smaller in severity.
Among these, Location 2 proved to be the most severe. Because the casting is poured with the flange down, Location 2 is the highest point in the mold. The hot metal reaching this point is already cooler than the metal that entered first, and there is no significant riser above it to provide gravity feeding. The initial attempts without any riser at Location 2 produced visible surface sinks and severe internal shrinkage porosity, as shown by sectioning the defective ductile iron casting.
We also observed a process-induced hot spot at Location 1. The ingate was originally placed at that boss, and the local concentration of heat from the flowing metal combined with the existing structural boss to create a hot spot that required a feeder. According to the modulus calculation, the boss at Location 1 had a relatively low modulus, but the ingate heat made it necessary to place a hot riser directly on the ingate to avoid shrinkage porosity. The final design at this location used a hot riser with a diameter of 55 mm and a height of 85 mm. The use of a hot riser (i.e., a riser that is filled directly with hot metal from the ingate) proved to be an effective solution.
2.3 Riser Design for the Upper Isolated Hot Spot
For Location 2, the challenge was greater. The modulus of the thick pad was calculated as:
$$M_{hot\ spot} = \frac{V_{pad}}{A_{pad}}$$
Using the casting geometry and local thickness, we obtained an approximate modulus of about 7.5 mm. According to common feeding guidelines for ductile iron, the riser modulus should be at least 1.2 times the hot-spot modulus, but for isolated hot spots in green sand, a safety factor of 1.5 to 2 is often required. We initially designed a cylindrical riser with a diameter \(D = 80\) mm and height \(H = 110\) mm. Its modulus is given by:
$$M_{riser} = \frac{D H}{4H + 2D}$$
Substituting \(D = 80\) mm and \(H = 110\) mm:
$$M_{riser} = \frac{80 \times 110}{4 \times 110 + 2 \times 80} = \frac{8800}{440 + 160} = \frac{8800}{600} = 14.67 \text{ mm}$$
This value was about 1.95 times the hot-spot modulus, which should have been sufficient in theory. However, because this riser was a “cold riser” — it was not directly fed by the ingate and the metal entering it was relatively cool — the riser itself solidified before the hot spot, thereby failing to provide the required liquid metal. As a result, about 20% of the ductile iron casting produced with this design still showed shrinkage cavities at Location 2.
We analyzed the solidification sequence and realized that the problem was not simply the modulus but the temperature gradient. The metal in the riser had a lower temperature than the metal in the hot spot because the riser was placed at the extremity of the filling path. In addition, the riser neck created a thermal center that interfered with the natural feeding direction. To overcome this issue, we introduced two improvements:
- We placed the riser directly inside the 125 mm diameter circular pocket, so that the riser neck was surrounded by a sharp sand edge.
- We placed an exothermic (heating) block inside the cold riser to increase the riser temperature and prolong its effective feeding time.
The sharp sand edge effect is particularly interesting. When a sharp sand corner is formed at the junction of the riser neck and the casting cavity, the local sand volume is smaller and the heat extraction is two-dimensional rather than three-dimensional. In practice, this reduces the cooling rate in the sand wedge and keeps the neck molten longer. Combined with the exothermic block, the riser remained liquid long enough to feed the isolated hot spot. After these modifications, the shrinkage defect at Location 2 disappeared completely.
3. Gating System Design
The sand box dimensions on our horizontal line are 700 mm × 600 mm × 250 mm on both cope and drag. We arranged one casting per mold. Based on the earlier trials, the final production gating system is shown schematically in the layout of the mold, with the following key elements:
- A vertical sprue of standard cylindrical cross-section;
- A ceramic foam filter of size 60 mm × 60 mm × 15 mm, with 10 PPI (pores per inch);
- A cross runner distributing metal to two ingates;
- One hot riser at the ingate near Location 1;
- One cold riser with exothermic block at Location 2.
We adopted a semi-open, semi-closed gating system (often called a semi-pressurized system) to balance flow velocity and dross trapping. The cross-sectional area ratios are given by:
$$\Sigma S_{sprue} : \Sigma S_{cross} : \Sigma S_{ingate} = 1 : 1.15 : 1.3$$
The actual areas are listed in Table 3.
| Gating element | Area ratio | Comment |
|---|---|---|
| Sprue exit area | 1.0 | Vertical sprue |
| Cross runner area | 1.15 | Distributes metal |
| Ingate area (total) | 1.3 | Two ingates |
The use of a ceramic foam filter was essential for ductile iron casting because the metal is highly reactive and easily forms dross. The filter smooths the flow and captures nonmetallic inclusions, which reduces the likelihood of inclusions interacting with the solidification front and creating secondary porosity. The two ingates were placed at symmetrical positions relative to the central axis to avoid mold misalignment and to disperse the thermal input. Each ingate was designed with a relatively low velocity, about 0.5 m/s, to prevent sand erosion in the green sand mold.
We also verified the filling time using the formula:
$$t_f = \frac{G}{\rho \cdot Q}$$
where \(G\) is the total poured weight (including gating system), \(\rho\) is the density of liquid iron (about 7.0 g/cm³), and \(Q\) is the average filling rate. For our casting, the filling time was about 12–15 seconds, which is appropriate for a ductile iron casting of 20 kg with thin walls.
4. Melt Composition and Process Control
Apart from geometry and gating, the metallurgical quality of the iron is equally critical in ductile iron casting. We melted with a charge of 35% return scrap and 65% steel scrap, and recarburized using graphite carbon raiser. To reduce the natural shrinkage tendency of the iron, we strictly controlled the magnesium residual and spheroidizing treatment. Excessive residual magnesium increases carbide stability and also increases shrinkage because of the higher solidification contraction. We also added 0.6% to 0.8% silicon carbide in the late stage of melting to improve graphite nucleation and reduce undercooling.
Inoculation was performed in three stages: ladle inoculation, spheroidization treatment, and stream inoculation during pouring. The total amount of inoculant was fixed at 1.0% of the liquid iron weight. The carbon equivalent was maintained at:
$$CE = C + \frac{Si + P}{3} = 4.55\% \pm 0.05\%$$
This relatively high carbon equivalent helps to promote graphite expansion, which partially compensates for the solidification shrinkage. However, too high a carbon equivalent can lead to graphite flotation and reduce mechanical properties. Therefore, the target composition was tightly controlled, as shown in Table 4.
| Element | Mass fraction (%) | Notes |
|---|---|---|
| C | 3.65 – 3.80 | Carbon equivalent control |
| Si | 2.55 – 2.80 | Silicon promotes ferrite |
| Mn | < 0.3 | Low manganese reduces shrinkage |
| P | < 0.06 | Phosphorus increases brittleness |
| S | < 0.02 | Low sulfur for spheroidization |
| RE | < 0.03 | Rare earth in nodulizer |
| Mg | 0.03 – 0.05 | Residual magnesium |
The pouring temperature was kept between 1380 °C and 1400 °C. Such a controlled temperature window was chosen to balance fluidity and shrinkage. If the pouring temperature is too high, the liquid shrinkage is larger and the resulting solidification shrinkage is more severe. If too low, the metal may not reach the farthest cavities or may cause cold shut defects. In our ductile iron casting production, we found that a pouring temperature in the range of 1380–1400 °C gave the best compromise.
We also paid attention to the spheroidization efficiency. The nodularity was checked on every heat using thermal analysis and microstructural samples. We required a minimum nodularity of 85% and a ferrite content above 80% to meet the QT400-15 specification. The control of residual magnesium was critical: above 0.05% increases shrinkage, while below 0.03% may lead to poor nodularity.
5. Defect Analysis and the Role of Solidification Simulation
In the course of optimizing this ductile iron casting, we performed several defect analyses using both physical sectioning and numerical solidification simulation. The simulation helped us visualize the solidification sequence and identify the last-to-solidify regions. Applying the modulus criterion, we can compute the solidification time distribution across the casting. The feeding demand can be expressed by the local shrinkage volume:
$$V_{shrink} = \alpha \cdot V_{liquid\ withdrawal}$$
where \(\alpha\) is the contraction coefficient of the ductile iron. For QT400-15, the eutectic graphitization expansion partially compensates for the liquid contraction, so the apparent shrinkage tendency is lower than that of steel but still high enough to cause defects if the mold is not rigid or if the feeding path is blocked.
The first simulations confirmed that Location 2 solidified last and was isolated from the other hot regions. Without an effective riser, the liquid inside this pocket was cut off from the rest of the mold, and the surrounding solidified shell formed a pressure drop. The resulting negative pressure caused the formation of an internal shrinkage cavity. The image below shows a typical defective ductile iron casting section from that early trial.
After the riser redesign, the simulation showed that the riser remained liquid for more time than the local hot spot. The sharp sand effect delayed the solidification of the riser neck, while the exothermic block raised the riser temperature by about 50–80 °C. This combination changed the temperature gradient so that the hot spot was fed by the riser until the end of solidification. The simulation predicted a sound casting, and the subsequent production trials confirmed this result.
6. Countermeasures and Production Validation
Based on our findings, we implemented a set of countermeasures that are summarized in Table 5. These countermeasures were integrated into the production control plan for every batch of ductile iron casting.
| Defect mechanism | Countermeasure | Key parameters |
|---|---|---|
| Inadequate feeding at upper isolated hot spot | Place cold riser inside the 125 mm pocket; use exothermic block | Riser Ø80 × 110 mm; Exothermic block type standard |
| Riser neck solidifies too early | Utilize sharp sand edge in riser neck | Neck geometry with sharp angle |
| Ingate creates process hot spot on flange boss | Use hot riser directly over ingate | Hot riser Ø55 × 85 mm |
| High shrinkage tendency of iron | Control Mg residual; add SiC; use high CE | Mg 0.03–0.05%; CE 4.55±0.05% |
| Excessive pouring temperature | Lower pouring temperature range | 1380–1400 °C |
| Dross and inclusions acting as nucleation sites | Use ceramic foam filter; smooth bottom gating | 10 PPI, 60×60×15 mm |
After implementing all of these measures, we ran a full production validation. A total of 500 main reducing shell covers were produced in a single campaign, with destructive testing performed on every tenth casting. The results were excellent:
- No external shrinkage sinks were observed on any casting.
- Sectioning of 50 castings revealed no internal shrinkage cavities.
- Radiographic inspection on a sample of 30 castings met the required class 2 quality level.
- The microstructure was consistently ferritic with 85–95% nodularity.
- Mechanical tests showed tensile strengths of 420–440 MPa and elongations of 18–22%, exceeding the minimum specification.
This level of stability was achieved without adding expensive chills and without changing the mold material. The main improvement came from a better understanding of the solidification sequence and the inventive use of an exothermic block in a cold riser, combined with a sharp sand edge. The final ductile iron casting not only met the quality requirements but also demonstrated a reproducible process that is now used in mass production.
7. The Role of Mold Rigidity and Green Sand Properties
Another factor that we investigated during this project was the rigidity of the green sand mold. In ductile iron casting, the eutectic graphitization expansion creates an internal pressure that can help to feed shrinkage if the mold is rigid. If the mold walls yield, the expansion is lost and shrinkage defects form more easily. On our horizontal molding line, the standard mold hardness was about 85–90 on the B-scale. We took care to maintain this hardness uniformly, particularly in the cope where the riser and the isolated hot spot are located.
We measured the compaction density of the sand at several positions and adjusted the squeezing pressure slightly for the mold containing the main reducing shell cover. This adjustment allowed us to minimize mold wall movement. The effect was noticeable: with a harder cope, the tendency for surface sinks on the upper pad decreased even before the riser modification. Therefore, in the final production process, we specify a mold hardness of 90±2 for both cope and drag.
We also controlled the moisture content of the green sand between 3.2% and 3.5%, because excessive moisture increases gas generation and also softens the mold. The permeability was kept above 100 to allow gases to escape during pouring. These parameters are important for any ductile iron casting, but they become even more critical when the casting has a large flat surface and a heavy hot spot.
8. Mathematical Model for Riser Sizing
To provide a more rigorous foundation for our ductile iron casting process design, I developed a simplified mathematical model for the feeding requirement. The total volumetric contraction can be expressed as:
$$\Delta V_{total} = \beta_l V_l + \beta_s V_s – \Delta V_{graphite}$$
where \(\beta_l\) is the liquid contraction coefficient, \(V_l\) is the liquid volume at the beginning of solidification, \(\beta_s\) is the solidification contraction coefficient, \(V_s\) is the volume that solidifies in the pasty zone, and \(\Delta V_{graphite}\) is the expansion due to graphite precipitation. For ductile iron, \(\Delta V_{graphite}\) is significant and can compensate for 30–50% of the total contraction. However, this compensation occurs only if the mold is rigid and if the graphite precipitates in time to feed the solidifying matrix.
The required riser volume \(V_r\) is then:
$$V_r = \frac{\Delta V_{shrink}}{\eta}$$
where \(\eta\) is the feeding efficiency of the riser (typically 0.1–0.2 for cold risers, and 0.3–0.5 for hot or exothermic risers). For our upper isolated hot spot, the shrinkage volume was estimated to be about 100 cm³. With a cold riser efficiency of 12%, the required riser volume would be:
$$V_r = \frac{100}{0.12} = 833 \text{ cm}^3$$
The actual riser volume was:
$$V_r = \frac{\pi}{4} (8.0 \text{ cm})^2 (11.0 \text{ cm}) = 553 \text{ cm}^3$$
This was smaller than the theoretical requirement, which explains why the cold riser failed in 20% of castings. With the exothermic block, the effective feeding efficiency increased significantly because the riser metal stayed liquid longer and did not produce a solidified cap. We estimated that the exothermic block raised the feeding efficiency to about 20%, making the effective feeding volume:
$$V_{effective} = 553 \times 0.20 + V_{exothermic} \approx 110 + 150 = 260 \text{ cm}^3$$
This was more than the 100 cm³ shrinkage volume demand, providing a comfortable safety margin. In practice, the combination of the exothermic block and the sharp sand neck converted the cold riser into a semi-exothermic riser, giving the required feeding action.
9. Lessons Learned and Guidelines
From this long development journey, I can draw several important lessons that apply to similar ductile iron casting components with isolated hot spots:
- Do not rely solely on riser modulus. The temperature gradient and the riser liquidus temperature are equally important. A large cold riser may fail if it is not effectively hot enough to feed the casting.
- Use the casting geometry to your advantage. By placing the riser inside an existing cavity, we created a natural sharp sand effect that prolonged the liquid state of the neck. This is a simple and inexpensive technique.
- Exothermic sleeves or blocks are not a luxury. In ductile iron casting, the cost of adding an exothermic block is far lower than the cost of scrap caused by shrinkage defects. It stabilized the process and eliminated the intermittent defect that we had seen in 20% of the castings.
- Control the process window tightly. Pouring temperature, residual magnesium, carbon equivalent, and inoculant addition all have a direct effect on shrinkage. A narrow process window is essential for reproducible quality.
- Simulation is a helpful guide, but not a substitute for experiments. We used simulation to understand the solidification sequence, but the final solution came from iterative physical trials combined with a good understanding of the physics.
10. Conclusion
In conclusion, the shrinkage defects observed in the main reducing shell cover produced by ductile iron casting were caused by a combination of an isolated structural hot spot, an inefficient cold riser, and a less-than-optimal temperature gradient. After systematic analysis, we solved the problem by:
- Selecting the flange-down pouring position with a bottom gating system for better surface quality;
- Using a hot riser at the ingate near the flange boss to counteract the process-induced hot spot;
- Placing a cold riser of Ø80 × 110 mm directly inside the upper circular pocket;
- Creating a sharp sand edge at the riser neck to delay neck solidification;
- Adding an exothermic block inside the riser to extend its feeding time;
- Maintaining a controlled chemical composition and a pouring temperature of 1380–1400 °C.
These actions eliminated the shrinkage cavities and allowed us to achieve a consistently high yield of sound ductile iron casting parts. The final process is now used for mass production, and the lessons learned have been transferred to other ductile iron casting products with similar structural features. The combination of careful casting design, metallurgical control, and the strategic use of exothermic materials proved to be the key to success.
I hope that sharing our experience helps other foundry engineers who face shrinkage problems in ductile iron casting. The path to a robust process may require several iterations, but by applying fundamental solidification principles and continuously measuring the results, it is possible to produce complex ductile iron casting components with zero internal defects.