I undertook the development of high-performance ductile iron castings for a high-power nuclear gearbox planetary carrier because the component must survive a design life of twenty years while transmitting very large dynamic and static loads. The planetary carrier is a critical structural body in the planetary transmission system, and its service reliability directly affects the gearbox, the circulating pump, and the entire nuclear power train. Previous ductile iron castings made in a 600-3 type grade showed insufficient strength and plasticity in service, and the resulting early failures forced me to redesign the chemistry, casting process, melting practice, spheroidizing treatment, inoculation, pouring, and post-treatment route. The internal development target was clear: tensile strength of at least 700 MPa and elongation of at least 5.0%, while maintaining a predominantly pearlitic matrix, a nodularity above 90%, and a sound internal structure in a very heavy section. My work therefore treated the planetary carrier not simply as a shaped part but as a system of solidification, feeding, microstructure, and property control problems.
I first analyzed the geometry, mass, section thickness, and loading condition of the planetary carrier. The outline diameter was approximately 1280 mm, the overall height was approximately 700 mm, and the main wall thickness ranged from 100 to 115 mm. Four columns contained local hot spots with a maximum thickness above 210 mm, and the total casting mass was approximately 2400 kg. These dimensions placed the part firmly in the category of heavy-section ductile iron castings. Because the hot spots were large and distributed around the component, I knew that isolated liquid regions could form easily, and that shrinkage porosity, graphite flotation, chunky graphite, spheroidal degeneration, intercellular carbides, and inverse chill were all credible risks. The geometry also made gating and riser design difficult because a single central feeder could not reach every hot spot without creating a long feeding path.
| Parameter | Value or condition | Consequence for ductile iron castings |
|---|---|---|
| Outline diameter | Approximately 1280 mm | Large thermal mass and long filling distance |
| Overall height | Approximately 700 mm | Need for stable vertical filling and venting |
| Main wall thickness | 100 to 115 mm | Slow cooling and coarse microstructure risk |
| Maximum column thickness | Greater than 210 mm | Strong hot spots and feeding difficulty |
| Casting mass | Approximately 2400 kg | High pouring load and large liquid shrinkage |
| Design life | 20 years | Requires high fatigue resistance and structural integrity |
| Load type | Dynamic and static transmission load | Requires high strength, stiffness, and plasticity |
| Required tensile strength | At least 700 MPa | Needs pearlitic matrix and effective alloying |
| Required elongation | At least 5.0% | Needs low segregation and low carbide content |
| Previous grade | 600-3 type ductile iron castings | Insufficient strength and plasticity in service |
| Observed failure mode | Short life and early failure | Material property margin was inadequate |
I used the modulus method to quantify the thermal condition of the casting. For a given casting volume \(V\) and cooling surface area \(A\), the cooling modulus \(M_c\) is:
$$M_c = \frac{V}{A}$$
For the法兰 and column hot spots, the calculated modulus was approximately \(M_c = 4.5\,\text{cm}\). The solidification time \(t_s\) can be approximated by Chvorinov’s rule:
$$t_s = K M_c^n$$
where \(K\) is a mold constant and \(n\) is an exponent usually close to 2 for heavy-section ductile iron castings. Because \(t_s\) increases strongly with modulus, the thick column regions remained liquid much longer than the thin walls. I therefore designed the feeding system so that each hot spot had a nearby riser with sufficient modulus and sufficient liquid metal. I also used the carbon equivalent to control graphite precipitation and expansion during eutectic solidification:
$$CE = C + \frac{Si + P}{3}$$
For heavy-section ductile iron castings, I targeted a near-eutectic carbon equivalent between 4.3% and 4.4%. This range promoted graphitization, reduced white iron tendency, improved magnesium recovery, increased graphite sphericity, and maintained good fluidity. At the same time, I avoided excessive carbon equivalent because high carbon and silicon can increase graphite flotation risk, especially in sections above 200 mm. The final carbon equivalent of the produced casting was calculated from the measured composition as:
$$CE = 3.65 + \frac{2.04 + 0.018}{3} = 4.336\%$$
This value sat inside the intended window and helped explain the good combination of strength and elongation obtained in the attached test block.
I designed the chemical composition according to the principle of high carbon, low silicon, low phosphorus, low sulfur, moderate magnesium, low rare earth, and controlled copper. The control range is summarized in Table 2. I treated the base iron carbon content and the post-treatment silicon content separately because the charge carbon and the final silicon after spheroidizing and inoculation have different effects on graphite flotation, matrix structure, and shrinkage behavior. For ductile iron castings with thick sections, small chemistry errors can produce large property losses, so I set a narrow window for every element.
| Element | Control range (wt.%) | Primary role in ductile iron castings |
|---|---|---|
| C in base iron | 3.5 to 3.7 | Promotes graphitization, fluidity, and eutectic graphite formation |
| Si after treatment | 2.0 to 2.3 | Promotes graphite formation and controls ferrite/pearlite balance |
| Mn | Less than 0.5 | Stabilizes pearlite but segregates and can form carbides |
| S | Less than 0.02 | Consumes magnesium and rare earth, causing spheroidal degeneration |
| P | Less than 0.03 | Forms phosphide eutectic and lowers plasticity and toughness |
| Mg residual | 0.04 to 0.06 | Main spheroidizing element for graphite nodules |
| RE residual | Less than 0.02 | Auxiliary spheroidizing and impurity neutralization, but excess causes carbides |
| Cu | 0.5 to 0.8 | Promotes and refines pearlite, strengthens matrix |
I controlled carbon in the base iron between 3.5% and 3.7% because this range gave enough graphite expansion to offset liquid shrinkage without causing severe graphite flotation. I controlled silicon after treatment between 2.0% and 2.3% because silicon strongly graphitizes the iron and affects the final matrix. If silicon is too low, free cementite can appear in heavy-section ductile iron castings; if silicon is too high, ferrite increases and tensile strength falls. I kept manganese below 0.5% because manganese stabilizes pearlite but has a strong segregation tendency. High manganese in thick sections can produce intercellular network carbides and reduce elongation. I kept sulfur below 0.02% because sulfur reacts with magnesium and rare earth to form MgS and RES inclusions. These inclusions not only consume spheroidizing elements but also increase slag, subsurface porosity, and hot cracking risk. I kept phosphorus below 0.03% because phosphorus has low solubility in ductile iron and forms phosphide eutectic at cell boundaries, which lowers strength, plasticity, and toughness.
I used magnesium as the main spheroidizing element and rare earth as an auxiliary element. The residual magnesium was controlled between 0.04% and 0.06%, and residual rare earth was kept below 0.02%. In heavy-section ductile iron castings, excessive magnesium and rare earth increase the tendency toward carbides and inverse chill. I therefore aimed at the lowest magnesium and rare earth levels that could still produce a nodularity above 90%. I added copper between 0.5% and 0.8% because copper promotes graphitization during eutectic solidification and refines pearlite during the eutectoid transformation. The copper addition strengthened the matrix without the severe segregation problems associated with high manganese. The final measured composition of the produced planetary carrier is given later, but the design logic was already fixed in this composition window.
I next designed the casting process using the large-orifice discharge principle and the balanced solidification concept. I selected a small-neck insulating riser arrangement to feed the four hot spots and to promote uniform temperature distribution. The gating system was designed with the following cross-sectional area ratio:
$$\sum A_{\text{ingate}} : \sum A_{\text{runner}} : \sum A_{\text{sprue}} = 1.00 : 2.00 : 1.56$$
I calculated the choke area by the choke-section design method and obtained:
$$A_{\text{choke}} = 38\,\text{cm}^2$$
I used a 100 mm diameter ceramic sprue tube, an 80 mm diameter ceramic runner tube, two ceramic filters of 150 mm × 30 mm with 10 PPI, and four 40 mm diameter ceramic ingates distributed relatively uniformly around the mold cavity. The purpose of this arrangement was to fill the mold quickly while keeping the flow smooth and to close the ingates in time after filling. If the ingates remain open during graphite expansion, iron can be pushed back into the runner and riser system, which can produce shrinkage porosity in ductile iron castings. The gating design therefore had to balance filling speed, slag entrapment prevention, and expansion-driven feeding behavior.
| Gating item | Design value | Reason for selection |
|---|---|---|
| Pouring time | 60 to 70 s | Fast enough to avoid cold shut, slow enough to avoid turbulence |
| Gating area ratio | 1.00 : 2.00 : 1.56 | Controlled filling and reduced slag entrapment |
| Choke area | 38 cm² | Established by choke-section calculation |
| Sprue | 100 mm ceramic tube | Stable vertical flow |
| Runner | 80 mm ceramic tube | Uniform distribution to ingates |
| Filters | Two 150 mm × 30 mm, 10 PPI | Removed slag and inclusions |
| Ingates | Four 40 mm ceramic tubes | Dispersed entry and timely closure |
For the hot spots, I placed four 200 mm necked insulating risers slightly away from the main thermal centers. The modulus relationship used for riser design was:
$$M_r = k M_c$$
where \(M_r\) is the riser modulus, \(M_c\) is the casting hot-spot modulus, and \(k\) is a safety factor. I selected \(k > 1.2\) for the heavy sections. I also placed 30 mm easy-break cores at the riser roots. These cores controlled the liquid feeding time and made riser removal easier after solidification. By moving the risers slightly away from the exact hot spots, I made the temperature field more uniform and encouraged simultaneous solidification around the法兰 and columns. This was important because ductile iron castings expand during graphitization, and if the feeding path remains open too long, the expansion can push iron back into the riser and create internal shrinkage. The easy-break core acted as a timed valve that closed the feeding channel after the useful liquid feeding stage.
| Riser parameter | Value | Function |
|---|---|---|
| Number of risers | 4 | One for each major hot spot |
| Riser diameter | 200 mm | Provided sufficient liquid volume and modulus |
| Riser type | Necked insulating riser | Reduced heat loss and improved feeding efficiency |
| Riser location | Slightly offset from hot spots | Uniform temperature and reduced local overheating |
| Easy-break core | 30 mm | Controlled feeding time and eased riser removal |
| Modulus factor | \(k > 1.2\) | Safety margin for heavy-section ductile iron castings |
I used three-dimensional modeling and casting simulation to verify the filling and solidification behavior before production. The simulation showed that the iron entered the mold cavity in a stable manner. During filling, I observed no violent splashing, no obvious surface turbulence, and no stagnant regions that would trap slag or sand. The liquid level rose uniformly, and the filters and gating channels distributed the metal to the four ingates without severe impingement. This was important for ductile iron castings because turbulent filling can entrain air and slag, while too slow filling can cause cold shut in thin sections. The selected pouring time of 60 to 70 s and the gating ratio produced a favorable balance.
| Filling stage | Simulated condition | Interpretation |
|---|---|---|
| Early filling, about 2% | Metal entered the runner and began to fill the ingate region | Flow was stable and controlled |
| Intermediate filling, about 8% | Metal distributed through filters and runners | Slag and inclusion risk was reduced |
| Late filling, about 62% | Mold cavity filled progressively from lower regions | No severe splashing or dead zones |
| Final filling | Complete filling of the cavity and riser necks | Good filling without cold shut |
During solidification, the simulation showed that the feeding channels remained open while liquid metal was still available. There was no isolated liquid region inside the casting. The riser necks supplied liquid metal to the hot spots and then closed at the correct time because of the easy-break cores. This behavior reduced the risk of shrinkage porosity. The simulation also indicated that the last solidified regions were located in the risers rather than in the load-bearing sections of the planetary carrier. I used this result to confirm that the gating and riser design could produce sound ductile iron castings with acceptable internal quality.
| Solidification stage | Simulated condition | Process meaning |
|---|---|---|
| About 25% solidified | Outer walls and thinner sections began to freeze first | Directional solidification started from the surface |
| About 50% solidified | Feeding paths remained connected to the hot spots | No isolated liquid pocket formed |
| About 75% solidified | Riser necks continued to feed the thick sections | Shrinkage risk was controlled |
| Final solidification | Risers solidified last | Defects were concentrated in the riser system |
For mold making, I used three-box molding with a sand box size of 1600 mm × 1600 mm × 320 mm. I selected sodium silicate sand to produce a high-strength and high-rigidity mold. A rigid mold is essential for heavy-section ductile iron castings because mold wall movement can enlarge the casting and change the feeding behavior. If the mold wall yields during graphite expansion, the expansion cannot be used effectively to compensate shrinkage. I arranged sixteen 20 mm diameter vents in the upper mold to allow gas to escape during pouring. I also coated the mold surface with a graphite alcohol-based coating at a thickness of about 1.5 mm. The coating improved surface quality, reduced sand adhesion, and made cleaning easier.
| Molding parameter | Value or method | Purpose |
|---|---|---|
| Molding method | Three-box molding | Good access for gating and risers |
| Sand box size | 1600 mm × 1600 mm × 320 mm | Accommodated the large casting and feeding system |
| Mold material | Sodium silicate sand | High strength and high rigidity |
| Venting | Sixteen 20 mm vents | Prevented gas defects and incomplete filling |
| Coating | Graphite alcohol-based coating | Improved surface quality and reduced sand burn-on |
| Coating thickness | About 1.5 mm | Balanced refractory protection and gas permeability |
For melting, I used a 3 t medium-frequency induction furnace and high-purity pig iron together with clean charge materials. I added copper in the early melting stage as an alloy so that its concentration in the liquid iron was uniform before spheroidizing. I used a direct-reading optical emission spectrometer to adjust the chemical composition before tapping. The melting practice aimed at high temperature, clean liquid iron, low sulfur, low phosphorus, and stable carbon. For ductile iron castings with thick sections, dirty liquid iron and high sulfur can destroy the spheroidizing effect even when the magnesium addition is high. I therefore controlled slag, removed dross, and avoided excessive holding time before treatment.
| Melting item | Practice | Reason |
|---|---|---|
| Furnace | 3 t medium-frequency induction furnace | Stable composition and temperature control |
| Charge material | High-purity pig iron and selected scrap | Low trace element content |
| Copper addition | Added early as alloy | Uniform copper distribution |
| Composition analysis | Optical emission spectrometer | Rapid adjustment before tapping |
| Target condition | High temperature and clean liquid iron | Improved spheroidizing and casting quality |
I used the tundish cover or sandwich method for spheroidizing and inoculation in the ladle. The treatment temperature was 1400 to 1430 °C. The spheroidizer addition was 1.3% of the liquid iron mass. The spheroidizer contained 5.68% Mg, 0.55% Al, 44.29% Si, 1.0% Ca, and 0.93% Re, with a particle size of 4 to 25 mm. The in-ladle inoculant was a 0.4% high-calcium barium inoculant containing 1.03% Al, 72.76% Si, 1.55% Ca, and 2.35% Ba, with a particle size of 3 to 8 mm. After spheroidizing and inoculation, I immediately skimmed the slag to prevent sulfur return and covered the liquid surface with a slag conditioner to reduce oxidation.
| Treatment agent | Addition (wt.%) | Main composition | Particle size |
|---|---|---|---|
| Spheroidizer | 1.3 | 5.68% Mg, 0.55% Al, 44.29% Si, 1.0% Ca, 0.93% Re | 4 to 25 mm |
| In-ladle inoculant | 0.4 | 1.03% Al, 72.76% Si, 1.55% Ca, 2.35% Ba | 3 to 8 mm |
| Late-stream inoculant | 0.1 | 73% Si, 1.0% Ca, 1.75% Ce, 1.0% Al, trace S and O | 0.2 to 0.7 mm |
To strengthen inoculation further, I added a 0.1% sulfur-oxygen inoculant during pouring as a late-stream inoculant. This fine-particle inoculant contained 73% Si, 1.0% Ca, 1.75% Ce, 1.0% Al, and trace sulfur and oxygen. The purpose of late-stream inoculation was to increase the number of heterogeneous nuclei, reduce carbide formation, improve graphite nodularity, and increase the feeding ability of the iron during final solidification. In heavy-section ductile iron castings, the spheroidizing and inoculation effect can fade during long holding and slow cooling. Late-stream inoculation helps restore the nucleation potential at the moment the iron enters the mold. The measured nodularity above 90% confirmed that the combination of in-ladle and late-stream inoculation was effective.
I baked the mold in hot air at 240 °C for more than 4 h before closing, so that the cavity temperature reached 50 to 60 °C. The pouring temperature was 1350 °C, and the pouring time was 61 s. I used a low-temperature fast-pouring strategy to reduce liquid shrinkage and to avoid excessive superheat. High pouring temperature increases liquid shrinkage and gas defects, while too low a pouring temperature can cause cold shut and incomplete filling. The selected temperature and time provided a safe window for these ductile iron castings. After pouring, I allowed the casting to cool in the mold until the temperature fell below 300 °C before shakeout. This controlled cooling step reduced residual stress and avoided cracking in the heavy sections.
| Pouring and cooling parameter | Value | Purpose |
|---|---|---|
| Mold baking | 240 °C for more than 4 h | Reduced moisture and gas defects |
| Cavity temperature | 50 to 60 °C | Improved filling stability |
| Pouring temperature | 1350 °C | Balanced fluidity and shrinkage |
| Pouring time | 61 s | Fast filling without severe turbulence |
| Shakeout temperature | Below 300 °C | Reduced residual stress and cracking risk |
After shakeout, I cleaned the casting, removed the risers and gates, ground the easy-break core marks, and inspected the surfaces. I did not use a high-temperature normalizing treatment because the target microstructure was an as-cast pearlitic matrix. A high-temperature treatment could coarsen the microstructure and reduce the benefit of copper alloying. Instead, I relied on controlled cooling, appropriate silicon, moderate manganese, and copper to obtain the desired pearlite content. This post-treatment route was intentionally simple but strict: controlled shakeout, controlled cooling, and careful cleaning. The attached test block was cast together with the planetary carrier so that the measured properties represented the actual solidification condition of the ductile iron castings.
The measured chemical composition of the attached test block is shown in Table 9. The carbon and silicon values were within the designed window, and the sulfur and phosphorus levels were low. The residual magnesium was 0.05%, which was sufficient for spheroidization without excessive carbide formation. The copper content was 0.56%, which promoted pearlite without making the matrix too hard or brittle. The low sulfur and phosphorus contents helped maintain elongation by reducing inclusions and phosphide eutectic at cell boundaries.
| Element | Measured value (wt.%) | Assessment |
|---|---|---|
| C | 3.65 | Within the base iron target |
| Si | 2.04 | Within the post-treatment target |
| Mn | 0.23 | Below the upper limit |
| P | 0.018 | Below the upper limit |
| S | 0.008 | Very low and favorable for spheroidization |
| Cu | 0.56 | Sufficient for pearlite strengthening |
| Mg residual | 0.05 | Good spheroidizing range |
The microstructure of the attached test block was mainly pearlitic. The nodularity was above 90%, which indicated that the spheroidizing and inoculation process produced well-formed graphite nodules. The matrix contained pearlite as the dominant phase, with a limited amount of ferrite. This microstructure explained the combination of high tensile strength and acceptable elongation. Pearlite increases strength and hardness, while a small amount of ferrite and a high nodularity help preserve plasticity. The graphite morphology was also important because irregular or chunky graphite can act as a stress concentrator and reduce both strength and elongation. I used the area fraction of nodular graphite to evaluate nodularity:
$$N = \frac{A_{\text{nodular}}}{A_{\text{total}}} \times 100\%$$
where \(N\) is the nodularity, \(A_{\text{nodular}}\) is the area of nodular graphite, and \(A_{\text{total}}\) is the total graphite area. The measured \(N\) was greater than 90%. The pearlite fraction can be expressed as:
$$f_P = \frac{A_P}{A_P + A_F} \times 100\%$$
where \(A_P\) and \(A_F\) are the pearlite and ferrite areas. For the produced ductile iron castings, \(f_P\) was high enough to meet the strength target while still leaving enough ferrite to meet the elongation target.
| Microstructural feature | Result | Effect on properties |
|---|---|---|
| Matrix type | Predominantly pearlite | High tensile strength and yield strength |
| Graphite nodularity | Above 90% | Reduced stress concentration |
| Graphite shape | Nodular | Improved elongation and fatigue resistance |
| Ferrite content | Limited | Preserved plasticity without sacrificing strength |
| Carbide content | Low | Avoided brittle fracture |
| Segregation | Controlled | Improved uniformity in heavy sections |
The tensile properties were measured using a 14 mm diameter tensile specimen and a universal testing machine. The results were a tensile strength of 797 MPa, a yield strength of 444 MPa, and an elongation of 7.0%. These values exceeded the internal target of at least 700 MPa tensile strength and at least 5.0% elongation. The yield ratio was calculated as:
$$YR = \frac{\sigma_y}{\sigma_t} = \frac{444}{797} = 0.557$$
where \(\sigma_y\) is the yield strength and \(\sigma_t\) is the tensile strength. The yield ratio was moderate, which indicated that the material had both strength and some deformation capacity. The strength margin over the target was:
$$\Delta \sigma_t = 797 – 700 = 97\,\text{MPa}$$
The elongation margin over the target was:
$$\Delta \delta = 7.0 – 5.0 = 2.0\%$$
I also calculated a simple quality index to summarize the strength-ductility balance:
$$QI = \sigma_t \times \delta = 797 \times 7.0 = 5579\,\text{MPa}\%$$
This quality index is not a standard specification, but it helped me compare the produced ductile iron castings with the previous 600-3 type material. The previous material had lower tensile strength and lower elongation, so the new process window produced a clear improvement in both properties. The measured properties are summarized in Table 11.
| Property | Measured value | Internal target | Margin |
|---|---|---|---|
| Tensile strength | 797 MPa | At least 700 MPa | +97 MPa |
| Yield strength | 444 MPa | Not specified by internal target | Good load-bearing capacity |
| Elongation | 7.0% | At least 5.0% | +2.0% |
| Nodularity | Above 90% | Above 90% | Met |
| Matrix | Mainly pearlite | Pearlitic matrix | Met |
| Yield ratio | 0.557 | Balanced strength and plasticity | Acceptable |
| Quality index | 5579 MPa% | Improved over 600-3 type | Improved |
I paid special attention to defect prevention because heavy-section ductile iron castings are sensitive to several solidification defects. Graphite flotation occurs when low-density graphite nodules rise through the liquid before the eutectic front immobilizes them. Chunky graphite forms in slow-cooling regions and reduces mechanical properties. Spheroidal degeneration occurs when magnesium and rare earth are consumed by sulfur and oxygen or when the cooling rate is too slow. Intercellular carbides form when segregation and insufficient inoculation leave carbon-rich liquid between cells. Inverse chill can appear in thick sections when the cooling rate and chemistry favor carbide rather than graphite. Shrinkage porosity forms when feeding is inadequate or when expansion is not properly used. I addressed each risk with a specific process measure, as summarized in Table 12.
| Defect risk | Mechanism | Countermeasure I used |
|---|---|---|
| Graphite flotation | Low-density graphite rises in slow-cooling liquid | Controlled carbon equivalent between 4.3% and 4.4%; limited silicon; fast pouring |
| Chunky graphite | Slow cooling and excess rare earth | Low RE residual; late-stream inoculation; controlled cooling |
| Spheroidal degeneration | Mg and RE consumed by S and O | Low sulfur; immediate skimming; protective covering |
| Intercellular carbides | Segregation and insufficient nucleation | Controlled Mn; high-calcium barium inoculant; sulfur-oxygen late inoculant |
| Inverse chill | Slow cooling and high carbide-forming tendency | Balanced Si and Cu; sufficient inoculation; controlled pouring temperature |
| Shrinkage porosity | Insufficient feeding or expansion loss | Four insulating risers; easy-break cores; rigid sodium silicate mold |
| Gas porosity | Moisture and gas trapped in mold | Hot-air baking; sixteen vents; graphite alcohol coating |
| Sand burn-on | High temperature reaction between metal and mold | Graphite alcohol coating; controlled pouring temperature |
I also established a quality inspection plan for the ductile iron castings. The plan included chemical composition analysis, microstructure examination, nodularity measurement, tensile testing, surface inspection, dimensional inspection, and internal defect inspection. The attached test block was used for destructive testing because it experienced a similar thermal history to the planetary carrier. The casting surface was checked for cold shut, sand inclusion, gas porosity, and cracks. The riser and gate removal areas were inspected for shrinkage defects. The dimensional check confirmed that the main outline, flange, and column positions were within the allowed tolerance. This combination of process control and inspection gave me confidence that the developed route could be transferred to production.
| Inspection item | Method | Purpose |
|---|---|---|
| Chemical composition | Optical emission spectrometry | Verify C, Si, Mn, P, S, Cu, Mg, and RE |
| Microstructure | Metallographic examination | Evaluate graphite shape, nodularity, and matrix |
| Nodularity | Image analysis | Confirm nodularity above 90% |
| Tensile properties | Universal testing machine | Measure tensile strength, yield strength, and elongation |
| Surface quality | Visual inspection | Detect cold shut, sand burn-on, and cracks |
| Dimensional accuracy | Measurement against drawing | Verify geometry and machining allowance |
| Internal soundness | Process simulation and sectioning when required | Confirm absence of harmful shrinkage |
I found that the most important process variables for these ductile iron castings were carbon equivalent, residual magnesium, residual rare earth, silicon content, inoculation practice, pouring temperature, and riser closure timing. A small change in any of these variables can shift the microstructure from a strong pearlitic matrix to a mixture containing carbides or excessive ferrite. I therefore used a narrow process window. For example, the carbon equivalent was kept near the eutectic value, but not so high that graphite flotation became likely. The residual magnesium was kept at 0.04% to 0.06%, but not higher, because excess magnesium increases carbide tendency. The copper content was kept at 0.5% to 0.8%, because copper strengthens pearlite but excessive copper can increase hardness and reduce machinability. The late-stream inoculant was kept at 0.1%, because too little inoculant allows carbides to form, while too much can cause dross and gas defects.
I also considered the effect of section size on cooling rate. The cooling rate \(CR\) can be approximated as:
$$CR \approx \frac{T_p – T_m}{t_s}$$
where \(T_p\) is the pouring temperature, \(T_m\) is the mold temperature, and \(t_s\) is the solidification time. In the thick columns, \(t_s\) was long, so \(CR\) was low. Low cooling rate promotes graphite growth, ferrite formation, and segregation. To compensate, I used copper to stabilize pearlite, controlled silicon to avoid excessive ferrite, and used late inoculation to keep graphite nuclei active. The resulting matrix was predominantly pearlite even in the heavy sections. This was a key reason why the tensile strength reached 797 MPa while elongation remained at 7.0%.
The relationship between matrix, graphite, and tensile strength can be qualitatively expressed as:
$$\sigma_t = \sigma_0 + k_P f_P^{1/2} + k_N N^{1/2} – k_C f_C – k_D D^{-1/2}$$
where \(\sigma_0\) is a base strength, \(f_P\) is the pearlite fraction, \(N\) is nodularity, \(f_C\) is the carbide fraction, \(D\) is the graphite nodule spacing or a microstructural scale parameter, and the \(k\) terms are positive or negative coefficients. This expression is not a precise predictive equation, but it summarizes the physical trend: high pearlite and high nodularity increase strength, while carbides and coarse microstructure reduce strength and plasticity. The produced ductile iron castings had high pearlite, high nodularity, low carbides, and a refined enough microstructure to meet the property targets.
| Microstructural variable | Desired direction | Effect on tensile strength | Effect on elongation |
|---|---|---|---|
| Pearlite fraction | High but not excessive | Increases strength | Can reduce elongation if too high |
| Ferrite fraction | Limited | Reduces strength | Increases elongation |
| Nodularity | Above 90% | Increases strength | Increases elongation |
| Carbide fraction | Very low | Reduces strength through cracking | Strongly reduces elongation |
| Graphite size | Fine and uniform | Improves strength | Improves elongation |
| Segregation | Low | Improves uniformity | Improves plasticity |
I compared the developed ductile iron castings with the previous 600-3 type material and with the internal target. The comparison is given in Table 15. The previous material was insufficient for the high-power nuclear gearbox planetary carrier because it did not provide enough strength and plasticity margin. The new material exceeded the target tensile strength by almost 100 MPa and exceeded the target elongation by 2.0 percentage points. The nodularity remained above 90%, and the matrix was predominantly pearlite. This combination is exactly what I wanted for a component that must resist both high contact loads and occasional overloads.
| Material condition | Tensile strength (MPa) | Yield strength (MPa) | Elongation (%) | Assessment |
|---|---|---|---|---|
| Previous 600-3 type ductile iron castings | About 600 | Not sufficient | About 3 | Early failure and short life |
| Internal target | At least 700 | Not specified | At least 5.0 | Required for nuclear gearbox service |
| Developed ductile iron castings | 797 | 444 | 7.0 | Exceeded target and improved reliability |
I also summarized the final process window in Table 16 so that the route could be repeated. The window includes the base iron carbon range, post-treatment silicon range, manganese limit, sulfur limit, phosphorus limit, residual magnesium range, residual rare earth limit, copper range, pouring temperature, pouring time, and shakeout temperature. The process window is intentionally narrow because the planetary carrier is a heavy-section safety-related component. A wider window might produce acceptable results in a small test casting but fail in the 2400 kg production casting. By keeping the window tight and monitoring each step, I reduced the risk of graphite flotation, chunky graphite, spheroidal degeneration, and shrinkage porosity.
| Process variable | Final window | Critical reason |
|---|---|---|
| Base iron carbon | 3.5% to 3.7% | Balanced graphitization and flotation risk |
| Post-treatment silicon | 2.0% to 2.3% | Controlled matrix and carbide tendency |
| Manganese | Less than 0.5% | Reduced segregation and carbides |
| Sulfur | Less than 0.02% | Protected magnesium and rare earth |
| Phosphorus | Less than 0.03% | Reduced phosphide eutectic |
| Residual magnesium | 0.04% to 0.06% | Ensured nodularity without excess carbides |
| Residual rare earth | Less than 0.02% | Avoided chunky graphite and carbides |
| Copper | 0.5% to 0.8% | Promoted pearlite and strength |
| Pouring temperature | 1350 °C | Balanced fluidity and shrinkage |
| Pouring time | 60 to 70 s | Controlled filling and turbulence |
| Shakeout temperature | Below 300 °C | Reduced residual stress |
I used the simulation results to guide the actual production trial. The filling simulation showed that the metal front advanced evenly and that the filters prevented large inclusions from entering the mold cavity. The solidification simulation showed that the risers remained liquid after the casting sections had begun to freeze. This meant that the feeding channels were still open during the period when liquid shrinkage was greatest. The easy-break cores then closed the channels before graphite expansion could push iron back into the risers. This sequence is critical for heavy-section ductile iron castings because the volume change during graphite precipitation can either help or harm the casting depending on whether the feeding path is open or closed. In my design, the path was open during liquid shrinkage and closed during expansion, which is the desired behavior.
I also considered the mechanical loading of the planetary carrier. The component transmits torque through the columns and法兰, so the load-bearing sections must have high stiffness and strength. The Young’s modulus of ductile iron is affected by graphite shape and matrix. Nodular graphite provides a higher modulus than flake graphite, and a pearlitic matrix provides higher stiffness than a ferritic matrix. The modulus can be approximated by a rule-of-mixtures type expression:
$$E = E_P f_P + E_F f_F – k_G f_G$$
where \(E_P\) and \(E_F\) are the moduli of pearlite and ferrite, \(f_P\) and \(f_F\) are their fractions, \(f_G\) is the graphite fraction, and \(k_G\) is a reduction coefficient for graphite. Although this expression is simplified, it shows why I wanted a predominantly pearlitic matrix with well-formed nodules. The measured properties confirmed that the microstructure provided both strength and stiffness for the nuclear gearbox application.
For fatigue resistance, the graphite nodule shape is especially important. A sharp graphite flake acts as an internal notch, but a nodular graphite particle distributes stress more uniformly. The nodularity above 90% and the low carbide content were therefore not only tensile property requirements but also fatigue life requirements. The planetary carrier experiences cyclic loading during operation, and a small internal defect or irregular graphite particle can initiate a fatigue crack. By controlling sulfur, phosphorus, magnesium, rare earth, and inoculation, I reduced the probability of such defects in the ductile iron castings. The low sulfur and phosphorus also reduced the amount of brittle intercellular phases, which improved crack initiation resistance.
I further analyzed the solidification and feeding behavior using the concept of expansion pressure. During eutectic solidification, graphite precipitation causes volume expansion. The expansion pressure \(P_e\) can be qualitatively written as:
$$P_e = K_e \frac{\Delta V}{V_0}$$
where \(K_e\) is an effective stiffness of the mold-metal system, \(\Delta V\) is the volume change from graphite precipitation, and \(V_0\) is the initial volume. If the mold is rigid and the feeding path is closed at the right time, the expansion can compensate for liquid and solidification shrinkage. If the mold is soft, the expansion is lost by mold wall movement. If the feeding path remains open, the expansion pushes iron back into the riser and creates a shortage in the casting. I used sodium silicate sand to keep the mold rigid, and I used easy-break cores to close the feeding path at the correct time. This combination was a central part of my process design for sound ductile iron castings.
| Physical phenomenon | Process consequence | My control method |
|---|---|---|
| Liquid shrinkage | Volume deficit before eutectic solidification | Risers and open feeding channels |
| Graphite expansion | Volume increase during eutectic solidification | Rigid mold and timely riser closure |
| Mold wall movement | Loss of expansion compensation | Sodium silicate sand and rigid flask |
| Feeding channel closure | Prevents iron from flowing back to risers | Easy-break cores at riser necks |
| Slow cooling | Coarse microstructure and segregation | Controlled silicon, copper, and inoculation |
| Fast cooling | Carbide and chill tendency | Hot mold and controlled pouring temperature |
I inspected the final casting after cleaning. The surface was free from major burn-on, cold shut, and visible cracks. The riser and gate removal areas were sound after grinding. The attached test block showed a uniform microstructure with well-distributed graphite nodules and a predominantly pearlitic matrix. The tensile test bars failed in a ductile manner, which was consistent with the 7.0% elongation. The tensile strength of 797 MPa and yield strength of 444 MPa indicated that the material had a good combination of elastic load capacity and plastic reserve. The nodularity above 90% and the low carbide content were also consistent with the measured elongation. I considered these results a clear validation of the composition design and casting process.
I also reviewed the sensitivity of the final properties to processing changes. A decrease in carbon equivalent can increase carbide tendency and reduce elongation. An increase in carbon equivalent can increase graphite flotation risk and reduce strength. An increase in silicon can increase ferrite and reduce tensile strength. A decrease in silicon can increase carbides and reduce elongation. An increase in manganese can increase pearlite but also increase segregation and carbides. A decrease in manganese may reduce strength if the pearlite content falls. An increase in residual magnesium can improve nodularity up to a point, but excess magnesium increases carbides and dross. An increase in residual rare earth can cause chunky graphite. An increase in copper can increase pearlite and strength, but too much copper can reduce elongation and machinability. The final process window was therefore a balance among these competing effects.
| Variable change | Likely microstructural effect | Likely property effect | My preferred control |
|---|---|---|---|
| Carbon equivalent too low | Carbide tendency increases | Elongation decreases | Keep CE near 4.3% to 4.4% |
| Carbon equivalent too high | Graphite flotation risk increases | Strength and uniformity decrease | Limit carbon and silicon |
| Silicon too low | Free cementite forms | Elongation decreases | Keep post-treatment Si above 2.0% |
| Silicon too high | Ferrite increases | Tensile strength decreases | Keep post-treatment Si below 2.3% |
| Manganese too high | Intercellular carbides and segregation | Elongation decreases | Keep Mn below 0.5% |
| Residual Mg too low | Nodularity decreases | Strength and elongation decrease | Keep Mg residual above 0.04% |
| Residual Mg too high | Carbides and dross increase | Elongation decreases | Keep Mg residual below 0.06% |
| Rare earth too high | Chunky graphite risk increases | Strength and elongation decrease | Keep RE residual below 0.02% |
| Copper too low | Pearlite fraction decreases | Tensile strength decreases | Keep Cu above 0.5% |
| Copper too high | Matrix becomes too hard | Elongation decreases | Keep Cu below 0.8% |
I used the final measured properties to back-calculate the achieved process capability. The tensile strength margin was 13.9% above the target:
$$\frac{797 – 700}{700} \times 100\% = 13.9\%$$
The elongation margin was 40.0% above the target:
$$\frac{7.0 – 5.0}{5.0} \times 100\% = 40.0\%$$
These margins are important for a nuclear gearbox component because production variability, section thickness variation, and trace element variation can shift properties. A material that just meets the target has little safety margin. The developed ductile iron castings had enough margin in both strength and elongation to support reliable service. The nodularity above 90% and the predominantly pearlitic matrix provided a microstructure that can be reproduced with the same process window.
| Performance index | Calculation | Result |
|---|---|---|
| Tensile strength margin | \((797 – 700)/700 \times 100\%\) | 13.9% |
| Elongation margin | \((7.0 – 5.0)/5.0 \times 100\%\) | 40.0% |
| Yield ratio | \(444/797\) | 0.557 |
| Quality index | \(797 \times 7.0\) | 5579 MPa% |
| Carbon equivalent | \(3.65 + (2.04 + 0.018)/3\) | 4.336% |
| Nodularity requirement | Measured graphite area ratio | Above 90% |
I concluded that the successful development of these ductile iron castings depended on integrating chemistry, melting, spheroidizing, inoculation, molding, gating, riser design, simulation, pouring, and post-treatment into one controlled route. The chemistry provided a near-eutectic carbon equivalent, low sulfur and phosphorus, controlled magnesium and rare earth, and copper-strengthened pearlite. The melting practice provided clean liquid iron with stable composition. The spheroidizing and inoculation practice provided nodularity above 90% and low carbide content. The gating and riser design provided smooth filling and effective feeding. The rigid mold and easy-break cores used graphite expansion correctly. The controlled pouring and shakeout reduced shrinkage and residual stress. The attached test block confirmed a tensile strength of 797 MPa, a yield strength of 444 MPa, and an elongation of 7.0%. These values exceeded the internal target and represented a significant improvement over the previous 600-3 type ductile iron castings.
For future production, I would keep the same chemistry window and process controls, but I would continue to monitor carbon equivalent, residual magnesium, nodularity, and pouring temperature as the most critical variables. I would also maintain the late-stream inoculation practice because it helps preserve nodularity and reduces carbide formation in the heavy sections. I would continue to use simulation before each significant geometry change because the hot-spot pattern controls riser placement and feeding behavior. The developed route is suitable for high-power nuclear gearbox planetary carriers and similar heavy-section ductile iron castings that require high strength, high plasticity, and long service life.
The main technical achievement is that the produced ductile iron castings met the demanding combination of tensile strength above 700 MPa and elongation above 5.0% while maintaining a nodularity above 90% and a predominantly pearlitic matrix. The measured values, 797 MPa tensile strength, 444 MPa yield strength, and 7.0% elongation, demonstrated that the designed composition and casting process were effective. I consider this development a practical example of how careful process integration can improve the reliability of heavy-section ductile iron castings for nuclear power transmission equipment.