In modern industrial manufacturing, rolls are critical consumable components in rolling mills, directly impacting production efficiency and cost. The maintenance and usage costs of rolls account for 5% to 15% of total production expenses, making their performance—especially in terms of wear resistance, strength, red hardness, and toughness—a key factor in operational success. Over time, roll materials have evolved from conventional ductile iron and alloyed indefinite chilled iron to advanced compositions like needle-structured ductile iron, high-alloy indefinite chilled iron, high-chromium iron, cemented carbides, semi-high-speed steel, and high-speed steel. Similarly, production methods have progressed from traditional casting to sophisticated techniques such as centrifugal composite casting (CF), continuous pouring process (CPC), electro-slag remelting (ESR), hot isostatic pressing (HIP), and spray casting (OSPREY). Among these, centrifugal composite casting stands out due to its simplicity, high efficiency, cost-effectiveness, and ability to produce dense microstructures, making it the preferred method for roll manufacturing. As a researcher focused on metal forming and finite element simulation, I have explored the centrifugal casting process for high nickel-chromium infinitely chilled ductile iron rolls through numerical modeling, aiming to optimize parameters and reduce defects like residual stress, poor metallurgical bonding, segregation, and porosity. This article details my approach using ProCAST software, with emphasis on flow and temperature field analysis, and validates findings through trial production. Throughout, the significance of ductile iron casting is highlighted, as it forms the core of this advanced roll technology.
Centrifugal casting involves pouring molten metal into a rapidly rotating mold, where centrifugal force drives the metal to fill the mold and solidify into a shaped component. Based on the rotation axis orientation, it can be classified into horizontal, vertical, or inclined types; roll production typically employs horizontal centrifugal casting. This method leverages centrifugal force to enhance density and minimize defects such as gas holes, sand inclusions, and slag entrapment. For high nickel-chromium infinitely chilled ductile iron composite rolls, the process combines centrifugal casting for the working layer and gravity casting for the core. The working layer, rich in nickel, chromium, and molybdenum, offers high hardness, wear resistance, and thermal crack resistance, making it ideal for hot strip rolling lines. However, the complexity of centrifugal composite casting demands precise control of parameters, which traditional trial-and-error methods fail to address efficiently. Hence, numerical simulation has become indispensable for visualizing stress, temperature, flow fields, and grain growth during casting. In my work, I utilized ProCAST to simulate the centrifugal casting of the working layer, enabling a deeper understanding of the process dynamics and facilitating optimized production.
The numerical simulation begins with model establishment and parameter setting. I selected a ø475 mm × 680 mm hot strip finishing centrifugal composite roll as the study object, with its working layer made of high nickel-chromium infinitely chilled ductile iron. The chemical composition is crucial for material behavior, and I defined it in ProCAST’s material library based on typical values, as summarized in Table 1. The alloy’s thermophysical properties were calculated, revealing a solidus temperature of 1051 °C and a liquidus temperature of 1239 °C, which guide the simulation of solidification.
| Element | C | Si | Mn | Cr | Ni | Mo | S | P |
|---|---|---|---|---|---|---|---|---|
| Content | 3.43 | 1.3 | 0.85 | 1.80 | 4.33 | 0.35 | 0.015 | 0.06 |
Using UG software, I created a 3D model of the mold and exported it as an STL file for import into ProCAST. The mesh was generated with a size of 10 mm, resulting in 432,787 elements to ensure computational accuracy. Key assumptions and boundary conditions were set to reflect real-world conditions: the mold material was 40CrMoV5 hot-work die steel, assumed rigid with no plastic deformation; the molten ductile iron was treated as an incompressible fluid; the mold preheat temperature was 180 °C, coated with resin-coated sand; rotation was constant around the Z-axis at 800 rpm; pouring temperature was 1340 °C; pouring rate was 25 kg/s; heat transfer coefficients were 25 W/(m²·K) for mold-air interface and 3000 W/(m²·K) for metal-mold interface. These parameters are essential for simulating the ductile iron casting process realistically.
The centrifugal casting process was divided into two stages for accurate simulation: gravity flow and centrifugal flow filling. In the gravity flow stage, molten iron exits the pouring basin and moves under gravity and inertia along a parabolic trajectory until contacting the mold inner wall. This stage is critical for initial metal distribution. In the centrifugal flow filling stage, upon contact, the iron is subjected to centrifugal force, gravity, and friction, causing it to spread circumferentially and axially along the mold wall. The flow behavior can be described by fundamental equations. For instance, the centrifugal force per unit volume is given by:
$$ F_c = \rho \omega^2 r $$
where \( \rho \) is the density of the molten iron, \( \omega \) is the angular velocity, and \( r \) is the radial distance from the rotation axis. The mold rotation speed is determined by the empirical formula:
$$ n = \beta \frac{55200}{\sqrt{r R}} $$
Here, \( n \) is the rotational speed in rpm, \( \beta \) is an adjustment coefficient (1.2 to 1.5), \( r \) is the specific weight of the alloy in N/m³, and \( R \) is the inner radius of the mold in mm. For this study, \( n \) was set to 800 rpm based on calculations. The flow field simulation revealed that molten iron initially moves toward the mold end closer to the sprue, then spreads axially to the opposite end, forming a layered structure. Table 2 summarizes key flow characteristics at different time intervals, illustrating the filling progression.
| Time (s) | Filling Percentage (%) | Flow Behavior Description |
|---|---|---|
| 1.1 | 4.3 | Iron contacts mold wall; begins circumferential motion |
| 8.0 | 31.1 | Layered flow along wall; axial spread initiated |
| 13.3 | 50.0 | Uniform circumferential distribution; thickness varies |
| 16.7 | 62.2 | Thickness homogenization; near-sprue region thicker |
| 23.1 | 85.2 | Full axial coverage; thickness becoming uniform |
| 27.1 | 99.3 | Nearly complete filling; smooth surface achieved |
The temperature field simulation provided insights into solidification patterns. Due to the mold’s lower preheat temperature, chilling occurred at the metal-mold interface, causing rapid cooling near the sprue area. The temperature distribution followed Fourier’s heat conduction law:
$$ \frac{\partial T}{\partial t} = \alpha \nabla^2 T $$
where \( T \) is temperature, \( t \) is time, and \( \alpha \) is thermal diffusivity. Cooling progressed symmetrically from both ends and surfaces toward the center, with axial ends solidifying before the middle and radial solidification proceeding from the outer surface inward. This directional cooling promotes columnar grain growth, which is desirable for wear resistance in ductile iron casting. The solidification sequence is quantified by the solid fraction, which evolved over time: 0% at 28.9 s, 51.2% at 466.9 s, 94.8% at 1136.9 s, and 100% at 1369.9 s. These results indicate that the outer surface solidifies first due to higher heat transfer, while the inner surface retains heat longer, leading to a temperature gradient conducive to dendritic growth.
To validate the simulation, I oversaw trial production of the high nickel-chromium infinitely chilled ductile iron roll. The centrifugal casting was performed on a horizontal centrifugal machine. The mold, coated with resin-coated sand, was heated to 300 °C and mounted on rollers at a pre-pouring temperature of 150–180 °C. Molten iron at 1340 °C was poured into the rotating mold at 800 rpm for 28 s, forming a 90 mm thick working layer. To prevent oxidation, a mixture of “O” type glass slag and N-B slag was added post-pouring. After cooling below the crystallization temperature, the machine was stopped, and the mold was vertically arranged for gravity casting of the ductile iron core. Inoculation with 0.5–3 mm 75% ferrosilicon was applied during core pouring. The roll was cooled for over 50 hours, then rough machined and inspected.
The produced roll showed excellent metallurgical bonding between the working layer and core, with no surface defects like gas holes or slag. The working layer hardness measured 76 HSD, meeting technical requirements. Microstructural analysis of samples taken 10 mm from the surface revealed a dendritic crystal structure, characteristic of centrifugal ductile iron casting. The graphite spheroidization was grade 1 (spheroidization rate ≥95%), and graphite size was grade 8 (≤0.015 mm). The matrix consisted of bainite, martensite (minor), retained austenite, graphite, and carbides, with pronounced directional solidification due to chilling and composition undercooling. This aligns with the simulation predictions, confirming the accuracy of the numerical model.
The dendritic growth in the working layer can be explained by the Mullins-Sekerka instability theory, where composition undercooling drives interface instability. For a binary alloy, the undercooling condition is:
$$ G < m C_0 (1 – k_0) / D $$
where \( G \) is the temperature gradient, \( m \) is the liquidus slope, \( C_0 \) is the initial composition, \( k_0 \) is the partition coefficient, and \( D \) is the diffusion coefficient. In high nickel-chromium ductile iron, elements like Cr, Mn, and Mo increase undercooling, promoting dendritic morphology. Additionally, centrifugal force enhances segregation control, as described by the Taylor-Proudman theorem for rotating fluids, which suppresses turbulence and ensures layered solidification. These principles underscore the complexity of ductile iron casting under centrifugal conditions.
Further analysis involved quantifying process parameters to optimize performance. Table 3 lists key simulation outputs and their implications for quality control. By correlating these with trial results, I refined the casting protocol to minimize defects and enhance reproducibility.
| Parameter | Simulated Value | Effect on Roll Quality |
|---|---|---|
| Total Filling Time | 27.3 s | Ensures complete mold filling without premature solidification |
| Peak Temperature Gradient | ~150 °C/cm | Promotes columnar grain growth for hardness |
| Solidification Time | ~1370 s | Adequate for stress relief and bonding |
| Maximum Flow Velocity | 1.2 m/s | Prevents erosion and ensures uniform thickness |
In conclusion, numerical simulation of centrifugal casting for high nickel-chromium infinitely chilled ductile iron rolls provides valuable insights into flow and temperature dynamics. The two-stage filling process—gravity flow followed by centrifugal flow—ensures uniform layer formation, while directional solidification yields a dendritic structure with enhanced properties. The ProCAST-based model accurately predicted outcomes, validated by successful trial production with a hardness of 76 HSD and defect-free surfaces. This approach underscores the importance of ductile iron casting simulation in modern manufacturing, enabling parameter optimization, cost reduction, and improved roll performance. Future work could explore advanced alloys or multi-physics coupling for even greater precision in centrifugal casting applications.