The production of high-quality rolls is a cornerstone of modern metal rolling operations. As a consumable component in direct contact with the workpiece under severe thermal and mechanical conditions, the roll’s performance dictates mill efficiency and product quality. Among various materials, high-nickel-chromium infinite chill nodular cast iron has emerged as a superior choice for hot strip mill finishing stands due to its exceptional combination of hardness, wear resistance, and thermal crack resistance. The centrifugal composite casting process is the predominant method for manufacturing such rolls, effectively combining a hard, wear-resistant working layer with a tough, ductile core. However, this process involves complex interactions between fluid flow, heat transfer, and solidification under high rotational speeds, making defect prediction and control challenging. Traditional trial-and-error methods are costly and time-consuming. Therefore, from my perspective as a researcher focused on metal forming and simulation, leveraging advanced numerical tools like finite element analysis (FEA) is not just beneficial but essential for modernizing and optimizing this critical manufacturing process. This article details a comprehensive study on simulating the centrifugal casting process for the working layer of such rolls, aiming to visualize and control the underlying physics to ensure defect-free production.
1. Introduction to Roll Manufacturing and Centrifugal Casting
Rolls are critical, consumable components in rolling mills, with their maintenance and replacement constituting 5% to 15% of production costs. Their performance, particularly regarding wear resistance, strength, hot hardness, and toughness, directly impacts mill utilization and product yield. The evolution of roll materials has progressed from ordinary nodular cast iron and alloy indefinite chill irons to advanced materials like high-nickel-chromium indefinite chill iron, high-chromium iron, and high-speed steels. Concurrently, manufacturing techniques have advanced from conventional static casting to sophisticated methods such as Centrifugal Casting (CF), Continuous Pouring Process for Cladding (CPC), and Electroslag Remelting (ESR). Centrifugal casting stands out due to its relative simplicity, high production efficiency, cost-effectiveness, and ability to produce dense, sound structures by utilizing centrifugal force to shape the molten metal.
Centrifugal casting involves pouring molten metal into a rapidly rotating mold. The metal is forced against the mold wall by centrifugal force, where it solidifies. For roll manufacturing, the horizontal centrifugal casting method is standard. The key advantages include:
- Improved Density: Centrifugal force helps feed molten metal during solidification, reducing shrinkage porosity.
- Reduced Inclusions: Lighter impurities and gases tend to migrate toward the inner bore, where they can be machined away.
- Directional Solidification: Favorable for achieving a columnar grain structure perpendicular to the mold wall, enhancing certain mechanical properties.
The composite roll process typically involves two stages: first, centrifugally casting the high-alloy working layer; second, after its solidification, assembling the mold with pre-heated core molds and gravity-pouring the ductile core iron to achieve a metallurgical bond.
2. Fundamentals of Numerical Simulation for Centrifugal Casting
The physics of centrifugal casting is governed by a coupled system of equations describing fluid flow, heat transfer, and solidification, all under a rotating reference frame. The primary forces acting on the molten nodular cast iron include centrifugal force, Coriolis force, gravity, and viscous forces. The general momentum equation (Navier-Stokes) in a rotating frame with angular velocity $\vec{\omega}$ is expressed as:
$$
\rho \left[ \frac{\partial \vec{v}}{\partial t} + (\vec{v} \cdot \nabla) \vec{v} \right] = -\nabla p + \mu \nabla^2 \vec{v} + \rho \vec{g} + \rho \left[ -2 \vec{\omega} \times \vec{v} – \vec{\omega} \times (\vec{\omega} \times \vec{r}) \right]
$$
where $\rho$ is density, $\vec{v}$ is velocity, $p$ is pressure, $\mu$ is dynamic viscosity, $\vec{g}$ is gravity, and $\vec{r}$ is the position vector. The terms $-2 \vec{\omega} \times \vec{v}$ and $-\vec{\omega} \times (\vec{\omega} \times \vec{r})$ represent the Coriolis and centrifugal accelerations, respectively. For high-speed rotation, the centrifugal force term dominates over gravity.
The energy equation, accounting for latent heat release during the solidification of the nodular cast iron melt, is:
$$
\rho c_p \frac{\partial T}{\partial t} + \rho c_p (\vec{v} \cdot \nabla T) = \nabla \cdot (k \nabla T) + \rho L \frac{\partial f_s}{\partial t}
$$
where $c_p$ is specific heat, $T$ is temperature, $k$ is thermal conductivity, $L$ is latent heat, and $f_s$ is the solid fraction. The simulation must track the evolving solid fraction, often modeled using relationships like the lever rule or Scheil equation for multicomponent alloys like high-Ni-Cr nodular cast iron.
The key process parameters that must be accurately defined in the simulation include:
| Parameter | Symbol | Role in Simulation |
|---|---|---|
| Mold Rotational Speed | $n$ | Determines centrifugal acceleration ($G = \omega^2 r / g$). Critical for proper mold filling and density. |
| Pouring Temperature | $T_{pour}$ | Initial condition for energy equation. Affects fluidity and solidification mode. |
| Mold Pre-heat Temperature | $T_{mold}$ | Boundary condition. Influences thermal shock, cooling rate, and initial solidification structure. |
| Interfacial Heat Transfer Coefficient (IHTC) | $h_{interface}$ | Governs heat flux between metal and mold. Highly variable, depends on air gap formation. |
| Pouring Rate / Time | $Q$, $t_{pour}$ | Affects the thermal history during filling and potential remelting of the initial solidified layer. |
3. Simulation Methodology and Model Setup
For this study, the focus was on simulating the centrifugal casting of the working layer for a Ø475 mm × 680 mm hot strip mill roll. The material was a high-Ni-Cr infinite chill nodular cast iron. Its nominal composition and key thermal properties, essential for simulation input, are summarized below.
| Element | C | Si | Mn | Cr | Ni | Mo |
|---|---|---|---|---|---|---|
| wt.% | 3.4 – 3.5 | 1.2 – 1.4 | 0.8 – 0.9 | 1.7 – 1.9 | 4.2 – 4.4 | 0.3 – 0.4 |
| Calculated Thermal Properties | ||||||
| Liquidus Temperature ($T_L$) | ~1240 °C | |||||
| Solidus Temperature ($T_S$) | ~1050 °C | |||||
A 3D CAD model of the cylindrical mold cavity was created and discretized into approximately 430,000 tetrahedral finite element cells. The following assumptions and boundary conditions were applied to make the complex problem computationally tractable:
- The molten nodular cast iron is a Newtonian, incompressible fluid.
- The mold (40CrMoV5 steel) is rigid and undergoes no plastic deformation.
- Mold rotation is constant around the horizontal (Z) axis at $n = 800$ rpm. This speed was derived from empirical formulas such as: $$ n = \beta \sqrt{\frac{55200 \gamma}{\rho R}} $$ where $\beta$ is an empirical factor (1.2-1.5), $\gamma$ is alloy specific weight, and $R$ is the inner mold radius.
- Initial temperatures: $T_{pour} = 1340°C$, $T_{mold} = 180°C$.
- Heat transfer coefficients: $h_{metal-mold} = 3000 \, \text{W/m}²\text{K}$ (during contact), $h_{mold-air} = 25 \, \text{W/m}²\text{K}$.
A critical aspect of simulating horizontal centrifugal casting is accurately modeling the two-stage filling process. The software’s standard centrifugal module often assumes the metal is instantly under centrifugal influence. In reality, Stage 1 involves gravity-driven flow from the pour cup down the sprue and a free-fall trajectory onto the rotating mold wall. Stage 2 begins upon impact, where the metal is dragged by the mold wall, forming a thin, spreading film under the combined action of centrifugal force, Coriolis force, and viscous friction. This two-stage approach was meticulously modeled to capture true filling patterns.
4. Simulation Results and Analysis
4.1 Fluid Flow and Filling Behavior
The simulation provided a clear visualization of the filling sequence. In the initial stage, the stream of molten nodular cast iron falls under gravity. Upon striking the rotating mold wall, it immediately experiences a high tangential drag force. The fluid then spreads axially along the mold wall in a layered fashion. The flow is not uniform from the start; the metal initially accumulates slightly more on the side closer to the pour point before progressively spreading to achieve a uniform thickness distribution along the entire mold length. The filling time was simulated to be approximately 27.3 seconds. The velocity vectors clearly showed the primary circumferential flow, with secondary axial flow components responsible for the uniform coverage. This detailed flow field analysis helps ensure that the entire mold cavity is filled progressively without excessive turbulence, which could lead to oxide entrapment.
4.2 Temperature Field and Solidification Evolution
The thermal analysis revealed critical insights into the solidification mechanism of the nodular cast iron layer. Immediately upon contact with the relatively cold mold (180°C), the metal at the interface experiences rapid chilling, causing a thin shell to solidify almost instantaneously. This creates a steep temperature gradient radially outward from the mold wall, a condition highly conducive to directional, columnar grain growth.
As filling continues, the incoming hot metal reheats the initially solidified shell to some extent, but the overall gradient is maintained. The cooling and solidification sequence can be summarized as follows:
- Radial Direction: Solidification progresses from the outer diameter (OD, mold wall) inwards towards the inner diameter (ID). The OD cools fastest due to direct mold contact, while the ID cools mainly through radiation and convection to the air inside the bore.
- Axial Direction: The two ends of the cylindrical sleeve solidify slightly earlier than the central region because they can lose heat through the mold end walls, creating a minor axial temperature gradient.
The evolution of the solid fraction ($f_s$) over time was tracked. The solidification time (from end of pour to complete solidification) was significantly longer than the fill time, taking over 1300 seconds in the simulation. The solidification front movement was largely planar from the OD to the ID, which is ideal for promoting the desired columnar structure and minimizing centerline shrinkage in the working layer. The thermal history at every point, calculated by the simulation, is the primary input for predicting the final microstructure, including graphite nodule count, carbide distribution, and matrix structure (e.g., amount of retained austenite, bainite, or martensite).
The significant temperature difference between the OD and ID during solidification can be quantified. Assuming simplified one-dimensional heat flow, the temperature distribution in the solidifying shell can be approximated by:
$$
T(r) = T_{mold} + (T_{L} – T_{mold}) \cdot \frac{\ln(r/R_{ID})}{\ln(R_{OD}/R_{ID})}
$$
for steady-state solidification in a cylindrical geometry, where $R_{OD}$ and $R_{ID}$ are the outer and inner radii of the solidified layer. This gradient drives the columnar growth.
5. Experimental Validation and Microstructural Correlation
Guided by the simulation insights, actual trial productions of the composite roll were conducted. The key process parameters—mold speed, pouring temperature, and mold pre-heat—were set as per the simulation inputs. After centrifugal casting of the working layer and subsequent gravity casting of the core, the rolls were heat treated and machined.
Non-destructive testing (Ultrasonic Examination) confirmed a sound metallurgical bond between the high-Ni-Cr nodular cast iron working layer and the ductile iron core, with no significant fusion defects, porosity, or inclusions at the interface or in the辊颈. The surface hardness of the working layer measured 76 HSD (approx. 58 HRC), meeting the technical specification. This successful trial validated the accuracy of the simulated thermal and flow parameters.
Metallographic samples were extracted from the working layer. The analysis confirmed the microstructure predicted by the simulation’s thermal history:
- Graphite Morphology: Well-formed, fine, spherical graphite nodules (Type I, size class 8) were uniformly dispersed. The high Ni and Si content, combined with effective inoculation during the pour, ensured excellent graphitization potential even in this alloy-rich nodular cast iron, countering the chilling tendency from Cr and Mo.
- Matrix Structure: The matrix consisted primarily of a fine bainitic structure with some martensite and retained austenite. The presence of alloy carbides (M₃C, M₇C₃) was also observed.
- Solidification Structure: Most significantly, a distinct dendritic/columnar crystal growth morphology was observed radiating from the OD towards the ID. This is a direct result of the high thermal gradient and constitutional supercooling present at the solidification front, as captured in the simulation. The equation for constitutional supercooling is:
$$
G_L / v < m_L C_0 (1-k_0) / (k_0 D_L)
$$
where $G_L$ is the temperature gradient in the liquid, $v$ is growth velocity, $m_L$ is the liquidus slope, $C_0$ is alloy composition, $k_0$ is the partition coefficient, and $D_L$ is the diffusion coefficient. The high gradient $G_L$ established by the centrifugal casting process promotes planar or cellular/dendritic growth, leading to the observed oriented structure. This directional solidification enhances properties like hot strength and thermal fatigue resistance along the critical radial direction.
| Simulation Prediction | Experimental Observation | Conclusion |
|---|---|---|
| Steep radial thermal gradient from OD to ID. | Columnar/dendritic grain growth observed from OD towards ID. | Simulation accurately predicted the driving force for directional solidification. |
| Uniform filling and absence of hot spots. | No major shrinkage cavities or segregation defects in the working layer. | Flow and thermal fields led to sound feeding and progressive solidification. |
| Appropriate thermal cycle for bonding. | Excellent metallurgical bond at the interface, confirmed by UT. | Simulated pre-heat and pouring parameters ensured proper fusion. |
6. Conclusion and Outlook
This integrated study demonstrates the powerful synergy between numerical simulation and practical manufacturing in the field of advanced roll production. The finite element simulation of the horizontal centrifugal casting process for high-nickel-chromium infinite chill nodular cast iron provided an unprecedented view into the transient phenomena of filling and solidification. Key findings include:
- The filling process is effectively divided into a gravity-dominated stage and a subsequent centrifugal-force-dominated spreading stage, which must be correctly modeled for accuracy.
- The process establishes an intense radial thermal gradient, forcing directional solidification that results in a columnar/dendritic microstructure in the nodular cast iron working layer. This structure is beneficial for the in-service performance of the roll.
- The simulation acts as a reliable digital prototype, allowing for the optimization of critical parameters—such as pour temperature, mold speed, and pre-heat—before any metal is melted, thereby reducing development cost, time, and material waste.
- The successful production of rolls meeting all quality specifications, based on simulation-guided parameters, provides strong validation for this virtual engineering approach.
Future work will focus on even more sophisticated multi-physics coupling, including:
– Modeling the stress evolution during solidification and cooling to predict residual stresses and distortion.
– Directly simulating microstructure evolution (e.g., graphite nodule count, phase fractions) using cellular automaton (CA) or phase-field methods coupled with the thermal-calculation results.
– Optimizing the gating system design for the initial gravity-fall stage to further minimize turbulence.
– Extending the simulation to the subsequent gravity casting of the core to analyze the remelting and bonding dynamics at the interface in greater detail.
In conclusion, the application of numerical simulation is transformative for the centrifugal casting of high-performance nodular cast iron rolls. It moves the industry from empirical reliance to a science-based, predictive manufacturing paradigm, ensuring the production of superior, reliable components for demanding industrial applications.