The advancement of CAD/CAM/CAE technologies has propelled casting simulation into a critical tool for modern foundry engineering. It offers a scientific, predictive approach that significantly reduces the reliance on costly and time-consuming trial-and-error methods. This is particularly vital for producing complex, high-performance castings like ball mill liners. These components endure severe abrasive wear, impact forces, and corrosive environments, necessitating superior metallurgical quality and soundness to ensure longevity. The transition from traditional sand casting to more advanced methods is often driven by the need to eliminate defects such as burn-on, excessive casting stress, cracking, and distortion. The lost foam casting process presents a compelling alternative, offering the potential for improved dimensional accuracy and surface finish. This study employs the numerical simulation software ProCAST to investigate the filling and solidification behavior of a high-manganese steel ball mill liner produced via the lost foam casting process. The focus is on evaluating the impact of different molding media and optimizing the feeding system design to mitigate shrinkage porosity, thereby enhancing the final casting quality and performance.
The core principle of the lost foam casting process involves embedding a foam pattern of the desired casting within unbonded sand. Molten metal is then poured directly onto the pattern, which vaporizes and is replaced by the metal, precisely replicating the pattern’s shape. A key advantage is the elimination of parting lines and cores, allowing for greater design freedom. The process parameters, however, are complex and interrelated. Computer simulation provides an invaluable window into the transient events during casting, such as the coupled phenomena of foam degradation, gas evolution, metal flow, heat transfer, and solidification.
The ball mill liner under investigation has dimensions of approximately 500 mm in length, 314 mm in width, and a thickness ranging from 40 to 50 mm, with a mass of about 45-55 kg. Its relatively thin-walled and flat geometry presents specific challenges for feeding. Adding conventional side risers directly onto the casting body is impractical as it can introduce high stresses, leading to distortion and creating potential weak points. Therefore, the initial casting design employs a top-gating system where the pouring basin also acts as the primary riser, aiming to utilize gravitational force and thermal gradients for directional solidification towards this feeder.
The primary material for the liner is a modified ZGMn13 high-manganese steel, renowned for its exceptional work-hardening capability and impact abrasion resistance. Its typical chemical composition is summarized in Table 1.
| Element | C | Mn | Si | Cr | Mo | W | S | P | Re |
|---|---|---|---|---|---|---|---|---|---|
| Wt. % | 1.0-1.2 | 12.0-14.0 | 0.4 | 1.0-2.0 | 0.4-0.6 | 1.0 | < 0.5 | < 0.5 | 0.1–0.3 |
The success of the lost foam casting process heavily depends on the properties of the molding sand. Two types were considered: conventional silica sand and iron sand. Their thermo-physical properties critically influence permeability (for evacuating foam pyrolysis products) and cooling rate. The pattern material was expanded polystyrene (EPS) with a density of 25 kg/m³. Its thermal properties, including degradation behavior, are essential inputs for an accurate simulation. Key thermal parameters for the EPS pattern include a thermal conductivity ($k$) of 0.15 W/(m·K), a specific heat capacity ($C_p$) of 3.7 kJ/(kg·K), and a latent heat of vaporization ($L_v$) of 100 kJ/kg. The pattern begins to gasify at temperatures between 330–350°C. The process is conducted under a vacuum of 0.03–0.05 MPa to aid in the removal of decomposition gases and to stabilize the sand mold.
The three-dimensional model of the liner, along with its gating system and sand mold, was created using UGS NX 6.0. This geometry was imported into ProCAST and processed in its meshing module, MeshCAST. A fine mesh was necessary to capture the details of the thin liner section adequately. The final finite element mesh consisted of 23,894 nodes and 117,409 volume elements. After meshing, the crucial step of assigning material properties, boundary conditions, and initial conditions was performed. The interfacial heat transfer coefficients (IHTC) between the metal, the decomposing foam, and the sand were defined based on standard values for the lost foam casting process. The initial conditions included setting the mold and pattern temperature to ambient (20°C) and defining the pouring temperature as a variable for study. The key process parameters for the simulation are outlined in Table 2.
| Parameter | Value / Specification |
|---|---|
| Casting Material | ZGMn13 High-Manganese Steel |
| Mold Material (Options) | Silica Sand, Iron Sand |
| Pouring Temperature Range | 1480°C – 1520°C |
| Mold Temperature | 20°C |
| Environmental Temperature | 20°C |
| Pouring Time (Estimated) | 10-20 seconds |
| Vacuum Pressure | 0.03 – 0.05 MPa |
The initial gating design, employing a simple top-pour basin/riser, was simulated first. The shrinkage porosity prediction for this configuration is illustrated in the results. A significant concentration of shrinkage defects was predicted at the junction between the liner and the downsprue, as well as within the downsprue itself. This defect formation can be explained by analyzing the solidification sequence. The thinner downsprue, surrounded by sand, solidifies rapidly. Once solidified, it severs the liquid feed path from the main riser/basin to the thicker sections of the liner that are still solidifying. Consequently, isolated liquid pools within the liner, particularly in its central working surface, form microporosity as they contract without a source of feed metal. This renders the initial design unacceptable for a wear part where surface integrity is paramount.
To address this, the gating/feeding system was redesigned. The key modification was the introduction of a well-defined riser neck connecting a larger, more thermally efficient riser to the liner. This neck is designed to remain open and molten longer than the casting section it feeds, ensuring a continuous feed path until the casting is fully solidified. The modulus (Volume/Surface Area ratio) of the riser neck and riser must be calculated to be greater than that of the casting to be effective. The improved design was re-simulated. The results showed a dramatic improvement: the shrinkage porosity was successfully relocated from the critical working face of the liner into the riser neck and riser itself, which are later removed. This confirms the principle of directional solidification in the lost foam casting process, where the thermal geometry can be manipulated to ensure sound castings.
A comparative study was conducted to evaluate the influence of molding sand and pouring temperature on the filling behavior within the lost foam casting process. Four simulation schemes were set up, as detailed in Table 3. The primary metric for comparison was the total filling time and the stability of the metal front advancement.
| Scheme | Pouring Temp. (°C) | Mold Sand | Filling Time (s) | Filling Characteristic |
|---|---|---|---|---|
| 1 | 1480 | Silica Sand | 5.905 | Intermittent, unstable front |
| 2 | 1480 | Iron Sand | 4.496 | Smooth and stable |
| 3 | 1520 | Silica Sand | 6.086 | Intermittent, unstable front |
| 4 | 1520 | Iron Sand | 4.658 | Smooth and stable |
The results clearly demonstrate the superiority of iron sand for this lost foam casting process application. Despite variations in pouring temperature, schemes using iron sand consistently resulted in shorter and smoother filling compared to silica sand. This can be attributed to the higher thermal conductivity and better permeability of iron sand. The higher thermal conductivity leads to more rapid heat extraction from the metal front, but more importantly, it facilitates faster and more complete pyrolysis of the foam pattern directly ahead of the advancing metal. The superior permeability ensures that the gaseous and liquid decomposition products are efficiently evacuated through the sand under the applied vacuum. This prevents back-pressure buildup that can slow down or even halt the metal front, causing mistruns. The intermittent filling observed with silica sand is a classic symptom of inadequate permeability in the lost foam casting process. Therefore, for this liner geometry and material, iron sand at a pouring temperature of 1480°C was selected as the optimal combination.
The filling sequence for the optimized process (Scheme 2) reveals the unique dynamics of the lost foam casting process. Unlike conventional empty-cavity filling, the metal progressively replaces the decomposing foam. The simulation snapshots show that filling begins steadily but accelerates in the later stages. The initial slower pace is due to the significant heat sink effect of the cold foam and sand; the metal front loses heat to vaporize the foam, which requires absorbing its latent heat of vaporization ($Q_{vap}$).
$$ Q_{vap} = m_{foam} \cdot L_v $$
Where $m_{foam}$ is the mass of foam being vaporized per unit time. Once the initial thermal barrier is overcome and the system heats up, the filling rate increases. The final areas to fill are the riser neck and the main riser, ensuring they are the hottest sections of the system, which is ideal for feeding.
The solidification analysis is critical for defect prediction. The solidification sequence is visualized through fractional solid plots and solidification time contours. The results confirm the desired directional solidification. The thin edges and corners of the liner solidify first, followed by the bulk of the plate. The thermal center, last to solidify, is located within the riser neck and riser, not in the casting body. This is perfectly aligned with the concept of a thermally efficient feeding path established by an appropriate riser neck design in the lost foam casting process. The solidification time ($t_s$) at any point can be related to the local modulus ($M$) and the heat transfer conditions, often approximated by Chvorinov’s rule:
$$ t_s = B \cdot \left( \frac{V}{A} \right)^n = B \cdot M^n $$
Where $B$ is a mold constant dependent on material properties and interfacial heat transfer, $V$ is volume, $A$ is cooling surface area, and $n$ is an exponent (typically ~2). The riser neck was designed to have a larger modulus than the adjacent casting section, ensuring it remains liquid longer.
The final shrinkage porosity prediction for the optimized design shows a negligible fraction of porosity in the casting itself, with the vast majority of the shrinkage volume concentrated in the riser. The quantitative measure of porosity fraction ($f_{porosity}$) predicted by the simulation uses a criterion function (e.g., Niyama criterion) that combines local thermal gradients ($G$) and solidification rates ($\dot{T}$):
$$ Niyama = \frac{G}{\sqrt{\dot{T}}} $$
Regions where this value falls below a critical threshold are flagged as potential shrinkage porosity sites. In the optimized lost foam casting process simulation, the critical areas flagged were successfully isolated to the feeder system.
In conclusion, this numerical simulation study demonstrates the powerful role of CAE tools in optimizing the lost foam casting process for critical components like ball mill liners. Key findings are:
1. The initial top-pour design without a dedicated riser neck led to unacceptable shrinkage porosity on the liner’s working face due to premature feeding path isolation.
2. The introduction of a properly sized riser neck successfully redirected the thermal center and the associated shrinkage into the feeder, ensuring a sound casting. This highlights the universal importance of controlled directional solidification, even in the lost foam casting process.
3. A comparative analysis of molding sands proved that iron sand, with its higher thermal conductivity and permeability, provides significantly smoother and faster filling compared to silica sand for this application, reducing the risk of filling-related defects.
4. The optimized process parameters—iron sand mold, a pouring temperature of 1480°C, and the redesigned gating/feeding system—were virtually validated to produce a liner free from internal shrinkage defects.
5. The ProCAST software effectively simulated the coupled phenomena of foam degradation, fluid flow, heat transfer, and solidification inherent to the lost foam casting process, providing deep insights that guide practical engineering decisions without the need for physical trials.
This workflow underscores how numerical simulation can drastically reduce development time and cost, minimize material waste, and enhance product quality and reliability in advanced casting processes like lost foam casting.