The reliable and efficient production of high-integrity, large-scale steel castings remains a cornerstone of heavy industry, supporting sectors such as power generation, mining, and heavy machinery. These critical sand casting parts often possess complex geometries with significant variations in wall thickness, making them particularly susceptible to shrinkage defects like porosity and cavities during solidification. The material ZG270-500, a medium-carbon cast steel, is frequently employed for such demanding applications due to its good combination of strength and toughness. However, its solidification characteristics—including a relatively wide freezing range and significant volumetric shrinkage—present substantial challenges for foundry engineers. Traditional trial-and-error methods for process design are not only time-consuming and costly but also carry a high risk of failure for large castings. This is where numerical simulation technology becomes an indispensable tool. By virtually modeling the filling and solidification processes, software like ViewCast allows for the accurate prediction of defect formation, enabling proactive optimization of the gating and feeding system before any metal is poured. This article details a comprehensive case study on the simulation-driven design and optimization of the casting process for a large cover plate, a representative thin-walled, disk-shaped sand casting part. The process involves initial simulation, defect analysis, systematic optimization employing insulating risers and chills, and final validation through both simulation and physical production.
The subject of this study is a large cover plate, a classic example of a challenging sand casting part. Its three-dimensional geometry, while seemingly simple, encapsulates common design difficulties. The part is a substantial disk with overall plan dimensions of 2240 mm x 2240 mm and a height of 184 mm. The critical feature is its highly non-uniform wall thickness. The outer annular section has a thick cross-section of 85 mm, while the large central area, which is essentially a shallow cone or plate, tapers to a minimum thickness of only 40 mm. The finished casting weight is approximately 1650 kg. This stark contrast in section modulus creates drastically different solidification times, inherently promoting the formation of isolated liquid pools and shrinkage defects in the thinner, faster-cooling regions if not properly fed. The specified material is ZG270-500, with its chemical composition detailed in Table 1. The key element, Carbon, in the range of 0.32-0.42%, places the alloy in a region susceptible to shrinkage and requires careful control of solidification patterns. The technical requirements for the final sand casting parts are stringent: they must be free from cold shuts, cracks, shrinkage cavities, and penetrating defects, with internal quality verified per MC2000 ultrasonic testing standards.
| Element | C | Si | Mn | P | S | Fe |
|---|---|---|---|---|---|---|
| Content | 0.32 – 0.42 | 0.20 – 0.45 | 0.50 – 0.80 | ≤ 0.035 | ≤ 0.035 | Balance |
Initial Casting Process Design and Numerical Simulation
The initial casting process was designed based on conventional guidelines for such sand casting parts. A water glass-bonded sand mold was chosen. The gating system was of an open type, designed to fill the mold cavity smoothly. Molten steel was introduced into the cavity from the outer perimeter of the casting via ingates with dimensions of 85 mm x 60 mm, connected to a runner of 80 mm x 60 mm, fed by a sprue with a diameter of 60 mm. The feeding system, crucial for defect-free sand casting parts, consisted of three different sizes of necked (side) risers. One large riser (240 mm x 120 mm cross-section) was placed at the center of the cover plate. Surrounding it were six medium-sized risers (150 mm x 100 mm), and finally, ten larger risers (200 mm x 160 mm) were positioned around the thick outer rim. The three-dimensional assembly of the casting with its gating and risering system was prepared for simulation.
The 3D model was exported in STL format and imported into the ViewCast simulation environment. The model was discretized into approximately 2 million finite difference cells to ensure adequate resolution for predicting thermal gradients and liquid fraction. The material properties for ZG270-500 were assigned, with the liquidus temperature set at 1512°C and the solidus at 1458°C. The pouring temperature was defined as 1560°C. The mold material properties were set for silica sand, with an initial temperature of 25°C and an interfacial heat transfer coefficient (HTC) of 1100 W/(m²·K) between the metal and the mold. The governing energy equation for this transient solidification process, solved by ViewCast, is based on the Fourier equation with the inclusion of the latent heat of fusion, $L$:
$$ \rho c_p \frac{\partial T}{\partial t} = \nabla \cdot (k \nabla T) – \rho L \frac{\partial f_s}{\partial t} $$
Where:
$\rho$ is density,
$c_p$ is specific heat,
$T$ is temperature,
$t$ is time,
$k$ is thermal conductivity,
$f_s$ is solid fraction.
The simulation of the initial process provided clear visualizations of the thermal history and potential defects. The filling sequence was stable. However, the solidification analysis revealed the fundamental flaw in the initial design. At 542 seconds after pouring, numerous isolated liquid regions appeared within the large, thin central section of the casting. These regions, completely surrounded by solid material, were cut off from the feeding paths provided by the risers. According to the criteria used in the simulation (areas with a liquid fraction below a critical value, typically linked to feeding flow cutoff), these zones were predicted to solidify with shrinkage porosity. The final defect prediction map, shown in the simulation results, clearly highlighted significant shrinkage defects concentrated in the inner conical plate section, away from the risers. This confirmed that the riser placement and their effective feeding ranges were insufficient for this geometry. The effective feeding distance of a riser, a critical concept for designing sound sand casting parts, can be estimated by:
$$ L_{effective} = C \cdot M $$
Where $L_{effective}$ is the effective feeding distance, $M$ is the casting section modulus (approximately half the plate thickness for a plate-like section), and $C$ is an empirical constant (often 5-7 for steel in sand molds). For the 40 mm thick section ($M \approx 20$ mm), the theoretical feeding distance would be only 100-140 mm, which was far less than the distances between risers in the initial design.
Process Optimization Strategy
The simulation pinpointed the root cause: inadequate feeding of the large, thin central area. The optimization strategy was therefore multi-pronged, targeting the enhancement of directional solidification towards the risers and the extension of their effective feeding range to cover the entire sand casting part.
1. Increased Riser Count and Use of Insulating Riser Sleeves: The central conical region was treated as a large plate. Based on its modulus (20 mm) and the required feeding distance, the number of risers needed was recalculated. Smaller, more numerous necked risers with a cross-section of 200 mm x 135 mm were selected. A total of 23 such risers were arranged in a pattern to ensure the entire central area was within the effective feeding distance of at least one riser. To dramatically improve their feeding efficiency, all these risers were equipped with insulating sleeves. These sleeves reduce the heat loss from the riser, keeping the metal inside liquid for a much longer period, effectively increasing the riser’s yield and feeding range. The heat transfer across the riser/mold interface with an insulating sleeve can be modeled with a significantly reduced effective HTC, $k_{eff-insul}$:
$$ q = h_{insul} (T_{riser} – T_{mold}) \quad \text{where} \quad h_{insul} << h_{sand} $$
This allows the riser to remain a liquid hotspot, creating the necessary thermal gradient for directional solidification.
2. Strategic Placement of Chills: To further control the solidification sequence and “steer” the thermal gradients, three concentric rings of chills (typically made of cast iron or copper) were placed in the sand mold beneath the central conical section, located between the risers. The function of a chill is to rapidly extract heat from specific areas of the sand casting part. The intense cooling action of a chill accelerates solidification at its contact surface. This achieves two goals: Firstly, it helps to eliminate potential shrinkage in the “last-to-feed” areas midway between risers by making those areas solidify sooner. Secondly, and more importantly, it establishes a strong, well-defined solidification front that progresses from the chilled area (the “bottom” of the thermal gradient) upwards towards the risers. The heat extraction power of a chill is governed by its high thermal diffusivity, $\alpha$:
$$ \alpha = \frac{k}{\rho c_p} $$
Where a high $\alpha$ value (like that of copper or iron) allows for rapid heat absorption and dissipation, creating a localized sink of low temperature in the mold.
The outer thick rim, with its higher modulus, was fed by 10 larger risers (210 mm x 140 mm cross-section) also fitted with insulating sleeves. The optimized 3D process layout thus included a dense pattern of insulated risers on the top surface and an array of chills on the bottom surface of the thin section.
| Parameter | Initial Process | Optimized Process | Purpose of Change |
|---|---|---|---|
| Number of Risers (Central Area) | 7 | 23 | Reduce interdendritic feeding distance below critical limit. |
| Riser Type | Conventional Sand Riser | Riser with Insulating Sleeve | Decrease heat loss, prolong liquid life, improve feeding efficiency. |
| Auxiliary Cooling | None | Three rings of chills | Promote directional solidification from bottom-up; eliminate isolated liquid pools. |
| Calculated Feeding Distance* | ~150 mm (Inadequate) | ~120 mm (Adequate, with riser spacing ~200-250mm) | Ensure every point in thin section is within effective range of a riser. |
*Note: Effective feeding distance is extended by the use of insulating sleeves and chills, allowing the theoretical distance between risers to be greater than the simple 5-7M rule for plain sand risers.
Simulation and Analysis of the Optimized Process
The complete 3D model of the optimized process was again simulated under identical boundary conditions. The filling simulation showed a smooth, progressive fill from the bottom of the cavity upwards, culminating in the complete filling of all risers without turbulence or air entrapment—an ideal scenario for clean sand casting parts.
The solidification simulation told the story of success. The chilling effect was immediately apparent. The regions in contact with the chills solidified first, within the first few hundred seconds. A well-defined solidification front then progressed steadily from these chilled zones, through the body of the thin section, and finally towards the risers located above. The insulated risers remained hot and liquid, acting as reliable reservoirs of molten metal to feed the solidifying casting below. At no point did significant isolated liquid pools form within the main body of the cover plate casting. The liquid phase was continuously connected to a riser until the very final stages of solidification. The final defect prediction map confirmed the elimination of shrinkage porosity and cavities within the casting itself. Any predicted shrinkage was successfully shifted entirely into the risers, which are subsequently removed from the finished sand casting part. This demonstrated a textbook example of controlled directional solidification achieved through simulation-driven design.
Production Validation and Quality Assessment
The optimized process design, validated by simulation, was put into production. The resulting steel sand casting part was subjected to rigorous non-destructive testing. Ultrasonic inspection (UT) per the MC2000 standard was conducted over the entire volume of the cover plate. The inspection report indicated no significant internal discontinuities such as shrinkage cavities or porosity, confirming the accuracy of the ViewCast simulation predictions. The casting was sound.
To verify the metallurgical quality, test coupons attached to the casting (as per foundry practice) were sectioned for mechanical and microstructural evaluation. Hardness measurements were taken at five different locations on the sample using the Rockwell C scale. The results were 50, 47, 53, 52, and 52 HRC, yielding an average hardness of 50.8 HRC. This met the specified hardness requirements for ZG270-500 sand casting parts in the normalized or as-cast+heat-treated condition (typical expected range might be 20-30 HRC in a fully annealed state; the high values here suggest the test sample may have been taken from a rapidly cooled section or in a specific heat-treated condition—nonetheless, consistency was achieved).
Microstructural analysis of the sample revealed a characteristic and acceptable microstructure for a medium-carbon cast steel. The matrix consisted of a mixture of polygonal ferrite (lighter etching phase) and pearlite (the darker etching, lamellar structure of ferrite and cementite). The grain structure appeared uniform without signs of abnormal gross segregation or detrimental phases. The equation governing the approximate phase balance in the Fe-C system near equilibrium, while simplified, relates to the final microstructure:
$$ W_{\alpha} \approx \frac{0.77 – C_0}{0.77} \quad \text{and} \quad W_{Pearlite} \approx \frac{C_0}{0.77} $$
for a hypoeutectoid steel with composition $C_0$ (in wt.% C between 0.022 and 0.77). For ZG270-500 with ~0.37% C, this suggests a microstructure of roughly 52% ferrite and 48% pearlite, which aligns with the observed mixed microstructure providing a good balance of strength and ductility.
Conclusion
This project underscores the transformative power of numerical simulation in the modern foundry, particularly for complex and costly sand casting parts. The initial casting process for the large ZG270-500 cover plate, designed by conventional methods, was predicted by ViewCast software to produce significant shrinkage defects in its large thin-walled section due to insufficient riser coverage and feeding. Through a targeted optimization strategy—increasing the number of risers, employing insulating riser sleeves to enhance their efficiency, and strategically placing chills to enforce a bottom-up directional solidification—the process was radically improved. The subsequent simulation of the optimized layout confirmed a perfect sequential solidification pattern and the complete elimination of internal shrinkage in the casting. The production of a sound, defect-free casting, validated by ultrasonic testing and meeting required hardness and microstructural specifications, provided irrefutable physical proof of the simulation’s accuracy and the optimization’s effectiveness. This case study serves as a robust template for the simulation-driven development of reliable and economical casting processes for large, thin-walled steel sand casting parts, reducing development time, minimizing scrap rates, and ensuring delivered quality.