In the field of mechanical engineering, the beam structure is a critical load-bearing component widely used in machine tools, cranes, and various heavy-duty equipment. As a structural part, the beam often experiences substantial loads under real working conditions, which imposes strict requirements on the quality of the cast component. Steel castings, particularly those made of medium carbon steel grades, offer an excellent combination of high strength, good plasticity, and toughness while maintaining relatively low production costs. These properties make steel castings highly suitable for such demanding applications. However, the production of steel castings is inherently challenging due to the high melting point of steel, limited superheat of the molten metal, wide solidification temperature range, and significant volumetric and linear shrinkage. Moreover, the fluidity of molten steel is relatively poor, which frequently leads to defects such as stress concentration, shrinkage cavities, shrinkage porosity, deformation, and hot cracking during the casting process. These defects can severely deteriorate the mechanical properties of steel castings. Therefore, the design of the casting process for steel castings demands rigorous engineering judgment and careful control.
Traditional casting process design relies heavily on the experience of design engineers and often requires multiple trial-and-error iterations to validate the rationality of the process. This approach is not only time-consuming but also costly, making it increasingly inadequate for modern production demands. With the advent of computational simulation tools, it is now possible to numerically model the flow field, temperature field, and stress field during the casting process. These simulations can predict the locations and extents of defects in steel castings, thereby providing valuable technical guidance for process improvement and enabling the production of high-quality castings. In this study, the View Cast software was employed to simulate the solidification process of a steel castings beam produced by a local foundry. The original process was analyzed, and based on the simulation results, an optimized process was developed to enhance the quality and yield of steel castings.
Initial Casting Process Design
Analysis of the Beam Casting
The three-dimensional solid model of the beam casting is relatively simple, with uniform wall thickness. The overall dimensions of the casting are approximately 1900 mm × 260 mm × 190 mm, classifying it as a medium-sized casting. Since the component is required in large quantities, sand casting with self-hardening sodium silicate sand was selected as the molding method. Because the casting is elongated, a two-up molding technique was adopted to improve production efficiency. The material of the casting is designated as ZG270-500 (analogous to ZG35), a medium carbon steel. Its chemical composition is listed in Table 1, with the balance being iron. During solidification, this steel undergoes a peritectic transformation, which is accompanied by considerable linear contraction and a high tendency for shrinkage porosity and shrinkage cavities. The casting is not permitted to have defects such as cold shuts, cracks, shrinkage cavities, or through-wall discontinuities. Furthermore, ultrasonic inspection according to the MC2000 standard is mandatory to ensure the internal quality of steel castings.
| Element | C | Si | Mn | P | S | Mo | Cr | Ni |
|---|---|---|---|---|---|---|---|---|
| Content | 0.32–0.42 | 0.20–0.45 | 0.50–0.80 | ≤0.04 | ≤0.04 | ≤0.20 | ≤0.35 | ≤0.30 |
Design of the Gating System
For steel castings, the solidification temperature range is relatively narrow, while the melting point is high, fluidity is poor, shrinkage is large, and oxidation tendency is strong. Therefore, the gating system must be simple in configuration, large in cross-sectional area, and capable of filling the mold quickly and smoothly with an appropriate rise in liquid level. A stopper ladle is commonly employed for pouring because of its excellent slag-removal capability. An open gating system was chosen for this purpose, as it does not require high dross-trapping ability but should fill the mold rapidly and steadily. The dimensions of the gating system components were determined by first selecting the nozzle diameter. Based on the mass of steel castings, a nozzle diameter of 45 mm was selected. According to the nozzle diameter, the cross-sectional areas of the gating system components were proportionally determined. The ratio of the total cross-sectional areas of the nozzle (ΣA孔), inner gates (ΣA内), runner (ΣA横), and sprue (ΣA直) was taken as:
$$ \Sigma A_{\text{孔}} : \Sigma A_{\text{内}} : \Sigma A_{\text{横}} : \Sigma A_{\text{直}} = 1 : (1.8 \sim 2.0) : (1.8 \sim 2.0) : (2.0 \sim 2.5) $$
The resulting cross-sectional shapes and dimensions of the runner, sprue, and inner gates are illustrated in the initial design. The schematic representation of the gating system is consistent with the dimensions mentioned above.
Riser Design
Risers should be placed at the highest and last-solidifying regions of the casting. Simultaneously, measures must be taken to establish a sequential solidification path from the casting toward the riser. Risers should avoid locations of stress concentration to prevent increasing thermal gradients that may cause deformation or cracking. Preferably, risers should be positioned on machined surfaces to minimize subsequent finishing work on steel castings.
Analyzing the structure of the beam, the left and right ends and the two reinforcing ribs are the thick sections where hot spots are likely to develop. These areas require risers to provide adequate feeding. Using the UG software, the volumes and surface areas of these sections were measured, and the modulus of the end regions was calculated as 1.9 cm, while the modulus at the reinforcing ribs was 2.4 cm. Based on the casting handbook, the riser dimensions were determined. For the ends, cylindrical risers with diameter d = 120 mm and height h = 180 mm were selected. At the reinforcing ribs, due to the thin wall, top risers were difficult to install; therefore, side risers were designed. Each side riser fed two castings simultaneously, with a diameter d = 140 mm and height h = 280 mm. The three-dimensional arrangement of the original process is shown in the initial model.
Numerical Simulation of the Initial Process
The solid model of the casting was created in UG and then exported as an STL file, which was imported into View Cast for mesh generation. Finer mesh yields higher computational accuracy, but overly fine meshes significantly increase computation time. Thus, a mesh count of 2 million elements was selected for the casting. Prior to calculation, the thermal properties and boundary conditions were defined: the pouring temperature was set to 1550°C, and the initial sand mold temperature was 20°C.
After configuring the parameters, the solidification process was numerically simulated. Figure 4 demonstrates the liquid fraction distribution during solidification. In this representation, the transparent regions indicate areas that have already solidified. At t = 172 s, the edges of the beam began to solidify. At t = 352 s, the temperature had decreased further, and the gating system fully solidified. At t = 712 s, the thin wall regions between the ends and the reinforcing ribs were almost completely solidified. At t = 981 s, isolated liquid pools appeared in the middle of the casting. At t = 1132 s, the four top risers were nearly solidified, with a small amount of liquid remaining at the riser roots. At t = 1342 s, the solidification process was essentially complete, but small amounts of metal remained unfrozen at the junctions between the side riser necks and the reinforcing ribs. At this point, the risers lost their feeding capability.
The predicted shrinkage defects under the initial process are shown in Figure 5. Overall, the risers provided some feeding effect. The edges of the casting solidified first, and the solidification zone propagated from the thin sections toward the thick sections. The riser locations, having the largest modulus, solidified last, which indicates a basically sequential solidification pattern. Nevertheless, several regions required improvement. On one hand, because the casting is relatively long, some regions exceeded the maximum feeding distance of the risers. Consequently, these remote areas were not adequately fed, leading to shrinkage porosity. On the other hand, shrinkage cavities were observed at the connection between the riser and the casting, suggesting that the risers did not fully feed the casting. Therefore, the process must be modified.
Process Optimization and Re-simulation
Optimization Measures
Although the initial design achieved a generally progressive solidification, the feeding channels were interrupted in the middle section and in the thin-walled areas between the reinforcing ribs and the ends. As a result, shrinkage porosity appeared in these locations. The cause was likely the excessive distance from the risers. To address this issue, chill irons were added to these regions. Additionally, the shrinkage cavities at the riser necks indicated insufficient feeding capacity. To enhance feeding efficiency, the conventional risers were replaced with insulating and exothermic risers. The optimized three-dimensional model is depicted in Figure 6.
Simulation Results of the Optimized Process
The optimized model was converted into STL format and imported into View Cast. Meshing and parameter settings were performed similarly to the initial simulation. The filling process was simulated, and the results are shown in Figure 7, where the dark regions represent the metal that has filled the mold cavity. At t = 0.7 s, the molten steel began to enter the mold cavity through the inner gates. At t = 8.1 s, the gating system and the casting body were completely filled, and the liquid started rising into the risers. At t = 12.7 s, the entire mold cavity, including the risers, was fully filled. The filling simulation indicated that the mold filling was smooth and stable, with complete filling achieved. This satisfies the requirement of pouring steel castings at a relatively low temperature with rapid filling, demonstrating the rationality and feasibility of the optimized process.
The solidification simulation results for the optimized process are presented in Figure 8. At t = 158 s, the edges of the casting had lower temperatures and began to solidify, with the regions near the chill irons solidifying faster. At t = 338 s, the gating system had completely solidified. At t = 458 s, the solidified zones expanded outward from the chill irons, and the casting was divided into several isolated liquid regions. At t = 1148 s, the thick end sections had solidified fully, leaving only a small amount of liquid metal in the top risers. At t = 1478 s, the top risers were completely solidified, while two isolated liquid pools remained in the side risers. At t = 1598 s, the side risers were almost fully solidified, and the final solidification regions were successfully transferred into the risers. This indicates that the optimized design promotes a well-established sequential solidification pattern, which is essential for producing sound steel castings.
The predicted shrinkage defects for the optimized process are shown in Figure 9. It is evident that the shrinkage cavities that existed in the original process have basically disappeared. A few minor defects remain on the bottom surface of the casting, but these can be easily removed by welding repair during subsequent machining. Therefore, the optimized process effectively minimizes shrinkage-related discontinuities and significantly improves the internal integrity of steel castings.
Quality Inspection and Microstructural Analysis
Following the optimized process, actual castings were produced in the foundry. The appearance of a representative production casting is shown in Figure 10. Ultrasonic inspection confirmed that no obvious shrinkage cavities or porosity were present in the castings. To evaluate the mechanical and microstructural properties, test coupons were cut from the castings attached blocks. Hardness measurements were performed using the Rockwell C scale (HRC). The measured hardness values were 50, 47, 53, 52, and 52, with an average value of 50.8 HRC. This satisfies the required hardness specification for the steel castings.
Metallographic specimens were prepared and etched with a 4% nital solution. The resulting microstructure is shown in Figure 11. The microstructure consists of a typical hypoeutectoid steel structure: the light regions are ferrite, while the darker regions are lamellar pearlite. Notably, some proeutectoid ferrite appears as needles penetrating into the pearlite, forming a distinctive Widmanstätten structure. This morphology is primarily attributed to the coarse as-cast grain structure and the relatively rapid cooling rate during the austenite-to-ferrite transformation. The presence of Widmanstätten ferrite can significantly degrade the mechanical properties of steel castings, particularly plasticity and impact toughness, and it also raises the ductile-to-brittle transition temperature. To eliminate this undesirable structure, heat treatments such as normalizing, annealing, or forging are typically adopted. In severe cases, double normalizing may be required. For the current production, the mechanical properties still meet the acceptance criteria, but further heat treatment could further improve the ductility and toughness of the steel castings.
Discussion on Solidification and Feeding of Steel Castings
The solidification behavior of steel castings is governed by the cooling rate, the temperature gradient, and the modulus of the casting regions. The modulus (M) is defined as the ratio of volume (V) to cooling surface area (A):
$$ M = \frac{V}{A} $$
For a steel casting, the solidification time (ts) can be estimated by Chvorinov’s rule:
$$ t_s = k \left( \frac{V}{A} \right)^2 = k M^2 $$
where k is a constant that depends on the mold material, the pouring temperature, and the thermal properties of the metal. This relationship underscores the importance of designing risers with a larger modulus than the casting section they feed. In the initial design, the modulus of the ends was 1.9 cm, while the side riser modulus was calculated from its dimensions. For a cylindrical side riser with d = 140 mm and h = 280 mm, considering the effective cooling surface, the modulus is approximately:
$$ M_{\text{riser}} = \frac{V_{\text{riser}}}{A_{\text{riser}}} = \frac{\pi r^2 h}{2\pi r h + 2\pi r^2} = \frac{r h}{2h + 2r} $$
Substituting r = 70 mm and h = 280 mm gives:
$$ M_{\text{riser}} = \frac{70 \times 280}{2(280 + 70)} = \frac{19600}{700} = 28 \text{ mm} = 2.8 \text{ cm} $$
This modulus is greater than that of the reinforcing ribs (2.4 cm), which is favorable for feeding. However, the feeding distance of a riser is limited. For steel castings, the maximum feeding distance (Lf) can be approximated by:
$$ L_f = 4.5 T_{\text{max}} + t $$
where Tmax is the maximum section thickness and t is the thickness of the thin section. For the long beam, the distance from the end riser to the middle of the casting may exceed this limit, leading to shrinkage porosity. In the optimized design, chill irons were strategically placed to increase the cooling rate and extend the effective feeding distance. Chill irons increase the local solidification rate, thereby modifying the temperature gradient and promoting directional solidification. This allows the riser to feed a longer section without forming shrinkage defects.
Another important parameter in the production of steel castings is the pouring temperature. The pouring temperature affects fluidity, solidification time, and defect formation. The actual pouring temperature Tp is related to the liquidus temperature TL and the superheat ΔT:
$$ T_p = T_L + \Delta T $$
For ZG270-500, the liquidus temperature is approximately 1515°C. With a pouring temperature of 1550°C, the superheat is 35°C, which is appropriate for steel castings to ensure adequate fluidity while avoiding excessive superheat that could increase shrinkage and hot tearing tendencies. The filling time tf can be estimated from the total weight of the casting and gating system (G) and the effective cross-sectional area of the gating system (Ae):
$$ t_f = \frac{G}{\rho A_e v} $$
where ρ is the density of molten steel and v is the average velocity through the gates. In the optimized process, the filling time of 12.7 s ensures a rapid but smooth fill, minimizing heat loss and oxidation.
Comparative Analysis of Initial and Optimized Processes
To better illustrate the improvements achieved through process optimization, a comparative summary is presented in Table 2. The key changes include the addition of chill irons and the replacement of ordinary risers with insulating/exothermic risers. These modifications significantly altered the solidification sequence and reduced the tendency for shrinkage defects.
| Parameter | Initial process | Optimized process |
|---|---|---|
| Riser type at ends | Ordinary cylindrical riser, d=120 mm, h=180 mm | Insulating/exothermic cylindrical riser, d=120 mm, h=180 mm |
| Riser type at ribs | Ordinary side riser, d=140 mm, h=280 mm | Insulating/exothermic side riser, d=140 mm, h=280 mm |
| Chill irons | None | Added at middle and thin-wall sections |
| Filling time (s) | Not measured | 12.7 |
| Filling behavior | Not simulated | Smooth and stable |
| Last solidifying region | Some residual liquid at side riser necks | Completely transferred into risers |
| Shrinkage cavities | Present at riser roots and middle sections | Essentially eliminated |
| Shrinkage porosity | Significant in thin-wall and middle areas | Minor surface defects only |
| Process yield | Not reported | 72% |
| Quality acceptance | Failed ultrasonic inspection | Passed ultrasonic inspection |
The numerical simulation of steel castings not only identified problematic regions but also provided quantitative insights into the temperature distribution and liquid fraction evolution. For instance, the temperature gradient in the casting can be calculated from the simulated thermal field. The gradient G can be expressed as:
$$ G = \nabla T = \frac{\partial T}{\partial x} \mathbf{i} + \frac{\partial T}{\partial y} \mathbf{j} + \frac{\partial T}{\partial z} \mathbf{k} $$
A higher temperature gradient promotes directional solidification and reduces the propensity for shrinkage porosity. In the optimized process, the chill irons increased the local gradient near the center of the beam, thus enabling the risers to feed the casting more effectively.
The solidification shrinkage of steel is approximately 3% by volume. The shrinkage cavity volume (Vsh) can be estimated from the total volume of the liquid metal (VL) and the solidification contraction coefficient β:
$$ V_{sh} = \beta V_L $$
For steel, β ≈ 0.03. In a 1900 mm long beam weighing, say, 300 kg, the volumetric shrinkage would be about 0.03 × (300,000/7.8) ≈ 1,150 cm³. This substantial volume must be compensated by risers. Therefore, riser dimensions must be large enough to provide sufficient liquid metal. In the original process, the riser volume was insufficient for the entire length, causing defects. The exothermic risers in the optimized process not only have a larger effective feeding volume because they maintain the metal liquid for a longer time, but also their insulating sleeves reduce heat loss. The exothermic riser generates heat from chemical reactions, which slows down solidification and enables better feeding.
Role of Numerical Simulation in Steel Castings Production
The use of View Cast in this study demonstrates the immense potential of casting simulation software in the manufacturing of steel castings. Simulation allows engineers to visualize the filling and solidification processes that are otherwise invisible. It can predict defects such as shrinkage cavities, porosity, cold shuts, and misruns. Moreover, it enables virtual experimentation, so multiple process alternatives can be evaluated without the cost and time of physical trial castings. By systematically varying parameters such as riser size, neck geometry, chill iron placement, and pouring temperature, a robust process can be developed. This approach aligns with the modern trend of digital manufacturing and intelligent foundry.
For steel castings, the key to soundness lies in achieving controlled directional solidification. The modulus of risers must be greater than that of the casting sections they feed. A common rule of thumb is:
$$ M_{\text{riser}} \geq 1.2 M_{\text{casing}} $$
In our optimized design, the side riser modulus was 2.8 cm, which is 1.17 times the rib modulus of 2.4 cm, slightly below the recommended 1.2. However, the use of exothermic risers effectively increases the freezing ratio, because the exothermic material reduces the rate of heat extraction from the riser, making it behave as if it had a larger modulus. Therefore, the actual feeding performance was sufficient, as evidenced by the simulation and production results.
Additionally, the gating system for steel castings must ensure that the mold cavity is filled without causing turbulence and entrapping gases or inclusions. The Reynolds number at the gates can be estimated as:
$$ Re = \frac{\rho v d}{\mu} $$
where ρ is the density of molten steel (≈7000 kg/m³), v is the velocity, d is the gate diameter, and μ is the dynamic viscosity (≈0.006 Pa·s). To avoid turbulent flow, Re should be kept below 10,000 ideally. In our system, the open gating design with multiple gates distributes the flow and reduces the velocity, contributing to smooth filling.
Microstructure and Mechanical Properties of the Produced Steel Castings
The final quality of the produced steel castings was assessed through ultrasonic testing, hardness measurements, and metallographic examination. The hardness of 50.8 HRC is suitable for the intended service. However, the presence of Widmanstätten microstructure indicates that the as-cast structure is not fully optimized. To improve the mechanical properties, a normalizing heat treatment is recommended. Normalizing involves heating the steel castings to a temperature above the upper critical temperature (Ac3) and then cooling in air. For ZG270-500, the normalizing temperature is typically 850–900°C. This process refines the grain size and eliminates the Widmanstätten ferrite, resulting in a more uniform distribution of ferrite and pearlite. The improvement in impact toughness can be significant. If the application demands high impact resistance, quenching and tempering may be considered, although this would increase production costs.
In many industrial applications, the as-cast condition might be acceptable if defects are absent. However, for critical structural components, heat treatment is often mandatory to reduce residual stress and improve ductility. In the present case, the castings already meet the specified hardness, and the absence of internal defects means that they are acceptable for delivery. Nevertheless, additional heat treatment could further enhance the reliability of steel castings under dynamic loading.
Conclusions
In this study, a comprehensive numerical simulation and process optimization approach was applied to the production of steel castings, specifically a beam component. The following conclusions can be drawn:
- The initial casting process designed by conventional methods exhibited significant shrinkage defects, which were accurately predicted by the View Cast numerical simulation. The defects were located in the middle section and the thin-wall areas between the reinforcing ribs and the ends, primarily due to insufficient feeding distance and inadequate riser capacity.
- By adding chill irons at the central and thin-wall regions and replacing ordinary risers with insulating/exothermic risers, the solidification pattern was markedly improved. The optimized process achieved smooth and stable mold filling, with a filling time of 12.7 s, and exhibited a well-directed sequential solidification. The final liquid regions were successfully transferred into the risers, effectively eliminating shrinkage cavities and reducing shrinkage porosity to only minor surface imperfections.
- The simulation-guided optimization significantly enhanced the quality of steel castings. The production castings passed ultrasonic inspection, and the hardness values averaged 50.8 HRC, meeting the material specifications. The process yield reached 72%, demonstrating both technical and economic benefits.
- The as-cast microstructure of the steel castings exhibited ferrite, pearlite, and Widmmanstätten ferrite. A normalizing heat treatment is recommended to refine the grains and improve the impact toughness if required by the service environment.
- The successful application of numerical simulation in this work confirms that casting simulation is an indispensable tool for the modern foundry industry, particularly for the production of high-quality steel castings. It reduces trial-and-error costs and accelerates process development, thus providing a competitive advantage in manufacturing.
In summary, the integration of numerical simulation with conventional casting design practices enables foundries to produce sound steel castings with high efficiency and reliability. The methodology described herein can be readily adapted to other steel castings with complex geometries and stringent quality requirements.