Lost foam casting (LFC), also widely referred to as expandable pattern casting (EPC), is a near-net-shape forming technology that has been increasingly recognized for its process design flexibility, low production cost, lower environmental pollution, and high product quality. Over the past six decades, this technology has been successfully implemented in various industrial sectors, especially for cast iron and special alloy steels. However, its application in low-carbon steel fields remains relatively limited, predominantly due to the persistent issue known as carbon defect or carbon pick-up. With the growing demand for low-carbon steel castings produced by lost foam castings, solving this carbon defect problem has become a critical technical breakthrough. In this thesis, the research object is a low-carbon steel casting used for bridge bearings. The investigation integrates computational numerical simulation through the ProCAST software, production verification, and experimental studies on carbon defects, with the aim of optimizing the casting process while maintaining high quality and low production costs.
The research began with a thorough structural analysis of the bridge bearing product. This particular component serves as a critical connector for transferring loads between the bridge superstructure and substructure. It must withstand significant vertical loads while allowing rotational movement. The selected low-carbon steel casting, designated as ZGS10CrNiCu, is a newly developed corrosion-resistant steel characterized by its excellent low-temperature toughness and atmospheric corrosion resistance. The product geometry, with an overall dimension of 700 mm × 445 mm × 155 mm and a weight of approximately 262 kg, presents a varying cross-sectional thickness ranging from a minimal 30 mm to a maximum of 155 mm. This non-uniform thickness distribution, combined with the presence of four square blind holes and an arc-shaped concave surface, classifies the product as a moderately complex steel casting, demanding high internal quality standards (meeting ultrasonic inspection Level II).
1. Process scheme design for bridge bearing steel castings
The selection of casting parameters is particularly significant in the lost foam casting process, as the filling behavior is fundamentally different from that of conventional cavity casting. In lost foam castings, the molten metal gradually replaces the foam pattern, which decomposes under high temperatures. The design principles for the pouring system, gating system, and riser system were detailed in this study. Two distinct pouring schemes were initially proposed: horizontal pouring and vertical pouring.
For the gating system design, a combined bottom-side injection was selected to optimize the filling pattern. This choice allows the molten metal to enter the mold cavity from the side near the bottom, promoting a diagonal advancement of the liquid front. This strategy helps concentrate the decomposition products from the foam pattern into the riser, thereby reducing carbon-related defects within the casting body. The gating system dimensions were designed according to the recommended cross-sectional area ratio of F直:F横:∑F内 = 1:1.15:1.42, with a hollow sprue diameter of 3 cm (area 28.26 cm²), a runner cross-section of 5 cm × 6.5 cm (area 32.5 cm²), and two ingates each sized at 4 cm × 5 cm (total ingate area 40 cm²).
The riser design was based on the modulus method, where the modulus M is defined as:
$$M = \frac{V}{A}$$
Here, V represents the cooling volume, A indicates the cooling surface area of the casting, and M (cm) corresponds to the solidification modulus. The optimum riser modulus was derived using the simplified modulus method:
$$M_R = f \cdot M_c$$
where MR is the riser modulus, Mc is the casting modulus, and f represents the safety factor (f=1.2 for open riser). The calculated casting modulus was 4.5 cm, resulting in an optimal riser modulus of 5.4 cm. For vertical pouring, a rectangular riser with dimensions of 240 mm × 320 mm × 360 mm was initially established, ensuring a modulus of 5.6 cm.
The steel’s thermophysical properties were calculated using the ProCAST software based on the actual chemical composition of ZGS10CrNiCu. The liquidus and solidus temperatures were estimated to be 1512°C and 1269°C, respectively. The temperature-dependent properties, including thermal conductivity, density, enthalpy, and viscosity, were all generated automatically and are displayed in the following figures.
2. Numerical simulation results and process optimization
The ProCAST software, a well-established numerical simulation code, was employed to simulate both the mold filling and solidification processes for the bridge bearing castings. The finite element meshes were generated with a grid size of 20 mm for the casting and gating system, while a coarser grid size of 50 mm was used for the sand mold. The material settings, including the “Lost Foam Casting” module, heat transfer module, and fluid flow module, were configured accordingly.
2.1 Comparison of horizontal and vertical pouring schemes
The simulation results of the horizontal pouring scheme indicated that the total filling time was approximately 56.4 seconds, with an average pouring rate of 9 kg/s. The filling pattern showed a relatively slow initial advance, followed by a rapid and slightly turbulent flow during the middle and later stages. The most critical issue was the long horizontal flow distance, which caused the leading edge of the molten metal to lose significant temperature. As a result, the foam pattern in the far-end corners of the casting was not fully gasified, leading to the formation of isolated regions of foam decomposition products. This created a higher risk of carbon defects, slag inclusion, and gas porosity in the edge sections of the casting. This prediction was experimentally confirmed through production trials where visible slag and gas pore defects appeared on the upper surface and corners of the castings.
In contrast, the vertical pouring scheme exhibited a more stable and controlled filling performance. The total filling time was approximately 101 seconds with an average pouring rate of 5 kg/s. The metal flow was smooth and steady, rising vertically without significant turbulence. The slow upward filling permitted better gas evacuation, and the shorter horizontal flow distance minimized the temperature drop at the metal front. Similarly, more complete foam gasification was achieved, and the decomposition residues were effectively directed toward the riser. Production trials of the vertical pouring scheme yielded castings free from visual surface and internal defects, meeting the stringent quality requirements after machining and ultrasonic inspection. This scheme was therefore selected as the preferred casting orientation for subsequent optimization.
2.2 Orthogonal experimental design for process parameters
To achieve the best casting quality and optimize the riser feeding efficiency, an orthogonal experiment was designed based on three primary process parameters: pouring temperature (A), vacuum degree (B), and pattern density (C). The orthogonal array L9(3³) was applied, using the safe height of the riser after feeding, defined as the vertical distance between the lowest shrinkage cavity region and the top of the casting, as the evaluation indicator. The selected factors and levels are listed in the following table.
| Level | Factor A: Pouring Temperature (°C) | Factor B: Vacuum Degree (MPa) | Factor C: Pattern Density (g/L) |
|---|---|---|---|
| 1 | 1720 | 0.06 | 20 |
| 2 | 1700 | 0.05 | 18 |
| 3 | 1680 | 0.04 | 16 |
The orthogonal table and the simulated safe height results for the nine experimental combinations were summarized. From the range analysis, the factor significance order was determined to be: pouring temperature > foam pattern density > vacuum degree. The range values were 2.79 for factor A, 1.18 for factor C, and 0.71 for factor B. The trend analysis revealed that the optimal combination was A2B2C1, corresponding to a pouring temperature of 1680°C, a vacuum degree of 0.05 MPa, and a pattern density of 20 g/L. This parameter set was used for further numerical validation. The simulation for the optimal combination provided a maximum safe height of 11.15 cm, confirming the effectiveness of the chosen parameters.
2.3 Secondary optimization and production verification
Based on the initial optimized parameters, a secondary process refinement was conducted to reduce production costs while maintaining the targeted casting quality. External chills were added at the bottom and side walls of the casting to enhance local solidification rates and promote directional solidification. The chill material was specified as 35# steel, and the interfacial heat transfer coefficient was set to 1000 W/(m²·K). Additionally, an exothermic covering agent was adopted to reduce heat loss from the open riser top, thereby extending the solidification time of the riser and improving its feeding efficiency. The modified riser dimensions were adjusted to 200 mm × 300 mm × 300 mm, improving the process yield from 55% to 65%.
Simulation results confirmed that the solidification sequence was remarkably improved by the addition of chills. The isolated liquid regions consistently remained inside the riser, with the final solidification occurring in the riser center. Shrinkage predictions indicated that all porosity was confined to the gating and riser systems, with the casting body fully dense. The measured safe height after this secondary optimization was 2.32 cm, which approached the minimum safety limit yet remained adequate for production robustness.
A small batch production trial was carried out with 50 castings. Following machining and non-destructive testing, 48 pieces met all required standards, achieving a quality qualification rate of 96%. Two rejected pieces were attributed to slag and gas porosity resulting from inclusions in the molten steel. The particle tracking simulation was applied to study the inclusion flow behavior, indicating that some inclusions remained trapped in the casting corners, probably due to the faster cooling rate induced by the chills in these zones. This suggests that further improvement of steel melt cleanliness during melting and pouring operations would be beneficial.
3. Experimental study on carbon defects
Carbon defects, or carbon pick-up, originate from the decomposition of the foam pattern during mold filling. The fundamental mechanism is the diffusion of carbon into the molten steel from the thermal degradation byproducts of the polymer foam. As the molten steel front advances, the foam undergoes pyrolysis, yielding gaseous, liquid, and solid products. The gaseous products can mostly escape through the permeable coating under the influence of negative pressure. However, solid carbon residue tends to adhere to the coating surface, causing severe surface carburization. On the other hand, the liquid and gaseous hydrocarbons condense within the coating and the surrounding sand, continuing to decompose as the casting cools, further aggravating carbon enrichment.
The carbon transfer equation can be expressed as:
$$C_w = K_{\sum} S (C_E – C_0) \tau \times 10^6$$
where Cw is the final carbon content of the casting (%), KΣ is the mass transfer coefficient (S·m⁻²), S is the contact area between the decomposition products and the molten metal (m²), CE is the carbon content of the decomposition products (%), C0 is the initial carbon content of the metal (%), and τ is the contact time (s).
These defects manifest in three predominant forms: surface carbon pick-up, volumetric carbon pick-up, and localized carbon pick-up. Surface carburization is mainly caused by the solid carbon residues deposited on the coating layer, while volumetric carburization results from the diffusion of carbon into the liquid steel at the leading edge. Localized defects arise from the turbulent flow that entrains liquid pyrolysis products into the metal, which subsequently decompose and generate carbon-rich zones in the solidified structure.
3.1 Experimental methodology
To quantify the influence of carbon defects on the final quality of the bridge bearing steel castings, controlled experiments were performed using two distinct foam materials: EPS (expandable polystyrene) and STMMA (a copolymer of methyl methacrylate and styrene). EPS has a carbon content as high as 92%, while STMMA has a lower carbon concentration of approximately 69.6% and contains oxygen atoms that facilitate the formation of CO and CO₂ gases during decomposition. This compositional difference was expected to significantly influence the extent of carbon pick-up.
For the experiments, six castings were produced using the optimized vertical pouring process: three with EPS foam and three with STMMA foam. All six castings were cast from the same heat of molten steel to ensure a consistent initial chemical composition. The initial carbon content was measured at 0.09%. Four sampling locations were designated for analysis: near the ingate (A), at the casting center (B), at the rib plate (C), and at the top of the casting adjacent to the riser (D). This arrangement allowed the evaluation of carbon pick-up as a function of distance from the ingate and solidification time.
3.2 Effect of carbon defects on chemical composition
The carbon content at the surface and at the center of each sample was measured using a spark emission spectrometer (Bruker Q4-130). The results for surface carbon content are presented in the following table.
| Foam Material | Casting No. | Location A (%) | Location B (%) | Location C (%) | Location D (%) |
|---|---|---|---|---|---|
| EPS | 1 | 0.18 | 0.24 | 0.27 | 0.33 |
| 2 | 0.20 | 0.29 | 0.33 | 0.38 | |
| 3 | 0.17 | 0.25 | 0.30 | 0.30 | |
| STMMA | 4 | 0.092 | 0.104 | 0.128 | 0.142 |
| 5 | 0.096 | 0.115 | 0.126 | 0.175 | |
| 6 | 0.098 | 0.108 | 0.138 | 0.157 | |
| Average Increase | EPS | 0.10 | 0.17 | 0.21 | 0.25 |
| STMMA | 0.005 | 0.019 | 0.039 | 0.070 |
From the surface carbon content data, several crucial observations were noted. For castings produced with EPS, the surface carbon pick-up ranged from 0.10% to 0.25%, with the maximum single measurement being as high as 0.3%. In contrast, for STMMA-produced castings, the surface carbon pick-up was consistently below 0.1%, with the highest average increase being 0.070%. This substantial difference highlights the beneficial effect of the lower carbon content and the presence of oxygen atoms in STMMA, which produce fewer solid carbon residues during pyrolysis.
Additionally, the carbon content trend across locations A through D exhibited a progressive increase along the metal flow path from the ingate to the riser region. Longer flow distances and later solidification times accumulated more carbon at the surface. These relationships are illustrated in the following data table.
| Foam Material | Location | Average Surface Carbon Pick-up (%) | Average Volumetric Carbon Pick-up (%) |
|---|---|---|---|
| EPS | A | 0.10 | 0.013 |
| B | 0.17 | 0.022 | |
| C | 0.21 | 0.023 | |
| D | 0.25 | 0.034 | |
| STMMA | A | 0.005 | 0.001 |
| B | 0.019 | 0.002 | |
| C | 0.039 | 0.002 | |
| D | 0.070 | 0.012 |
The volumetric carbon pick-up was much lower than the surface pick-up in all cases. With EPS, the volumetric carbon pick-up ranged from 0.013% to 0.034%, with a maximum below 0.04%; for STMMA, the volumetric values ranged from 0.001% to 0.012%, all below 0.015%. These findings suggest that the primary carbon contamination mechanism is strongly dominated by surface reactions. The restrained volumetric pick-up may be attributed to the presence of carbide-forming elements (Cr, Mn) that partially inhibit carbon diffusion in the steel matrix, even though their total content is less than 2%.
3.3 Effect of carbon defects on mechanical performance
The mechanical performance of the castings was evaluated through tensile testing and impact testing in accordance with national standards. The specimens were extracted following the same sampling scheme (locations A, B, C, D) and subjected to a normalization heat treatment at 930°C for 3 h, followed by air cooling. The tensile properties, including yield strength (Rel), ultimate tensile strength (Rm), elongation (A), and reduction of area (Z), are presented in the following table.
| Foam Material | Location | Rm (MPa) | Rel (MPa) | A (%) | Z (%) |
|---|---|---|---|---|---|
| EPS | A | 549 | 309 | 31 | 52 |
| B | 558 | 326 | 30 | 51 | |
| C | 561 | 330 | 28 | 48 | |
| D | 574 | 344 | 24 | 46 | |
| STMMA | A | 542 | 304 | 34 | 62 |
| B | 543 | 308 | 33 | 63 | |
| C | 545 | 309 | 31 | 56 | |
| D | 555 | 312 | 30 | 52 |
From the tensile test results, a clear trend emerged. As the carbon content increased from location A to location D, the tensile strength and yield strength also increased. The EPS castings, which experienced a higher degree of carbon pick-up, exhibited notably higher strength values compared to STMMA castings at the same location. Conversely, the ductility indicators, elongation and reduction of area, showed a slight decreasing trend with increasing carbon content. Despite these differences, all tensile properties remained within the required specification ranges for the bridge bearing material.
Impact toughness tests were carried out at room temperature, -20°C, and -40°C to assess the material’s performance under low-temperature service conditions. The results are summarized in the following table.
| Foam Material | Temperature | Location A (J) | Location B (J) | Location C (J) | Location D (J) |
|---|---|---|---|---|---|
| EPS | RT | 93 | 87 | 86 | 82 |
| -20°C | 41 | 36 | 26 | 18 | |
| -40°C | 25 | 22 | 12 | 8 | |
| STMMA | RT | 98 | 96 | 95 | 89 |
| -20°C | 55 | 55 | 53 | 45 | |
| -40°C | 33 | 30 | 28 | 27 |
The impact test results demonstrated a dramatic decline in impact toughness with increasing carbon content at all testing temperatures. This reduction was especially pronounced at lower temperatures. For EPS castings, the -40°C impact toughness values fell well below the required minimum of 15 J, indicating an unacceptable level of embrittlement for bridge bearing applications. In contrast, STMMA castings consistently met the specification criteria at both room temperature and low-temperature conditions. This sharp difference is attributed to the presence of a higher proportion of pearlite in the EPS castings, which possesses a body-centered cubic lattice structure with a ductile-to-brittle transition temperature within the service environment.
3.4 Effect of carbon defects on the microstructure
The microstructure of the castings was examined using optical microscopy after etching with 4% nital solution. The samples were taken from location D, which exhibited the highest carbon content. In both EPS and STMMA castings, the microstructure consisted of ferrite and pearlite, consistent with the hypoeutectoid nature of the steel. However, distinct differences were observed. The EPS castings displayed an inhomogeneous distribution of pearlite, with some regions showing clustered pearlite colonies and sharp-edged grain boundaries. This non-uniform structure contributes to localized stress concentrations and is likely responsible for the inferior impact toughness. On the contrary, the STMMA castings exhibited a uniform and well-distributed ferrite-pearlite structure with clear and smooth grain boundaries. This homogeneous microstructure is beneficial for achieving stable mechanical properties and enhanced impact resistance.
4. Conclusions
This research provided a comprehensive investigation into the lost foam casting process for manufacturing low-carbon bridge bearing steel castings, integrating numerical simulation, production verification, and experimental analysis of carbon defects. The principal conclusions can be summarized as follows:
(1) For the ZGS10CrNiCu steel bearing, the combination of side-bottom gating and vertical pouring is optimal. This scheme ensures more stable mold filling, faster liquid level rise, more complete gasification of the foam pattern, and significantly reduced carbon-related defects. The vertical pouring arrangement produced castings with superior quality and integrity.
(2) The orthogonal experimental analysis established the influence order of the three process parameters on the riser feeding performance as: pouring temperature > pattern density > vacuum degree. The optimum process parameters were determined to be a pouring temperature of 1680°C, a pattern density of 20 g/L, and a vacuum degree of 0.05 MPa.
(3) The secondary optimization, through the addition of external chills, the application of an exothermic covering agent on the riser, and the reduction of the riser dimensions, successfully increased the process yield rate to 65%. Small-batch production trials confirmed a high quality qualification rate of 96%.
(4) The experimental investigation of carbon defects revealed a distinct distribution rule: the severity of carbon pick-up escalates with the distance from the ingate and with delayed solidification time. This phenomenon was consistently observed in both surface and volumetric carbon content measurements.
(5) EPS-produced castings exhibited surface carbon pick-up values between 0.1% and 0.3%, and volumetric carbon pick-up values between 0.01% and 0.04%. The resulting chemical composition and impact toughness failed to satisfy the bridge bearing quality standards. In contrast, STMMA-produced castings exhibited a surface carbon pick-up below 0.1% and volumetric carbon pick-up below 0.015%, with all measured chemical and mechanical properties conforming to the quality requirements.
(6) The microstructural analysis confirmed that the EPS-produced castings had a non-uniform pearlite distribution with sharp grain boundaries, which is detrimental to mechanical reliability. The STMMA-produced castings showed a uniform ferrite-pearlite matrix with smooth boundaries, which is favorable for ensuring robust mechanical performance. Therefore, STMMA is recommended as the preferred foam material for low-carbon steel industries using the lost foam castings process.
These findings have been successfully implemented in a steel foundry, enabling the production of high-quality bridge bearing steel castings with both improved product quality and enhanced economic efficiency. The proposed process methodology provides valuable guidance for extending the application of the lost foam castings technology to other low-carbon steel components with similar structural and material requirements.