Lost foam casting (LFC), also known as expendable pattern casting (EPC), is a near-net-shape manufacturing technology that has attracted worldwide attention due to its design flexibility, low production cost, reduced environmental pollution, and high product quality. The fundamental principle of lost foam casting involves using a foam pattern that is geometrically similar to the final casting, coating it with refractory paint, drying it, embedding it in dry sand with vibration compaction, and then pouring molten metal under vacuum. The molten metal vaporizes the foam pattern and occupies its space, solidifying to form the intended component. This process has been widely applied in the production of cast iron, aluminum alloys, and special steels. However, its application in the field of low-carbon steel has been severely limited by a persistent and challenging defect known as carbon pickup or carburization. The increasing demand for low-carbon steel castings, particularly for critical structural applications such as bridge bearings, has made the resolution of carbon defects a top priority for both researchers and industrial practitioners.
This thesis presents a comprehensive study on the lost foam casting of a specific bridge bearing steel casting, designated as a steel damping bearing. The research integrates numerical simulation using ProCAST software, orthogonal experimental design, production validation, and systematic experimental investigation of carbon defects. The primary objectives are to determine an optimal casting process that ensures high product quality, to minimize carbon-related defects, and to improve the overall yield and economic efficiency for the manufacturing enterprise. The work is organized around the design of two alternative pouring schemes, the numerical simulation of mold filling and solidification, the optimization of process parameters, and a detailed comparative study of two foam pattern materials—EPS and STMMA—and their influence on carbon defect formation and mechanical properties.
1. Research Background and Significance
Bridge bearings are essential structural components that connect the superstructure and substructure of a bridge. They are designed to transmit vertical loads, accommodate rotational movements, and provide displacement capabilities in one or multiple directions. Depending on the structural type, bridge bearings can be classified as plate rubber bearings, pot bearings, spherical steel bearings, and cylindrical steel bearings. The selection of bearing type and specification must consider factors such as bridge span, support reaction, architectural height constraints, displacement requirements, and seismic isolation demands. Because bridge bearing steel castings are load-bearing components, their reliability directly affects the safety and durability of the entire bridge. Consequently, the quality requirements for these castings are extremely stringent. They must be free from shrinkage porosity, cracks, gas porosity, inclusions, and other casting defects. Moreover, critical load-bearing areas must pass ultrasonic inspection to at least Class II standards according to relevant specifications.
Common steel grades used for bridge bearings include ZG230-450, ZG270-500, and ZG20Mn, all of which belong to the family of medium-carbon or low-carbon alloy steels. The chemical composition limits for typical bridge bearing steels are summarized in Table 1. Traditional production methods for these components often rely on water glass sand casting, but this process faces significant environmental challenges, including poor sand reclamation and pollution from waste sand. Lost foam casting presents an environmentally friendly and clean alternative. However, the application of lost foam casting to low-carbon steels has been hindered by the carbon pickup defect, which occurs when the foam pattern decomposes under high heat, releasing free carbon that diffuses into the molten steel. This carbon contamination can lead to unacceptable chemical composition deviations and deteriorated mechanical properties, particularly impact toughness at low temperatures.
| Steel grade | C | Si | Mn | P | S | Ni |
|---|---|---|---|---|---|---|
| ZG230-450 | 0.22–0.26 | ≤0.60 | 0.60–0.90 | ≤0.030 | ≤0.030 | ≤0.40 |
| ZG270-500 | 0.26–0.32 | ≤0.60 | 0.60–0.90 | ≤0.030 | ≤0.030 | ≤0.40 |
| ZG20Mn | 0.17–0.23 | ≤0.80 | 1.00–1.30 | ≤0.025 | ≤0.025 | ≤0.80 |
The significance of this research lies in its potential to expand the industrial application of lost foam casting for high-quality low-carbon steel components. By systematically investigating the process through simulation and experiment, this work provides practical solutions to overcome the longstanding carbon defect problem, thereby enabling foundries to produce bridge bearing castings with high precision, low cost, and acceptable environmental performance.
2. Casting Process Design for the Bridge Bearing Component
The research object is a steel damping bearing casting with a rectangular base and complex wall thickness variations. The overall dimensions are 700 mm × 445 mm × 155 mm, and the casting mass is approximately 262 kg. The structure features symmetric side walls with square blind holes, while the central region contains a curved concave surface. The minimum wall thickness is 30 mm at the ribbed sections, and the maximum thickness is 155 mm at the central boss. This geometry represents a relatively complex bridge bearing casting, requiring high internal integrity and dimensional accuracy.
Designing the gating system for lost foam casting differs significantly from conventional sand casting because the molten metal must displace and vaporize the foam pattern during filling. The filling behavior is influenced by the decomposition products, coating permeability, vacuum pressure, and gating design. For this component, two pouring positions were considered: horizontal pouring and vertical pouring. Each configuration has distinct advantages and disadvantages, which were evaluated through numerical simulation and production trials.
2.1 Gating System Design
The gating system adopted a side-bottom injection approach, where the ingates were positioned on one side near the bottom of the pattern. This configuration promotes directional filling along a diagonal path, concentrating decomposition products and inclusions into the riser. The gating system proportions were selected as:
$$ F_{\text{direct}} : F_{\text{runner}} : \sum F_{\text{ingate}} = 1 : (1.1\text{–}1.3) : (1.2\text{–}1.5) $$
where \(F\) denotes the cross-sectional area. For this design, the direct sprue was a hollow cylinder with a diameter of 60 mm, giving an area of 28.26 cm². The runner had a rectangular cross-section of 50 mm × 65 mm, with an area of 32.5 cm². Two ingates were used, each with a cross-section of 40 mm × 50 mm, providing a total ingate area of 40 cm². Thus, the actual ratio was \(1 : 1.15 : 1.42\), which falls within the recommended range for an open gating system. The hollow sprue design effectively prevents mold splashing and stabilizes the metal front.
2.2 Riser Design
The riser (feeder) was designed using the modulus method. The modulus \(M\) is defined as the ratio of volume \(V\) to cooling surface area \(A\):
$$ M = \frac{V}{A} $$
To ensure adequate feeding, the riser modulus \(M_R\) should be larger than the casting modulus \(M_C\) multiplied by a safety factor \(f\):
$$ M_R = f \cdot M_C $$
For an open riser, \(f = 1.2\). The modulus of the casting was calculated to be \(M_C = 4.5\) cm, leading to a required riser modulus of \(M_R = 5.4\) cm. For horizontal pouring, a cylindrical open riser with dimensions of 340 mm diameter and 340 mm height was chosen, giving a modulus of 5.66 cm. The riser was verified using the liquid mass balance criterion:
$$ \varepsilon (V_{\text{riser}} + V_{\text{casting}}) \leq \eta V_{\text{riser}} $$
where \(\varepsilon\) is the volumetric solidification shrinkage (about 5% for low-carbon steel poured above 1600°C), and \(\eta\) is the feeding efficiency (assumed 15% with insulating covering). The calculation confirmed that the riser provided sufficient feed metal. For vertical pouring, a rectangular (elongated) open riser of 240 mm × 320 mm × 360 mm was selected, with a modulus of 5.6 cm and adequate feeding capacity.
2.3 Pattern Materials and Properties
Two types of foam pattern materials were used in this study: EPS (expandable polystyrene) and STMMA (a copolymer of methyl methacrylate and styrene). EPS has a carbon content of approximately 92%, while STMMA has a significantly lower carbon content of about 69.6%. The molecular structures differ: EPS contains benzene rings that produce abundant free carbon during decomposition, whereas STMMA has a more linear structure with oxygen atoms that promote gaseous decomposition products, reducing the amount of residual carbon. Table 2 lists the thermal properties of both materials.
| Property | STMMA | EPS |
|---|---|---|
| Density (g/L) | 16–22 | 16–22 |
| Thermal conductivity (W/(m·K)) | 0.15 | 0.15 |
| Specific heat (kJ/(kg·K)) | 3.5 | 3.7 |
| Latent heat (kJ/kg) | 80 | 100 |
| Liquidus temperature (°C) | 330 | 350 |
| Solidus temperature (°C) | 300 | 330 |
2.4 Sand, Coating, and Pouring Parameters
Based on the requirements for good flowability and low thermal expansion, a ceramsite sand (also known as baozhu sand) with 20/40 mesh was selected. This sand exhibits high refractoriness (above 1800°C), low thermal expansion, and excellent flowability, which is critical for filling the blind holes and recesses of the bearing casting. A commercial refractory coating was applied in three layers, with intermediate drying at 40–60°C for 12–16 hours per coat, achieving a final coating thickness of 3–4 mm. The pouring temperature was initially set at 1680°C, which is 50–80°C higher than conventional sand casting to compensate for the energy consumed by foam decomposition. The vacuum degree was initially set at 0.05 MPa, a typical value for steel casting in lost foam.
Using Creo software, three-dimensional solid models of the casting, gating system, and risers were created and assembled for both horizontal and vertical pouring configurations. These models were then imported into ProCAST for meshing and simulation.
3. Numerical Simulation and Process Optimization
ProCAST, a well-established finite-element simulation package, was employed to model the lost foam casting process. The software can simulate the complex filling and solidification behavior, including the interaction between the molten metal and the decomposing foam pattern. The simulation setup requires specific parameters for the foam material, sand permeability, vacuum pressure, and heat transfer coefficients.
3.1 Simulation Setup for Horizontal Pouring
For the horizontal pouring scheme, the finite-element mesh was generated with a maximum element size of 20 mm for the casting and gating system, and 50 mm for the sand mold. The module was set to “Lost Foam Casting,” and the alloy was defined as ZGS10CrNiCu, a low-carbon corrosion-resistant steel. The alloy’s thermal properties were calculated automatically by ProCAST based on its chemical composition (C 0.07–0.12%, Si ≤0.6%, Mn 0.6–1.4%, Cr 0.8–1.2%, Ni 0.5–1.0%, Cu 0.3–0.5%). The liquidus temperature was computed as 1512°C, and the solidus temperature as 1269°C. Figures 2–5 in the original thesis illustrate the temperature-dependent thermal conductivity, density, enthalpy, and viscosity, but are not reproduced here.
The boundary conditions included a vacuum pressure of 0.05 MPa applied to the sand mold exterior, while the pouring cup was at atmospheric pressure. The heat transfer coefficients were set to 500 W/(m²·K) between mold and casting, and 10 W/(m²·K) between insulation and mold. Key lost-foam parameters included a “FOAMHTC” of 0.02, “FOAMHTCMAX” of 0.25, and a burn zone width of 2.5 cm.
3.2 Simulation Results for Horizontal Pouring
The filling simulation revealed a relatively fast but turbulent metal front advance. The total filling time was 56.4 seconds, corresponding to an average filling rate of about 9 kg/s. Early filling was slow due to the backpressure from foam decomposition, but as the contact area increased, the metal velocity increased, leading to potential turbulence and possible entrapment of decomposition products. The solidification analysis indicated a generally progressive solidification from the thin walls toward the thick central sections and riser. However, a localized isolated liquid region was observed in a corner away from the ingate, which could lead to shrinkage porosity or carbon concentration. The predicted shrinkage porosity was confined to the riser and gating system, but the safe feeding height of the riser was only 7 cm, which is above the minimum requirement but suggests room for optimization.
Production trials using the horizontal pouring configuration confirmed the simulation findings. The castings exhibited surface defects, including gas and slag porosity, particularly on the top surfaces and at the corners. After machining, inclusions were visible at the edges, attributed to foam decomposition residues and molten slag. This configuration was deemed suboptimal for this geometry.
3.3 Simulation Results for Vertical Pouring
In the vertical pouring scheme, the filling pattern was significantly more stable. The metal ascended smoothly from the bottom, with a total filling time of 101 seconds and an average filling rate of about 5 kg/s. The vertical arrangement enabled a shorter horizontal flow distance, promoting complete filling of the thin sections before the metal reached the riser. The solidification sequence followed a proper directional pattern, with no isolated liquid regions detected. Shrinkage porosity was predicted only within the riser, and the safe feeding height was measured at 9 cm. The filling time contour showed that the last region to fill was the riser, which acted as an effective collector for slag and decomposition products.
Trials with vertical pouring produced castings with excellent as-cast surfaces and no visible defects. Machining and ultrasonic inspection of six castings from the same heat confirmed internal soundness. Based on these results, vertical pouring was selected as the preferred casting orientation.
3.4 Orthogonal Experiment for Process Parameter Optimization
To further optimize the process, a three-factor, three-level orthogonal experiment was designed. The selected factors were pouring temperature (A), vacuum degree (B), and foam pattern density (C). The response variable was the safe feeding height of the riser after solidification, a direct measure of feeding efficiency. The factor levels are listed in Table 3, and the orthogonal array \(L_9(3^4)\) is shown in Table 4.
| Level | A: Pouring temperature (°C) | B: Vacuum degree (MPa) | C: Pattern density (g/L) |
|---|---|---|---|
| 1 | 1720 | 0.06 | 20 |
| 2 | 1700 | 0.05 | 18 |
| 3 | 1680 | 0.04 | 16 |
| Run | A | B | C | Safe feeding height (cm) |
|---|---|---|---|---|
| 1 | 1 | 1 | 1 | 8.40 |
| 2 | 1 | 2 | 2 | 8.16 |
| 3 | 1 | 3 | 3 | 6.42 |
| 4 | 2 | 1 | 2 | 10.52 |
| 5 | 2 | 2 | 3 | 10.48 |
| 6 | 2 | 3 | 1 | 10.35 |
| 7 | 3 | 1 | 3 | 8.71 |
| 8 | 3 | 2 | 1 | 10.40 |
| 9 | 3 | 3 | 2 | 10.16 |
The range analysis (Table 5) provides the average values \(k_1\), \(k_2\), \(k_3\) for each factor level and the range \(R\), which indicates the significance of each factor. The results show that pouring temperature has the largest influence (R=2.79), followed by pattern density (R=1.18), while vacuum degree has the smallest effect (R=0.71). The optimal combination was determined as \(A_2B_2C_1\), i.e., pouring temperature 1680°C, vacuum degree 0.05 MPa, and pattern density 20 g/L. This combination was then verified through simulation, yielding a safe feeding height of 11.15 cm, the highest among all tested configurations.
| Statistical parameter | A | B | C |
|---|---|---|---|
| \(K_1\) | 22.98 | 27.63 | 29.15 |
| \(K_2\) | 31.35 | 29.04 | 28.84 |
| \(K_3\) | 29.27 | 26.93 | 25.61 |
| \(k_1\) | 7.66 | 9.21 | 9.72 |
| \(k_2\) | 10.45 | 9.68 | 9.61 |
| \(k_3\) | 9.76 | 8.98 | 8.54 |
| Range \(R\) | 2.79 | 0.71 | 1.18 |
3.5 Secondary Process Optimization and Production Validation
Although the optimal parameter combination markedly improved riser feeding, the overall casting yield was still only about 55% due to the large riser size. To increase efficiency and reduce costs, the riser was redesigned with a reduced height of 300 mm (from 360 mm), resulting in a yield of 60% while maintaining a safe feeding height of 5 cm. Further improvements were achieved by adding external chills to accelerate local solidification at critical areas, thereby enhancing directional solidification. Three chills were designed: a large chill at the bottom measuring 270 mm × 120 mm × 150 mm, and two side chills each measuring 110 mm × 100 mm × 50 mm. These chills were made of carbon steel and modeled as direct chills with an interface heat transfer coefficient of 1000 W/(m²·K). Additionally, an exothermic covering compound was applied on the riser top to prolong liquid metal retention. The final riser dimensions became 200 mm × 300 mm × 300 mm, increasing the casting yield to 65%.
Simulation of the optimized design demonstrated excellent solidification behavior. The chills accelerated cooling at the intended regions, and the liquid metal was confined to the riser at a solidification fraction of 93.5%. Shrinkage porosity remained entirely within the riser and gating system. The measured safe feeding height was 2.32 cm, which, while lower than the previous 5 cm, still exceeded the minimum requirement of 3.4 cm? It actually was close to the theoretical limit and fulfilled the feeding requirements. Production validation was carried out by manufacturing 50 bridge bearing castings using the optimized process. After machining and ultrasonic testing, 48 castings met all quality requirements, giving a pass rate of 96%. Two castings failed due to slag and gas pores originating from inclusions in the molten steel, which were unrelated to the foam pattern. A particle tracking simulation confirmed that inclusions originating from the molten steel tend to accumulate near the casting corners because of the rapid freezing induced by chills. This issue can be mitigated by improving steel melt cleanliness and refining practices.
4. Experimental Study of Carbon Defects
Carbon pickup is the most critical obstacle in lost foam casting of low-carbon steels. The defect arises from the thermal decomposition of the foam pattern, which produces free carbon, hydrogen, and other gaseous products. The carbon diffuses into the molten steel at the metal front, and the extent of pickup depends on factors such as the contact area, contact time, and the carbon concentration gradient between the decomposition products and the steel. The governing equation can be expressed as:
$$ C_w = K_{\Sigma} S (C_E – C_0) \tau \cdot 10^6 $$
where \(C_w\) is the carbon content of the casting (%), \(K_{\Sigma}\) is the mass transfer coefficient (s·m⁻²), \(S\) is the contact area (m²), \(C_E\) is the carbon content of the decomposition products, \(C_0\) is the initial carbon content of the steel, and \(\tau\) is the contact time (s). This equation confirms that larger contact areas, longer contact times, and larger concentration gradients all increase the degree of carbon pickup.
Carbon defects manifest in three primary forms: surface carbon pickup, volumetric (bulk) carbon pickup, and localized carbon pickup. Surface carburization occurs when solid carbon particles are deposited on the coating walls and are absorbed by the solidifying skin. Volume carburization results from carbon diffusion into the bulk liquid steel, especially near the solidification front. Localized carburization is caused by turbulent flow that entraps liquid decomposition products, which later decompose and release carbon.
4.1 Experimental Methodology
To quantify the carbon defects and their effect on the mechanical properties, six test castings were produced using the optimized vertical pouring process. Three castings were made with EPS patterns (designated 1, 2, 3) and three with STMMA patterns (designated 4, 5, 6). All castings were poured from the same heat of ZGS10CrNiCu steel, which had an initial carbon content of 0.09%. After cooling, samples were extracted from four positions as illustrated in Figure 4-7 of the original thesis: (A) near the ingate, (B) center of the casting, (C) at the rib section, and (D) at the top of the casting near the riser. Chemical analysis was performed using an optical emission spectrometer on both the surface (after light grinding to about 0.5–1 mm below surface) and the center of the cross-section to evaluate surface and volumetric carbon pickup, respectively.
4.2 Carbon Content Results and Analysis
Surface carbon content results are presented in Table 6 for each group of castings. The data show that EPS patterns cause substantially more surface carburization than STMMA. For EPS, the average surface carbon content ranged from 0.19% at position A to 0.34% at position D, corresponding to carbon pickup values of 0.10% to 0.25%. In contrast, STMMA castings exhibited surface carbon contents from 0.095% to 0.160%, with carbon pickup limited to 0.005% to 0.070%. The maximum single measurement for STMMA was below 0.1%, confirming the benefit of the copolymer material.
| Foam material | Casting No. | A | B | C | D |
|---|---|---|---|---|---|
| EPS | 1 | 0.18,0.18,0.16 | 0.24,0.26,0.25 | 0.27,0.28,0.27 | 0.33,0.32,0.34 |
| 2 | 0.20,0.22,0.22 | 0.29,0.27,0.28 | 0.33,0.33,0.32 | 0.38,0.37,0.39 | |
| 3 | 0.17,0.20,0.18 | 0.25,0.25,0.27 | 0.30,0.29,0.31 | 0.30,0.32,0.31 | |
| Average | 0.19 | 0.26 | 0.30 | 0.34 | |
| Avg. pickup | 0.10 | 0.17 | 0.21 | 0.25 | |
| STMMA | 4 | 0.092,0.095,0.093 | 0.104,0.107,0.106 | 0.128,0.122,0.123 | 0.142,0.153,0.146 |
| 5 | 0.096,0.097,0.097 | 0.115,0.113,0.113 | 0.126,0.129,0.121 | 0.175,0.172,0.173 | |
| 6 | 0.098,0.097,0.095 | 0.108,0.110,0.107 | 0.138,0.139,0.136 | 0.157,0.161,0.159 | |
| Average | 0.095 | 0.109 | 0.129 | 0.160 | |
| Avg. pickup | 0.005 | 0.019 | 0.039 | 0.070 |
The volumetric (bulk) carbon content measured at the center of the samples is summarized in Table 7. For EPS castings, the average volumetric carbon pickup ranged from 0.013% to 0.034%, with all values below 0.04%. For STMMA castings, the pickup was only 0.001% to 0.012%, with a maximum of 0.015%. This demonstrates that volumetric carbonization is significantly less severe than surface carbonization. The lower volumetric pickup is attributed to the alloying elements in the steel, such as chromium and manganese, which are carbide formers and hinder the diffusion of carbon into the bulk liquid. Nevertheless, with EPS, the volumetric carbon content can still cause local chemical composition deviations beyond acceptable limits. With STMMA, the variations are negligible and all sampled locations meet the specification.
| Foam material | Casting No. | A | B | C | D |
|---|---|---|---|---|---|
| EPS | 1 | 0.102,0.104,0.110 | 0.112,0.112,0.114 | 0.114,0.114,0.112 | 0.128,0.125,0.127 |
| 2 | 0.105,0.102,0.104 | 0.108,0.106,0.107 | 0.117,0.115,0.116 | 0.119,0.118,0.120 | |
| 3 | 0.098,0.102,0.101 | 0.117,0.115,0.117 | 0.110,0.108,0.108 | 0.125,0.125,0.124 | |
| Average | 0.103 | 0.112 | 0.113 | 0.124 | |
| Avg. pickup | 0.013 | 0.022 | 0.023 | 0.034 | |
| STMMA | 4 | 0.090,0.091,0.091 | 0.091,0.091,0.092 | 0.091,0.092,0.091 | 0.101,0.103,0.102 |
| 5 | 0.092,0.092,0.090 | 0.093,0.093,0.093 | 0.094,0.092,0.092 | 0.102,0.102,0.102 | |
| 6 | 0.092,0.090,0.091 | 0.092,0.094,0.093 | 0.093,0.092,0.093 | 0.103,0.104,0.104 | |
| Average | 0.091 | 0.092 | 0.092 | 0.102 | |
| Avg. pickup | 0.001 | 0.002 | 0.002 | 0.012 |
The distribution pattern of carbon pickup clearly increases along the flow direction of the metal from the ingate to the farthest point. Position D, which solidifies last, exhibits the highest carbon content. This is consistent with the mechanism that the advancing metal front continuously contacts fresh foam, accumulating decomposition products, and prolonged contact time during solidification further enhances diffusion.
4.3 Effect of Carbon Pickup on Mechanical Properties
Since bridge bearing components are almost entirely machined, the surface carbon layer is removed during processing. Therefore, volumetric carbon content is the primary factor influencing the final mechanical properties. Standard tensile and impact specimens were prepared from heat-treated samples obtained from each sampling position. The heat treatment consisted of normalizing at 930 ± 10°C for 3 hours followed by air cooling. Table 8 lists the tensile properties measured at the four positions for both foam materials.
| Foam material | Position | \(R_m\) (MPa) | \(R_{eL}\) (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 |
The results show a clear increasing trend in tensile strength and yield strength from position A to D, corresponding to increasing carbon content. This is expected because carbon is a strengthening element. However, ductility, as indicated by elongation and reduction of area, tends to decrease slightly with carbon increase. All measured tensile properties, even for EPS castings at position D, still meet the required specifications for ZGS10CrNiCu: \(R_{eL} \geq 270\) MPa, \(R_m \geq 500\) MPa, \(A \geq 18\%\), \(Z \geq 40\%\). The STMMA castings exhibit more consistent tensile data across positions, indicating better homogeneity.
Impact toughness was evaluated at room temperature, -20°C, and -40°C. Table 9 presents the absorbed energy \(AK_V\) results. The influence of carbon pickup on impact toughness is dramatic. For EPS castings, the room-temperature impact energy decreased from 93 J at position A to 82 J at position D. At -40°C, the values dropped from 25 J at A to only 8 J at D. These low-temperature values are far below the required minimum of 15 J at -40°C. In contrast, STMMA castings maintained acceptable impact energies at all positions: from 98 J at room temperature to 33 J at -40°C at position A, and still 27 J at -40°C at position D. Even the lowest recorded value (25 J at -40°C) meets the specified minimum of 15 J. Thus, EPS castings fail the impact toughness requirement, particularly at low temperatures, while STMMA castings pass all criteria.
| Foam material | Test temp. | Position | |||
|---|---|---|---|---|---|
| A | B | C | D | ||
| 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 | |
4.4 Influence on Microstructure
Metallographic examination was conducted on samples taken from position D of both EPS and STMMA castings. The microstructure consisted of ferrite and pearlite, as expected for a hypoeutectoid steel. However, significant differences were observed. In the EPS casting, the pearlite content was higher and distributed unevenly, with some regions exhibiting clustered pearlite and sharp grain boundaries. This non-uniform microstructure is attributed to the local carbon enrichment caused by the carbon pickup, leading to inconsistent mechanical properties and poor low-temperature toughness. In contrast, the STMMA casting showed a uniform distribution of ferrite and pearlite with clear, smooth grain boundaries. The uniformity is beneficial for achieving stable mechanical properties, particularly impact resistance at low temperatures.
5. Discussion and Industrial Implications
The experimental results unequivocally demonstrate that the choice of foam pattern material is the most critical factor in controlling carbon defects in lost foam casting of low-carbon steels. STMMA, with its lower carbon content and oxygen-containing molecular structure, effectively reduces both surface and volumetric carbon pickup to levels that meet the stringent chemical and mechanical requirements for bridge bearing applications. Although EPS is more economical, its high carbon content produces excessive carbon pickup, leading to failure of impact toughness specifications, especially at low service temperatures. Therefore, for high-integrity low-carbon steel castings, STMMA is the recommended pattern material despite its higher material cost, because it significantly reduces scrap rates and ensures reliable performance.
The optimized casting process developed in this study, which includes vertical pouring, a side-bottom gating system, external chills, an exothermic covering compound, and the optimal process parameters (pouring temperature 1680°C, vacuum degree 0.05 MPa, pattern density 20 g/L), has been successfully implemented in an industrial foundry. The process achieved a 96% acceptance rate in small-batch production, demonstrating its practical viability. The casting yield was improved from an initial 55% to 65%, offering substantial economic benefits by reducing metal consumption and machining allowance.
6. Conclusions
This thesis presents a systematic study of the lost foam casting process for a bridge bearing steel casting, combining numerical simulation, experimental design, and industrial validation. The main conclusions are as follows:
- Vertical pouring with a side-bottom gating system provides smoother mold filling, faster rise of the liquid level, more complete foam gasification, and fewer defects compared to horizontal pouring. It is therefore the preferred orientation for box-like bearing castings.
- The influence of key process parameters on riser feeding efficiency ranks as: pouring temperature > pattern density > vacuum degree. The optimal combination is 1680°C pouring temperature, 20 g/L pattern density, and 0.05 MPa vacuum degree.
- Adding external chills and using an exothermic covering compound allowed the riser size to be reduced, increasing the casting yield from 55% to 65% without compromising quality. Small-batch production achieved a 96% pass rate.
- Carbon defect distribution follows the pattern of increasing severity with distance from the ingate and with later solidification time. Surface carbon pickup is much more severe than volumetric carbon pickup.
- Using EPS patterns leads to surface carbon pickup of 0.1%–0.3% and volumetric carbon pickup of 0.01%–0.04%, causing chemical composition deviation and unacceptable low-temperature impact toughness. STMMA patterns limit surface pickup to below 0.1% and volumetric pickup to below 0.015%, ensuring all chemical and mechanical properties meet the requirements for bridge bearing castings.
- Microstructural examination reveals that EPS-produced castings suffer from non-uniform pearlite distribution and sharp grain boundaries, while STMMA-produced castings exhibit uniform ferrite-pearlite structures with smooth boundaries, which is conducive to reliable mechanical performance.
The research findings have been successfully applied in an industrial foundry to produce high-quality bridge bearing castings with improved yield and cost-effectiveness. The developed methodology and process design can be extended to other low-carbon steel castings with similar structures, offering significant practical value and promotion potential in the field of lost foam casting.