In our investigation of the lost foam casting process for manufacturing oil pans, we focused on addressing common defects such as deformation, metal penetration, dross, cold shuts, shrinkage porosity, and substandard metallographic structure. The lost foam casting technique offers significant advantages in producing complex, thin-walled components with minimal machining allowances, but it requires precise control over multiple parameters to achieve high-quality castings. This article details our first-hand experiences and improvements in optimizing the lost foam casting process for oil pans, incorporating tables and formulas to summarize key findings and recommendations. Throughout this discussion, the term “lost foam casting” will be emphasized to highlight its centrality in our work.
The lost foam casting process begins with the creation of a foam pattern, which is then coated with a refractory material and embedded in dry sand. Upon pouring molten metal, the foam vaporizes, allowing the metal to take its shape. For oil pans, which are typically made of HT250 gray iron with wall thicknesses ranging from 6 to 30 mm and overall dimensions of approximately 601.5 mm in length, 32 mm in width, and 163 mm in height, this method reduces environmental impact and production costs compared to traditional sand casting. However, challenges like deformation often arise due to the thin-walled nature of the component. In our practice, we observed that deformation primarily occurred along the wide edges, leading to insufficient machining allowances and a high scrap rate initially. To mitigate this, we implemented strict controls on pattern making, drying, assembly, coating, and molding processes.
One critical aspect of lost foam casting is the pattern production. We used H-S expandable polystyrene beads with a density controlled between 22 and 26 g/L. The beads were heated at 40–50°C for 3 minutes in a molding machine to ensure proper expansion and curing. After demolding, the foam patterns were checked for dimensional accuracy on a calibration platform, allowing tolerances of -1 mm to +3 mm for key dimensions. To prevent deformation during handling and drying, we developed定型卡板 (fixed templates) and used fiber rods for support, as illustrated in our bonding configuration. The patterns were dried flat with multiple support points to ensure even contraction, and humidity was monitored to meet process requirements before proceeding to coating.
Coating plays a vital role in lost foam casting by providing a barrier between the foam and sand, ensuring surface quality. We applied a water-based refractory coating with additions of 3% bentonite, 15% graphite powder, and 15% quartz powder to enhance strength and fluidity. The coating thickness was maintained at a minimum of 1.6 mm, applied through a combination of dipping and brushing to cover all surfaces uniformly. Drying parameters were optimized to prevent distortion: for the first coat, drying occurred at 45°C for 10 hours; for the second coat, at 50°C for 14 hours; and for any touch-ups, at 45°C for 8 hours. This systematic approach helped reduce coating-related defects like metal penetration and improved overall casting integrity.
Molding involved using dry silica sand with a particle size of 0.4 to 0.8 mm, which was essential for achieving proper compaction and minimizing metal penetration. We adopted a strategy of embedding six patterns per sandbox, arranging them in a staggered layout to avoid overlapping surfaces that could affect vacuum distribution. The sand was added in layers while vibrating the flask at frequencies between 40 and 45 Hz for a total time of at least 360 seconds. This ensured adequate compaction, especially in hard-to-reach areas like corners, where manual sand ramming was employed. The vacuum pressure during casting was maintained at 0.04 to 0.05 MPa to prevent mold collapse and reduce defects. Our experiments showed that these measures significantly lowered the incidence of metal penetration, a common issue in lost foam casting.
Melting and pouring parameters were rigorously controlled to address defects such as dross, cold shuts, and shrinkage. The charge composition for HT250 iron included scrap steel, returns, and alloys like ferrosilicon and ferromanganese, with low-sulfur carburizers added in stages to achieve the desired carbon equivalent. The chemical composition was closely monitored, as summarized in Table 1. Pouring temperature was critical; we set the tapping temperature at 1520 ± 20°C and the pouring temperature at 1420–1460°C to balance fluidity and minimal gas evolution. The gating system was designed as a pressurized type with a cylindrical sprue to facilitate smooth metal flow and reduce turbulence, which helped minimize dross formation. The cross-sectional area ratio of the gating system was optimized to sprue:runner:ingate = 7:1:0.4, and the pouring sequence followed a “slow-fast-slow” pattern to ensure complete filling without cold shuts.
| Element | Composition (%) |
|---|---|
| C | 2.9–3.1 |
| Si | 1.7–1.9 |
| Mn | 0.7–0.9 |
| P | ≤ 0.09 |
| S | ≤ 0.055 |
| Cr | 0.015–0.035 |
| Mg | 0.2–0.3 |
| Other | ≤ 0.12 |
Defect analysis revealed that cold shuts and shrinkage porosity often resulted from inadequate feeding and rapid solidification in thin sections. To combat this, we incorporated blind risers at locations prone to cold shuts, as shown in our gating layout. The pouring time per mold was limited to 25 seconds, with a total box pouring time under 3 minutes, followed by a 3-minute pressure hold to ensure complete solidification. The carbon equivalent (CE) was calculated using the formula: $$ CE = C + \frac{Si}{3} $$, and we aimed for a CE between 3.8% and 4.1% to promote sound microstructure. Additionally, the Si/C ratio was controlled between 0.6 and 0.7 to stabilize pearlite content in the matrix, addressing metallographic issues. Inoculation was performed during tapping and in the ladle, with a holding time of at least 5 minutes to enhance nucleation and reduce undercooling.
Dross and slag inclusions were mitigated through improved fluxing and slag removal practices. We used large-particle fluxing agents and implemented a four-stage slag removal process: two stages under high power in the furnace and two during tapping and ladle transfer. The ladle was preheated and repaired regularly to maintain a clean pouring stream. During pouring, fiber blankets were used to cover the metal surface, leaving only a small opening for flow, which reduced slag entrainment. These steps, combined with the closed gating system, effectively minimized dross defects in the final castings.
To quantify the improvements, we conducted multiple production trials with 32 oil pans per batch. The results demonstrated a significant reduction in defects, as outlined in Table 2. For instance, deformation-related scrap rates dropped from over 50% to below 3%, and overall product yield reached 96%. This success underscores the importance of integrated process control in lost foam casting, from pattern making to pouring. The use of fiber rods for support, optimized sand compaction, and precise temperature management were key factors in achieving dimensional stability and surface quality.
| Defect Type | Initial Scrap Rate (%) | Improved Scrap Rate (%) | Key Improvement Measures |
|---|---|---|---|
| Deformation | 50 | 3 | Use of fiber rods, fixed templates, and controlled drying |
| Metal Penetration | 20 | 2 | Sand size 0.4–0.8 mm, vibration ≥40 Hz, vacuum ≥0.04 MPa |
| Dross | 15 | 1 | Large-particle flux, multi-stage slag removal, closed gating |
| Cold Shuts | 10 | 1 | Blind risers, cylindrical sprue, pouring temperature 1420–1460°C |
| Shrinkage Porosity | 12 | 2 | Optimized gating ratio, CE control, inoculation |
| Substandard Metallography | 18 | 2 | Si/C ratio 0.6–0.7, controlled alloy addition, holding time |
The relationship between process parameters and casting quality can be expressed through empirical formulas. For example, the pouring time \( t \) in seconds for a thin-walled casting like an oil pan can be estimated as: $$ t = \frac{V}{A \cdot v} $$ where \( V \) is the volume of the casting, \( A \) is the cross-sectional area of the ingate, and \( v \) is the flow velocity. In our case, with a volume of approximately 0.001 m³ and an ingate area of 0.0004 m², the velocity was maintained at 0.5 m/s to achieve a pouring time of 25 seconds. Similarly, the thermal gradient during solidification influences shrinkage, and we used Chvorinov’s rule to approximate solidification time: $$ t_s = k \left( \frac{V}{A} \right)^2 $$ where \( t_s \) is the solidification time, and \( k \) is a constant dependent on mold material. By optimizing these factors, we minimized shrinkage defects in the lost foam casting process.
In conclusion, our hands-on experience with the lost foam casting process for oil pans highlights the necessity of a holistic approach to defect prevention. Through systematic improvements in pattern support, coating application, sand selection, gating design, and melting practices, we achieved a high yield of quality castings. The lost foam casting method, when properly executed, offers tremendous benefits for producing complex components, and our findings provide a practical framework for similar applications. Future work could focus on automating certain steps to further enhance consistency and reduce human error in the lost foam casting process.