In industrial applications, vacuum pump housings are critical components that endure harsh conditions, including continuous exposure to water and abrasive particles. Traditional welded steel structures often fail due to leakage, leading to reduced system pressure and compromised efficiency. To address this, we developed a method to produce vacuum pump housings using ductile iron castings, which offer superior durability and longevity. This article details the lost foam casting process employed to manufacture these thick-walled ductile iron castings, focusing on design, simulation, and production controls to achieve leak-free components.
The vacuum pump housing casting has a large cylindrical structure with a maximum outer dimension of φ864 mm × 637 mm and a wall thickness of 35 mm. The single casting weight is approximately 425 kg, and the material specification requires QT600-3 grade ductile iron, with strict standards against leakage. To meet these demands, we designed a top-gating system with risers for effective feeding, ensuring dense microstructure and sound castings. The mold box dimensions were set at 1200 mm × 1000 mm × 1300 mm, accommodating one casting per box. The total weight per box during pouring was 560 kg, using a 700 kg ladle for iron handling. We implemented a cored wire injection process for nodularization, with a pouring temperature range of 1430–1450°C and a vacuum pressure of -0.05 to -0.06 MPa. After pouring, the pressure was maintained for one hour to prevent deformation.
The gating system was designed as a semi-open top-pouring arrangement, with a gating ratio of ΣFsprue : ΣFrunner : ΣFingate = 1 : 1.4 : 1.2. The sprue cross-section was φ50 mm, the runner measured 75 mm × 75 mm, and the ingate was 40 mm × 140 mm. Six risers, each 120 mm × 120 mm × 140 mm, were evenly distributed on the top of the casting to facilitate directional solidification and compensate for shrinkage. Two of these risers served as hot risers, where molten iron entered the cavity, while the others filled as the metal level rose, collectively providing feeding during solidification. This approach avoided the complexities of external chills or internal chills, which could lead to fusion issues in lost foam casting.
To validate the casting design, we conducted a comprehensive simulation using MAGMA software, analyzing the filling and solidification processes. The results demonstrated that the top-gating system allowed molten iron to enter the cavity smoothly, descending to the bottom and rising progressively without turbulence. This ensured that cooler metal at the bottom was replenished by hotter metal from above, while the risers provided adequate feeding to eliminate shrinkage defects. The simulation outputs, as shown in the filling and solidification sequences, confirmed the absence of major isolated liquid zones, with minor areas mitigated through graphite expansion. The governing equations for solidification time and thermal gradients were considered, such as Chvorinov’s rule for solidification time: $$ t = B \left( \frac{V}{A} \right)^2 $$ where \( t \) is the solidification time, \( B \) is the mold constant, \( V \) is the volume, and \( A \) is the surface area. Additionally, the carbon equivalent (CE) for ductile iron castings was calculated to ensure proper composition: $$ CE = C + \frac{1}{3}(Si + P) $$ This helped in optimizing the iron chemistry for reduced shrinkage tendency.
The lost foam pattern process began with creating the white pattern, which was segmented due to the large size of the ductile iron castings. Using CAD software, the housing was divided into multiple sections—12 pieces for the cylindrical body and 8 for the flanges—with fillet radii of R10 mm. These segments were cut on a fully automated CNC cutting platform and assembled manually on a glass panel aligned to the center. Fiber rods were used for temporary fixation, followed by sealing gaps with adhesive and tape to prevent coating infiltration. Wooden strips (15 mm × 15 mm) were attached to the top and bottom for reinforcement during coating. The risers were secured with fiber rods to withstand handling stresses. This meticulous assembly ensured high dimensional accuracy for the ductile iron castings, suitable for small-batch production.
For coating, the assembled pattern was immersed in a refractory coating with a Baume degree controlled between 69–71 °Bé. Operators carefully rotated the pattern in the coating bath and used smaller containers to reach inaccessible areas, ensuring uniform coverage. The pattern was coated four times to withstand the erosive forces during pouring. After each coating, it was dried in a dedicated oven to remove moisture. The coated pattern, now referred to as the yellow pattern, was ready for molding. In the molding stage, the pattern was placed in a sandbox on a leveled sand bed, and the box was filled with sand using manual gentle sand addition. Vibration was applied for 90 seconds to compact the sand around the pattern, ensuring stability during pouring.
The melting process was critical for achieving the desired properties in the ductile iron castings. The target chemical composition is summarized in the table below:
| Element | Content (wt%) |
|---|---|
| C | 3.7 |
| Si (after nodularization) | 2.4 |
| Mn | 0.5 |
| S | 0.015 |
| P | 0.05 |
| Residual Mg | 0.05 |
| Sn | 0.06 |
We employed a covered ladle dual-wire injection method for nodularization, using FeSiMg25Re3 cored wire with an addition rate of 0.7%. In-mold inoculation involved 0.3% preconditioner and 0.2% 75FeSi composite inoculant. The wire injection, controlled by PLC for length and speed, delivered the alloy directly to the ladle bottom, minimizing oxidation and enhancing absorption. After nodularization, slag was removed promptly, and the ladle was transported to the pouring station. To prevent shrinkage, the pouring temperature was maintained between 1430–1450°C; if it exceeded this range, clean cold iron pieces were added to cool the metal. After pouring, the cavity was topped up with additional iron, and vacuum pressure was held at -0.05 to -0.06 MPa for one hour to ensure integrity.
The production outcomes were evaluated through metallurgical and mechanical tests on attached test specimens. The results are presented in the following table:
| Property | Value |
|---|---|
| Nodularity Grade | 2 |
| Graphite Size Grade | 7 |
| Pearlite Volume Fraction (%) | 65 |
| Cementite Volume Fraction (%) | ≤1.0 |
| Phosphide Eutectic Volume Fraction (%) | ≤0.5 |
| Tensile Strength (MPa) | 641 |
| Elongation (%) | 3.5 |
Microstructural analysis revealed a fine graphite distribution in a pearlitic matrix, with no significant carbides or phosphides, confirming the quality of the ductile iron castings. After shakeout, the castings were cleaned using a 5810 shot blasting unit, and the risers and gating systems were removed. Examination of the riser necks showed no shrinkage porosity or voids. Dimensional checks aligned with design specifications, and subsequent machining revealed no defects such as shrinkage or sand inclusions, meeting the operational requirements for vacuum pumps.
In conclusion, the lost foam process proved highly effective for producing thick-walled ductile iron castings like vacuum pump housings. The segmentation of white patterns via CNC cutting and manual assembly ensured precision, making it viable for limited production runs. The covered ladle dual-wire injection method provided consistent nodularization, fulfilling the material specifications for ductile iron castings. The top-gating system with riser feeding promoted orderly filling and directional solidification, resulting in dense, leak-free components. This approach underscores the potential of lost foam casting for complex ductile iron castings, with future work focusing on optimizing riser design and expanding to other applications.
Throughout this project, we emphasized the importance of process control in achieving high-quality ductile iron castings. The integration of simulation, precise pattern making, and controlled melting parameters contributed to the success. For instance, the feeding efficiency of risers can be modeled using the feeding distance equation: $$ L = \frac{(T_{\text{pour}} – T_{\text{solidus}})}{G} $$ where \( L \) is the feeding distance, \( T_{\text{pour}} \) is the pouring temperature, \( T_{\text{solidus}} \) is the solidus temperature, and \( G \) is the thermal gradient. Such formulations aid in refining the process for even larger ductile iron castings. Overall, this methodology demonstrates a reliable pathway for enhancing the performance and longevity of industrial components through advanced ductile iron castings.