We are a foundry team dedicated to improving the casting quality and efficiency of complex iron castings. In recent years, we have focused on the hydraulic transmission locomotive valve body, a critical component made of gray iron with the material grade HT200. The valve body has a net weight of 24.1 kg, overall dimensions of 300 mm × 250 mm × 150 mm, and a main wall thickness of only 6 mm. Its internal structure is extremely intricate, containing two independent oil passage cavities that require precise dimensional control. Traditionally, we relied on manual molding and core making, which led to numerous problems: high scrap rates, excessive sand-to-iron ratio, poor dimensional accuracy, and frequent defects such as cold shuts, core shift, and penetration. To overcome these challenges, we introduced 3D printing casting technology, replacing conventional handmade molds and cores. This paper describes our development of a 3D printing casting process for the valve body, focusing on core design, gating system optimization, and production validation. Through this work, we have achieved significant improvements in casting quality, productivity, and cost reduction.
In the conventional manual molding process, the internal cavity of the valve body was so complex that we had to split the core into eight individual sand cores, requiring six separate core boxes. Each core needed precise positioning, and the assembly process involved more than ten sand cores. The tooling cost was high, and the manual core-setting procedure was labor-intensive and time-consuming. During mold closing, we used custom-made gauges to measure the wall thickness between cores and the mold, then adjusted the core positions manually to control the oil passage dimensions. This operation was cumbersome and introduced cumulative positioning errors that often exceeded the dimensional tolerance of the casting. As a result, the castings frequently suffered from core shift, uneven wall thickness, and penetration defects. Many small cores were fragile and broke easily, forcing us to produce additional cores, which increased both material and labor costs. The manual molding process also limited the number of cavities per mold: each flask measured 700 mm × 700 mm × 550 mm and could hold only one casting per box. The sand-to-iron ratio was as high as 15.6, meaning we used 15.6 kg of sand for every kilogram of iron poured. For such thin-walled castings, we had to pour at high temperature to avoid cold shuts and misruns, but even then we could only pour six boxes (six castings) per session. This resulted in significant waste of molding sand, core sand, and molten iron. The scrap rate under this process reached 11%, and the dimensional accuracy could only achieve GB/T 6414 CT11 grade.
Faced with these persistent problems, we decided to implement 3D printing casting technology. We acquired a 3D sand printer that allows us to fabricate sand molds and cores directly from digital models, bypassing the need for physical pattern equipment. The core concept of 3D printing casting is to build sand layers one by one, bonding them with a resin binder, producing complex geometries that are impossible or prohibitively expensive to achieve with conventional tooling. This technology gives us design freedom that was previously unimaginable. For the valve body, we redesigned the entire core and mold architecture to take full advantage of 3D printing casting capabilities.
The first and most critical step was to determine the core split plan. The valve body’s internal oil passage cavities are relatively independent, but they are interconnected through thin walls. In conventional core making, we would have to split the internal cores into many small pieces to allow core removal and assembly. However, with 3D printing casting, we can consolidate multiple core features into a single monolithic core, as long as we can design sufficient draft angles and ensure that the core can be cleaned of loose sand after printing. We evaluated three key factors: core strength, dimensional accuracy, and sand removal.
For core strength, the valve body has several thin core sections with roots as small as 18 mm in diameter. During pouring, these delicate cores experience buoyancy forces from the liquid iron, which can cause them to break or float, leading to penetration defects. To prevent this, we integrated the slender oil passage cores with the thicker main body cores. Specifically, we designed the cores so that part of the core is connected to the drag (bottom) mold and part to the cope (top) mold, ensuring that the cores are firmly anchored. The resulting core assembly provides adequate tensile, compressive, and flexural strength to withstand the ferrostatic pressure.
For dimensional accuracy, we aimed to minimize the number of cores to reduce alignment errors. In 3D printing casting, we can print complex core geometries as a single piece, eliminating multiple joints. For the valve body, we reduced the number of cores from eight to two: one integrated core for the drag side and one for the cope side. This simplification dramatically reduced the cumulative tolerance stack-up. The main wall thickness of 6 mm is now controlled directly by the printed mold cavity, and the internal oil passage walls are formed by the two cores that fit together precisely. By eliminating intermediate core positioning, we achieved a dimensional accuracy of GB/T 6414 CT9 grade, a two-level improvement over the conventional CT11 grade.
Sand removal is a practical concern in 3D printing casting. After printing, the cores contain loose unbonded sand that must be evacuated. We designed the core cavities with generous access openings and avoided any blind pockets where sand could become trapped. The 3D printing casting process produces cores with smooth surfaces, but we still need to ensure that the coating (if applied) can drain freely without pooling. In our final design, all internal cavities are visible and accessible for cleaning, so no sand remains after the core is blown out.
Table 1 summarizes the differences between the conventional manual process and the new 3D printing casting process for the valve body.
| Parameter | Conventional Manual Molding | 3D Printing Casting |
|---|---|---|
| Number of sand cores per casting | 8 | 2 |
| Number of core boxes | 6 | 0 (direct printing) |
| Cavities per flask | 1 | 8 |
| Sand-to-iron ratio | 15.6 | 1.3 |
| Dimensional accuracy (GB/T 6414) | CT11 | CT9 |
| Scrap rate | 11% | 5% |
| Tooling cost | High (6 core boxes) | Negligible (digital only) |
| Labor intensity | High (manual core setting) | Low (automated assembly) |
After establishing the core design, we developed the gating system for 3D printing casting. The valve body is a thin-walled casting, so we chose a bottom-gating, open-riser system to ensure smooth filling and avoid cold shuts. The gating ratio was set as follows:
$$ \frac{\sum F_{\text{sprue}}}{\sum F_{\text{runner}}}{\frac{\sum F_{\text{ingate}}}{\sum F_{\text{ingate}}}} = 1:2.1:2.3 $$
where \(F\) represents the cross-sectional area. Each casting receives two ingates from the bottom, allowing the molten iron to rise steadily from the base, filling the cavity with minimal turbulence. This reduces the risk of sand erosion and inclusion defects. We controlled the pouring temperature between 1350 and 1370 °C, with a pouring time of 15 to 25 seconds. The higher temperature helps the iron flow through thin sections without freezing prematurely. The open gating system also prevents aspiration and keeps the mold cavity clean.
Another important aspect of 3D printing casting for thin-walled castings is the addition of metal pads (feeders) in critical areas. The valve body has two corners in the internal cavity where the wall thickness is only 2.9 mm, which is extremely thin. In conventional casting, such thin sections are prone to penetration (iron bleeding through the core) or cold shuts. With 3D printing casting, we can easily add a metal pad (a local thickness increase) to these corners, raising the wall thickness to 6.5 mm. This small modification, printed directly into the sand core, eliminates the risk of penetration and ensures sound metal. The pad is later removed by machining. Figure 1 illustrates the thin-wall location and the added pad.
In the 3D printing casting process, the drag and cope molds are printed separately along with their integrated cores. After printing, we clean the loose sand, apply a refractory coating if needed, and then assemble the two halves. The assembly is secured with bolts, and the mold can be transported directly to the pouring station. This entire sequence is much faster than conventional core setting and mold closing. With a flask size of 700 mm × 700 mm × 550 mm, we can now place eight valve body cavities in one box. The total sand weight per mold is 254 kg, giving a sand-to-iron ratio of only 1.3 (since the total iron weight is 24.1 kg × 8 = 192.8 kg, sand weight is 254 kg, ratio = 254/192.8 ≈ 1.3). This is a dramatic reduction from the previous ratio of 15.6, leading to substantial savings in sand, resin, and disposal costs.
We conducted a production trial with 80 pieces using the 3D printing casting process. The results were very positive. The external surfaces showed excellent finish, with minimal flash that was easy to remove. Dimensional inspection by CMM confirmed that the castings met CT9 tolerance. The internal oil passage walls were uniform, with no evidence of core shift. Of the 80 pieces, we found no cold shuts or penetration defects. Four castings were scrapped due to slag inclusions, giving a scrap rate of 5%. This is a significant improvement over the 11% scrap rate of the conventional process. Table 2 summarizes the defect analysis.
| Defect type | Number of defective castings | Percentage |
|---|---|---|
| Cold shut | 0 | 0% |
| Penetration (iron bleed) | 0 | 0% |
| Core shift | 0 | 0% |
| Slag inclusion | 4 | 5% |
| Other (dimensional, etc.) | 0 | 0% |
| Total scrap | 4 | 5% |
The success of 3D printing casting for this valve body can be attributed to several factors. First, the ability to print complex internal cavities as one piece eliminates the need for multiple core joints, which are the primary source of dimensional errors and core-shift defects. Second, the bottom-gating open system ensures stable filling, preventing cold shuts and erosion. Third, the addition of local metal pads in thin sections, made possible by 3D printing casting, directly addresses the penetration risk. Finally, the low sand-to-iron ratio reduces the thermal mass of the mold, which can improve solidification characteristics and reduce the tendency for shrinkage defects.
We also quantified the cost benefits. The tooling for conventional core boxes cost tens of thousands of dollars and had long lead times. With 3D printing casting, we eliminated all hard tooling. The digital design is directly transferred to the printer, and modifications can be made at no additional cost. The labor for core setting was reduced by approximately 70% because we only handle two large cores instead of eight small ones. Production throughput increased from 6 castings per pour (conventional) to 8 castings per pour, and the per-casting sand usage dropped by about 92%. Table 3 compares key economic indicators.
| Item | Conventional Manual | 3D Printing Casting |
|---|---|---|
| Number of pours required | 17 (6 castings each) | 13 (8 castings each) |
| Total sand consumption (kg) | 15.6×100×24.1 ≈ 37,596 | 1.3×100×24.1 ≈ 3,133 |
| Core box cost (US$) | ~15,000 | 0 |
| Labor hours for core setting | ~200 h | ~60 h |
| Scrap cost (at $5/kg scrap value) | 11%×100×24.1×5 = $1,326 | 5%×100×24.1×5 = $603 |
Furthermore, 3D printing casting enabled us to shorten the product development cycle. In conventional casting, any design change would require new core boxes, taking weeks. With 3D printing casting, we can modify the CAD file and print a new mold overnight. This agility is invaluable for a foundry that must respond quickly to customer requirements.
We have now established 3D printing casting as the standard process for this valve body. The lessons we learned are applicable to other complex iron castings. For example, any component with intricate internal oil or water passages, thin walls, and stringent dimensional tolerances can benefit from 3D printing casting. The key is to rethink the core design from the ground up, leveraging the freedom of additive manufacturing. We also recommend using simulation software in conjunction with 3D printing casting to predict filling and solidification, ensuring a robust process before printing the first mold.
In conclusion, the development of the 3D printing casting process for the hydraulic transmission locomotive valve body has yielded outstanding results. The technology eliminated the need for expensive core boxes, reduced the number of sand cores from eight to two, improved dimensional accuracy from CT11 to CT9, lowered the sand-to-iron ratio from 15.6 to 1.3, and cut the scrap rate from 11% to 5%. The castings exhibit excellent surface quality, uniform wall thickness, and no cold shuts or penetration defects. Production efficiency increased significantly, and labor costs dropped. 3D printing casting has proven to be a game-changing method for producing complex gray iron castings, and we are actively applying it to other products in our foundry.