In the realm of modern manufacturing, the integration of sand casting and 3D printing has revolutionized the way we approach complex metal component production. Our recent work on a redesigned motor frame casting exemplifies how selective laser sintering (SLS) 3D printing technology can accelerate the development cycle while reducing costs and risks. Drawing from our firsthand experience, we present a comprehensive study where casting process simulation and 3D-printed sand molds were employed to produce a high-quality motor frame in record time. This article details the entire workflow—from casting design and numerical simulation to mold fabrication and final casting—highlighting the transformative role of sand casting facilitated by 3D printing.
Our methodology begins with a thorough analysis of the motor frame geometry, followed by the design of a robust gating and risering system. Using ProCAST simulation, we validated the process before committing to physical production. The sand molds were then directly manufactured using SLS 3D printing, eliminating the need for traditional pattern equipment. The final casting met all quality requirements, demonstrating the effectiveness of combining sand casting and 3D printing for rapid prototyping and low-volume production. This approach not only shortened the development timeline from one month to under one week but also provided unparalleled flexibility for design iterations.
1. Introduction
The synergy between sand casting and 3D printing has emerged as a game-changer in the foundry industry. Traditional sand casting relies on patterns, core boxes, and dedicated tooling, which incur high costs and long lead times—especially for new product development. SLS 3D printing, a subset of additive manufacturing, enables the direct fabrication of sand molds and cores from digital models. This technology, first conceptualized by researchers at the University of Texas at Austin and later commercialized by DTM Corporation (now part of 3D Systems), uses a laser to selectively sinter resin-coated sand particles layer by layer. The resulting sand molds exhibit sufficient strength (up to 1.3 MPa in our case) to withstand the metallostatic pressure during pouring. By adopting SLS 3D printing for sand casting, we bypass the need for expensive and time-consuming pattern-making, allowing rapid design changes and efficient production of complex geometries such as the closely spaced cooling fins on the motor frame.
Over the past decade, numerous studies have demonstrated the viability of 3D-printed sand molds for ferrous and non-ferrous castings. For instance, researchers at Huazhong University of Science and Technology developed a dual-hopper feeding system that significantly reduces sintering time. Similarly, teams at Nanjing University of Aeronautics and Astronautics successfully printed integrally bladed rotors and gear sand molds using SLS. Our work builds on these advancements by applying SLS 3D printing to a large (1.45-ton) motor frame casting with intricate features, thereby validating its industrial applicability. In the following sections, we elaborate on the casting design, simulation, and actual production process, emphasizing the pivotal role of sand casting combined with 3D printing.
2. Motor Frame Structure and Challenges
The motor frame under investigation is a modified design intended for an explosion-proof motor. Its overall dimensions are 1623 mm × 852 mm × 852 mm, with a nominal wall thickness of 25.5 mm. The casting weighs approximately 1.45 tons. A distinctive feature is the dense array of cooling fins, each only 4 mm thick, which poses significant challenges for conventional sand casting. The fins require excellent mold surface finish and dimensional accuracy to ensure proper heat dissipation. Additionally, the frame must exhibit high strength and stiffness to withstand operational vibrations and thermal loads.
Table 1 summarizes the geometric parameters of the motor frame.
| Parameter | Value |
|---|---|
| Overall dimensions (L×W×H) | 1623 mm × 852 mm × 852 mm |
| Nominal wall thickness | 25.5 mm |
| Minimum fin thickness | 4 mm |
| Maximum local thickness (bosses) | 45 mm |
| Casting weight | ~1.45 t |
| Material | HT250 gray cast iron |
The casting’s symmetrical design facilitates a four-part mold layout. However, the cooling fins create deep, narrow cavities that are difficult to fill with traditional sand molds. If we had used conventional wooden or metal patterns, the time required to machine the fin impressions would have been weeks, and any design modification would necessitate entirely new patterns. By turning to 3D printing, we could directly produce the sand mold with the fin geometry embedded in the digital model, thereby eliminating pattern costs and enabling rapid iteration.
3. Casting Process Design
We adopted a four-part mold configuration with a vertical parting plane, as illustrated conceptually. The gating system was designed as a “rain” style (shower gate) bottom fill system, which ensures smooth metal flow and minimizes turbulence. Four risers were placed on the top flange to compensate for shrinkage, and four additional vents were provided for gas escape. The gating ratio was calculated based on the chvorinov’s rule and the modulus of the casting sections.
The modulus (M) of a casting section is defined as the ratio of its volume (V) to its cooling surface area (A):
$$
M = \frac{V}{A}
$$
For gray cast iron, the solidification time is proportional to the square of the modulus. Using the modulus values computed from the 3D model, we positioned risers at locations where the modulus exceeded that of the adjacent casting region by at least 1.2 times. The main body of the motor frame has a modulus of approximately 1.5 cm, while the thick bosses (45 mm) have a modulus of about 2.2 cm. Therefore, the risers were sized with a modulus of 2.5 cm to ensure directional solidification.
Table 2 lists the key process parameters used in the design.
| Parameter | Value |
|---|---|
| Mold configuration | 4-part (vertical parting) |
| Gating system | Rain (shower) bottom fill |
| Number of risers | 4 (top flange) |
| Number of vents | 4 (top) |
| Pouring temperature | 1390 °C |
| Pouring time | 92 s |
| Average mold filling velocity | ~1 m/s |
| Mold material (3D printed) | Furan resin-coated sand (SLS) |
| Sand mold strength (after printing) | 1.3 MPa |
One of the critical aspects of the design was ensuring that the thin cooling fins (4 mm) would fill completely without cold shuts. The rain gating system, with multiple small in-gates distributed along the bottom, allows the metal to rise uniformly, reducing the risk of premature solidification in the fins. The pouring rate was controlled so that the metal level in the mold rises at an average velocity of 1 m/s, which is within the recommended range for gray iron to avoid turbulent flow and gas entrapment.
4. Numerical Simulation Validation
Before committing to physical production, we performed a full-scale numerical simulation using ProCAST software. The simulation model represented half of the casting (due to symmetry) to reduce computation time. The geometry was meshed with approximately 1.2 million tetrahedral elements. The material properties for HT250 gray cast iron were obtained from ProCAST’s material database, with the specific composition given in Table 3.
| Element | C | Si | Mn | S | P |
|---|---|---|---|---|---|
| Content | 3.0–3.2 | 1.9–2.1 | 0.90 | 0.0483 | 0.031 |
Other simulation boundary conditions were: mold initial temperature 25 °C, interface heat transfer coefficient (metal/mold) 750 W/(m²·K), and mold/ambient coefficient 10 W/(m²·K). The simulation was terminated when the casting temperature dropped below 500 °C.
4.1 Filling Behavior
The flow simulation results indicated that the liquid metal enters the cavity through the bottom in-gates and rises uniformly. The maximum velocity during filling was about 5 m/s, occurring only locally where the metal jets impinge on the already filled pool. Importantly, these jets did not directly strike the mold walls, thus avoiding mold erosion. The overall filling pattern was smooth and laminar, as evidenced by the relatively even velocity distribution.
We can characterize the flow regime using the Reynolds number:
$$
Re = \frac{\rho v D_h}{\mu}
$$
where \(\rho\) is the melt density (7000 kg/m³), \(v\) the flow velocity (1 m/s), \(D_h\) the hydraulic diameter of the gate (~0.02 m), and \(\mu\) the dynamic viscosity (0.006 Pa·s). This gives \(Re \approx 2.3 \times 10^4\), which indicates turbulent flow at the gate exit. However, because the liquid quickly decelerates upon entering the larger cavity, the bulk of the mold filling remains within acceptable limits. The simulation showed no distinct recirculation zones or surface waves that would entrap gas or slag.
4.2 Solidification and Shrinkage Prediction
The solidification simulation revealed that the thick sections (bosses) solidify last, as expected. However, because HT250 undergoes graphite expansion during eutectic solidification, the risk of shrinkage porosity is inherently lower than that of ductile iron or steel. In ProCAST, the porosity criterion is given by the Niyama criterion:
$$
Niyama = \frac{G}{\sqrt{R}}
$$
where \(G\) is the thermal gradient and \(R\) is the cooling rate. For gray iron, a Niyama value below 0.7 typically indicates a high risk of microporosity. However, our foundry experience suggests that for resin sand molds, as long as the calculated porosity (a ProCAST output) is below 10%, the casting will be sound. The simulation showed that the maximum porosity in any region was less than 8%, predominantly in the thick bosses. The risers were designed with enough volume to feed these regions, and the final simulated shrinkage was confined to the risers themselves. Figure 1 (from simulation) confirms that the critical areas all have porosity below 10%.
Table 4 summarizes the simulation results for key defect-sensitive locations.
| Region | Maximum Porosity (%) | Niyama Value (K·s⁰·⁵/mm) | Defect Risk |
|---|---|---|---|
| Main body (wall) | 2.1 | 1.3 | Low |
| Cooling fins | 0.5 | 2.4 | Very low |
| Bosses (thick sections) | 7.8 | 0.8 | Moderate (fed by riser) |
| Riser neck | 12.5 | 0.6 | Acceptable (riser) |
The simulation gave us confidence that the designed process was sound. If any adjustments were needed—for example, increasing riser size or modifying gating—we could have altered the 3D model and re-simulated within hours. This iterative capability is a major advantage of combining sand casting simulation with 3D printing.
5. 3D Printing of Sand Mold
After finalizing the casting design, we proceeded to manufacture the sand mold using SLS 3D printing. The machine used was a commercial SLS printer (similar to EOSINT P/M/S series) with a maximum build volume of 800 mm × 800 mm × 800 mm. Because our motor frame is larger than the build volume, the four-part mold was printed in segments and then assembled. Each segment was printed using furan resin-coated sand with an average particle size of 150 μm. The key printing parameters are given in Table 5.
| Parameter | Value |
|---|---|
| Laser power | 100 W |
| Laser scan speed | 4000 mm/s |
| Layer thickness | 0.2 mm |
| Scan spacing (hatch) | 0.15 mm |
| Resin type | Furan (acid-catalyzed) |
| Resin content (by weight of sand) | 1.8% |
| Post-curing temperature | 150 °C |
| Post-curing time | 2 hours |
The printed sand mold exhibited a compressive strength of 1.3 MPa, which is more than adequate for the metallostatic pressure exerted by a 1.5 m high column of iron (approximately 0.12 MPa). The mold surface finish was excellent, with no noticeable stair-stepping on the cooling fins. The total printing time for all four mold segments was approximately 48 hours, which is far less than the weeks required for traditional pattern making. After printing, the segments were inspected for dimensional accuracy using a 3D scanner and found to be within ±0.3 mm of the CAD model.
The figure above shows one of the 3D-printed sand mold segments prior to assembly. The cooling fin pattern is clearly visible. The mold was then sprayed with a zircon-based wash to improve surface quality and reduce metal penetration. Assembly involved applying a thin layer of furan sand paste on the joint faces and clamping the four parts together with steel bands. Core prints were provided for any internal cores (in this casting, no cores were required as the interior is hollow).
6. Production and Validation
The assembled mold was placed in a flask and supported with loose sand. The pouring was carried out at the foundry using 1390 °C HT250 melt. The pouring time was controlled to 92 s as per the simulation. After pouring, the mold was allowed to cool for 24 hours before shakeout. The casting was then cleaned and subjected to inspection.
Non-destructive testing (ultrasonic and dye penetrant) on the critical fins and bosses revealed no internal defects. The measured hardness on the main body averaged 190 HB, which is within specification for HT250. A dimensional check using a coordinate measuring machine confirmed that all critical features, including the bearing housing bores and mounting flanges, were within tolerance. The cooling fins were fully formed with no mistuns or cold shuts. The total weight of the finished casting was 1.43 t, very close to the target 1.45 t, indicating minimal shrinkage.
Table 6 compares the actual results with the simulated predictions.
| Parameter | Simulated | Actual | Agreement |
|---|---|---|---|
| Porosity in main body | < 3% | No visible porosity on cut section | Excellent |
| Fin filling | 100% | All fins fully formed | Excellent |
| Shrinkage defects | Confined to risers | Risers cleanly separated, no internal defects | Perfect |
| Dimensional accuracy | ±0.3 mm (from CAD) | ±0.4 mm on critical features | Within tolerance |
| Mold strength | 1.3 MPa | No mold deformation observed | Confirmed |
The first article passed all acceptance criteria, and the casting was approved for use in the motor assembly. The entire development cycle—from design to finished casting—took only 5 working days, compared to the typical 30 days required when using traditional pattern equipment. This dramatic reduction in lead time underscores the value of integrating sand casting with 3D printing for new product introduction.
7. Conclusions
Our successful development of the motor frame casting demonstrates that the combination of sand casting and 3D printing offers a powerful solution for rapid casting manufacturing. By using SLS 3D printing to produce the sand mold directly from the CAD model, we eliminated the need for patterns, core boxes, and associated tooling. This not only reduced the upfront cost but also enabled us to iterate the casting design quickly based on simulation feedback. The numerical simulation (ProCAST) accurately predicted the filling and solidification behavior, guiding us to a robust process with no defects.
Key takeaways from this project include:
- The use of 3D printing in sand casting allows production of complex geometries (such as thin cooling fins) that would be difficult or expensive with conventional methods.
- The combination of simulation and 3D printing creates a digital thread from design to physical part, minimizing trial-and-error.
- The time to first functional casting was reduced from one month to less than one week, significantly lowering the risk of new product development.
- The printed sand mold strength (1.3 MPa) was sufficient for casting a 1.45-ton gray iron part.
We believe that as 3D printing technologies continue to advance—with higher build speeds, larger build volumes, and lower material costs—the adoption of 3D-printed sand molds for sand casting will become mainstream for both prototyping and low-volume production. This case study provides a practical blueprint for foundries and product developers seeking to leverage additive manufacturing for metal casting.