As a graduate student specializing in mechanical engineering, I have devoted my master’s thesis research to the rapid sand casting process of engine cylinder heads using numerical simulation. My work combines three-dimensional modeling, casting process CAD design, experimental measurement of interfacial heat transfer coefficients, and finite-element based casting simulation with ProCAST. The entire study is presented here from my personal perspective, emphasizing the methodology, experimental findings, simulation results, and practical production trials. Throughout the text, I use the term sand casting foundry repeatedly because it represents the core technology under investigation. The combination of rapid prototyping and traditional sand casting foundry techniques offers a powerful route for producing complex aluminum alloy castings such as cylinder heads with short lead times and reduced tooling costs.
1. Introduction and Background
The automotive industry has experienced rapid growth in the past two decades, especially in China. Engine cylinder heads are among the most complex and critical components in an internal combustion engine. They must withstand high thermal and mechanical loads while maintaining precise dimensional tolerances. Traditional development of cylinder head castings relies on wooden or metal patterns, which are expensive and time-consuming to produce. A typical development cycle for a cylinder head using conventional sand casting foundry methods can take six to eight months, with gravity die tooling costs exceeding one million yuan per set. This approach is no longer suitable for the modern demands of rapid product innovation and small-batch production.
Rapid sand casting, also known as fast sand casting, is an innovative manufacturing process that integrates rapid prototyping with conventional sand casting foundry technology. In this process, a stereolithography (SLA) pattern replaces the traditional wooden or metal pattern. The SLA pattern is used to produce sand molds and sand cores, which are then assembled and poured with molten metal. This technique significantly reduces development time and cost while maintaining high dimensional accuracy. It is particularly suitable for complex castings that require multiple sand cores, such as cylinder heads, cylinder blocks, and intake manifolds.
The application of numerical simulation in sand casting foundry has become an essential tool for process optimization. Simulation software can predict filling behavior, solidification patterns, and the formation of defects such as shrinkage porosity, gas porosity, and cold shuts. By virtually testing different gating system designs and process parameters, engineers can identify optimal conditions before physical trials. This not only saves material and labor but also improves the quality of the final casting.
The objective of my research was to develop a complete rapid sand casting process for a ZL105 aluminum alloy cylinder head. I used Unigraphics NX 8.0 for three-dimensional solid modeling and a laboratory-developed rapid sand casting process CAD system for casting design. I then employed ProCAST, a professional casting simulation software from ESI Group, to simulate the mold filling and solidification processes for three different gating system designs. I also experimentally determined the interfacial heat transfer coefficient between ZL105 aluminum alloy and phenolic urethane resin sand, which is a crucial boundary condition for accurate simulation. Finally, I designed and manufactured the sand mold tooling using SLA patterns and successfully produced trial castings.
2. Literature Review and Current Status
2.1 Aluminum Alloys in Automotive Applications
Aluminum alloys have become the dominant material for many automotive components due to their low density, high specific strength, excellent thermal conductivity, and good corrosion resistance. Table 1 summarizes the typical applications of aluminum castings in automobiles.
| Component | Usage Rate | Characteristics | Example |
|---|---|---|---|
| Aluminum cylinder block | >45% | Low cost, good lightweight effect | General Motors |
| Aluminum cylinder head | >90% | High strength, good heat dissipation | Most vehicle models |
| Aluminum radiator | >90% | Excellent durability and heat transfer | Almost all vehicles |
| Aluminum wheel | 80% in foreign, 50% in China | Good balance, various shapes | First used in 1964 |
| Chassis components | — | Lightweight, long service life | GM, Ford |
| Aluminum bumper | — | Lightweight, reinforcing effect | — |
The table shows that aluminum has become a key material in the automotive industry. By 2020, it was estimated that there would be 72 million passenger cars in China, and the use of aluminum alloys in vehicle manufacturing is expected to continue increasing.
2.2 Rapid Sand Casting Research Status
Rapid prototyping (RP) technologies such as SLA, LOM, SLS, and FDM have been developed since the late 1980s. SLA is particularly attractive for pattern making because it offers high dimensional accuracy (within ±0.1 mm), good surface finish (Ra 3.25 μm), and fast build speeds. When SLA patterns are used as patterns in sand casting, they can be reused multiple times, making the process economical for small-batch production. Combining SLA with resin sand molding creates a rapid sand casting process that is especially beneficial for casting complex geometries like cylinder heads.
International research on rapid casting has been active since the early 1990s. In the United States, 3D Systems Corporation introduced the “Quick Cast” method for investment casting. Similar developments occurred in Europe and Japan. In China, institutions such as Tsinghua University, Xi’an Jiaotong University, Beijing Longyuan Company, and Shandong University have made significant contributions. Despite these advances, the integration of SLA patterns with resin sand molds for large automotive castings remains a challenging area. The design of the mold assembly requires deep casting knowledge and careful consideration of pattern extraction, core placement, gating, and risering.
2.3 Casting Numerical Simulation
The history of casting simulation dates back to the 1940s, when Paschkis from Columbia University used early computers to simulate heat transfer in solidification. In 1962, Forsund in Denmark applied finite difference methods to calculate temperature fields during solidification. In 1965, Henzel and Keveria simulated a nine-ton turbine casing for General Electric. In the 1970s, Flemings and his group at MIT developed models for predicting shrinkage defects based on temperature field calculations. Since then, many commercial software packages have become available, including ProCAST, MAGMAsoft, AnyCAST, and AutoCAST. These tools enable engineers to visualize filling, solidification, and defect formation in detail.
Numerical simulation of mold filling involves solving the Navier-Stokes equations for incompressible viscous flow with a free surface. The governing equations include continuity, momentum, and energy conservation. For solidification, the heat conduction equation is solved with latent heat release. Defect prediction models use criteria such as critical solid fraction, temperature gradient, and cooling rate. With accurate material properties and boundary conditions, simulation results can closely match experimental observations.
3. Process Design and Pre-Processing for Simulation
3.1 Workpiece Description
The engine cylinder head used in my research is made of ZL105 aluminum alloy. Its overall dimensions are 426 mm × 200 mm × 131 mm, with a volume of approximately 4581 cm³ and a weight of about 12.4 kg. The component has thin walls with a minimum thickness of 4 mm. It contains multiple internal passages for water, oil, and gas, requiring several sand cores. The casting tolerance grade is CT10 for manual mold making.
3.2 CAD Modeling
I created the three-dimensional model of the cylinder head using Unigraphics NX 8.0. This software provides powerful parametric solid modeling capabilities, allowing for complex surface and solid construction. The final CAD model is shown in my thesis. The model includes all external features as well as internal cavities that will be formed by sand cores.
3.3 Rapid Sand Casting Process CAD System
The casting process design was carried out using a laboratory-developed rapid sand casting CAD system. This system was built on the Pro/E platform with VC++2005, Microsoft Access 2003, and Pro/ToolKit. It includes modules for casting process analysis, parting surface design, machining allowance, shrinkage allowance, sand core design, gating system design, riser design, and exhaust system design. The workflow of the CAD system is illustrated in a flow chart in my original research.
3.4 Sand Core Design and Parting
Based on the cylinder head geometry, I divided the internal cores into six separate sand cores:
- Lower left oil gallery core
- Lower right oil gallery core
- Water jacket core
- Upper oil gallery core
- Intake port core
- Exhaust port core
Each core was designed with appropriate core prints for precise positioning. The lower left and right oil gallery cores were placed on the bottom core. The intake and exhaust port cores were located using side core prints. The water jacket core was connected to the lower oil gallery cores, forming an integral lower assembly. The upper oil gallery core was positioned with three locating points. To ensure sufficient strength, each core included internal reinforcement ribs or core wires where necessary. A three-dimensional exploded view of the sand mold assembly is provided at the end of the thesis.
3.5 Gating System Design
For the gravity sand casting of aluminum alloy, I selected an open gating system with a choke in the sprue. This system ensures smooth filling and minimizes oxidation of the molten metal. The design followed the empirical cross-sectional area ratio Asprue : Arunner : Aingate = 1 : 2 : 4. The ingate area was calculated using the equation:
$$A_{内} = \frac{G_L}{\mu \rho_L t \sqrt{2 g h_p}}$$
where GL is the mass of liquid metal passing through the choking area, ρL is the density of the metal, μ is the discharge coefficient, t is the pouring time, and hp is the average static pressure head. The pouring time was estimated by:
$$t = 3\sqrt{G_{件}} + \frac{G_{件}}{3\delta}$$
where δ is the average wall thickness. Flattened ingates were used to distribute the metal evenly and to avoid sand erosion. The runner was trapezoidal in cross-section to reduce heat loss and facilitate mold machining.
Three different gating system configurations were considered:
| Option | Position of Ingates | Advantages | Disadvantages |
|---|---|---|---|
| Option 1 | Unilateral top injection | Short filling time, good sequential solidification, simple structure | High impact, turbulence, risk of oxidation |
| Option 2 | Symmetrical top injection | Even metal distribution, relatively smooth filling | Possible flow merging, turbulence |
| Option 3 | Unilateral bottom injection | No direct impact on sand, steady filling | May require higher static pressure for complete filling |
3.6 Riser Design
Risers were designed to compensate for solidification shrinkage and to provide a path for gas escape. I applied the modulus method, using Chvorinov’s rule:
$$t_c = \left(\frac{M_c}{K_c}\right)^2, \quad t_r = \left(\frac{M_r}{K_r}\right)^2$$
The modulus M is defined as V/A, where V is the volume and A is the cooling surface area. The condition for adequate feeding is tr ≥ tc. Two standard top blind risers were placed on the upper part of the cylinder head. The riser dimensions were selected based on the geometry of the hot spots.
3.7 Simulation Mesh Generation
The CAD models were exported from UG NX in IGES format and imported into ProCAST’s MeshCAST module. I used tetrahedral elements with a target element size of about half to one-third of the minimum wall thickness. The mesh statistics for the three options are shown below:
| Option | Number of Nodes | Number of Elements |
|---|---|---|
| Option 1 (unilateral top) | 107,385 | 471,675 |
| Option 2 (symmetrical top) | 107,035 | 470,289 |
| Option 3 (unilateral bottom) | 115,990 | 508,256 |
The mesh quality was checked by evaluating the aspect ratio, radius ratio, and dihedral angle. After smoothing and optimization, all elements met the recommended criteria.
4. Experimental Determination of Interfacial Heat Transfer Coefficient
4.1 Motivation
Accurate boundary conditions are essential for reliable casting simulation. The interfacial heat transfer coefficient (IHTC) between the casting and the mold is a critical parameter. However, it is difficult to measure directly because it depends on temperature, contact pressure, surface roughness, and gap formation. I therefore used an inverse approach based on temperature measurements.
4.2 Experimental Setup
A simplified cylindrical casting geometry was used to minimize computation time. The casting was 300 mm in diameter and 200 mm high. The mold was made of phenolic urethane resin sand. Five K-type thermocouples were placed at different positions: one in the casting near the interface (TC1) and four in the sand mold at distances of 5, 10, 15, and 20 mm from the interface (TC2, TC3, TC4, TC5). The schematic of the experiment is shown in Figure 3-1 of my thesis. The bottom and top of the mold were insulated with ceramic fiber blankets to ensure one-dimensional heat flow.
ZL105 aluminum alloy was melted and poured at 690 °C into the mold. The temperature data were recorded at a frequency of 1 Hz using an 8-channel data acquisition system. I waited until the casting temperature dropped below 400 °C before stopping the recording.
4.3 Material Properties
The thermal properties of ZL105 aluminum alloy and the resin sand mold were temperature dependent. The solid fraction curve shows a solidus temperature of 536 °C and a liquidus temperature of 622 °C. The density, enthalpy, and thermal conductivity of the alloy as functions of temperature are plotted in Figure 3-2 of the thesis. For the resin sand, the thermal conductivity was approximately 0.5 W/(m·K), and the density was about 1500 kg/m³.
4.4 Mathematical Model for Inverse Heat Conduction
I simplified the heat transfer to one-dimensional cylindrical coordinates. The governing equation is:
$$\frac{\partial T}{\partial t} = \alpha \left( \frac{\partial^2 T}{\partial r^2} + \frac{1}{r} \frac{\partial T}{\partial r} \right)$$
The heat flux at the interface was estimated using Beck’s nonlinear inverse method. The temperature at node i at time p+1 was calculated using the finite difference scheme:
$$T_i^{p+1} = T_i^p (1 – 2 F_0) + F_0 \left(1 – \frac{\Delta r}{2r_i}\right) T_{i-1}^p + F_0 \left(1 + \frac{\Delta r}{2r_i}\right) T_{i+1}^p$$
where F0 is the Fourier number, Δr is the grid spacing, and Δt is the time step. The inverse algorithm iteratively updated the heat flux until the calculated temperatures matched the measured temperatures within a tolerance.
4.5 Results of the Inverse Calculation
The measured cooling curves for the casting and mold are shown in Figure 3-5. The casting center temperature remained high for a long time, while the mold surface temperature increased rapidly during the first 60 seconds. The calculated heat flux reached a maximum of about 256,000 W/m² early in the solidification process, then decreased rapidly. The IHTC was then computed using:
$$h(t) = \frac{q(t)}{T_{cast} – T_{mold}}$$
The resulting IHTC as a function of casting surface temperature is plotted in Figure 3-7. I found the following values:
| Temperature Range | IHTC (W/(m²K)) |
|---|---|
| Above liquidus (622 °C) | 811 |
| Liquid-solid region (622 °C – 536 °C) | Decreasing from 790 to 310 |
| Below solidus (536 °C) | Approximately 230 |
These values are physically reasonable. The high IHTC in the liquid state is due to perfect contact between the liquid metal and the mold. As solidification proceeds, the metal shrinks, creating an air gap and reducing heat transfer. The IHTC eventually levels off at about 230 W/(m²K).
4.6 Validation of the IHTC
To verify the inverse result, I performed a direct simulation of the experiment using ProCAST with the calculated IHTC as the boundary condition. Figure 3-9 compares the simulated and measured temperatures at the casting center and at a point 5 mm into the mold. The maximum deviation was approximately 11 °C, which is acceptable for engineering purposes. This confirms that the IHTC values can be used for simulation of the actual cylinder head casting.
5. Numerical Simulation of the Three Gating System Options
5.1 Simulation Setup
Using the experimental IHTC and other boundary conditions, I set up simulations for the three gating system options in ProCAST’s PreCAST module. A virtual mold was defined as a rectangular box surrounding the casting. The mold material was phenolic urethane resin sand, and its initial temperature was 25 °C. The casting material was ZL105 alloy. I set the pouring temperature at 690 °C and the pouring speed at 0.22 m/s (which corresponds to a fill rate of about 0.75 to 1.5 kg/s, depending on the cross-section). Gravity was set to 9.8 m/s². The interface heat transfer coefficient between UD and the top of the cylinder head was assigned according to the inverse results. The run parameters were specified as shown in Figure 2-14 of the thesis.
5.2 Option 1: Unilateral Top Gating System
5.2.1 Filling Process
The filling process of the cylinder head with a unilateral top gating system is shown in Figure 4-1 of my thesis. The molten metal entered through a sprue, then through a runner, and finally through four ingates. The total filling time was about 15.7 seconds. During the initial stage, the metal spread across the bottom of the cavity. After 8 seconds, the free surface became nearly horizontal, indicating stable filling. There was no evidence of severe turbulence or air entrapment. The fill time distribution (Figure 4-2) shows that the region near the ingates filled slightly earlier, but overall the filling was uniform.
5.2.2 Solidification Process
The solidification temperature field is shown in Figure 4-3. At 25 seconds after filling, a temperature gradient existed from the riser toward the bottom. The ingates remained hot, providing feeding. At 60 seconds, solidification started at the thin walls. At 135 seconds, the risers began to solidify, and isolated liquid regions appeared in the lower thick sections. The solid fraction distribution in Figure 4-4 clearly shows that by 125 seconds, the metal around the ingates had reached 85% solid, while the interior of the thick sections was only 45% solid. This formed isolated liquid pockets, leading to shrinkage porosity as shown in Figure 4-5.
5.3 Option 2: Symmetrical Top Gating System
5.3.1 Filling Process
In the second option, eight ingates were distributed symmetrically on both sides of the top runner. The filling time was about 14 seconds. The fill time distribution (Figure 4-7) shows that the left and right sides filled unevenly, because the metal from both sides met in the middle. This merging of two flow fronts can cause oxide inclusions and gas entrapment. The surface profile of the free surface was less uniform compared to Option 1.
5.3.2 Solidification Process
The temperature field evolution for Option 2 is shown in Figure 4-8. The solidification pattern was similar to Option 1, but the localized hot spots were smaller because the ingates were more numerous. However, isolated liquid regions still appeared in the lower central thick section. The solid fraction in Figure 4-9 indicates that at 255 seconds, a large region remained partially liquid. The predicted shrinkage porosity (Figure 4-10) was concentrated in one large area near the bottom center. The overall volume of defects was slightly less than in Option 1.
5.4 Option 3: Unilateral Bottom Gating System
5.4.1 Filling Process
The unilateral bottom injection system delivered metal from a single side at the bottom of the casting. The filling time was 15.4 seconds. Because the metal entered from below, the free surface rose smoothly and steadily. The fill time distribution (Figure 4-12) shows that after the initial 10% of the volume, all points within the same horizontal level were filled almost simultaneously. This is the most favorable condition for avoiding oxidation and entrainment. No sign of sand erosion was observed.
5.4.2 Solidification Process
The solidification temperature fields are shown in Figure 4-13. After filling, the liquid metal in the ingates and the risers remained hot, providing a continuous feeding path. At 46 seconds, the risers were actively feeding the top sections. At 126 seconds, the thick lower sections had solidified while the risers still contained liquid metal. At 236 seconds, only a few isolated liquid islands remained, and they were small. The solid fraction distribution in Figure 4-14 confirms that no large isolated liquid regions formed. The shrinkage porosity prediction in Figure 4-15 shows only small, dispersed microporosity, which is acceptable for this type of casting.
5.5 Comparison and Selection
I compared the three options based on filling smoothness, tendency to form defects, and degree of sequential solidification. The table below summarizes my evaluation:
| Criteria | Option 1 | Option 2 | Option 3 |
|---|---|---|---|
| Filling time (s) | 15.69 | 13.98 | 15.43 |
| Filling stability | Moderate | Moderate, with flow merging | Excellent |
| Turbulence / air entrapment | Possible | Possible at merge zone | Minimal |
| Sequential solidification | Fair | Fair | Good |
| Predicted shrinkage porosity | Concentrated in three areas | One large area | Dispersed micro-porosity |
| Overall process reliability | Low | Medium | High |
Based on these results, I selected Option 3, the unilateral bottom gating system, as the optimal design for trial production. This configuration provides the smoothest filling, minimizes defect formation, and ensures proper feeding through both the risers and the ingates.
6. Rapid Sand Casting Mold Design and Production
6.1 Mold Assembly
Using the selected gating system, I designed the complete sand mold assembly in UG NX. The assembly included the bottom plate, lower sand core, oil gallery cores, water jacket core, intake/exhaust port cores, upper oil gallery core, and the top core. The outer mold was formed by ramming resin sand around the SLA patterns of the two halves. The SLA patterns were made from a photosensitive resin with a layer thickness of 0.1 mm. The patterns were coated with a release agent before molding.
The three-dimensional exploded view of the mold assembly is shown in Figure 5-1. The mold design considered core print clearances, assembly sequence, and extraction direction. All cores were designed with sufficient draft angles (approximately 3 degrees) to allow easy removal. Venting channels were drilled in the mold and cores to allow gas escape during pouring.
6.2 Selection of Mold Material
For the rapid sand casting foundry process, I selected phenolic urethane resin sand as the molding material. This material offers high strength, excellent dimensional stability, good collapsibility, and fast hardening at room temperature. The properties of the base sand are critical. I used a natural washed silica sand with a grain fineness number of about 50 (A.F.S. 50-60). The specifications are listed below:
| Property | Requirement |
|---|---|
| SiO₂ content | > 85% |
| Moisture content | < 0.2% |
| Clay content | < 0.5% |
| Sand temperature | 20–35 °C |
| Grain shape | Round or sub-round |
6.3 Optimization of Resin Ratios
I conducted a series of tensile strength tests to determine the optimal ratio of the two resin components. The experiment was a two-factor, four-level orthogonal design. The components were mixed with the sand in a roller mixer for two minutes. The tensile strength of the cured sand specimens was measured after 24 hours. The results are summarized in Figure 5-2 of the thesis. The main findings are:
- If either component was less than 0.8% by mass of the sand, the compressive strength was insufficient.
- When the proportion of component I was between 0.8% and 1% and the ratio of component I to component II was 1:1, the tensile strength reached approximately 1.1 MPa, which is adequate for molding and core making.
- Increasing the resin content beyond 2% did not significantly improve strength and increased cost, as well as making sand removal more difficult.
Thus, I selected a ratio of component I : component II = 1:1, with component I at 0.8% to 1% of the sand mass. This gave a good balance between strength, collapsibility, and cost for the rapid sand casting foundry process.
6.4 Trial Production
After preparing the sand molds and cores, I assembled them in the correct sequence. The cores were placed using core prints as guides. Surface coatings were applied to the cores to improve surface finish and prevent metal penetration. The mold was closed and clamped. The ZL105 alloy was melted in a resistance furnace and degassed with hexachloroethane tablets. The molten metal was poured at 690 °C with a pouring speed of about 0.22 m/s. After solidification and cooling, the sand was broken out using a vibration table. The casting was cleaned and inspected.
Figure 5-4 shows the production process: (a) installation of intake and exhaust port cores, (b) installation of water jacket core, and (c) the final aluminum cylinder head casting. Visual inspection revealed a complete filling of the thin walls and internal passages. No obvious shrinkage cavities or cracks were observed. The casting was tested for air tightness using a water pressure test, and it passed the required standard. The dimensional measurements showed that the casting was within the tolerance range of CT10. This confirmed that the rapid sand casting process, combined with numerical simulation, is a viable method for producing complex aluminum cylinder heads.
7. Discussion
The successful trial production validates the numerical simulation results. The bottom gating system was found to be superior because it eliminates the direct impact of the falling metal stream, reduces turbulence, and provides a calm, rising liquid surface. This is particularly important for aluminum alloys, which easily form oxide films that can become trapped in the casting. The experimental IHTC values significantly improved the accuracy of the simulation, as verified by the temperature validation.
The use of SLA patterns instead of traditional metal patterns reduced the development time from months to weeks. The SLA patterns were accurate enough for sand molding, and their slightly lower strength compared to metal was not a problem because the sand compaction forces are moderate. Moreover, the SLA patterns could be repaired or modified easily if design changes were required.
The rapid sand casting foundry process is not only limited to cylinder heads. It can be applied to other complex castings such as intake manifolds, exhaust manifolds, and transmission housings. The simulation approach used here can be extended to optimize feeding systems, chill placement, and cooling channel designs. Future work should include stress analysis to predict hot tearing and residual stresses, as well as the coupling of microstructure evolution models to predict mechanical properties.
8. Conclusions
From my research, I draw the following conclusions:
- A complete rapid sand casting process for an aluminum cylinder head was successfully developed. The process combines SLA pattern making, phenolic urethane resin sand molding, and conventional gravity casting.
- The interfacial heat transfer coefficient between ZL105 aluminum alloy and phenolic urethane resin sand was experimentally determined using an inverse method. The values are: 811 W/(m²K) above the liquidus temperature, decreasing to 310 W/(m²K) at the solidus temperature, and stabilizing at about 230 W/(m²K) below the solidus.
- Numerical simulation using ProCAST showed that the unilateral bottom gating system produced the best filling and solidification behavior. This system minimized turbulence, avoided isolated liquid regions, and yielded only dispersed microporosity.
- The optimal pouring temperature was 690 °C and the optimal pouring speed was 0.22 m/s for the selected gating system.
- For the phenolic urethane resin sand, the optimal ratio of component I to component II was 1:1, with component I constituting 0.8% to 1% of the total sand mass. This provided adequate tensile strength of approximately 1.1 MPa for sand molding.
- The trial production yielded sound castings that met dimensional and pressure tightness requirements, confirming the feasibility of the rapid sand casting foundry approach for complex automotive components.
9. Future Prospects
Looking ahead, my work can be extended in several directions. First, the accuracy of the simulation can be further improved by incorporating additional phenomena such as oxide film formation, gas evolution from the sand, and the effect of mold permeability. Second, coupled thermal-stress analysis should be performed to predict distortion and hot tears in the casting. This would enable the optimization of mold rigidity and the design of more complex core geometries. Third, the rapid sand casting process could be combined with automated core-making and mold assembly systems to increase productivity for medium-batch production. Finally, machine learning algorithms could be applied to the simulation results to create surrogate models for rapid process optimization, thereby reducing the time required for iterative simulation studies.
In addition, the method of measuring interfacial heat transfer coefficients could be refined by using multiple thermocouples and by considering the effect of a protective coating on the mold surface. Different coatings may alter the heat transfer behavior, and this information is valuable for practical foundry engineers. With these improvements, the rapid sand casting foundry will become an even more powerful tool for the efficient development of high-quality metal castings.
In summary, this thesis demonstrates that the integration of CAD, rapid prototyping, experimental heat transfer measurement, and numerical simulation provides a robust engineering platform for developing new casting processes. The sand casting foundry industry can greatly benefit from these methods, especially in the context of increasing demand for lightweight automotive components and faster product innovation cycles. My research has laid a solid foundation for further studies in this exciting field.