In this thesis, I systematically investigate the rapid sand casting process and sand mold technology for a machine tool spindle box. The work combines casting process design, numerical simulation, multi-objective optimization, and experimental verification. I focus on minimizing casting defects such as shrinkage porosity and gas porosity through a comprehensive methodology that integrates ProCAST simulation, response surface methodology (RSM), and genetic algorithm (GA) optimization. In addition, I study the influence of furan resin and curing agent contents on the performance of 3D-printed sand molds. The results show that the optimized pouring parameters significantly reduce casting defects, and the recommended sand mold composition provides excellent mechanical properties and permeability. This study offers a reliable technical reference for the application of sand mold rapid prototyping in the production of high-quality machine tool components.
1. Introduction
Castings play an essential role in modern manufacturing, particularly in automotive, aerospace, and industrial machinery. With the global demand for lightweight structures, high precision, and shorter development cycles, traditional casting methods face increasing challenges. The machine tool spindle box is a critical component that determines the accuracy and rigidity of a machine tool. Its quality directly affects the performance of the whole machine. However, the spindle box has complex internal structures, significant wall thickness variations, and several hot spots, which make it prone to casting defects such as shrinkage, gas porosities, and cold shut. Therefore, I need to design a robust casting process and optimize the sand mold technology to achieve defect-free castings.
Sand mold rapid prototyping technology has emerged as a promising alternative to conventional pattern-based mold making. It allows the direct fabrication of sand molds from CAD models using additive manufacturing (3D printing) or numerically controlled machining. This technology eliminates the need for expensive and time-consuming patterns, enables design changes with minimal cost, and improves the accuracy of mold dimensions. In the past decades, many researchers have studied the effects of process parameters on sand mold quality and casting integrity. However, the application of sand mold rapid prototyping to large machine tool castings is still not fully exploited. I aim to fill this gap by developing a complete workflow for a machine tool spindle box, from mold design to casting verification.
In this thesis, I address several key aspects. First, I analyze the structural features and casting difficulties of the spindle box, and design a sand mold split scheme suitable for rapid prototyping. Second, I design two different gating systems (bottom-gating and parting-gating) and evaluate their performance using ProCAST numerical simulation. I then optimize the gating system with chills and insulating risers to reduce casting defects. Third, I carry out a single-factor study and response surface design to explore the influence of pouring temperature, pouring time, and mold temperature on porosity. A regression model is established and a genetic algorithm is used to find the optimal process parameters. Finally, I experimentally study the effects of resin and curing agent content on gas evolution, mechanical strength, and permeability of 3D-printed sand molds, and verify the final process by manufacturing a real spindle box casting.
2. Casting Process and Sand Mold Design
2.1 Structural Analysis and Difficulties
The machine tool spindle box considered in this study weighs 217 kg and has overall dimensions of 670 mm × 600 mm × 331 mm. The maximum wall thickness is 75 mm while the minimum is 20 mm. The material is HT300 gray cast iron, which offers good wear resistance and damping capacity. Table 1 lists the chemical composition of HT300.
| Element | C (%) | S (%) | P (%) | Mn (%) | Si (%) |
|---|---|---|---|---|---|
| Content | 2.9–3.2 | ≤0.12 | ≤0.15 | 0.5–1.4 | 1.0–2.5 |
The casting defects of this component are likely to appear in several regions. The internal bores for the main spindle and gear shafts intersect with the outer walls, creating multiple hot spots. Moreover, the guide rails have a maximum thickness of 75 mm, while the side walls are only 20 mm thick. This large difference in wall thickness leads to inappropriate solidification patterns, resulting in shrinkage defects. I therefore need a careful casting design to minimize these casting defects.
2.2 Sand Mold Split Design
Rapid sand mold prototyping imposes specific requirements on the mold splitting. In my design, I select the top surface of the guide rails as the parting plane. This avoids interfering with the critical functional surfaces and facilitates cleaning. I set the casting shrinkage rate to 1%, and for machining allowance I apply 4 mm on surfaces that require subsequent machining. The shrinkage rate is calculated as
$$ K = \frac{L_0 – L_1}{L_0} \times 100\% $$
where \(K\) is the casting shrinkage, \(L_0\) is the mold dimension, and \(L_1\) is the casting dimension. The sand mold is divided into an upper mold, a lower mold, and a set of cores. Using the surface parting method in NX 12.0, I obtain the three-dimensional sand mold model. The upper mold has a relatively simple shape and is suitable for CNC machining, while the lower mold and cores contain undercuts that are better produced by sand 3D printing. This hybrid approach reduces manufacturing time and improves mold accuracy.
2.3 Gating System Design
Two types of gating systems are designed: a parting-line (middle gating) system and a bottom gating system. The gating system must ensure smooth filling, prevent oxide inclusions, and allow proper gas evacuation. I used the following empirical formula to estimate the pouring time:
$$ \tau = \sqrt{G} + G $$
where \( \tau \) is the pouring time (s) and \( G \) is the total pouring mass (kg). With \(G \approx 230\) kg, I obtain \( \tau \approx 21\) s. To validate this value, I check the liquid metal rising speed using
$$ v = \frac{C}{\tau} $$
where \(C = 331\) mm is the casting height. Thus, \(v \approx 15.8\) mm/s, which is within the recommended range for cast iron. For the parting-line gating system, the hydrostatic head \(H_0\) is 229 mm, the cavity height \(P\) above the gating line is 141 mm, and the average pressure head is
$$ H_p = H_0 – \frac{P^2}{2C} $$
Substituting values gives \(H_p \approx 199\) mm. The choke area can be calculated from the hydraulic formula
$$ S = \frac{m}{\rho \mu \tau \sqrt{2g H_p}} $$
where \(m\) is the total metal mass, \(\rho\) is the density of molten iron, \(\mu\) is the flow coefficient, and \(g\) is the gravitational acceleration. With \(\mu = 0.47\), I obtain \(S \approx 15.25\) cm² for the ingate. The cross-sectional area ratio for the gating system is chosen as \(S_{\text{ingate}}:S_{\text{runner}}:S_{\text{sprue}} = 1:1.5:1.2\). Therefore, the runner area is 22.88 cm² and the sprue area is 18.30 cm². The corresponding dimensions are visualized in the three-dimensional model. For the bottom gating system, \(P = C\), and the average pressure head becomes
$$ H_p = H_0 – \frac{P}{2} $$
With \(H_0 = 359\) mm and \(P = 331\) mm, I get \(H_p \approx 193.5\) mm. The flow coefficient is taken as 0.52, giving an ingate area of 13.98 cm², a runner area of 20.85 cm², and a sprue area of 16.68 cm². Both systems are equipped with risers to feed the solidification shrinkage and to allow gas to escape. The initial design predicts that casting defects may appear in thick sections due to insufficient feeding, which motivates the numerical simulation in the next section.
3. Numerical Simulation and Optimization
3.1 Simulation Setup and Mathematical Models
I use the finite-element based ProCAST software to simulate the mold filling and solidification of the spindle box. The simulation of the filling process is governed by the continuity equation
$$ \frac{\partial \rho}{\partial t} + \nabla \cdot \left( \rho \mathbf{V} \right) = 0 $$
and the Navier–Stokes equation for incompressible fluids
$$ \rho \frac{D \mathbf{V}}{Dt} = \mu \nabla^2 \mathbf{V} – \nabla P + \rho \mathbf{g} $$
where \(\rho\) is the density, \(\mathbf{V}\) the velocity vector, \(\mu\) the dynamic viscosity, \(P\) the pressure, and \(\mathbf{g}\) the gravitational acceleration. The energy conservation during filling is written as
$$ \rho c_p \left( \frac{\partial T}{\partial t} + u \frac{\partial T}{\partial x} + v \frac{\partial T}{\partial y} + w \frac{\partial T}{\partial z} \right) = \nabla \cdot \left( \lambda \nabla T \right) $$
where \(T\) is the temperature, \(c_p\) the specific heat, and \(\lambda\) the thermal conductivity. To track the free surface, I employ the volume-of-fluid equation
$$ \frac{\partial F}{\partial t} + u \frac{\partial F}{\partial x} + v \frac{\partial F}{\partial y} + w \frac{\partial F}{\partial z} = 0 $$
Solidification is modeled using latent heat release and the Niyama criterion to predict micro-porosity. The Niyama parameter is given by
$$ G / \sqrt{v_c} < C_{\text{Niyama}} $$
where \(G\) is the temperature gradient, \(v_c\) is the cooling rate, and \(C_{\text{Niyama}}\) is the critical value. When the Niyama parameter falls below the critical value, shrinkage porosity (a common casting defect) is expected.
3.2 Mesh Independence Study
Before finalizing the simulation, I perform a mesh independence analysis to balance computational cost and accuracy. The mold model is discretized with four different mesh sizes. The results are summarized in Table 2.
| Case | Sand max size (mm) | Cast max size (mm) | Surface elements | Volume elements | Porosity (%) |
|---|---|---|---|---|---|
| 1 | 10 | 8 | 168785 | 2429733 | 37.053 |
| 2 | 8 | 6 | 277501 | 5017050 | 37.631 |
| 3 | 6 | 4 | 449216 | 7203941 | 38.828 |
| 4 | 4 | 2 | 653113 | 11867570 | 38.862 |
From case 3 to case 4, the predicted porosity changes only slightly, while the mesh size increases dramatically. Therefore, I choose the mesh parameters of case 3 for all subsequent simulations.
3.3 Initial Simulation Results and Casting Defects
I simulate both gating systems with initial conditions: pouring temperature 1340 °C, pouring time 21 s, mold temperature 24 °C, and air cooling at the mold outer surface. The mold material is silica sand, except for the lower mold in the bottom gating case where I use chromite sand to enhance cooling at thick sections. The heat transfer coefficients are summarized in Table 3.
| Interface | Heat transfer coefficient (W/(m²·K)) |
|---|---|
| Casting – silica sand | 500 |
| Casting – chromite sand | 900 |
| Casting – chill | 2000 |
| Chill – silica/chromite | 300 |
| Insulating riser – sand | 100 |
For the parting-line gating system, the filling process is fairly stable, but after solidification starts, the thin outer walls solidify earlier than the thick internal bores. This creates isolated liquid pools, and the risers freeze too early to feed the casting defects. The simulation predicts shrinkage porosity in the guide rails and at the junctions between bores and walls. The maximum porosity reaches 2.3%, which exceeds the acceptable threshold for machine tool castings.
For the bottom gating system, the initial results show a similarly unsatisfactory solidification sequence. The lower part solidifies faster due to the higher heat transfer coefficient of chromite sand, but the upper part solidifies quickly because of the thin wall thickness. As a result, feeding channels to the thick rail regions are closed, causing casting defects near the rails and internal bores.
3.4 Optimized Gating System
To eliminate these casting defects, I place insulating risers at the last solidifying locations. In addition, I attach chills on the internal bore surfaces and guide rails to increase local cooling rates and keep the feeding path open. For the optimized parting-line system, four insulating risers are added on the top, and two additional risers serve the hot spots. Chills are positioned around the bores to accelerate cooling. The improved simulation shows that the solidification proceeds from bottom to top. The temperature profiles at five characteristic points confirm that point 5 (near the riser) remains hot until late in the solidification process. After optimization, the maximum porosity is reduced to 1.358%, which is below the 2% industry limit. The porosity is mainly confined to the risers, which are subsequently removed.
For the optimized bottom gating system, chills are placed on the thick rail sections and the bores, and the original risers are replaced with insulating risers. The simulation shows a sound solidification sequence without isolated liquid pools. The maximum porosity after optimization is 1.685%, still below 2%, but higher than the optimized parting-line system. Therefore, I select the optimized parting-line gating system as the baseline for further process parameter optimization.
4. RSM and GA based Process Parameter Optimization
4.1 Single-Factor Analysis
Using the optimized parting-line system, I investigate the influence of casting process parameters on the porosity of the casting, because porosity is a critical casting defect indicator. I first perform a single-factor study by varying one parameter at a time while keeping the others constant.
I set the pouring temperature to 1380 °C and pouring time to 22 s, and vary the mold temperature from 22 °C to 28 °C. The porosity results are listed in Table 4. As the mold temperature increases from 22 °C to 26 °C, the porosity decreases due to better filling of narrow gaps. The minimum porosity of 1.343% occurs at 26 °C. Above 26 °C, the porosity increases because the metal stays liquid longer and traps gas.
| Mold temperature (°C) | 22 | 23 | 24 | 25 | 26 | 27 | 28 |
|---|---|---|---|---|---|---|---|
| Porosity (%) | 1.386 | 1.362 | 1.353 | 1.345 | 1.343 | 1.353 | 1.359 |
Next, I fix the mold temperature at 26 °C and the pouring time at 22 s, and vary the pouring temperature from 1340 °C to 1400 °C. Table 5 shows that the porosity decreases initially, reaching a minimum of 1.316% at 1360 °C, after which it increases due to higher gas solubility and turbulence. The optimal pouring temperature is near 1360 °C.
| Pouring temperature (°C) | 1340 | 1350 | 1360 | 1370 | 1380 | 1390 | 1400 |
|---|---|---|---|---|---|---|---|
| Porosity (%) | 1.366 | 1.346 | 1.316 | 1.321 | 1.343 | 1.352 | 1.358 |
Finally, I fix the pouring temperature at 1360 °C and the mold temperature at 26 °C, and vary the pouring time from 20 s to 26 s. The results are given in Table 6. The lowest porosity of 1.308% appears at 23 s. Short pouring times cause turbulent filling and gas entrapment, while long pouring times lead to premature metal cooling and cold shut; both conditions produce casting defects.
| Pouring time (s) | 20 | 21 | 22 | 23 | 24 | 25 | 26 |
|---|---|---|---|---|---|---|---|
| Porosity (%) | 1.342 | 1.336 | 1.316 | 1.308 | 1.323 | 1.328 | 1.338 |
4.2 Response Surface Design
Based on the single-factor results, I select a narrow range around the best values: pouring temperature between 1340 °C and 1380 °C, pouring time between 21 s and 25 s, and mold temperature between 24 °C and 28 °C. Using a Box–Behnken design with three factors and three levels, I construct 17 simulation experiments. The factors are coded as \(A\) (pouring temperature), \(B\) (pouring time), and \(C\) (mold temperature). The response is the porosity \(\eta\). The experimental design and simulated porosity values are shown in Table 7.
| Run | A: Pouring temperature (°C) | B: Pouring time (s) | C: Mold temperature (°C) | Porosity \(\eta\) (%) |
|---|---|---|---|---|
| 1 | 1360 | 21 | 24 | 1.310 |
| 2 | 1360 | 23 | 26 | 1.307 |
| 3 | 1360 | 23 | 26 | 1.306 |
| 4 | 1340 | 21 | 26 | 1.346 |
| 5 | 1360 | 23 | 26 | 1.305 |
| 6 | 1360 | 23 | 26 | 1.311 |
| 7 | 1360 | 25 | 28 | 1.325 |
| 8 | 1340 | 25 | 26 | 1.350 |
| 9 | 1380 | 21 | 26 | 1.371 |
| 10 | 1380 | 23 | 28 | 1.360 |
| 11 | 1380 | 25 | 26 | 1.352 |
| 12 | 1340 | 23 | 24 | 1.341 |
| 13 | 1360 | 25 | 24 | 1.315 |
| 14 | 1360 | 21 | 28 | 1.338 |
| 15 | 1360 | 23 | 26 | 1.308 |
| 16 | 1340 | 23 | 28 | 1.342 |
| 17 | 1380 | 23 | 24 | 1.332 |
By fitting a quadratic polynomial to these data, I obtain the regression equation:
$$ \eta = 1.31 + 0.0045 A – 0.0029 B + 0.0084 C – 0.0058 AB + 0.0067 AC – 0.0045 BC + 0.0346 A^2 + 0.0129 B^2 $$
where \(A\), \(B\), and \(C\) are coded variables. The analysis of variance (ANOVA) is summarized in Table 8.
| Source | Sum of squares | df | Mean square | F-value | P-value |
|---|---|---|---|---|---|
| Model | 0.0072 | 8 | 0.0009 | 78.42 | <0.0001 |
| A | 0.0002 | 1 | 0.0002 | 14.15 | 0.0055 |
| B | 0.0001 | 1 | 0.0001 | 5.78 | 0.0430 |
| C | 0.0006 | 1 | 0.0006 | 49.01 | 0.0001 |
| AB | 0.0001 | 1 | 0.0001 | 11.55 | 0.0094 |
| AC | 0.0002 | 1 | 0.0002 | 15.92 | 0.0040 |
| BC | 0.0001 | 1 | 0.0001 | 7.07 | 0.0288 |
| A² | 0.0051 | 1 | 0.0051 | 442.64 | <0.0001 |
| B² | 0.0007 | 1 | 0.0007 | 61.32 | <0.0001 |
| Residual | 0.0001 | 8 | 0.0000 | ||
| Lack of fit | 0.0001 | 4 | 0.0000 | 3.32 | 0.1359 |
| Pure error | 0.0000 | 4 | 5.3×10⁻⁶ | ||
| Total | 0.0073 | 16 |
The model P-value is less than 0.0001, indicating that the regression model is highly significant. The lack-of-fit P-value is 0.1359, meaning the model adequately fits the data. The coefficient of determination \(R^2\) is 0.9874, the adjusted \(R^2_{\text{Adj}}\) is 0.9748, and the predicted \(R^2_{\text{Pre}}\) is 0.9023. The signal-to-noise ratio is 24.64, well above 4, which confirms excellent model accuracy. The F-values in Table 8 reveal the influence order of process parameters: mold temperature (C) has the greatest effect, followed by pouring temperature (A), and then pouring time (B). Furthermore, among the interaction terms, the interaction between pouring temperature and mold temperature (AC) is more significant than AB and BC.
To visualize the effect of the parameters and their interactions, I generate response surface plots. The interaction between A and B shows that the minimum porosity occurs around A=0 and B=0, which corresponds to 1360 °C and 23 s. The interaction between A and C similarly indicates a concave response surface. The interaction between B and C confirms that both parameters should be kept near their central values. In all cases, the porosity increases when the parameters move away from the central point, indicating that the optimal region lies within the selected range.
4.3 Genetic Algorithm Optimization
Since the regression equation is nonlinear, I use a genetic algorithm to find the global optimum. The optimization problem is formulated as
$$ \min \eta(A, B, C) $$
subject to the constraints
$$ 1340 \leq A \leq 1380, \quad 21 \leq B \leq 25, \quad 24 \leq C \leq 28 $$
I implement the genetic algorithm with a population size of 50, single-point crossover with probability 0.8, and non-uniform mutation with probability 0.01. The algorithm converges after 56 iterations. The convergence curve shows that the fitness function steadily decreases and eventually stabilizes at the global minimum. The optimal parameters obtained by GA are:
- Pouring temperature \(A = 1360\) °C
- Pouring time \(B = 23\) s
- Mold temperature \(C = 24\) °C
- Minimum predicted porosity \(\eta = 1.301\)%
I then run a verification simulation with these optimal parameters. The simulated porosity is 1.306%, which is very close to the predicted value. The relative error is only 0.38%. Compared with the initial unoptimized process (porosity 1.358%), the optimized process reduces the casting defect level by about 4.2%. This confirms that the combination of RSM and GA is effective for determining the optimal casting process parameters and minimizing casting defects.
5. Sand Mold Performance Study
5.1 Experimental Materials and Equipment
In addition to the casting process parameters, the properties of the sand mold itself strongly influence the formation of casting defects. For sand mold rapid prototyping, the binder and curing agent contents determine the mold strength, gas evolution, and permeability. I use a 70/140 mesh silica sand as the base material. The furan resin and sulfonic acid curing agent are chosen as the binder system. The properties of the resin and curing agent are listed in Tables 9 and 10.
| Density (g/cm³) | Viscosity (mPa·s) | Nitrogen content (%) | pH | Particle size (μm) |
|---|---|---|---|---|
| 1.1–1.2 | 9.5–12.5 | ≤0.5 | 6.0–7.5 | ≤0.5 |
| Density (g/cm³) | Viscosity (mPa·s) | Total acidity (%) | Free acid (%) |
|---|---|---|---|
| 1.2–1.3 | ≤15 | 18–20 | ≤1.5 |
I prepare standard test specimens using a laboratory sand mixer and specimen molding machine. The specimens are used to measure gas evolution, tensile strength, bending strength, and permeability.
5.2 Gas Evolution Test
Gas evolution is measured with an intelligent gas evolution tester. The furnace temperature is set to 850 °C, and 1 g of mixed sand is tested. The results are presented in Table 11 for different resin and curing agent contents. The gas evolution increases almost linearly with both resin and curing agent contents. For gray iron castings, the acceptable gas evolution is generally below 20 mL/g. Therefore, I need to control the resin content in the range of 1.7%–2.4% and the curing agent content in the range of 0.2%–0.4%.
| Resin content | Curing agent content | ||||
|---|---|---|---|---|---|
| 0.2% | 0.3% | 0.4% | 0.5% | 0.6% | |
| 1.7% | 13.73 | 14.63 | 15.65 | 16.64 | 18.64 |
| 2.4% | 18.52 | 19.18 | 19.58 | 20.07 | 20.56 |
| 3.1% | 20.69 | 21.61 | 22.05 | 22.39 | 23.19 |
| 3.8% | 22.82 | 23.47 | 24.46 | 24.89 | 26.01 |
5.3 Mechanical Strength Tests
I perform tensile strength tests on “8-shaped” specimens and bending strength tests on cylindrical specimens (φ50 mm × 50 mm). Table 12 lists the tensile strength results. For molding sand, the tensile strength should be greater than 1.0 MPa. At a resin content of 2.4%, all curing agent contents meet this requirement. The maximum tensile strength of 1.72 MPa occurs at 0.4% curing agent, but the strength at 0.3% is also acceptable (1.62 MPa).
| Resin content | Curing agent content | ||||
|---|---|---|---|---|---|
| 0.2% | 0.3% | 0.4% | 0.5% | 0.6% | |
| 1.7% | 0.84 | 0.97 | 1.14 | 0.81 | 0.74 |
| 2.4% | 1.35 | 1.62 | 1.72 | 1.46 | 1.41 |
| 3.1% | 1.92 | 2.13 | 2.12 | 2.32 | 1.79 |
| 3.8% | 2.03 | 2.56 | 2.61 | 2.78 | 2.15 |
The bending strength results are shown in Table 13. Similar to tensile strength, the bending strength increases with resin and curing agent contents. At a resin content of 2.4%, the bending strength reaches a maximum of 3.93 MPa at 0.4% curing agent; the value at 0.3% is 3.87 MPa, which is only 1.5% lower. Considering the requirements for minimal gas evolution and sufficient strength, a curing agent content of 0.3% is preferred.
| Resin content | Curing agent content | ||||
|---|---|---|---|---|---|
| 0.2% | 0.3% | 0.4% | 0.5% | 0.6% | |
| 1.7% | 2.23 | 2.56 | 2.67 | 2.33 | 2.08 |
| 2.4% | 3.58 | 3.87 | 3.93 | 3.71 | 3.53 |
| 3.1% | 3.97 | 4.08 | 4.16 | 4.31 | 3.72 |
| 3.8% | 4.13 | 4.34 | 4.42 | 4.54 | 4.05 |
Moreover, I notice that a curing agent content of 0.4% tends to produce laminations during 3D printing. The higher amount of curing agent increases the agglomeration of sand particles, which reduces the flowability of the sand and creates voids during spreading. Therefore, I select a resin content of 2.4% and a curing agent content of 0.3% as the optimal formula for the 3D-printed sand mold.
5.4 Permeability Test
Using the optimal formula, I measure the permeability of five cylindrical specimens with an electric permeability tester. The results are given in Table 14. The average permeability is 159.9 cm³/min·kPa, which is significantly higher than the typical value of 100 cm³ for conventional sand molds. High permeability promotes the efficient escape of gas generated during pouring, thus reducing gas-related casting defects.
| Test number | 1 | 2 | 3 | 4 | 5 | Average |
|---|---|---|---|---|---|---|
| Permeability (cm³) | 160.0 | 158.5 | 160.1 | 159.7 | 160.5 | 159.9 |
6. Sand Mold Fabrication and Verification
6.1 3D Printing of the Sand Mold
Based on the optimized gating system and the selected sand formula, I prepare the final sand mold design. The lower mold and cores are fabricated using a digital sand mold 3D printing machine. The printing process includes model slicing, parameter setup, and layer-by-layer deposition. The layer thickness is 0.4 mm, the image resolution is 360 DPI, and the nozzle width is set to 1000 pixels. I also apply a contour scaling of -0.25 mm to compensate for printing tolerances. During printing, the room temperature is controlled at 22 °C and the relative humidity at 40% to stabilize the curing reaction. The upper mold, which has no undercuts, is machined from a compacted sand blank using a CNC sand milling machine. The processing code is generated with NX 12.0 CAM module. A 50 mm flat-end mill is used for the outer shape, a 16 mm flat-end mill for rough cavity machining, and an 8 mm ball-end mill for finishing fine features.
6.2 Assembly and Pouring
The printed and machined sand parts are assembled as shown in the experimental setup. I set a 1 mm core print allowance between the core and the mold to ensure smooth core setting. The cores are inserted without difficulty, and the upper and lower molds match perfectly. The final mold is placed on the foundry floor, and the ingates are connected to the pouring cup. According to the optimized process parameters, I set the pouring temperature to 1360 °C and the pouring time to approximately 23 s. The mold temperature is 24 °C. The molten HT300 gray iron is poured manually. During pouring, the metal rises calmly and the gas is smoothly vented through the risers. After solidification and cooling, the mold is shaken out and the casting is cleaned.
6.3 Casting Quality
The resulting spindle box casting is complete and free of visible surface defects. The dimensions are measured as 680 mm × 610 mm × 340 mm, which is in good agreement with the theoretical dimensions after allowing for shrinkage and machining allowances. The machined surfaces show a dense structure without visible shrinkage cavities. No cracks, cold shuts, or gas porosity are observed macroscopically. The experimental casting confirms that the optimized gating system and process parameters effectively suppress casting defects. Figure below shows an example of a complex cast iron engine cylinder block that can benefit from the same rapid sand casting technology.
7. Conclusions
In this thesis, I have developed a complete rapid sand casting process for a machine tool spindle box and verified it through numerical simulation and experiments. The main conclusions are as follows:
- The structural characteristics of the spindle box make it susceptible to casting defects in thick sections and hot spots. The proposed sand mold design, which divides the mold into an upper machined part and a lower printed part, is both feasible and efficient.
- ProCAST simulation is a powerful tool for predicting casting defects. Initially, both parting-line and bottom gating systems produced shrinkage defects. After adding chills and insulating risers, the solidification sequence is optimized, and the maximum porosity is reduced to 1.358% for the parting-line system and 1.685% for the bottom gating system. The parting-line system is ultimately selected because it yields lower porosity.
- The single-factor and response surface analyses show that the casting process parameters significantly affect porosity. The regression model has high accuracy and reliability. The optimal parameters are a pouring temperature of 1360 °C, a pouring time of 23 s, and a mold temperature of 24 °C, with a predicted porosity of 1.301%. Experimental verification gives 1.306%, a deviation of only 0.38%.
- The sand mold formula affects gas evolution, mechanical strength, and permeability. A furan resin content of 2.4% and a curing agent content of 0.3% provide a tensile strength of 1.62 MPa, a bending strength of 3.87 MPa, a gas evolution of 19.18 mL/g, and a permeability of 159.9 cm³. This formula is suitable for 3D printing because it avoids delamination and ensures adequate gas escape.
- The final trial casting demonstrates that the combined methodology of simulation-based design, statistical optimization, and sand mold performance evaluation is effective for producing high-quality machine tool spindle box castings. The proposed framework can be extended to other complex castings to reduce casting defects and improve production efficiency.