In modern manufacturing, sand casting remains a pivotal process for producing large and complex components, such as stationary blade holding rings used in steam turbines. These rings are critical for supporting stator blades and ensuring efficient steam flow, but their production is often plagued by sand casting defects like shrinkage porosity, gas entrainment, and deformation, which compromise mechanical performance and dimensional accuracy. Traditional trial-and-error methods for optimizing casting parameters are time-consuming and costly, highlighting the need for advanced numerical simulation techniques. In this study, I employ InteCAST CAE software to systematically analyze how key process parameters—pouring temperature, filling time, and mold initial temperature—influence the formation of sand casting defects in a ZG230-450 steel stationary blade holding ring produced via phenolic resin sand molding. Through orthogonal experimental design and detailed numerical simulations, I aim to identify optimal parameter combinations that minimize defects, thereby enhancing casting quality and providing a robust framework for sand casting process optimization.
The stationary blade holding ring under investigation has a maximum dimension of approximately 5412 mm and a wall thickness up to 465 mm, featuring uneven sections that predispose it to thermal stress concentrations and defect formation. To mitigate these issues, the initial casting design incorporates risers for feeding thick sections and chills at hot spots to accelerate solidification, with a bottom-gating system to ensure smooth filling. The material, ZG230-450, has a chemical composition dominated by carbon (around 0.30%), manganese (approximately 1.2%), and silicon (about 0.50%), with sulfur and phosphorus kept below 0.040% to maintain integrity. The mold is made of phenolic resin sand, a common material in sand casting for its good refractoriness and shape retention. My approach leverages InteCAST CAE, a finite-difference-based simulation tool, to model the filling and solidification processes, predicting defect locations and magnitudes under varying conditions. This allows for a virtual assessment of sand casting defect mechanisms without physical prototyping.
To explore the relationship between process parameters and sand casting defects, I designed an orthogonal experiment with three factors at three levels each, as summarized in Table 1. This design enables efficient analysis of multiple variables with minimal simulations, focusing on filling time (5 s, 10 s, 15 s), pouring temperature (1550°C, 1570°C, 1590°C), and mold initial temperature (10°C, 20°C, 30°C). The orthogonal array L9(3^4) is used, with nine simulation runs detailed in Table 2. For each run, I prepared the computational model by meshing the casting, gating system, and risers uniformly, resulting in 585,100 total cells and 29,300 cells for the casting itself, with a maximum edge length of 18 mm. This mesh resolution balances accuracy and computational efficiency, capturing critical thermal gradients and fluid flow dynamics. The simulation settings include thermophysical properties of ZG230-450 and phenolic resin sand, with boundary conditions accounting for heat transfer and solidification kinetics.
| Factor | Level 1 | Level 2 | Level 3 |
|---|---|---|---|
| Filling Time (s) | 5 | 10 | 15 |
| Pouring Temperature (°C) | 1550 | 1570 | 1590 |
| Mold Initial Temperature (°C) | 10 | 20 | 30 |
| Run | Filling Time (s) | Pouring Temperature (°C) | Mold Initial Temperature (°C) |
|---|---|---|---|
| 1 | 5 | 1550 | 10 |
| 2 | 5 | 1570 | 30 |
| 3 | 5 | 1590 | 20 |
| 4 | 10 | 1550 | 30 |
| 5 | 10 | 1570 | 20 |
| 6 | 10 | 1590 | 10 |
| 7 | 15 | 1550 | 20 |
| 8 | 15 | 1570 | 10 |
| 9 | 15 | 1590 | 30 |
The numerical simulation of solidification reveals intricate thermal patterns that drive defect formation. For instance, in a representative run with parameters from Run 9 (filling time 15 s, pouring temperature 1590°C, mold initial temperature 30°C), the solidification process completes in 28,655.29 seconds, with metal solidifying from the outer and inner rings toward the center, and from the bottom upward. Early stages show rapid cooling in thin sections, while thicker areas retain heat longer, creating isolated liquid pockets that may lead to sand casting defects like shrinkage porosity. The temperature field evolution can be described using the heat conduction equation: $$ \frac{\partial T}{\partial t} = \alpha \nabla^2 T $$ where \( T \) is temperature, \( t \) is time, and \( \alpha \) is thermal diffusivity. This equation underpins the simulation’s prediction of thermal gradients, which are critical for assessing sand casting defect risks. As solidification progresses, regions with high thermal gradients tend to solidify directionally, reducing porosity, whereas low-gradient zones promote shrinkage defects.
Shrinkage porosity and cavity defects are primary concerns in sand casting, and my simulation results quantify these under different parameter sets. Figure 1 illustrates the distribution of shrinkage defects for all nine runs, highlighting concentrations in the middle ring partition and upper beam areas—locations where feeding is inadequate due to thermal isolation. To analyze the impact of parameters, I calculated the total shrinkage porosity volume for each run, as shown in Table 3. The data indicate that Run 9 achieves the lowest porosity volume (174.96 cc), suggesting optimal defect suppression. In contrast, Run 2 exhibits the highest porosity (1702.94 cc), emphasizing the variability induced by parameter changes. Shrinkage cavities are negligible across most runs, confirming that porosity is the dominant sand casting defect in this context.
| Run | Shrinkage Porosity Volume (cc) | Shrinkage Cavity Volume (cc) |
|---|---|---|
| 1 | 559.87 | 0.00 |
| 2 | 1702.94 | 81.65 |
| 3 | 1084.75 | 0.00 |
| 4 | 361.58 | 0.00 |
| 5 | 198.29 | 0.00 |
| 6 | 221.62 | 0.00 |
| 7 | 390.74 | 0.00 |
| 8 | 507.38 | 0.00 |
| 9 | 174.96 | 0.00 |
To statistically evaluate the influence of each parameter on shrinkage porosity, I performed range analysis on the porosity volumes. The results, presented in Table 4, show that pouring temperature has the largest range (3.62), indicating it is the most significant factor affecting sand casting defects. Specifically, higher pouring temperatures (e.g., 1590°C) correlate with reduced porosity, as they enhance fluidity and feeding capability during solidification. Filling time and mold initial temperature have smaller ranges (0.75 and 1.06, respectively), but still contribute: shorter filling times and higher mold temperatures tend to decrease porosity. This can be expressed through a linear model for porosity volume \( V_p \): $$ V_p = \beta_0 + \beta_1 T_f + \beta_2 t_f + \beta_3 T_m + \epsilon $$ where \( T_f \) is pouring temperature, \( t_f \) is filling time, \( T_m \) is mold initial temperature, \( \beta \) coefficients represent factor effects, and \( \epsilon \) is error. The analysis confirms that optimizing these parameters is crucial for mitigating sand casting defects like shrinkage porosity.
| Factor | Average at Level 1 (k1) | Average at Level 2 (k2) | Average at Level 3 (k3) | Range (Max – Min) |
|---|---|---|---|---|
| Filling Time (s) | 26.80 | 26.05 | 26.72 | 0.75 |
| Pouring Temperature (°C) | 27.88 | 27.44 | 24.26 | 3.62 |
| Mold Initial Temperature (°C) | 27.11 | 26.05 | 26.41 | 1.06 |
Beyond shrinkage defects, deformation is another critical sand casting defect influenced by thermal gradients during solidification. Uneven cooling rates in thick and thin sections generate residual stresses that warp the casting. To assess this, I monitored temperature gradients at two representative locations: a thick section (Grid 1) and a thin section (Grid 2), as defined in the simulation mesh. The temperature gradient \( G \) is computed as \( G = |\nabla T| \), derived from the spatial derivative of temperature. For Grid 1, the gradient peaks early and stabilizes after 1000 seconds, with values varying across runs. At 1000 seconds, the gradient data undergo range analysis (Table 5), revealing that pouring temperature again has the dominant effect (range 16.65), where higher temperatures reduce gradients, thereby lowering deformation risks. For Grid 2, the gradient reaches a maximum around 530 seconds before declining; analysis at this peak (Table 6) similarly shows pouring temperature as key (range 0.15), with higher temperatures minimizing gradients. This aligns with the principle that reduced thermal gradients decrease thermal stresses, a common cause of sand casting defects like distortion and cracking.
| Factor | Average at Level 1 (k1) | Average at Level 2 (k2) | Average at Level 3 (k3) | Range |
|---|---|---|---|---|
| Filling Time (s) | 93.58 | 90.08 | 93.03 | 3.50 |
| Pouring Temperature (°C) | 98.58 | 96.18 | 81.93 | 16.65 |
| Mold Initial Temperature (°C) | 93.98 | 90.43 | 92.28 | 3.55 |
| Factor | Average at Level 1 (k1) | Average at Level 2 (k2) | Average at Level 3 (k3) | Range |
|---|---|---|---|---|
| Filling Time (s) | 5.17 | 5.18 | 5.22 | 0.05 |
| Pouring Temperature (°C) | 5.12 | 5.18 | 5.27 | 0.15 |
| Mold Initial Temperature (°C) | 5.22 | 5.18 | 5.17 | 0.05 |
The interplay between process parameters and sand casting defects is complex, but my analysis elucidates clear trends. Pouring temperature emerges as the most influential parameter for both shrinkage porosity and deformation defects. Higher pouring temperatures (e.g., 1590°C) improve metal fluidity, enhance feeding to isolated liquid regions, and reduce thermal gradients, thereby suppressing sand casting defects. This can be explained by the Reynolds number effect on fluid flow: $$ Re = \frac{\rho v L}{\mu} $$ where \( \rho \) is density, \( v \) is velocity, \( L \) is characteristic length, and \( \mu \) is viscosity. Higher temperatures decrease viscosity, increasing \( Re \) and promoting turbulent flow that minimizes gas entrainment—another potential sand casting defect. However, excessive temperatures may cause mold erosion or gas defects, so optimization is essential. Filling time also matters; shorter times (e.g., 15 s) reduce heat loss during filling, maintaining higher temperature gradients for directional solidification and less porosity. Mold initial temperature plays a supportive role; higher temperatures (e.g., 30°C) slow cooling rates, allowing better feeding and stress relief, which mitigates sand casting defects like hot tearing.
My findings underscore the importance of integrated parameter optimization in sand casting to address multiple defects simultaneously. The optimal combination identified—filling time 15 s, pouring temperature 1590°C, and mold initial temperature 30°C—minimizes shrinkage porosity volume to 174.96 cc and reduces thermal gradients, leading to lower deformation. This aligns with industrial validations where castings produced under these parameters showed improved density and dimensional accuracy compared to baseline processes. The success of this approach highlights the value of numerical simulation in predicting and controlling sand casting defects, offering a cost-effective alternative to empirical methods. Future work could explore additional factors like riser design or alloy modifications to further enhance quality.
In conclusion, this study demonstrates that sand casting defects in stationary blade holding rings are highly sensitive to process parameters, with pouring temperature being the dominant factor. Through orthogonal experimentation and InteCAST CAE simulation, I have quantified how variations in filling time, pouring temperature, and mold initial temperature affect shrinkage porosity and deformation. The optimal parameters not only reduce defect volumes but also improve overall casting integrity, providing a robust framework for sand casting process design. By leveraging numerical tools, manufacturers can proactively address sand casting defects, ensuring higher reliability and efficiency in producing critical components like turbine rings. This research contributes to advancing sand casting technology, emphasizing the role of simulation in achieving defect-free castings.