Finite Element Modeling of Sand Casting Defects in Modified Gray Cast Iron Brake Drums
We present a comprehensive study on the prediction and control of sand casting defects in high-performance modified gray cast iron (MGCI) brake drums. The brake drum is a critical safety component in heavy-duty trucks, and its performance is directly linked to the reliability of the casting process. With the increasing demand for higher payloads and more severe service conditions, traditional gray cast iron materials often fail due to thermal fatigue and wear. To address this, a modified gray cast iron composition was developed by adjusting the content of alloying elements such as Ti and B, while reducing Ni and Cu. However, material modification inevitably alters the thermophysical properties of the melt, which can lead to unexpected sand casting defects such as shrinkage porosity if the process parameters are not properly optimized.
The primary objective of this research was to develop a reliable finite element model to simulate the mold filling and solidification behavior of the MGCI brake drum, predict the formation of sand casting defects, and optimize the casting process to minimize these defects. We integrated material property calculations, numerical simulation, and experimental validation to achieve this goal.
Material Property Calculation and Model Establishment
We utilized JMatPro software to calculate the thermophysical and fluid dynamic properties of the modified gray cast iron over a temperature range of 20 to 1,500 °C. The Back Diffusion model was selected to account for the finite diffusion of solute elements in the solid phase during solidification. A comparison between the original HT250 and the MGCI revealed significant changes in key properties, particularly the solid fraction and density as functions of temperature. The solidus and liquidus temperatures of MGCI decreased to 1,072 °C and 1,190 °C, respectively, compared to 1,078 °C and 1,198 °C for HT250. The density of MGCI was slightly lower than that of HT250 across the entire temperature range, implying a reduced volumetric shrinkage during solidification. These property changes directly influence the susceptibility to sand casting defects like shrinkage porosity.
The finite element model of the brake drum casting was established using Procast software. The brake drum has an inner diameter of 320 mm, an outer diameter of 335 mm, and a height of 190 mm. The casting process was designed as a sand casting with a two-cavity mold. The geometry was discretized using a progressive tetrahedral mesh, with finer elements (1-3 mm) in the casting, gating system, and riser regions, and coarser elements (10 mm) in the outer sand mold. This resulted in a mesh with 198,506 surface elements, 4,872,325 volume elements, and 833,831 nodes. The heat transfer coefficients between the casting and the sand mold, and between the sand mold and the ambient air (20 °C), were set to 500 W/(m²·K) and 20 W/(m²·K), respectively. The initial pouring temperature was set to 1,350 °C with a pouring time of 30 seconds, reflecting the original production parameters.
Simulation Results and Analysis of Sand Casting Defects
The mold filling simulation showed that the molten metal followed a stable bottom-to-top filling sequence without any splashing or confluence of flow fronts. This indicates that the initial gating system design was adequate for avoiding gas entrapment and inclusions, which are common sand casting defects. The temperature field at 98% fill showed that the temperature at the ingate and riser was 1,341 °C and 1,272 °C, respectively, both above the liquidus temperature of 1,190 °C, suggesting that the melt still had good fluidity for feeding.
During the solidification phase, we observed a critical phenomenon leading to sand casting defects. The simulation of the solid fraction evolution revealed isolated liquid pools forming in the reinforcing ribs of the brake drum adjacent to the ingate at the final stages of solidification. This occurred because the gating system solidified before the thick reinforcing rib sections. The casting’s upper flange and lower reinforcing ribs solidified first, while the drum wall and the thick rib areas solidified later. The solidification time distribution map showed that the riser and the area near the ingate solidified around 670 seconds, which was significantly later than the drum body (480-590 seconds) but the gating system solidified much earlier (220-550 seconds). This resulted in a violation of the desired directional solidification principle, where the casting should solidify before the riser and gating system to allow for effective feeding. Consequently, the isolated liquid pools could not be fed by the already solidified gating system, leading to severe shrinkage porosity. The simulation predicted that the average porosity of the defective elements was 3.53%, concentrated in the reinforcing ribs near the ingate.
Experimental Validation of the Sand Casting Defect Model
We conducted a casting experiment using the same process parameters as the simulation to validate the finite element model. The melt was prepared in a medium-frequency induction furnace. The cast brake drum showed a good external appearance without surface defects, confirming the stable filling process predicted by the simulation. Crucially, upon sectioning the casting, we observed pronounced shrinkage porosity at the exact location predicted by the model—the reinforcing ribs near the ingate. Scanning electron microscopy of the defect area revealed a typical interdendritic morphology characteristic of shrinkage porosity, confirming the nature of the sand casting defects. This excellent agreement between the simulation and experiment validated the accuracy and reliability of our finite element model for predicting sand casting defects in the MGCI brake drum.
Process Optimization for Minimizing Sand Casting Defects
1. Gating System Optimization
Based on the simulation and experimental findings, the root cause of the sand casting defects was identified as the premature solidification of the gating system, which failed to provide adequate feeding to the casting during the final stages of solidification. To rectify this, we modified the gating system structure by increasing the width of the ingate near the sprue and thickening the base of the sprue well. The goal was to keep the gating system hot for a longer period to delay its solidification and enhance its feeding capability.
After the modification, a new simulation was performed. The results showed a significant reduction in the volume and severity of sand casting defects. The total volume of defective elements decreased, and the severe shrinkage porosity was relocated from the casting’s reinforcing ribs to the gating system itself (specifically, the ingate and sprue well platform). The overall average porosity of the casting body was reduced to 2.89%, indicating a substantial improvement in the feeding efficiency.
2. Process Parameter Optimization
We further optimized the casting process parameters, specifically the pouring temperature and pouring time, using a full factorial design of experiments. The objective function was to minimize the shrinkage porosity in the casting. The parameters investigated are summarized in the following table.
| Parameter | Level 1 | Level 2 | Level 3 | Level 4 |
|---|---|---|---|---|
| Pouring Temperature (°C) | 1,350 | 1,375 | 1,400 | 1,425 |
| Pouring Time (s) | 25 | 30 | 35 | – |
We can define the defect severity, or shrinkage porosity (P), as a function of the volume of defective elements (Vdef) relative to the total volume of the casting (Vtotal), but more practically, we evaluate the average porosity of the defective regions. The effect of process parameters on the volume of defective elements (Vdef) and average porosity is described by the following qualitative relationships derived from our simulations:
$$ P \propto f(T_{pour}, t_{pour}) $$
where Tpour is the pouring temperature and tpour is the pouring time.
The results from the full factorial simulations are summarized in the table below, which quantifies the impact of each parameter combination on the formation of sand casting defects.
| Pouring Temperature (°C) | Pouring Time (s) | Volume of Defective Elements (cm³) | Average Shrinkage Porosity (%) |
|---|---|---|---|
| 1,350 | 25 | > 25 | 2.19 |
| 30 | > 25 | ~ 3.0 | |
| 35 | > 25 | 4.17 | |
| 1,375 | 25 | ~ 20 | ~ 1.30 |
| 30 | ~ 20 | ~ 1.30 | |
| 35 | 24 | 2.06 | |
| 1,400 | 25 | ~ 20 | ~ 1.30 |
| 30 | ~ 20 | ~ 1.30 | |
| 35 | 24 | 2.06 | |
| 1,425 | 25 | ~ 18 | 0.86 |
| 30 | ~ 18 | 0.86 | |
| 35 | 25 | 3.20 |
At low pouring temperatures (1,350-1,375 °C), the melt viscosity was high, and the fluidity was low. This led to premature solidification in the mold, blocking feeding paths and resulting in a high volume of sand casting defects with high average porosity. Longer pouring times exacerbated this issue by allowing the melt to cool further during filling.
As the pouring temperature increased to 1,400 °C, the fluidity improved, and the feeding capability was enhanced. The volume of defective elements and porosity decreased significantly for shorter pouring times. However, a longer pouring time of 35 seconds still led to a noticeable increase in defects due to increased heat loss during the prolonged filling process.
The best results were obtained at the highest pouring temperature of 1,425 °C. At this temperature, with pouring times of 25 and 30 seconds, the volume of defective elements was minimal (around 18 cm³), and the average shrinkage porosity was drastically reduced to just 0.86%. This represents a 76% reduction in porosity compared to the original process (3.53%). At the longest pouring time of 35 seconds, even at this high temperature, the defects increased again due to excessive cooling. We also observed that a very fast pour (25 s) could lead to temperature gradients that are not ideal for directional solidification, although the porosity results were similar. Therefore, a pouring time of 30 seconds was deemed the best balance for ensuring uniform temperature distribution and stable filling.
Based on a comprehensive analysis of the simulation data, the optimal process parameters for minimizing sand casting defects in the MGCI brake drum were determined to be a pouring temperature of 1,425 °C and a pouring time of 30 seconds.
Experimental Verification of the Optimized Process
We conducted a production trial using the optimized gating system and the determined optimal process parameters (1,425 °C pouring temperature, 30 seconds pouring time). The cast brake drums showed excellent surface quality. Subsequent sectioning and ultrasonic inspection revealed no internal sand casting defects. The microstructure was uniform and dense. The final machined brake drums exhibited a high-quality finish with no visible porosity. This successful trial confirmed the effectiveness of our integrated approach of material property calculation, finite element simulation, and experimental validation for predicting and controlling sand casting defects.
Conclusion
We successfully developed a finite element model to predict sand casting defects in a high-performance modified gray cast iron brake drum. The model was validated through experimental trials. The key findings are as follows:
- The modification of gray cast iron composition decreased the solidus and liquidus temperatures, altering the solidification behavior and increasing the susceptibility to certain types of sand casting defects like isolated liquid pools.
- The primary cause of shrinkage porosity in the initial design was the premature solidification of the gating system relative to the casting, which violated the principle of directional solidification and led to the formation of isolated liquid pools in the reinforcing ribs.
- Optimizing the gating system structure to delay its solidification was an effective first step in reducing sand casting defects, successfully shifting the defect location away from the casting.
- The pouring temperature and pouring time have a strong coupled effect on the formation of sand casting defects. The optimal parameters for this MGCI brake drum were found to be a pouring temperature of 1,425 °C and a pouring time of 30 seconds, which reduced the shrinkage porosity to a minimal 0.86%.
- The integration of material property calculation, numerical simulation, and experimental testing provides a robust and reliable methodology for optimizing casting processes and eliminating sand casting defects, leading to higher quality and more reliable components.