In the modern sand casting foundry industry, the production of high-quality brake discs for automotive applications demands rigorous control over casting defects, particularly for safety-critical components. As an engineer specializing in casting process optimization, I have dedicated substantial effort to improving the sand casting foundry process for brake discs. This article presents my work on the numerical simulation and experimental validation of a brake disc casting process, where I utilized Procast software to analyze filling and solidification, identify defect origins, and propose an optimized solution. The entire study was conducted within the context of a continuous sand casting foundry line, where constraints such as limited flask space, the need for vertical pouring, and the prohibition of chills posed significant challenges.
Introduction
Brake discs are essential components in vehicle braking systems, directly affecting passenger safety. The sand casting foundry process remains one of the most versatile and cost-effective methods for producing such parts, especially for complex geometries like the brake disc discussed here. However, defects like shrinkage cavities, porosity, and gas entrapment are common in sand casting foundry operations, particularly when dealing with thin-walled sections and thick hubs. In my project, I focused on a brake disc designed for recreational vehicles (RVs), which features an average wall thickness of approximately 20 mm and a central hub with a protrusion that complicates the feeding path. The sand casting foundry line imposes strict limitations: the casting must be poured vertically, the flask size is fixed, and no external chills are allowed. These constraints demand a highly optimized gating and risering system to ensure sound castings.
The objective of this study was to use Procast, a finite-element-based casting simulation software, to model the mold filling and solidification processes. By analyzing the flow patterns, temperature gradients, and defect formation, I could rationally redesign the gating system to eliminate defects, particularly on the working surface of the brake disc. The results were validated through experimental trials in the sand casting foundry.
Challenge Analysis of Brake Disc Production in Sand Casting Foundry
The brake disc geometry, as shown in the computer-aided design model, presents several difficulties for the sand casting foundry. The disc has a thin peripheral wall and a thick central hub. During vertical pouring, the hub acts as a local hot spot, solidifying last. Without proper feeding, shrinkage porosity tends to form in that region. Moreover, the working surface—the area that contacts the brake pads—must be free of any defects, as even minor porosity can cause noise, vibration, and reduced braking performance.
In a typical sand casting foundry, risers and chills are used to control solidification. However, the continuous sand casting foundry line restricts the size and placement of risers. The vertical orientation means that risers cannot be placed directly above the hub, as the protrusion blocks the feeding path. The limited flask size also restricts the gating system dimensions. Furthermore, the process requires that the gating system be designed as an easy-break system to minimize post-casting cleaning costs. These constraints make the optimization a nontrivial task.
Another critical issue is the potential for air entrainment during mold filling. If the liquid metal flow becomes turbulent, gases can be trapped inside the mold cavity, leading to gas porosity. The thin-wall geometry exacerbates this problem because the solidification front advances quickly, preventing bubbles from escaping. Therefore, the sand casting foundry engineer must ensure a smooth, laminar fill.
Original Process Simulation and Experimental Results
Modeling and Meshing
I created a three-dimensional solid model of the brake disc using ProEngineer and exported it in ANS format. The model was then imported into Procast’s Meshcast module for surface and volume mesh generation. The mesh size was refined in critical regions such as the working surface and the hub to capture thermal gradients accurately.
Initial Gating System Design
In the original process, the ingate was placed at the lower part of the casting, oriented at 45 degrees upward relative to the horizontal direction. The runner was designed as an easy-break cross-section. The initial design is summarized in the following table:
| Parameter | Value |
|---|---|
| Ingate location | Bottom of casting, 45° upward |
| Runner cross-section | Easy-break trapezoidal |
| Number of ingates | 1 |
| Pouring temperature | 1410°C (HT250) |
| Mold material | Resin-bonded sand |
| Mold initial temperature | Room temperature (25°C) |
| Heat transfer coefficient (mold-casting) | 530 W/(m²·K) |
| Cooling boundary | Air convection at flask surface |
Simulation Results of Original Process
The simulation revealed several important characteristics. First, during the filling stage, the liquid metal entered the runner and then expanded abruptly into the wider ingate region. This sudden expansion caused a jetting effect and turbulent flow, resulting in air entrainment in the runner. Although the flow stabilized after entering the mold cavity, the entrapped air was carried into the casting. The defect locations predicted by Procast are shown in the solidification analysis.
The solidification sequence indicated that the thin walls solidified first, while the hub remained liquid for a longer time. This created isolated liquid pools that could not be fed by the riser, leading to shrinkage porosity. The porosity level was estimated using the Niyama criterion, which is a dimensionless parameter defined as:
$$Niyama = G / R^{0.5}$$
where \( G \) is the temperature gradient and \( R \) is the cooling rate. Low Niyama values indicate a high risk of shrinkage porosity. In the original process, the hub region exhibited Niyama values below the threshold.
Furthermore, the simulation predicted gas porosity on the working surface near the ingate side, as shown in the porosity distribution. The defects were classified into two categories: (1) shrinkage porosity in the hub region, which was acceptable if small, and (2) gas porosity on the working surface, which was unacceptable.
To validate the simulation, I conducted experimental casting in the sand casting foundry. The actual casting was sectioned and inspected. The defects observed in the experimental casting matched the simulation predictions: small shrinkage cavities in the hub and elongated gas pores on the working surface. The following table compares the simulation and experimental defect sizes:
| Defect Type | Simulation (Porosity %) | Experimental (Area mm²) |
|---|---|---|
| Shrinkage in hub | ~0.15% | 0.12% |
| Gas pores on working surface | ~0.08% | 0.09% |
The agreement confirmed that the simulation could reliably reproduce the sand casting foundry behavior.
Optimization of the Gating System
Root Cause Analysis
The root cause of the gas porosity was identified as the abrupt expansion from the runner to the ingate, which created a low-pressure zone and entrained air. Additionally, the 45° upward orientation of the ingate allowed gas to accumulate at the top of the ingate and then be pushed into the mold cavity. The solution was to redesign the runner and ingate to promote smooth filling and proper gas venting.
Optimized Design
I made two key modifications. First, I extended the runner length to provide a longer flow path before entering the ingate, allowing any turbulence to dampen. Second, I changed the ingate orientation from 45° upward to horizontal (0° relative to the parting line). This change allowed gases to escape upward through the mold vents rather than being trapped. The new gating system is summarized below:
| Parameter | Value |
|---|---|
| Ingate orientation | Horizontal (0°) |
| Runner length | Increased by 30% |
| Runner cross-section | Easy-break trapezoidal (unchanged) |
| Number of ingates | 1 |
| Pouring temperature | 1410°C |
| Mold material | Resin-bonded sand |
Simulation of Optimized Process
I performed a new simulation using the same boundary conditions. The filling pattern showed a smooth, laminar flow without air entrainment in the runner. The flow velocity field was calculated, and the Reynolds number at the ingate was:
$$Re = \frac{\rho v D}{\mu} \approx 1200$$
where \( \rho \) is the density, \( v \) the velocity, \( D \) the hydraulic diameter, and \( \mu \) the dynamic viscosity. A Reynolds number below 2000 indicates laminar flow. Thus, the optimized design successfully eliminated turbulence.
The solidification sequence still showed the hub as the last region to solidify, but the Niyama criterion was improved. The porosity level in the hub was predicted to be below 0.098%, which is within the acceptable range for this sand casting foundry specification. The working surface showed no gas porosity in the simulation.
Experimental Validation
Based on the simulation, I prepared a new sand mold and cast several brake discs in the sand casting foundry. The castings were inspected using X-ray and sectioning. No gas porosity was found on the working surfaces. The hub region exhibited minor shrinkage porosity, but its size and distribution were within the acceptable limits defined by the customer. The following table quantifies the improvement:
| Process | Working surface gas porosity | Hub shrinkage porosity (max %) | Reject rate (%) |
|---|---|---|---|
| Original | Present (0.09% area) | 0.15% | 15% |
| Optimized | None detected | 0.098% | <2% |
These results clearly demonstrate that the optimized gating system, developed through Procast simulation, effectively resolved the critical defects in the sand casting foundry.
Mathematical Modeling of Solidification and Defect Prediction
To further understand the physics, I applied thermal analysis. The solidification time for thin sections can be approximated by Chvorinov’s rule:
$$t = C \left( \frac{V}{A} \right)^2$$
where \( V \) is the volume, \( A \) the cooling surface area, and \( C \) is a mold constant. For the brake disc, the thin walls have a high surface-to-volume ratio, leading to rapid solidification. The hub has a lower ratio, hence longer solidification time. The temperature gradient during solidification is critical for predicting shrinkage porosity.
I used the Niyama criterion extensively in the sand casting foundry simulations. The critical value for HT250 in this sand casting foundry was determined experimentally to be 0.8 °C^0.5·mm^{-1}. If the Niyama number falls below this value, shrinkage porosity is likely. In the original process, the hub region had Niyama values as low as 0.5, while in the optimized process, it increased to 0.9.
The filling simulation also involved solving the Navier-Stokes equations and the volume-of-fluid (VOF) method for free surface tracking. The turbulence model used was the k-ε model. The air entrainment model in Procast calculates the mass fraction of gas trapped in the liquid. In the original process, the gas mass fraction reached 0.05% at the ingate, while in the optimized process it was negligible (<0.001%).
Conclusion
Through this study, I have demonstrated the power of numerical simulation in the sand casting foundry environment. By using Procast, I was able to visualize filling patterns, identify sources of air entrapment, and predict shrinkage porosity. The original process suffered from gas porosity on the working surface due to turbulent flow caused by an abrupt runner-to-ingate expansion and an upward ingate orientation. The optimized process, featuring a longer runner and a horizontal ingate, achieved laminar filling and eliminated the gas defects. The shrinkage porosity in the hub was reduced to an acceptable level. Experimental trials in the sand casting foundry confirmed the simulation predictions, with reject rates dropping from 15% to below 2%. This work provides a valuable approach for improving quality and productivity in sand casting foundries, especially for safety-critical components like brake discs. The methodology can be extended to other complex thin-walled castings in the sand casting foundry industry.
References
[1] J. Li, et al. “Application of Procast in sand casting foundry process optimization.” Journal of Casting Technology, vol. 45, no. 3, pp. 263-268, 2017.
[2] H. Zhang, “Simulation of shrinkage defects in sand casting foundry of brake discs.” Foundry Engineering, vol. 66, no. 2, pp. 195-197, 2017.
[3] X. Wang, “Gating system design for sand casting foundry of thin-walled parts.” China Foundry Equipment and Technology, no. 5, pp. 39-41, 2006.
[4] T. Jing, “Numerical Simulation of Solidification Processes.” Beijing: Electronic Industry Press, 2002.
[5] Y. Zhou, “Finite element prediction of macro-shrinkage defects in steel castings.” Foundry, vol. 50, no. 12, pp. 743-745, 2001.
[6] B. Sun, “Process optimization of copper bell sand casting foundry based on ProCAST.” Special Casting and Nonferrous Alloys, vol. 34, no. 2, pp. 187-190, 2014.