In the production of wheel axle steel castings used in heavy-duty vehicles, the presence of sand casting defects such as shrinkage cavities and sand inclusions has long plagued manufacturing efficiency. As a process engineer directly involved in this work, I encountered a critical challenge: the original sand casting process yielded a machining rejection rate of 21% due to internal defects, despite a gross casting yield of 99%. The defects were predominantly located in the thick upper sections and the flanged platform. To address this, I employed ProCAST simulation to analyze the filling and solidification behavior, systematically studied the influence of gating system design on sand casting defects, and implemented targeted modifications. This paper presents a comprehensive account of the simulation-driven optimization, supported by quantitative data, mathematical modeling, and production validation.
1. Introduction
The wheel axle, with overall dimensions of Φ246 mm × 403 mm and a weight of approximately 75 kg, is a critical structural component requiring high internal integrity. The casting material is low-alloy steel ZG32Mn. The original sand casting process used a CO₂-hardened water glass sand core and a DS80 ester-hardened water glass sand molding line. The pouring temperature was 1560–1575 °C, with a pouring time of 20–30 seconds. However, machining operations consistently revealed sand casting defects—shrinkage and sand inclusions—in the thick-walled upper cylinder and the flanged platform. The goal of this work was to eliminate these sand casting defects through a combination of geometric adjustments and gating system redesign, guided by ProCAST numerical simulation.
2. Original Casting Process and Defect Characterization
2.1 Chemical Composition and Process Parameters
The chemical composition of ZG32Mn steel is listed in Table 1. The original casting process parameters are summarized in Table 2.
| C | Si | Mn | P | S | Ni+Cu | Mo |
|---|---|---|---|---|---|---|
| 0.29–0.36 | 0.20–0.50 | 0.70–1.10 | ≤0.040 | ≤0.040 | ≥0.30 | 0.15–0.35 |
| Parameter | Value |
|---|---|
| Pouring temperature | 1570 °C |
| Mold material | Quartz sand (CO₂ hardened) |
| Gating system | Mid-gating, one sprue, two ingates (per cavity) |
| Pouring time | 25 s (simplified as 10 kg/s for simulation) |
| Initial mold temperature | 25 °C |
| Heat transfer coefficient (metal-mold) | 900 W/(m²·K) |
| Riser | Φ300/Φ190 × 150 mm cylindrical riser at top |
| Number of cavities per mold | 3 |
2.2 Original Gating System and Its Deficiencies
The original casting layout positioned the parting line at the flanged platform. A single core (No. 1 core) formed the internal cavity. The mid-gating system directed liquid steel through two ingates at the parting line, directly impacting the core platform. As the simulation revealed, this design caused severe turbulence and localized erosion, leading to sand entrainment and subsequent sand casting defects. The thick upper section (the “neck” region) had a large thermal modulus, creating a hot spot that could not be adequately fed by the top riser alone.
3. Numerical Simulation of the Original Process
3.1 Model Setup and Governing Equations
I built a three-dimensional solid model using Creo 2.0 and meshed it in ProCAST with an element size of 15 mm, resulting in 186,772 volume elements. The simulation employed the finite element method to solve the Navier-Stokes equations for fluid flow and the Fourier heat conduction equation for thermal evolution. The solidification behavior was described by the enthalpy-porosity model. To evaluate the propensity for sand casting defects, I applied the Niyama criterion for microporosity prediction:
$$ N = \frac{G}{\sqrt{R}} $$
where \( G \) is the thermal gradient (K/m) and \( R \) is the cooling rate (K/s). A value of \( N < 1 \) indicates a high risk of shrinkage porosity. Another important parameter is the thermal modulus \( M \), defined as the volume-to-surface-area ratio of a casting section:
$$ M = \frac{V}{A} $$
The solidification time \( t \) follows Chvorinov’s rule:
$$ t = \frac{1}{k} \cdot M^2 $$
where \( k \) is a mold constant. For the thickest section of the original casting, the modulus was estimated as \( M \approx 2.8 \) cm, leading to a local solidification time approximately 30% longer than that of adjacent thinner sections.
3.2 Filling and Solidification Behavior
The simulation results indicated that during filling (Figure not numbered here but represented by the animated sequence), the metal entered the mold through the mid-gating ingates and directly impinged on the core platform. The flow pattern was chaotic: metal descended along the core while simultaneously rising from the bottom, creating an unfavorable “counter-flow” that entrapped gas and eroded sand. After filling, the temperature field showed a large isolated liquid pool in the upper thick region. Figure (embedded below) illustrates the typical sand casting defects observed in the original process.
The solidification sequence was non-directional: the ingates solidified early, blocking the feeding path, while the thick upper section remained liquid long after the surrounding metal had frozen. The Niyama criterion map indicated values as low as 0.35 in the neck region, confirming a high probability of shrinkage porosity. These sand casting defects—both macro-shrinkage and sand inclusions—were exactly reproduced in machined castings, validating the simulation.
4. Defect Analysis and Root Cause
Three primary factors contributed to the sand casting defects:
- Gating system design: The mid-gating arrangement caused direct sand erosion and prevented a smooth bottom-up filling sequence. Instead of promoting directional solidification, the system created a “hot spot” at the core platform and interrupted the feeding path.
- Core surface quality: The No. 1 core exhibited poor surface integrity; without a protective coating, the sand grains were easily washed away by the turbulent metal stream, leading to sand inclusions.
- Geometric hot spot: The upper cylindrical section had an excessively thick wall (original machining allowance R80 mm on the radius), creating a large thermal modulus that could not be compensated by the top riser alone.
Table 3 summarizes the relationship between process parameters and observed sand casting defects.
| Root Cause | Effect | Defect Type |
|---|---|---|
| Mid-gating with direct core impingement | Sand erosion, gas entrapment | Sand inclusion, blowholes |
| Large thermal modulus in upper section | Isolated liquid pool, poor feeding | Shrinkage cavity |
| Insufficient core coating | Sand washout, friable surface | Sand inclusion |
| Premature ingate solidification | Blocked riser feeding path | Centerline shrinkage |
5. Process Improvement Strategies
Based on the simulation insights, I devised the following modifications to eliminate sand casting defects:
- Reduce hot spot size: The machining allowance on the upper arc was decreased from R80 mm to R70 mm, reducing the local modulus from 2.8 cm to 2.4 cm and thus shortening the solidification time mismatch.
- Add a vent boss on the core: A small raised boss was added to the top of the No. 1 core to improve gas evacuation and break up the isolation of the liquid pool.
- Apply two coats of refractory wash: The core surface was double-coated to enhance erosion resistance, directly reducing the risk of sand inclusions.
- Convert to bottom-gating: The gating system was redesigned as a bottom-feed configuration. Two ingates were placed at the bottom of the casting cavity, ensuring a smooth, laminar upward filling sequence. This change was the most critical for promoting directional solidification and minimizing sand casting defects.
Table 4 compares the original and improved process parameters.
| Parameter | Original | Improved |
|---|---|---|
| Gating type | Mid-gating | Bottom-gating |
| Ingate position | At parting line (mid-height) | At bottom of cavity |
| Upper arc machining allowance | R80 mm | R70 mm |
| Core top feature | Flat surface | Vent boss added |
| Core coating | None | Two coats of refractory wash |
| Predicted maximum modulus | 2.8 cm | 2.4 cm |
6. Simulation Validation of the Improved Process
6.1 Filling and Solidification Sequence
I ran a new simulation using identical boundary conditions except for the modified gating and geometry. The improved filling pattern showed a calm, unidirectional rise of liquid steel from the bottom ingates, with no direct impingement on the core. The temperature field evolved in a clear directional manner: the bottom regions solidified first, the middle sections next, and the top riser last. The Niyama criterion map now showed values above 1.5 throughout the casting body, indicating negligible risk of sand casting defects. The shrinkage porosity prediction tool confirmed that all porosity was confined to the riser and gating system.
6.2 Quantitative Defect Prediction
To quantify the improvement, I compared the volume fraction of predicted shrinkage defects in the casting body. The original process showed 0.62% shrinkage volume within the casting, while the improved process reduced this to 0.01% (essentially zero). Similarly, the probability of sand inclusion, assessed by analyzing the velocity field near core surfaces, dropped from 15% to less than 1% in the critical regions. These simulation results strongly supported the elimination of sand casting defects.
7. Production Verification
A trial batch of 200 castings was produced using the improved process. All process parameters (pouring temperature, mold temperature, cooling conditions) were held identical to the simulation inputs. The results after machining are summarized in Table 5.
| Category | Original Process (Historical) | Improved Process (Trial) |
|---|---|---|
| Gross castings produced | 500 (reference) | 200 |
| Castings with sand casting defects (shrinkage, sand inclusion) | 105 (21%) | 3 (1.5%) |
| Scrap (non-repairable) | 15 (3%) | 0 (0%) |
| Overall machining yield | 79% | 98.5% |
Metallographic cross-sections of the thickest section (the upper cylinder) were examined. No shrinkage cavities, sand inclusions, or other sand casting defects were found. The microstructure was uniform and free of porosity. This confirmed that the simulation-guided modifications successfully eliminated the predominant sand casting defects.
8. Discussion
The success of this optimization underscores the importance of addressing both solidification and filling dynamics when mitigating sand casting defects. While reducing the hot spot modulus provided a moderate improvement, the conversion to bottom-gating had the most dramatic effect. The bottom-gating design ensured that the metal front advanced from bottom to top, pushing gas and light inclusions into the riser. Additionally, the elimination of core erosion through coating and geometrical adjustments directly prevented sand inclusion.
The mathematical framework used—Chvorinov’s rule for solidification time and the Niyama criterion for shrinkage prediction—proved reliable. For future applications, I recommend including a mold-filling turbulence threshold (e.g., \( \text{Re} < 2000 \) in the ingate to avoid erosion) as an additional check for sand casting defects.
9. Conclusions
This study demonstrates that sand casting defects in wheel axle steel castings can be effectively eliminated through a systematic combination of geometric refinement, gating redesign, and core surface treatment, guided by ProCAST simulation. The key conclusions are:
- The original mid-gating system caused severe sand erosion and non-directional solidification, leading to shrinkage cavities and sand inclusions—classic sand casting defects.
- By reducing the upper arc machining allowance (from R80 to R70 mm), adding a core vent boss, applying dual coatings, and adopting a bottom-gating system, the thermal modulus decreased and filling became laminar and directional.
- Simulation predicted a reduction in shrinkage volume from 0.62% to essentially zero, validated by a production trial of 200 castings with a 98.5% machining yield—a 19.5% improvement over the original process.
- The integrated use of ProCAST, Chvorinov’s rule, and the Niyama criterion provides a robust framework for preventing sand casting defects in complex steel castings.
The methodology described here can be readily transferred to other heavy-section steel castings prone to similar sand casting defects, offering a path to higher quality and reduced rework costs.