I began this study because the divided axle used in a tractor load-bearing axle is a safety-critical component that must survive high impact loads, rough terrain, saline-alkali soils, deserts, and heavily loaded service conditions. The axle is produced by sand mould casting, and the original production route suffered from a very high rate of internal shrinkage cavity and shrinkage porosity. In practice, the defect rate was reported to be as high as 70%, causing oil seepage and oil leakage around the inner half-shaft oil seal seat. The economic loss was significant every year. My objective was therefore to combine practical sand mould casting experience with finite element simulation, using Procast, to design and compare several gating systems, predict filling and solidification behavior, optimize risers and cold irons, and validate the final process through sectioning and air-tightness tests.
I treated the problem as a complete sand mould casting process design task rather than as an isolated defect repair. The work included structural analysis of the divided axle, material selection, raw material and auxiliary material control, determination of pouring position, parting surface, pouring temperature, shrinkage rate, sand core design, draft angle, gating system calculation, numerical simulation of flow and temperature fields, shrinkage defect prediction, riser and cold iron design, and experimental validation. The final goal was to obtain a robust sand mould casting process that could reduce the failure rate from about 70% to approximately 3% and provide a reliable basis for production.
Background and significance of the sand mould casting problem
The divided axle is used in a tractor axle assembly. It is a complex steel casting with a relatively thin wall and several local thick sections. The original sand mould casting process used a bottom-pouring gating system, a three-box moulding method, and a pouring temperature of about 1550 °C. That process was difficult to control. Because the liquid steel had poor fluidity and high shrinkage, the isolated liquid zones that formed during solidification could not be fed properly. As a result, shrinkage cavity and shrinkage porosity appeared in the inner half-shaft oil seal seat, especially in the concave area below the oil seal seat. Oil leakage then occurred during service. Visual inspection and conventional inspection methods could not reliably detect these internal defects, so the problem often escaped detection until the axle was assembled and used.
The significance of this study is practical. If I could predict the defect location before casting, and if I could adjust the sand mould casting process to promote sequential solidification, then I could reduce trial-and-error, shorten the development cycle, lower production cost, and improve the sealing reliability of the axle. I also wanted to demonstrate how numerical simulation can support sand mould casting process optimization in a heavy vehicle component.
Review of casting simulation and sand mould casting development
Casting simulation has developed rapidly from early heat-flow analyzers to modern three-dimensional finite element and finite difference software. In sand mould casting, simulation is now used to study mould filling, solidification, cooling, stress distribution, and defect formation. The main simulation tasks include flow field simulation, temperature field simulation, coupled flow and heat transfer simulation, microstructure simulation, and stress field simulation. Each of these has value for sand mould casting. Flow field simulation helps avoid sand erosion, incomplete filling, gas entrapment, and oxidation inclusions. Temperature field simulation helps predict shrinkage cavity and shrinkage porosity. Coupled simulation provides initial temperature conditions for solidification analysis. Microstructure simulation helps predict grain structure and mechanical properties. Stress simulation helps predict hot cracks, cold cracks, and deformation.
In my work, I focused on the flow field and temperature field because the dominant defect was shrinkage porosity and shrinkage cavity. I used Procast because it is widely applied in sand mould casting and can handle complex geometry, mould materials, chill blocks, risers, and filling conditions. I also used Creo for three-dimensional modeling and Visual-mesh for mesh generation. The numerical model was based on mass conservation, energy conservation, and momentum conservation. For the temperature field, the governing equation I used can be written as:
$$\rho C \frac{\partial T}{\partial t} = \lambda \left( \frac{\partial^2 T}{\partial x^2} + \frac{\partial^2 T}{\partial y^2} + \frac{\partial^2 T}{\partial z^2} \right) + w$$
where \(T\) is temperature, \(t\) is time, \(\rho\) is density, \(C\) is specific heat capacity, \(\lambda\) is thermal conductivity, and \(w\) is the latent heat released during solidification. This equation allowed me to simulate heat transfer and solidification in the sand mould casting process.
Main quality problems in the original sand mould casting process
The original sand mould casting route produced three main categories of defects: shrinkage cavity, lack of material, and air-tightness failure. Shrinkage cavity occurred because the feeding pressure was insufficient at the critical wall thickness. The cavity was irregular, rough inside, and often located in the inner half-shaft oil seal seat. Shrinkage porosity produced a less dense structure, which allowed oil to seep through the casting under pressure. Air-tightness failure was the final consequence. These defects were difficult to detect because they were internal. The leak location was consistently found below the inner half-shaft oil seal seat, where the wall thickness changed and where an isolated liquid zone formed during solidification.
| Defect type | Typical location | Cause in sand mould casting | Service effect |
|---|---|---|---|
| Shrinkage cavity | Inner half-shaft oil seal seat, concave region | Insufficient feeding, isolated liquid zone, late solidification | Oil leakage, reduced load capacity |
| Shrinkage porosity | Thick wall transitions, hot spots | Slow solidification without riser or chill | Seepage, pressure loss, poor fatigue life |
| Lack of material | Thin walls, far ends of mould cavity | Poor fluidity, low pouring temperature, incomplete filling | Scrap, weak section |
| Air-tightness failure | Seal seat and internal cavity | Internal cavity or porosity connected to surface | Oil leakage, customer complaint |
Structural analysis of the divided axle
I modeled the divided axle from its two-dimensional drawing using Creo. The overall dimensions were approximately 452 mm × 340 mm × 396 mm. The average wall thickness was about 25 mm, with a maximum of 29 mm and a minimum of 15 mm. The structure was generally symmetric, but it contained many ribs, bosses, flanges, and internal cavities. The wall thickness variation was therefore significant. The most important functional surfaces were the flange face and the inner cylindrical cavity. These regions required high dimensional accuracy and structural density. Because the oil seal seat was a critical sealing surface, any shrinkage porosity in that area could cause leakage. I therefore placed the important cylindrical cavity downward during pouring so that the lower part of the mould would be filled smoothly and the critical surfaces would be less likely to contain gas or inclusions.
The material selected for the axle was ZG40Mn, a low-alloy cast steel. It has good strength and toughness after normalizing and tempering. Its chemical composition and mechanical properties are summarized in the tables below. I controlled the composition because carbon, manganese, silicon, sulfur, and phosphorus all affect fluidity, shrinkage, and crack sensitivity in sand mould casting.
| Element | C | Mn | Si | S | P |
|---|---|---|---|---|---|
| ZG40Mn range, wt.% | 0.35–0.45 | 1.0–1.5 | 0.35–0.45 | ≤0.03 | ≤0.03 |
| Material | Heat treatment | Yield strength \(\sigma_s\) | Tensile strength \(\sigma_b\) | Elongation \(\delta_5\) | Reduction of area \(\psi\) | Hardness HB |
|---|---|---|---|---|---|---|
| ZG40Mn | Normalizing + tempering | ≥295 MPa | ≥640 MPa | ≥12% | ≥30% | 163–200 |
Raw materials and auxiliary materials for sand mould casting
For sand mould casting of ZG40Mn, the main raw materials were scrap steel, ferrosilicon, high-carbon ferromanganese, and silicon-aluminum-iron deoxidizer. The main auxiliary material was silica sand. I specified strict inspection because raw material quality directly influences shrinkage, porosity, and surface quality. Scrap steel was checked for shape, size, thickness, oil, mud, closed vessels, explosive materials, and radioactive materials. Ferrosilicon was checked for silicon and aluminum content and for slag or sand inclusions. High-carbon ferromanganese was checked for manganese content, sulfur and phosphorus content, and surface cleanliness. The silicon-aluminum-iron alloy was checked for aluminum content and non-metallic inclusions. Silica sand was checked for grain shape, grain size distribution, clay content, and foreign matter.
| Material | Main function | Key control points |
|---|---|---|
| Scrap steel | Base charge for steel melting | Clean surface, correct size, no closed vessels, no oil or mud |
| Ferrosilicon | Adjust silicon content and deoxidize | Si 72–80%, Al ≤1%, S and P ≤0.02% |
| High-carbon ferromanganese | Adjust manganese content | Mn 62–75%, S ≤0.03%, clean surface |
| Silicon-aluminum-iron | Deoxidizer and inclusion control | Al 35–55%, no bubbles or non-metallic inclusions |
| Silica sand | Moulding and core sand | Round or angular grains, low clay, no foreign matter |
Pouring position and parting surface
I compared two pouring positions. In the horizontal position, the axle lay horizontally, and the ingate was placed at the middle parting plane. This arrangement reduced sand erosion, allowed smoother metal flow, and protected the important flange and inner cavity. It also made moulding, core setting, and pattern drawing easier. In the vertical position, the ingate was placed at the bottom flange. That arrangement caused more sand erosion and gas entrapment, and it was less effective for the critical surfaces. I therefore selected the horizontal pouring position. The parting surface was set at the same level to avoid core movement, sand drop, and running fire. The upper and lower moulds were arranged so that the critical cylindrical cavity was positioned downward, ensuring a cleaner and denser casting surface.
Pouring temperature determination
The pouring temperature for ZG40Mn in sand mould casting must balance fluidity and defect risk. A higher temperature improves fluidity, but too high a temperature increases oxidation, sand burn-on, grain coarsening, hot tearing, and gas porosity. A lower temperature reduces shrinkage risk but may cause incomplete filling. I simulated three temperatures: 1540 °C, 1560 °C, and 1580 °C. The results showed that 1580 °C gave faster filling in the early stage and slower solidification in the later stage, which helped avoid hot cracks and reduced stress concentration. I therefore selected 1580 °C as the pouring temperature. The pouring time was set to 10 s based on the casting weight and gating calculation.
| Pouring temperature | Filling behavior | Solidification behavior | Decision |
|---|---|---|---|
| 1540 °C | Lower fluidity, higher risk of incomplete filling | Faster cooling, possible cold shut | Not selected |
| 1560 °C | Moderate fluidity | Moderate solidification | Not selected |
| 1580 °C | Good fluidity, rapid early filling | Slower late solidification, lower stress concentration | Selected |
Process parameters: tolerance, shrinkage, sand core, and draft
The dimensional tolerance class for sand mould casting was selected as CT12, which is appropriate for water-glass sand moulding. The machining allowance was chosen according to standard practice. The linear shrinkage of cast steel is a critical parameter because it directly affects the final dimensions of the axle. I used the following expression for casting shrinkage:
$$K = \frac{L_M – L_J}{L_J} \times 100\%$$
where \(L_M\) is the mould working surface dimension and \(L_J\) is the casting dimension. For ZG40Mn, the free linear shrinkage range is about 1.6–2.0%, and the restrained shrinkage range is about 1.3–1.7%. Considering the complex structure and restrained contraction of the axle, I selected a casting shrinkage rate of 2%.
| Alloy type | Restrained shrinkage | Free shrinkage | Selected value |
|---|---|---|---|
| Carbon steel and low-alloy steel | 1.3–1.7% | 1.6–2.0% | 2.0% |
The axle contained one interconnected internal cavity, so I designed a single complex sand core with both horizontal and vertical core prints. The core was made by a cold-box process using resin-coated sand. The horizontal core print length was 50 mm on each side, with a taper of 3°. The vertical upper core print length was 50 mm, the lower core print length was 30 mm, the lower taper was 1°, and the upper taper was 4°. The core print structure provided accurate positioning and prevented movement or rotation during pouring. The core also had good collapsibility so that it could be removed easily after solidification.
The draft angle was determined from the pattern height. For the axle ribs and walls, I selected a draft angle of 3°. This value ensured that the pattern could be withdrawn from the sand mould without damaging the mould cavity. The draft angle was kept within the wall thickness tolerance so that it did not affect the final casting dimensions.
| Pattern height range | Draft angle for manual moulding | Selected draft angle |
|---|---|---|
| <20 mm | 3°–5° | 3° |
| 20–50 mm | 2°–3° | 3° |
| >50–100 mm | 1°30′–2° | 3° |
| >100–200 mm | 0°45′–1°30′ | 3° |
Original sand mould casting process deficiency
The original process used a bottom-pouring gating system and a three-box moulding method. It produced a large temperature difference inside the mould cavity. The inner half-shaft oil seal seat formed an isolated liquid zone during solidification. Because that zone could not be fed by liquid metal, shrinkage cavity and shrinkage porosity formed. The original pouring temperature was 1550 °C, the pouring time was 25 s, and the shakeout temperature was below 200 °C. The long pouring time and bottom-pouring design increased the risk of cold shut and insufficient feeding. I therefore decided to redesign the gating system using a middle-pouring approach and to compare several gating ratios.
Gating system design theory
I designed the gating system based on fluid mechanics and sand mould casting practice. The gating system must fill the mould smoothly, avoid sand erosion, trap slag, and promote sequential solidification. I considered top pouring, middle pouring, bottom pouring, and step pouring. The advantages and disadvantages are summarized below. I selected middle pouring because it combines some advantages of top and bottom pouring, reduces splashing, and is suitable for the parting surface of this axle.
| Ingate position | Advantages | Disadvantages |
|---|---|---|
| Top pouring | Strong filling ability, good feeding | Splashing, oxidation, gas entrapment, slag entrapment |
| Middle pouring | Balanced filling, suitable for complex parting surface | Requires careful gating design |
| Bottom pouring | Smooth filling, less oxidation, gas escapes upward | Poor feeding at top, lower filling ability |
| Step pouring | Strong filling, smooth flow, good feeding | Complex structure, high metal consumption |
The minimum residual head was calculated to ensure that the highest point of the axle could be filled. The formula is:
$$h_M = L \tan \alpha$$
where \(L\) is the horizontal distance from the highest point of the casting to the sprue centerline, and \(\alpha\) is the pressure angle. I used \(L = 320\) mm and \(\alpha = 10^\circ\). Therefore:
$$h_M = 320 \times \tan 10^\circ \approx 56\ \text{mm}$$
Adding the height from the ingate to the highest point, the minimum sprue height was about 246 mm. This ensured that the metal head was sufficient to fill the mould completely.
Gating system types
I compared three gating system types: closed, open, and semi-closed. The gating ratio is defined by the cross-sectional areas of the ingate, runner, and sprue. The advantages and disadvantages are listed in the table below.
| Gating type | Advantages | Disadvantages |
|---|---|---|
| Closed gating system | Easy to fill, good slag trapping, low metal consumption, easy cleaning | High flow speed, strong erosion, possible splashing |
| Open gating system | Smooth filling, low erosion on mould cavity | Poor slag trapping, high metal consumption, difficult cleaning |
| Semi-closed gating system | Certain slag trapping ability, smoother than closed, lower erosion | Slightly later filling of gating system than closed |
I used the pouring speed method to determine the choke area. The formula was:
$$A = \frac{G_L}{t k S’}$$
where \(A\) is the choke cross-sectional area, \(G_L\) is the weight of liquid metal passing through the choke, \(t\) is pouring time, \(k\) is pouring specific speed, and \(S’\) is the fluidity coefficient. I used \(G_L = 108\) kg, \(t = 10\) s, \(k = 0.6\ \text{kg cm}^{-2}\text{s}^{-1}\), and \(S’ = 0.8\). Thus:
$$A = \frac{108}{10 \times 0.6 \times 0.8} = 22.5\ \text{cm}^2$$
The cylindrical ingate diameter was calculated as:
$$d = \sqrt{\frac{4A}{\pi}} = \sqrt{\frac{4 \times 22.5}{\pi}} \approx 5.35\ \text{cm} = 53.5\ \text{mm}$$
I selected an ingate diameter of 54 mm. The runner and sprue dimensions were then determined from the gating ratio for each scheme.
| Scheme | Gating type | Ingate area | Runner area | Sprue area | Gating ratio \(\sum A_{\text{ingate}} : \sum A_{\text{runner}} : \sum A_{\text{sprue}}\) |
|---|---|---|---|---|---|
| Scheme 1 | Closed | 22.5 cm² | 29.25 cm² | 33.75 cm² | 1 : 1.3 : 1.5 |
| Scheme 2 | Open | 22.5 cm² | 21 cm² | 18.8 cm² | 1.2 : 1.1 : 1 |
| Scheme 3 | Semi-closed | 22.5 cm² | 18 cm² | 27 cm² | 1 : 0.8 : 1.2 |
I also placed a silicon carbide foam ceramic filter at the bottom of the sprue. The filter area was about 30 cm². This filter helped remove large inclusions, adsorb small inclusions, and stabilize the metal flow. In sand mould casting, such a filter is useful for reducing turbulence and inclusion-related defects.
Numerical simulation setup for sand mould casting
I created three-dimensional models of the three gating systems in Creo and saved them as .stp files. I imported them into Procast’s Visual-mesh environment. I checked and repaired surfaces, performed Boolean operations to generate the sand core, chills, runners, and mould cavity, and then generated surface and volume meshes. The axle mesh size was 6 mm, and the runner mesh size was 15 mm. I used a virtual sand box to represent the mould. The mesh quality was checked before simulation.
The thermal properties of ZG40Mn were important for accurate simulation. I used the Lever algorithm in Procast for thermal conductivity because it matched the theoretical curve better than the Scheil algorithm. The thermal conductivity values at different temperatures are shown below. The sand mould material was represented by resin-bonded sand because it has properties similar to the water-glass sand used in production. The interfacial heat transfer coefficients were set according to Procast experience values. The casting-mould interface coefficient was 500 W/(m²·K). The initial conditions were: pouring temperature 1580 °C, sand and core temperature 30 °C, gravity 9.8 m/s², and pouring time 10 s.
| Temperature (°C) | 0 | 100 | 200 | 300 | 400 | 500 | 600 | 700 | 800 | 900 | 1000 | 1100 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Thermal conductivity W/(m·K) | 51.9 | 50.7 | 48.1 | 45.6 | 41.9 | 38.1 | 33.9 | 23.0 | 24.3 | 25.9 | 27.2 | 28.8 |
| Interface | Heat transfer coefficient W/(m²·K) |
|---|---|
| Casting–sand mould | 350–500 |
| Casting–air | 350–500 |
| Casting–chromite sand | 500–700 |
| Casting–insulating sleeve | 50–150 |
| Casting–cold iron | 1000–3000 |
Filling simulation of the three sand mould casting schemes
I analyzed the filling process by observing the velocity field, filling percentage, and solid fraction. The purpose was to determine whether the gating system filled the mould smoothly, whether the metal remained continuous, and whether turbulence, gas entrapment, or sand erosion could occur.
For Scheme 1, the closed gating system, the metal filled the sprue first and then entered the runner and ingate. At 25% filling, the solid fraction was 0.0%. At 30.6% filling, the solid fraction was still 0.0%. At 64% filling, the solid fraction was 2.5%. At 95.9% filling, the solid fraction was 10.4%. The flow was generally stable, and the gating system remained full. However, there was some dispersed metal in the early stage, and the slag trapping effect was only moderate.
For Scheme 2, the open gating system, the metal flow was less stable. At 25% filling, the solid fraction was 0.0%. At 30.3% filling, the solid fraction was 0.1%. At 64% filling, the solid fraction was 10.5%. At 96.4% filling, the solid fraction was 18.8%. Isolated liquid metal appeared, which increased the risk of cold shut and cold shot. The gating system was not always full, so the slag trapping effect was poor.
For Scheme 3, the semi-closed gating system, the metal flow was the most stable. At 25.3% filling, the solid fraction was 0.0%. At 30.6% filling, the solid fraction was 0.0%. At 64% filling, the solid fraction was 4%. At 96.3% filling, the solid fraction was 12.7%. The gating system remained full, and there was no broken or isolated liquid metal. The runner acted as a slag barrier, and the filling process was smooth.
| Scheme | Solid fraction at 25% filling | Solid fraction at 30% filling | Solid fraction at 64% filling | Solid fraction at 96% filling | Filling conclusion |
|---|---|---|---|---|---|
| Scheme 1 closed | 0 | 0 | 2.5% | 10.4% | Stable flow, moderate slag trapping |
| Scheme 2 open | 0 | 0.1% | 10.5% | 18.8% | Turbulence, isolated liquid, high cold shut risk |
| Scheme 3 semi-closed | 0 | 0 | 4% | 12.7% | Most stable flow, good slag trapping |
Temperature field and solidification simulation
I then simulated the temperature field and solidification process. The purpose was to identify isolated liquid zones and predict where shrinkage cavity and shrinkage porosity would form. In the temperature field plots, the solid fraction indicated how much of the casting had solidified. A higher solid fraction meant that less liquid metal remained to feed the hot spots.
For Scheme 1, at 8.4 s the solid fraction was 16%, and the flange end and sharp edges cooled first. At 16.7 s, the solid fraction was 26%, and cooling spread through the casting. At 96.7 s, the solid fraction was 74.4%, and isolated liquid zones began to appear. At 117.7 s, the solid fraction was 86.6%, and isolated liquid zones were scattered in the casting. These zones could not be fed and could become shrinkage cavities.
For Scheme 2, at 11.7 s the solid fraction was 30%, and cooling started at the outer edges. At 31.3 s, the solid fraction was 50.7%, and the flange face began to solidify. At 71.3 s, the solid fraction was 78.9%, and isolated liquid zones appeared. At 121.3 s, the solid fraction was 92.6%, and large isolated liquid zones remained. This scheme had the highest risk of volume deficiency and shrinkage cavity.
For Scheme 3, at 12.6 s the solid fraction was 24%, and cooling started at the flange and outer edges. At 82 s, the solid fraction was 69.9%, and cooling spread through the casting. At 102 s, the solid fraction was 77.3%, and a large isolated liquid zone appeared. At 122 s, the solid fraction was 82.7%, and the isolated liquid zones were smaller and more scattered. Compared with Scheme 2, Scheme 3 had the smallest isolated liquid zone and the lowest shrinkage tendency.
| Scheme | Early solidification | Middle solidification | Late solidification | Final solidification | Isolated liquid zone |
|---|---|---|---|---|---|
| Scheme 1 closed | 16% | 26% | 74.4% | 86.6% | Small isolated zones |
| Scheme 2 open | 30% | 50.7% | 78.9% | 92.6% | Large isolated zones, high shrinkage risk |
| Scheme 3 semi-closed | 24% | 69.9% | 77.3% | 82.7% | Smallest isolated zones |
Shrinkage cavity and porosity prediction
The simulation predicted shrinkage defects in all three schemes, but the severity differed. I classified the shrinkage defects as hot spot shrinkage, riser neck shrinkage, ingate shrinkage, axis shrinkage, and reentrant corner shrinkage. The hot spot shrinkage was the most relevant because the axle had thick sections that solidified last. If those sections were not fed by a riser or chilled by a cold iron, the volume contraction produced internal cavities.
| Shrinkage type | Formation mechanism | Relevance to the axle |
|---|---|---|
| Hot spot shrinkage | Last solidifying thick section without feeding | Main cause at oil seal seat and flange transitions |
| Riser neck shrinkage | Shrinkage penetrates through riser neck into casting | Occurs if riser design is poor |
| Ingate shrinkage | Ingate solidifies slowly and lacks feeding | Possible at ingate contact |
| Axis shrinkage | Dispersed small cavities along centerline | Can occur in thick sections |
| Reentrant corner shrinkage | Sharp internal corner creates hot spot and gas entrapment | Possible at rib intersections |
Scheme 2 produced the largest and most numerous shrinkage cavities, so I eliminated it. Scheme 1 and Scheme 3 were selected for further optimization with risers and cold irons.
Riser design for sand mould casting
I used the modulus method to design the risers. The modulus is defined as the ratio of volume to heat-dissipating surface area:
$$M = \frac{V}{A}$$
where \(M\) is the modulus, \(V\) is volume, and \(A\) is the heat-dissipating surface area. The axle modulus was calculated to be about 9.7 mm. The riser modulus must be larger than the casting modulus so that the riser solidifies later and continues to feed the casting. I used the following empirical formulas for riser diameter and height:
$$D_r = (1.2 \sim 2)\delta_c$$
$$H_r = (1.0 \sim 1.5)D_r$$
where \(\delta_c\) is the wall thickness at the hot spot. For a hot spot thickness of about 22 mm, the calculated riser diameter was 44 mm and the height was 66 mm. Considering practical production and feeding capacity, I selected a riser diameter of 80 mm and a riser height of 150 mm. The riser modulus was about 18.75 mm, which is greater than the axle modulus of 9.7 mm, so the riser can feed the casting effectively.
I used both blind risers and open risers. The blind riser was placed at the ingate to collect slag and feed the nearby hot spot. The open risers were placed on the upper surface of the axle to feed the thick sections and allow gas to escape. This arrangement promoted sequential solidification from the thin sections toward the risers.
| Parameter | Value | Purpose |
|---|---|---|
| Axle modulus | 9.7 mm | Reference for riser design |
| Calculated riser diameter | 44 mm | Minimum diameter from empirical formula |
| Calculated riser height | 66 mm | Minimum height from empirical formula |
| Selected riser diameter | 80 mm | Improved feeding capacity |
| Selected riser height | 150 mm | Improved feeding and gas escape |
| Riser modulus | 18.75 mm | Greater than axle modulus, effective feeding |
Cold iron design and optimization
Cold irons were placed in regions where a riser could not be installed. The cold iron increased the local solidification rate and created an artificial end zone. This helped establish sequential solidification and reduced shrinkage porosity. I used cast steel cold irons with dimensions of about 30 mm × 50 mm × 60 mm. The cold irons were cleaned and derusted before use. They could be reused multiple times. In the simulation, I placed internal cold irons in the lower part of the internal cavity and external cold irons at thick wall sections. The cold irons reduced the local hot spot and shifted the last solidifying region toward the riser.
I then re-modeled Scheme 1 and Scheme 3 with risers and cold irons and re-ran the Procast simulation. The optimized Scheme 1 still showed local shrinkage porosity near the cylindrical inner cavity close to the flange. The shrinkage was reduced but not eliminated. The optimized Scheme 3, however, showed that the shrinkage cavity and shrinkage porosity disappeared from the axle and transferred to the riser. The hot spot shrinkage was almost eliminated. This demonstrated that the semi-closed gating system, together with properly placed risers and cold irons, could achieve sequential solidification and produce a sound sand mould casting.
| Scheme after optimization | Flow field result | Late solidification fraction | Defect prediction | Result |
|---|---|---|---|---|
| Scheme 1 closed + risers + cold irons | Stable flow, moderate slag trapping | 86% | Local shrinkage remained | Not selected |
| Scheme 2 open | Poor flow, isolated liquid | 92% | Large shrinkage area | Eliminated |
| Scheme 3 semi-closed + risers + cold irons | Most stable flow, good slag trapping | 82% | Shrinkage eliminated from casting | Best scheme |
Final sand mould casting process
The final process I selected was the optimized semi-closed gating system. The axle material remained ZG40Mn. The pouring temperature was 1580 °C, and the pouring time was 10 s. The moulding method was one part per mould. The gating ratio was \(\sum A_{\text{ingate}} : \sum A_{\text{runner}} : \sum A_{\text{sprue}} = 1 : 0.8 : 1.2\). The ingate cross-sectional area was 22.5 cm², the runner area was 18 cm², and the sprue area was 27 cm². The gating system had one sprue, one runner, and one ingate. A blind riser was placed at the ingate to feed and collect slag. Open risers were placed on the upper surface of the axle. Internal cold irons were placed in the axle cavity, and external cold irons were placed at thick sections. A silicon carbide foam ceramic filter was placed at the bottom of the sprue to reduce inclusions and stabilize the flow. This configuration promoted sequential solidification and minimized shrinkage cavity and shrinkage porosity in the sand mould casting.
| Final parameter | Value |
|---|---|
| Material | ZG40Mn |
| Pouring temperature | 1580 °C |
| Pouring time | 10 s |
| Moulding method | One part per mould |
| Gating system | Semi-closed |
| Gating ratio | 1 : 0.8 : 1.2 |
| Ingate area | 22.5 cm² |
| Runner area | 18 cm² |
| Sprue area | 27 cm² |
| Riser arrangement | Blind riser at ingate, open risers on top surface |
| Cold iron arrangement | Internal cavity and thick wall sections |
| Filter | Silicon carbide foam ceramic filter, 30 cm² |
Experimental validation by sectioning
To verify the optimized sand mould casting process, I produced sample axles and sectioned them at intervals of 10 cm. In the first two batches, I sectioned 2 axles. In the next batch, I sectioned 8 axles. I inspected the original defect locations and critical areas with a 10× magnifying glass. I found no visible cracks and no holes larger than 3 mm. Over five production batches, I produced 630 axles. I selected 2 axles per hundred for sectioning, and I sectioned 31 axles in total. No shrinkage cavity or shrinkage porosity was found in the sectioned samples. This confirmed that the optimized gating system, risers, and cold irons were effective in sand mould casting.
Air-tightness test and production verification
I then performed air-tightness tests on the machined axles. The purpose was to check for internal defects that could cause oil leakage. The test used a direct pressure method. The axle was sealed with tooling, submerged in water, and pressurized with compressed air. The pressure was 0.5 MPa, and the holding time was at least 10 minutes. I observed whether bubbles appeared. If no bubbles appeared, the axle was considered qualified. The water contained a water-based rust inhibitor mixed at a ratio of 1:10. Rubber gaskets, combined washers, and O-rings were used to seal the openings. The test procedure was as follows: clean the axle surface, install the sealing tooling, connect compressed air, lower the axle into water to a depth of 10–20 cm, pressurize slowly to 0.5 MPa, hold for 10 minutes, observe bubbles, release pressure, remove the tooling, and mark any leaking location.
| Sample group | Number of samples | Test pressure | Holding time | Leakage | Result |
|---|---|---|---|---|---|
| Batch A | 8 | 0.5 MPa | 10 min | None | Qualified |
| Batch B | 12 | 0.5 MPa | 10 min | None | Qualified |
| Total tested | 20 | 0.5 MPa | 10 min | 0 | 100% qualified |
The air-tightness test results showed that the optimized semi-closed sand mould casting process effectively reduced shrinkage cavity and shrinkage porosity around the inner half-shaft oil seal seat. No leakage occurred in the tested samples. By the end of the production period, more than 6,000 axles had been produced with the new process. Only 174 parts were found to be nonconforming, giving a failure rate below 3%. This was a major improvement from the original failure rate of about 70%.
| Production stage | Original process | Optimized sand mould casting process |
|---|---|---|
| Failure rate | About 70% | Below 3% |
| Main defect | Shrinkage cavity and porosity at oil seal seat | Defects transferred to riser, casting sound |
| Leakage | Frequent oil seepage and leakage | No leakage in tested samples |
| Economic loss | More than 1 million RMB per year | Substantially reduced |
Conclusions from my study
I designed and optimized a sand mould casting process for a tractor divided axle. I analyzed the structure and material, determined the pouring position, parting surface, pouring temperature, shrinkage rate, sand core design, and draft angle. I calculated the gating system using the pouring speed method and designed three gating systems: closed, open, and semi-closed. I used Procast to simulate the filling and solidification processes of all three schemes. The open gating system produced unstable flow, isolated liquid metal, and the largest shrinkage defects, so I eliminated it. The closed gating system performed better but still produced local shrinkage after riser and cold iron optimization. The semi-closed gating system produced the smoothest filling, the best slag trapping, and the smallest isolated liquid zone. After adding risers and cold irons, the semi-closed system eliminated shrinkage cavity and shrinkage porosity from the axle and transferred the defects to the riser.
The final optimized sand mould casting process used a semi-closed gating system with a gating ratio of 1 : 0.8 : 1.2. The ingate area was 22.5 cm², the runner area was 18 cm², and the sprue area was 27 cm². The pouring temperature was 1580 °C, and the pouring time was 10 s. A blind riser was placed at the ingate, open risers were placed on the upper surface, internal cold irons were placed in the cavity, and external cold irons were placed at thick sections. A silicon carbide foam ceramic filter was placed at the bottom of the sprue. Sectioning and air-tightness tests confirmed that the optimized process produced sound castings with a failure rate below 3%, down from about 70%.
Innovation and practical value
The main innovation of my work was the combined use of Procast simulation, temperature field analysis, and experimental validation to determine the solidification sequence and shrinkage locations in a complex sand mould casting. I used risers to promote sequential solidification and cold irons to create artificial end zones. This reduced shrinkage cavity and shrinkage porosity without relying only on trial-and-error. The work also showed how different gating ratios affect flow stability, slag trapping, and defect formation in sand mould casting. The final process is practical for production and can be applied to similar heavy vehicle axle castings.
Future work
In future work, I would extend the simulation to a fully coupled flow–temperature–stress analysis. This would improve the accuracy of hot crack and deformation prediction. I would also measure the actual thermophysical properties of ZG40Mn and the water-glass sand used in production, rather than relying mainly on the software database. More accurate material data would improve the reliability of the sand mould casting simulation. I would also study microstructure evolution and mechanical properties after normalizing and tempering, because these properties are important for the service life of the divided axle. Finally, I would investigate the use of different riser shapes, insulating sleeves, and chill materials to further improve yield and reduce cleaning cost in sand mould casting.
Summary of key equations and parameters
| Item | Equation or value | Role in sand mould casting |
|---|---|---|
| Casting shrinkage | \(K = \frac{L_M – L_J}{L_J} \times 100\%\) | Compensates for linear contraction |
| Minimum residual head | \(h_M = L \tan \alpha\) | Ensures complete filling |
| Choke area | \(A = \frac{G_L}{t k S’}\) | Determines gating dimensions |
| Ingate diameter | \(d = \sqrt{\frac{4A}{\pi}}\) | Sizes cylindrical ingate |
| Modulus | \(M = \frac{V}{A}\) | Designs risers for feeding |
| Riser diameter | \(D_r = (1.2 \sim 2)\delta_c\) | Ensures riser feeding capacity |
| Riser height | \(H_r = (1.0 \sim 1.5)D_r\) | Provides sufficient metal head |
| Heat conduction | \(\rho C \frac{\partial T}{\partial t} = \lambda \nabla^2 T + w\) | Predicts solidification temperature field |
I conclude that the optimized semi-closed sand mould casting process is the best route for the tractor divided axle. It provides smooth filling, effective slag trapping, sequential solidification, and reliable feeding. The simulation results agree with the sectioning and air-tightness tests. The process reduces the failure rate from about 70% to below 3% and has clear practical value for production.