This work summarizes my research on the production of railway steel castings using the green moulding clay sand process. The aim is to achieve high-quality castings at low cost while keeping the process energy-efficient and capable of mass production. I systematically studied the raw materials, designed and verified sand recipes, used computer simulation to optimize the casting process, and analyzed the main categories of sand foundry defects encountered in trial and batch production. The resulting process enabled the foundry to produce No.17 coupler yokes and No.16 coupler knuckles that satisfied the technical requirements of railway component standards.
The green moulding clay sand process is widely used because of its high productivity, short production cycle, low energy consumption and suitability for mechanized continuous production. However, this process is also sensitive to several variables. The sand mixture contains water, bentonite and starch additives, and its properties can be affected by temperature and humidity. For steel castings, the high pouring temperature of steel, the large volume contraction, and the relatively low strength of a green sand mould make it difficult to avoid typical sand foundry defects such as blowholes, scabs, sand inclusions, metal penetration, shrinkage porosity and cracks. Therefore, the core task of this work is to control the sand mixture and the casting process precisely enough to reduce the probability of sand foundry defects while retaining the economic benefits of green moulding.
This article is organized as follows. First, I describe the fundamentals of green moulding clay sand materials and performance requirements. Then I present the experimental design and results for face sand and backing sand. After that, I explain how ProCAST simulation was used to optimize the gating and risering system of the coupler yoke and knuckle. Finally, I discuss the product trial, the identification and prevention of sand foundry defects, and the verification status of mass production.
Fundamentals of Green Moulding Clay Sand for Steel Castings
Green moulding clay sand consists mainly of silica sand, bentonite, starch-based additives and water. The sand is used in the as-mixed wet state without baking. This gives the mould good collapsibility and ease of shakeout, but it also limits the mould strength. During pouring, water near the mould surface vaporizes and migrates inward, so the surface layer may expand, lose strength, or crack. These phenomena are responsible for many sand foundry defects.
For steel castings, the mould must have high refractoriness. Silica sand should contain at least 94% SiO2, low clay content, and a controlled particle size distribution. The recommended grain size for steel casting is often 50/100 mesh or 70/140 mesh. Sodium- or sodium-modified bentonite is generally preferred because it provides better thermal stability, higher hot wet tensile strength, and more resistance to sand expansion defects. Starch-based additives such as pre-gelatinized starch and dextrin are added in small amounts to improve toughness, prevent brittle edges, increase surface stability, and provide sacrificial carbonaceous material that helps to reduce certain sand foundry defects.
The important performance parameters of green moulding clay sand include the following:
- Compactability: an indicator of the relative wetness of the sand and its behavior under ramming.
- Permeability: the ability of the mould to allow gas to escape during pouring.
- Green compression strength: the load-carrying ability of the sand in the wet state.
- Hot wet tensile strength: the strength of the sand at the wet layer beneath the hot surface, closely related to scabbing and expansion defects.
- Shatter index: a measure of toughness and deformation resistance.
- Surface stability: the ability of the sand surface to resist erosion by molten metal.
I used the following fundamental equations to describe some of these properties. The water content is calculated as:
$$
w = \frac{m_{w}}{m_{s}} \times 100\%
$$
where \(w\) is moisture content, \(m_{w}\) is the mass of water, and \(m_{s}\) is the mass of dry sand. Compactability can be defined by the decrease in height of a sand specimen after a standardized ramming action:
$$
C = \frac{H_{0} – H}{H_{0}} \times 100\%
$$
where \(H_{0}\) is the original height of sand in the cylinder and \(H\) is the height after ramming. When using the permeability meter, the permeability number is commonly expressed as:
$$
P = \frac{V \cdot H}{A \cdot p \cdot t}
$$
where \(V\) is the volume of air passing through the specimen, \(H\) is the specimen height, \(A\) is the cross-sectional area, \(p\) is the air pressure, and \(t\) is the time. A more quantitative understanding of these properties is essential because they directly control the tendency for various sand foundry defects.
Raw Material Testing and Sand Characterization
Before designing the sand recipes, I tested the new sand and the return sand used by the foundry. The new sand was a washed silica sand supplied from the Jinjiang deposit, classified as 40/70 mesh. The return sand came from the moulding line after repeated usage. The particle size distributions are summarized in the table below.
| Sieve fraction | 6 | 12 | 20 | 30 | 40 | 50 | 70 | 100 | 140 | 200 | 270 | >270 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| New sand / % | 0 | 0 | 0.14 | 1.02 | 53.86 | 38.3 | 3.27 | 2.73 | 0.39 | 0.38 | 0.07 | 0.04 |
| Return sand / % | 0 | 0 | 1.85 | 0.78 | 24.31 | 71.71 | 0.15 | 0.38 | 0.26 | 0.30 | 0.26 | 0 |
I also measured the return sand properties. The return sand contained 6.80% mud, 0.70% moisture, and 5.05% active bentonite determined by the methylene blue titration method. The standard adsorption curve of bentonite on new sand was established using known additions of 2%, 4%, 6%, 8% and 10%. The relationship between bentonite addition and methylene blue adsorption was linear in the tested range. This calibration was used to calculate the active clay content in return sand. The active clay content is calculated as:
$$
A_{b} = \frac{M_{B}}{m_{s}} \times 100\%
$$
where \(M_{B}\) is the equivalent mass of methylene blue adsorbed and \(m_{s}\) is the mass of the sand sample. Accurate measurement of active bentonite is important because too little active clay leads to low strength and erosion, while too much dead clay increases water demand and causes various sand foundry defects.
Face Sand Experiments
For the face sand, I used new sand as the base aggregate. The face sand directly contacts the molten steel and therefore must provide high refractoriness and good surface quality. I first carried out single-factor experiments to determine the reasonable ranges of bentonite, starch-based additive and water. A baseline mixture without additive was also tested. The following properties were measured: compactability, permeability, green compression strength, hot wet tensile strength, shatter index and surface stability.
In the baseline test using 100% new sand, 8.0% bentonite and 3.3% water, the compactability was 55%, permeability was 396, green compression strength was 86 kPa, hot wet tensile strength was 3.87 kPa, shatter index was 70%, and surface stability was 95.9%. The addition of starch-based additives improved the hot wet tensile strength and surface stability. The results showed that a total additive content of about 0.5% was suitable. I also tested different ratios of pre-gelatinized starch to dextrin. Based on the combined properties, the starch-to-dextrin ratio of 3:2 was selected as the additive system.
After the single-factor experiments, I designed an L9(33) orthogonal array. The three factors were bentonite content, additive content and water content. The levels are shown below.
| Factor | Level 1 / % | Level 2 / % | Level 3 / % |
|---|---|---|---|
| Bentonite | 7.5 | 8.0 | 8.5 |
| Starch-based additive | 0.1 | 0.3 | 0.5 |
| Water | 3.0 | 3.3 | 3.6 |
The orthogonal test results are summarized below.
| Run | Bentonite / % | Additive / % | Water / % | Compactability / % | Permeability | Green compression / kPa | Hot wet tensile / kPa | Shatter index / % | Surface stability / % |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 7.5 | 0.1 | 3.0 | 53 | 435 | 85 | 4.17 | 67.5 | 95.6 |
| 2 | 7.5 | 0.3 | 3.3 | 56 | 440 | 82 | 4.38 | 70.2 | 96.7 |
| 3 | 7.5 | 0.5 | 3.6 | 55 | 445 | 79 | 4.15 | 74.1 | 96.8 |
| 4 | 8.0 | 0.1 | 3.6 | 57 | 431 | 82 | 4.19 | 71.0 | 97.0 |
| 5 | 8.0 | 0.3 | 3.0 | 51 | 442 | 86 | 4.18 | 70.4 | 97.0 |
| 6 | 8.0 | 0.5 | 3.3 | 51 | 444 | 84 | 4.42 | 71.5 | 97.1 |
| 7 | 8.5 | 0.1 | 3.3 | 52 | 449 | 96 | 3.98 | 67.2 | 96.3 |
| 8 | 8.5 | 0.3 | 3.6 | 54 | 449 | 94 | 4.44 | 70.1 | 96.9 |
| 9 | 8.5 | 0.5 | 3.0 | 44 | 460 | 103 | 4.62 | 66.3 | 97.0 |
Using range analysis, I calculated the influence of each factor on each property. The range for factor \(j\) is:
$$
R_{j} = \max_{i}(K_{ji}) – \min_{i}(K_{ji})
$$
where \(K_{ji}\) are the average values of the property at the \(i\)-th level of factor \(j\). A larger range means that the factor has a stronger effect on that property. The results showed that water content was the main factor affecting compactability. Permeability was mainly affected by micro-powder content, which was related to bentonite and additives. Green compression strength was dominated by bentonite content. Hot wet tensile strength was mainly improved by starch-based additive content. Shatter index and surface stability were also strongly affected by additive content.
According to the orthogonal analysis and verification tests, the optimized face sand recipe is:
| Component | Recommended value / % |
|---|---|
| New silica sand | 100 |
| Bentonite | 8.0 – 8.5 |
| Starch-based additive | 0.3 – 0.5 |
| Water | 3.0 – 3.6 |
The sand properties achieved with this recipe are:
| Property | Value range |
|---|---|
| Compactability / % | 51 – 54 |
| Permeability | 440 – 450 |
| Green compression strength / kPa | 85 – 95 |
| Hot wet tensile strength / kPa | 4.1 – 4.5 |
| Shatter index / % | 70 – 72 |
| Surface stability / % | 97.0 – 97.5 |
This face sand met the foundry requirements. The water content should be adjusted according to sand temperature, environment temperature and humidity. In summer, a higher water content is needed; in winter, a lower value is suitable.
Backing Sand Experiments
The backing sand is used behind the face sand. It does not directly contact molten steel, so its refractoriness requirement is lower. The backing sand was based on return sand with small additions of bentonite, starch-based additive and water. A baseline test using 100% return sand and 3.3% total water gave a compactability of 56%, permeability of 205, green compression strength of 75 kPa, hot wet tensile strength of 3.41 kPa, shatter index of 65.93% and surface stability of 97.1%. Although this baseline already satisfied the foundry performance range, repeated use of sand would degrade its quality. Therefore, small additions of bentonite and additive were necessary.
I designed a second L9(33) orthogonal experiment for backing sand. The factor levels are shown below.
| Factor | Level 1 / % | Level 2 / % | Level 3 / % |
|---|---|---|---|
| Bentonite | 0.1 | 0.2 | 0.3 |
| Starch-based additive | 0.1 | 0.2 | 0.3 |
| Water | 3.0 | 3.3 | 3.6 |
The backing sand test results are shown in the following table.
| Run | Bentonite / % | Additive / % | Water / % | Compactability / % | Permeability | Green compression / kPa | Hot wet tensile / kPa | Shatter index / % | Surface stability / % |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 0.1 | 0.1 | 3.0 | 45 | 209 | 105 | 3.89 | 63.7 | 93.4 |
| 2 | 0.1 | 0.2 | 3.3 | 57 | 205 | 94 | 3.96 | 69.4 | 97.8 |
| 3 | 0.1 | 0.3 | 3.6 | 58 | 203 | 92 | 4.21 | 73.9 | 98.5 |
| 4 | 0.2 | 0.1 | 3.6 | 58 | 203 | 94 | 3.99 | 73.6 | 98.2 |
| 5 | 0.2 | 0.2 | 3.0 | 47 | 211 | 107 | 3.79 | 64.9 | 93.9 |
| 6 | 0.2 | 0.3 | 3.3 | 57 | 206 | 95 | 4.50 | 72.6 | 97.6 |
| 7 | 0.3 | 0.1 | 3.3 | 58 | 208 | 96 | 4.09 | 73.7 | 98.2 |
| 8 | 0.3 | 0.2 | 3.6 | 59 | 204 | 95 | 4.63 | 75.9 | 98.0 |
| 9 | 0.3 | 0.3 | 3.0 | 44 | 213 | 107 | 4.01 | 64.4 | 94.0 |
The optimized backing sand recipe is:
| Component | Recommended value / % |
|---|---|
| Return sand | 100 |
| Bentonite | 0.2 – 0.3 |
| Starch-based additive | 0.1 – 0.3 |
| Water | 3.3 – 3.6 |
The achieved backing sand properties are:
| Property | Value range |
|---|---|
| Compactability / % | 55 – 60 |
| Permeability | 200 – 210 |
| Green compression strength / kPa | 90 – 95 |
| Hot wet tensile strength / kPa | 3.8 – 4.5 |
| Shatter index / % | 72 – 75 |
| Surface stability / % | 97.5 – 98.0 |
For the backing sand, the range analysis showed that water content was the dominant factor for most properties. This is because return sand contains a large amount of fine particles and dead clay. The fines increase the specific surface area and water demand. Consequently, even a small change in water content can greatly alter the compactability, permeability and strength. Therefore, precise control of moisture is critical to prevent many sand foundry defects caused by overly wet or overly dry backing sand.
Foundry Verification of the Sand Recipes
The proposed face sand and backing sand formulas were verified during later batch production. The foundry used an advanced sand preparation system with online measurement of moisture, compactability, green strength and temperature. In production, the actual water and additive additions were automatically adjusted according to the online measurements. A typical production day in summer, with workshop temperature around 34°C and humidity around 80%, used the following sand formulations:
| Sand type | New sand / % | Return sand / % | Bentonite in return sand / % | Added bentonite / % | Starch additive / % | Water / % |
|---|---|---|---|---|---|---|
| Face sand | 100 | 0 | 0 | 8.0 | 0.3 | 3.3 |
| Backing sand | 0 | 100 | 5.1 | 0.1 | 0.2 | 3.5 |
The measured sand properties in the factory were close to the laboratory optimized values. The face sand showed compactability 55%, permeability 465, green compression strength 102 kPa, hot wet tensile strength 4.2 kPa, shatter index 74% and surface stability 98%. The backing sand showed compactability 58%, permeability 225, green compression strength 90 kPa, hot wet tensile strength 3.8 kPa, shatter index 72% and surface stability 97.7%. This confirmed that the laboratory-derived sand recipes could be implemented successfully in production.
Process Simulation and Optimization of Railway Steel Castings
Simulation is very useful for reducing the difficulty of trial production and for reducing the probability of sand foundry defects before the first real pour. I used ProCAST, a finite-element based casting simulation software, to analyze the filling, solidification and stress fields of the No.17 coupler yoke and No.16 coupler knuckle. The cast material was ZG25MnCrNiMo, a high-strength low-alloy E-grade steel. The important thermal parameters are listed below.
| Parameter | Value |
|---|---|
| Liquidus temperature / °C | 1505 |
| Solidus temperature / °C | 1433 |
| Latent heat / kJ·kg−1 | 65 |
| Pouring temperature / °C | 1585 |
| Mould/casting interface heat transfer coefficient / W·m−2·K−1 | 500 |
| Insulating riser/casting interface heat transfer coefficient / W·m−2·K−1 | 100 |
For the solidification process, the governing equation used in the simulation is:
$$
\rho c_{p} \frac{\partial T}{\partial t} = \nabla \cdot (k \nabla T) + \rho L \frac{\partial f_{s}}{\partial t}
$$
where \(\rho\) is density, \(c_{p}\) is specific heat, \(T\) is temperature, \(t\) is time, \(k\) is thermal conductivity, \(L\) is latent heat and \(f_{s}\) is solid fraction. For the liquid flow, the simulation solves the continuity and momentum equations:
$$
\nabla \cdot \vec{v} = 0
$$
$$
\rho \left( \frac{\partial \vec{v}}{\partial t} + \vec{v} \cdot \nabla \vec{v} \right) = -\nabla p + \mu \nabla^{2} \vec{v} + \rho \vec{g}
$$
where \(\vec{v}\) is velocity, \(p\) is pressure, \(\mu\) is viscosity and \(\vec{g}\) is gravitational acceleration. The predicted flow field allowed me to assess erosion and unwanted gas entrapment, while the temperature field enabled prediction of shrinkage porosity and hot spots.
Simulation of the Original Coupler Yoke Process
The No.17 coupler yoke is an important railway component with complex geometry. Its overall dimensions are approximately 975 mm × 574 mm × 496 mm and its mass is about 190 kg. The original process used one mould with two yokes, horizontal pouring, ordinary sand risers at the head and tail, and chills in crack-sensitive regions. The simulation predicted the following issues:
- High liquid steel velocity in the region opposite the ingate, especially near the tail pin hole, causing erosion of the mould and core and increasing the risk of sand inclusions.
- A large flat surface on the yoke body, where liquid steel flowed for a long distance and possible gas entrapment could create blowholes.
- Independent liquid pools at inner and outer corners of the hook head, at the root of the traction seat inside the hook cavity, and at the transition between the hook head and body, indicating insufficient feeding and a high risk of shrinkage porosity.
- High stress levels at the riser roots, tail pin hole area, hook shoulder and complex inner surfaces, with a tendency for hot tearing.
These simulation results corresponded well with the defects observed in real trial castings. The original process produced gas holes on the yoke body, sand inclusion and rough surfaces at the tail pin hole, and hot cracks at the riser roots and hook corners. I therefore proposed an optimized process for the yoke. The main changes were:
- Change the ingate position from the tail region to the side of the yoke body, using a preformed refractory runner to reduce erosion.
- Use three cylindrical insulating risers at the hook head and one additional insulating riser above the yoke body.
- Add a small blind riser near the ingate to improve feeding of the gating region.
- Stagger the mould vent holes and riser vent holes to prevent sand dropping into the cavity.
- Add shaped chills at hot spots to promote either simultaneous solidification or directional solidification toward the risers.
The optimized process simulation showed lower velocity in the tail pin hole area, no retained gas on the yoke body, improved feeding through the risers, and fewer hot cracks. The stress level was considerably reduced. The optimized yoke process was used as the final trial process and produced cast parts with fewer defects. Remaining minor defects could be handled by subsequent grinding and welding repair.
Simulation of the Original Coupler Knuckle Process
The No.16 coupler knuckle is a smaller but still critical steel casting. Its dimensions are about 280 mm × 297 mm × 220 mm and its mass is about 40 kg. The original process used one mould with eight knuckles. Four knuckles were arranged around a central blind riser. The simulation predicted the following issues:
- High metal velocity near the knuckle pin hole because the ingate directly faced that sensitive core, causing erosion and possible sand inclusion.
- Gas entrapment at the tail stop region because this area was filled last.
- Shrinkage porosity in the inner wrist surface and near the guard protrusion because the local wall thickness was large and the feeding path from the central riser was insufficient.
- High stress and hot tearing tendency at the traction lug, tail stop corner, and inner wrist corners.
In the real trial production, more than half of the knuckles produced by the original process had obvious erosion and inclusion defects around the pin hole. Magnetic particle inspection found micro-cracks at the gate root, the middle upper part of the inner wrist, the head opening and the guard protrusion. Therefore, I optimized the knuckle process as follows:
- Change the mould layout to one mould with four knuckles, arranged symmetrically in pairs, still with horizontal pouring.
- Move the ingate from the pin-hole side to the back side of the knuckle between the pin hole and the tail, thus avoiding direct erosion of the pin-hole core.
- Remove the central large riser and place a small top insulating blind riser above the pressure boss in the upper mould, which is a thick region requiring extra feeding.
The simulation of the optimized knuckle process showed more stable filling and lower velocity around the pin-hole core. The top blind riser fed the inner wrist and pressure boss better than the central riser. Shrinkage porosity in the guard protrusion and traction lug region shifted toward the middle of the knuckle, away from the most critical load-bearing areas. The stress calculation showed that the overall stress level was slightly higher than in the original process, but the important inner wrist surface remained acceptable. The hot tearing tendency in the tail stop and inner corner regions was reduced. Because the optimized process required a new pattern, which took time to manufacture, the early batch production of knuckles still used a modified version of the original process. The process modifications were later implemented step by step.
Foundry Layout and Production Verification
The partner foundry established a complete production line for the No.17 coupler yoke and No.16 coupler knuckle. The line included sand treatment, moulding, core making, mould assembly, melting, pouring, shakeout, primary cleaning, pre-normalizing, magnetic particle inspection, welding repair, quenching and tempering, hardness testing, final machining, assembly and function testing. Steel was melted in a 10-ton electric arc furnace and refined in a 20-ton ladle furnace. This double-refining scheme improved steel cleanliness and reduced gas content, which is helpful for minimizing sand foundry defects associated with dissolved gases and non-metallic inclusions.
One of the key production steps was the automatic pouring line used in batch production. The pouring operation directly influences mould erosion, gas entrapment and slag entrainment, all of which are important sources of sand foundry defects.
During trial production, the process parameters were continuously adjusted. The final sand recipes were used together with optimized gating, risering and cooling practices. After several adjustments, the foundry obtained qualified castings. The products passed dimensional inspection, surface inspection, magnetic particle inspection, and specified-section density examination. The chemical composition and mechanical properties met the requirements for E-grade steel. The product also passed the static tensile test, dynamic tear test and hardenability test. This confirmed that the foundry had the ability to begin mass production.
Analysis of Key Casting Defects
Although the final products were qualified, the trial and early batch casting still exposed some key defects. I analyzed these defects according to their appearance, location and formation mechanism. The main defects belonged to the broad category of sand foundry defects, including gas holes, sand scab, sand inclusion, metal penetration, shrinkage porosity and cracks.
For the coupler yoke, the main defects were cracks, gas holes, sand scab and mechanical penetration. The cracks mainly appeared at the roots of gates and risers, the hook shoulder, and complicated inner surface corners. These cracks were mostly hot cracks, which are irregular, intergranular and oxidized on the fracture surface. Cold cracks appeared occasionally at gate roots or repaired welds and were caused by excessive mechanical forces or poor cutting practice.
Gas holes in the yoke were found near the upper surface of the yoke body and under the surface, often appearing as smooth round voids. They were caused by excessive moisture, poor venting, or turbulent filling. Sand scab was found inside the hook head on the upper complex surfaces. The scab formed because the mould surface expanded and spalled under the heat of steel, allowing metal to enter behind the sand layer. Mechanical penetration occurred inside the tail pin hole, where the sand core suffered high thermal and mechanical attack. The surface became rough and showed a sponge-like appearance after cleaning.
For the coupler knuckle, the main defects were hot cracks, sand scab and mechanical penetration. Cracks appeared at the gate root, the middle upper part of the inner wrist, the head opening and the upper part of the guard protrusion. Sand scab occurred around the pin hole and on the upper surface near the knuckle head. Mechanical penetration was mainly found on the inner wrist surface.
Defect Formation Mechanisms
The formation of these defects is related to the special nature of the green moulding clay sand process and the complex geometry of railway steel castings. During pouring, the mould surface is heated rapidly and the water in the surface layer migrates inward. This creates a weak wet layer beneath the surface. At the same time, silica sand expands. If the expansion stress exceeds the strength of the wet layer, the mould surface can delaminate. This is the classic mechanism for scabbing and sand inclusions. The compactability and hot wet tensile strength of the sand mixture are therefore important parameters for controlling sand foundry defects.
Gas holes are influenced by the moisture content of the sand, the gas permeability of the mould, and the filling conditions. The amount of gas generated from the sand can be estimated as:
$$
G = G_{0} \cdot (1 – f_{s})
$$
where \(G_{0}\) is the initial gas amount at the molten-metal/sand interface and \(f_{s}\) is the solid fraction. If the gas cannot escape through the pores of the sand because the permeability is too low or the venting is blocked, the gas pressure rises and may enter the solidifying metal, forming blowholes. Therefore, stable pouring, adequate venting and controlled moisture are essential for reducing this group of sand foundry defects.
Cracks in steel castings are often caused by the combination of thermal stress, solidification shrinkage and mould constraint. The thermal stress can be expressed by:
$$
\sigma_{th} = E \alpha \Delta T
$$
where \(E\) is the elastic modulus, \(\alpha\) is the linear thermal expansion coefficient, and \(\Delta T\) is the temperature difference. For steel castings, the high pouring temperature and the great temperature difference between thick and thin sections generate large stresses. If the stress exceeds the high-temperature strength of the steel, hot cracks form. Cold cracks may form later during shakeout, cutting, grinding or heat treatment, when residual stresses are released suddenly.
Preventive Measures for Sand Foundry Defects
Based on the analysis, I proposed a series of preventive measures that were later implemented in batch production.
For hot cracks:
- Increase the fillet radius at the connection between gates, risers and castings.
- Add anti-crack ribs at critical corners to carry tensile stress in the early stage of solidification.
- Optimize the placement of chills so that the casting either cools simultaneously or solidifies directionally toward risers.
- Apply insulating coatings near the connection corners to balance cooling speed.
- Control sulfur and phosphorus content in steel because these elements increase hot-cracking susceptibility.
- Avoid too early shakeout and rapid cooling. Allow the casting to cool uniformly before removal.
- Control pre-normalizing and tempering cycles so that heating and cooling are uniform and moderate.
For cold cracks:
- Control the shakeout and handling procedures to avoid impact and excessive force.
- Improve the cutting technique for gates and risers. Preheat the root before cutting, cut step by step, and leave a small residual length before final grinding.
- After cutting, place the casting in a warm furnace or cover it with insulating material. Do not place it in a draft or cool it with water.
For welding cracks:
- Preheat the repair area when the ambient temperature is low or when the local section is thick.
- Control the welding heat input and use proper welding consumables.
- Slowly cool the casting after welding to allow hydrogen to diffuse out of the weld metal.
For gas holes:
- Strictly control the moisture content of the sand mixture.
- Ensure the mould and core venting system is open and sufficient.
- Pour the metal smoothly without interruption. Avoid excessively high pouring speed.
- Use clean and dry charge materials to reduce dissolved gases in the steel.
For scabs and sand inclusions:
- Add an appropriate amount of starch-based additive to provide sacrificial space for sand expansion and to improve the hot wet tensile strength.
- Keep the degree of ramming uniform, especially in complex corners.
- Use high-quality coatings in critical areas. For severe conditions, use zircon sand or chromite sand in the mould face layer.
- Ensure proper mould assembly and avoid loose sand in the cavity.
For metal penetration:
- Improve the coating quality on cores and mould surfaces. Apply a sufficiently thick and uniform coating.
- Control the pouring temperature and the pouring height so that the ferrostatic pressure is not excessively high.
- Maintain adequate refractoriness by controlling the amount of fine powder and returning sand dust in the face sand.
These measures greatly reduced the frequency and severity of common sand foundry defects in the subsequent batch production.
Batch Production Quality Analysis
After the trial production and process optimization, the foundry started formal batch production. During this stage, I monitored the product quality by collecting inspection data. The first inspection after shakeout was a visual check for obvious surface defects such as sand inclusion, gas holes and cracks. The results are summarized below.
| Product | Inspected quantity | Defective quantity | Defect rate / % | Scrap reason |
|---|---|---|---|---|
| Coupler yoke | 154 | 8 | 5.2 | Penetrating sand inclusion through wall |
| Coupler knuckle | 223 | 6 | 2.7 | Penetrating sand inclusion through wall |
The scrap defects were caused by severe sand inclusions that penetrated through the whole wall thickness. The main reason was insufficient mould strength at the ingate and complex corners, or improperly assembled moulds that allowed loose sand to be washed into the cavity by the molten steel. Although this defect was not frequent, it caused complete rejection of the castings. Therefore, stricter control of mould ramming and assembly was required.
After pre-normalizing and primary welding repair, magnetic particle inspection was performed in the foundry shop. For the yoke, the inspection positions included the front face, side faces, neck, tail and gate/riser locations. The total number of detected cracks decreased significantly compared with the trial stage, but cracks still occurred. The distribution of cracks in the foundry shop inspection is shown below.
| Position | Number of cracked yokes |
|---|---|
| Front and side faces | 36 |
| Hook neck | 9 |
| Tail | 21 |
| Gate/riser roots and other positions | 112 |
| Total | 179 |
The total number is larger than the number of inspected yokes because one casting could have cracks at several positions. Cracks at the gate and riser roots represented the largest share. For the knuckle, the foundry shop inspection found 21 cracks on the inner wrist surface and 8 cracks at other positions in a batch of 178 inspected knuckles. These cracks were shallow surface cracks and could be removed by air-arc gouging and welding repair.
After quenching and tempering, the parts were inspected again in the finishing shop. The results are shown below.
| Product | Inspected quantity | Cracked quantity | Crack rate / % | Main positions |
|---|---|---|---|---|
| Coupler yoke | 690 | 77 | 11.2 | Hook ear, body, neck, tail, gate/riser roots |
| Coupler knuckle | 500 | 15 | 3.0 | Inner wrist, riser/gate roots |
The cracks in the finishing shop were fewer than in the foundry shop. Many of them were recurrence cracks at positions that had been previously welded, indicating that the welding procedure and post-weld heat treatment must be strictly controlled. Some cracks were caused by quenching and tempering, especially at stress concentration points. Therefore, I recommended additional worker training, stricter heat treatment control, and improved grinding and gouging procedures.
In general, the batch production results showed that the optimized process effectively reduced the majority of sand foundry defects. The main residual problem was cracks, not only the original casting cracks but also cracks introduced during cutting, welding and heat treatment. Since E-grade steel has high hardenability and limited weldability, the post-casting process is as important as the casting process itself for controlling defects. The solutions I proposed were successfully applied in the production line, and the remaining cracks were within the range that could be repaired by standard grinding and welding procedures.
Major Achievements
This work produced several important achievements. First, I developed laboratory-optimized face sand and backing sand recipes that were suitable for railway steel casting production. The face sand recipe used 100% new sand, 8.0% to 8.5% bentonite, 0.3% to 0.5% starch additive, and 3.0% to 3.6% water. The backing sand recipe used 100% return sand, 0.2% to 0.3% bentonite, 0.1% to 0.3% starch additive, and 3.3% to 3.6% water. These recipes met the required sand properties and were validated in mass production.
Second, I used ProCAST simulation to identify the high-risk regions for sand foundry defects in the original coupler yoke and knuckle processes. The optimized gating and risering designs reduced the predicted erosion, gas entrapment, shrinkage porosity and hot cracking tendency. The actual trial casting confirmed the validity of the simulation results.
Third, I analyzed the key defects in trial and batch production, including gas holes, scab, sand inclusion, metal penetration, hot cracks, cold cracks, heat treatment cracks and welding cracks. I proposed concrete preventive measures that were subsequently applied in production. The final products passed all required inspections, including chemical composition, mechanical properties, hardness, magnetic particle inspection, dimensional inspection and section density inspection. The foundry obtained permission for mass production and successfully supplied qualified coupler yokes and knuckles.
Future Development Direction
The green moulding clay sand process will continue to be important for railway steel castings because of its economic and environmental advantages. However, the process still depends on careful control of sand properties and environmental conditions. In the future, more attention should be paid to the online measurement and closed-loop control of sand quality. The relationship between sand composition, sand temperature and sand performance should be further quantified. New automatic control systems can reduce human error and improve stability.
For the simulation part, the current model used fixed interface heat transfer coefficients. A more accurate model should include the effect of sand moisture, gas generation and the thermal degradation of bentonite. Multi-physics simulation coupling fluid flow, heat transfer, stress and microstructure evolution would provide more reliable predictions of sand foundry defects.
For post-casting processing, future work should establish a complete set of standardized procedures for cutting, grinding, welding, heat treatment and stress-relief treatment. Since E-grade steel has high cracking sensitivity during welding and heat treatment, better process specifications and improved worker training are necessary. This will further reduce the repair workload and increase production efficiency.
In conclusion, this study successfully demonstrated that railway steel castings can be produced economically with the green moulding clay sand process, provided that the sand recipe, casting design and post-processing procedures are properly controlled. The knowledge gained from this work can help foundries reduce sand foundry defects, reduce scrappage, save energy, and improve product reliability. The optimized process has been verified by mass production, and the results provide a valuable reference for the railway foundry industry.