This dissertation is grounded in my extensive research and practical engagement with the foundry industry, specifically focusing on the production of high-quality railway steel castings using the wet-type clay sand molding process. The imperative for this research stems from the dual pressures of industrial efficiency and rigorous quality standards. The railway sector, particularly for components like couplers and knuckles, demands exceptional reliability due to the severe service conditions involving high impact, fatigue, and wear. My work addresses the inherent challenges of the green sand process, a method celebrated for its low cost and high productivity but equally criticized for its difficulty in controlling casting quality, particularly for steel. The central goal is to establish a robust framework for minimizing ‘sand foundry defect’ occurrences through systematic experimentation, advanced numerical simulation, and meticulous analysis of production data.
In the following chapters, I delineate the research trajectory. Firstly, I detail the comprehensive experimental design and analysis to develop optimized facing sand and backing sand formulations. Secondly, I describe the application of the ProCAST simulation software to analyze the filling and solidification processes for the No.17 coupler yoke and No.16 coupler knuckle, leading to the proposal and verification of improved process designs. Thirdly, I investigate the critical casting defects encountered during trial and mass production, particularly focusing on the root cause analysis of cracks, and propose effective prevention strategies. The quantitative findings from this research have not only facilitated the successful mass production of these critical railway components but also provided a significant theoretical and practical reference for similar manufacturing contexts. The outcomes confirm that a meticulously controlled wet-type clay sand process can indeed yield steel castings of sufficient quality to meet the stringent demands of the railway industry, thus offering a compelling alternative to more expensive or environmentally burdensome processes.
1. Research Background and Engineering Significance
The rapid advancement of railway transportation, characterized by higher speeds and heavier axle loads, places unprecedented demands on the mechanical properties and service life of railway components. Among these, the coupler and its buffer system are paramount, as they transmit traction and impact forces between vehicles. The working environment of these components is extremely harsh, involving complex alternating stresses, making them susceptible to failures such as fatigue fracture, wear, and deformation. In China, strategic initiatives have prioritized the development of high-speed and heavy-haul railways, which necessitates a leap in the quality standards for all railway vehicle parts. The 17-type coupler yoke and 16-type coupler knuckle are exemplary of such critical load-bearing parts. Traditionally, methods like water glass sand and resin sand have been used for their production. However, my research focuses on the potential of the green moulding clay sand process, a technology with significant economic and environmental advantages, provided that its technological challenges can be effectively managed.
The primary technological hurdles for green sand steel casting include: high demand on raw material quality (e.g., high silica sand and premium bentonite), significant property variations due to environmental factors, and a high susceptibility to defects like blowholes, sand scabbing, and hot tears. The cost of failure is immense, not just in terms of scrap rate but also the potential for catastrophic in-service failures. Therefore, a meticulous, science-based approach is required to ensure that every step, from raw material selection to final heat treatment, is optimized to prevent ‘sand foundry defect’. The central hypothesis of my work is that by integrating advanced experimentation with computer-aided process design and rigorous defect analysis, the inherent disadvantages of wet-type clay sand can be mitigated, positioning it as a viable and superior process for railway steel castings.
This research was undertaken as part of a collaborative project with a major railway rolling stock manufacturer. The company, aiming to transform its production line to be more energy-efficient and environmentally sustainable, sought to replace its existing water glass sand process with the wet-type clay sand one. My specific mandate was threefold: (i) to develop and validate the optimal sand formulations and performance parameters; (ii) to employ computational simulation to optimize the casting process design, minimizing the risk of common ‘sand foundry defect’ types; and (iii) to analyze the defects arising during trial production and propose effective preventive solutions to ensure a stable and high-quality mass production process. The successful execution of these objectives forms the core of this dissertation, contributing not only to the specific enterprise but also to the broader knowledge base of green sand steel casting technology.
2. Foundational Theory of Wet-Type Clay Sand
Wet-type clay sand is a composite material consisting of silica sand, bentonite clay, starch-based additives, and water. Its functionality is governed by the physical and chemical interactions between these constituents. The high moisture content, a key identifier of the “green” state, imparts strength and plasticity to the mold but also introduces difficulties in controlling gas evolution and thermal expansion. To programmatically and effectively prevent ‘sand foundry defect’, a deep understanding of each component’s role is crucial.
Silica Sand: This is the refractory base of the mold. For casting steel, which is poured at temperatures exceeding 1500°C, the sand must possess high refractoriness. Therefore, the SiO₂ content must be exceptionally high (typically ≥94%), with strict control over impurities that lower the melting point. The grain size and distribution significantly influence permeability (the ability for gases to escape) and surface finish. For steel castings, a fine, uniform grain size (e.g., 50/100 or 70/140 mesh) is generally preferred to prevent severe mechanical sand penetration, a primary ‘sand foundry defect’.
Bentonite: This is the primary binder, giving the sand its green strength. It functions by forming a colloidal clay-water film on the sand grains, which creates bond bridges. Sodium-based bentonite, either natural or artificially modified, offers superior thermal stability and resistance to sand expansion defects, making it the preferred choice for steel castings. The quality of bentonite, determined by its montmorillonite content, directly influences the hot wet tensile strength, which is a critical indicator of a sand’s ability to resist scabbing and erosion during pouring.
Starch-Based Additives: These include materials like α-starch and dextrin. They enhance the toughness and plasticity of the mould, reducing the brittleness of the sand. During pouring, these organic additives burn out, creating micro-voids in the sand which provide space for the quartz sand to expand. This “sacrificial” behavior is instrumental in preventing scabbing defects, another common ‘sand foundry defect’ in steel castings. Furthermore, they improve the surface stability of the mold, preventing sand from being washed away by the turbulent flow of molten steel.
Water: Water content is the most critical and dynamically controlled variable. It activates the bentonite, enabling it to create the bond bridges. The moisture level must be precisely balanced; too little water leads to a weak, brittle mold that crumbles, while too much water reduces permeability and strength, leading to blowholes and gas-related ‘sand foundry defect’. The relationship between water content and sand properties is the core of sand control, often managed through the parameter of “compatibility” or “compactability.”
The performance of this mixture is evaluated using several standard tests, each providing insights into its suitability for casting. The table below summarizes the key properties and their significance in the context of steel casting and defect prevention.
| Property | Description | Relevance to ‘Sand Foundry Defect’ Prevention |
|---|---|---|
| Compactability (%) | A measure of the volume reduction of sand under a defined ramming effort. It’s a functional indicator of moisture content. | Indirectly controls strength and density. Low compactability means dry sand, prone to erosion; high values mean wet sand, leading to blowholes. |
| Green Compressive Strength (kPa) | The maximum compressive stress the sand can withstand in its green state. | A measure of mold strength, ensuring it can withstand the static and dynamic pressure of molten metal without deformation, preventing dimensional issues and mold wall movement. |
| Hot Wet Tensile Strength (kPa) | The tensile strength of a thin layer of sand measured under conditions resembling rapid heating by molten metal. | The single most important indicator of a sand’s resistance to scabbing. High values indicate the ability to withstand the spalling caused by steam condensation and quartz expansion. |
| Permeability (unit) | The ability of the sand to allow gases to pass through it. | Crucial for venting the steam and gases generated during pouring. Low permeability is a direct cause of blowholes and misruns, common ‘sand foundry defect’ types. |
| Shatter Index (%) | A measure of the sand’s toughness and plasticity. | High toughness prevents cracking, crumbling, and breakage of mold edges during handling, coring, and closure, thus preventing sand wash and inclusions. |
| Surface Stability Index (%) | The ability of the mold surface to resist erosion by flowing molten metal. | High surface stability minimizes the risk of sand erosion and the subsequent formation of sand inclusions (sand holes) within the casting. |
3. Experimental Development and Analysis of Molding Sand
My experimental methodology was designed to systematically determine the optimal formulation for both facing sand and backing sand, addressing separate functional requirements. Facing sand, which is in direct contact with the molten metal, must have high refractoriness and surface stability. Backing sand, which fills the rest of the flask, primarily provides structural integrity and bulk. The research flow began with a characterization of raw materials, followed by single-factor, orthogonal, and validation experiments.
3.1 Raw Material Characterization
The first step involved a thorough analysis of the incoming materials: new sand and recycled old sand. The new sand was identified as Fujian Jinjiang sand (40/70 mesh). The grain fineness and distribution of the sand were determined by sieve analysis, and the results showed a specific distribution which was crucial for later calculations. The recycled sand was analyzed for its fineness, contamination level, and residual binder content. I found that the old sand system had a mean grain size distribution different from the new, as shown in the accompanying table of results. Furthermore, I measured the mud content, moisture level, and active clay content of the reconditioned sand, which were crucial for determining what additions were necessary to restore its binding potential and prevent ‘sand foundry defect’. The findings were clear: with a moisture content of 0.7%, mud content of 6.8%, and active clay content of 5.05%, the old sand required careful reconstitution.
| Property | New Sand Value | Recycled Sand Value |
|---|---|---|
| Grain Size Range | Primary fraction at 40-70 mesh | Primary fraction at 50 mesh, minor fraction at 40 mesh |
| Moisture Content | <1% (typical) | 0.7% |
| Mud Content | <1% (specified) | 6.8% |
| Active Clay Content (by Methylene Blue) | — | 5.05% |
3.2 Facing Sand Experiments and Analysis
The optimization of facing sand was conducted in stages. The initial stage involved single-factor experiments to define the effect of each component on the final sand properties. I systematically varied the content of starch additives, bentonite, and water to understand their influence and to narrow down feasible ranges for a more intricate orthogonal experiment.
My experiments revealed that starch additives, specifically a mix of α-starch and dextrin, were essential for improving properties like toughness and surface stability. A critical finding was the optimal ratio, with a mixture of α-starch to dextrin in a 3:2 ratio at a total content of 0.5% produced arguably the best overall balance of properties, especially in terms of hot wet tensile strength. The bentonite content was varied between 7.0% and 8.5% to assess its impact on green compressive strength. As expected, increasing the bentonite content increased the strength, but an excessive amount led to a situation where water was absorbed by the clay, causing the sand to become dry and brittle, thereby increasing the risk of erosion. The water content experiments demonstrated a clear correlation: water increased compactability and toughness but decreased permeability and strength. These experiments were instrumental in establishing the factor levels for the subsequent orthogonal array, adhering to the factory’s target property windows.
The next stage employed a standard L9 (3³) orthogonal array to evaluate the combined effects and interactions of the three main factors: bentonite content (B), starch additive content (S), and water content (W). Each factor was set at three levels, based on the single-factor experiments. The table below shows the L9 orthogonal array and the experimental results for key ‘sand foundry defect’ resistance parameters.
| Exp. No. | Bentonite (%) | Additives (%) | Water (%) | Compactability (%) | Permeability | Green Compressive Strength (kPa) | Hot Wet Tensile Strength (kPa) | Shatter Index (%) | Surface Stability Index (%) |
|---|---|---|---|---|---|---|---|---|---|
| 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 | 3.44 | 70.1 | 96.9 |
| 9 | 8.5 | 0.5 | 3.0 | 44 | 460 | 103 | 4.62 | 66.3 | 97.0 |
The orthogonal analysis was then performed using the range (R) method to determine the primary influencing factor for each property. The range is the difference between the maximum and minimum average result for each factor level. A larger range indicates a more significant influence. Through this analysis, I identified the key control parameters for preventing each type of ‘sand foundry defect’. The results are summarized in the following table.
| Sand Property | Primary Influencing Factor | Secondary Influencing Factor | Physical Reasoning and Defect Link |
|---|---|---|---|
| Compactability | Water (Range = 13) | Additive (Range = 5) | Water is the main plasticizer. Too high compactability lowers permeability, increasing blowhole risk. |
| Permeability | Bentonite (Range = 14) | Water (Range = 6) | Bentonite fines and water both fill the voids between sand grains. Lower permeability increases gas-related defect risk. |
| Green Compressive Strength | Bentonite (Range = 16) | Bentonite (Range = 6) | Bentonite forms the binding bridges. Higher strength can prevent mold deformation and erosion, but too high can reduce collapsibility. |
| Hot Wet Tensile Strength | Additive (Range = 0.29) | Bentonite (Range = 0.12) | Additives burn out, providing space for sand expansion, which is critical for preventing scabbing. Bentonite improves the bond at high temperature. |
| Shatter Index | Water (Range = 3.6) | Bentonite (Range = 3.1) | Water and bentonite improve toughness; however, excess can make the mold too soft. Toughness prevents cracking and handling defects. |
| Surface Stability Index | Additive (Range = 0.7) | Bentonite (Range = 0.6) | Additives provide a cushioning effect against expansion and improve the cohesive strength of the surface layer, reducing erosion and sand wash defects. |
Based on the orthogonal analysis, experiments No. 5, 6, and 8 yielded the best overall balance of properties, making them ideal candidates as optimized formulations. To confirm these results and ensure their robustness, I performed a validation experiment for each of the selected formulations. The validation tests confirmed that the properties were indeed reproducible. The final recommended facing sand formulation was `100% Jinjiang sand + (8.0-8.5%) bentonite + (0.3-0.5%) starch additives + (3.0-3.6%) water`. This formulation was predicted to reliably achieve the target performance ranges shown in the table below.
| Property | Predicted Performance Range |
|---|---|
| Compactability (%) | 51 – 54 |
| Permeability | 440 – 450 |
| Green Compressive Strength (kPa) | 85 – 95 |
| Hot Wet Tensile Strength (kPa) | 4.1 – 4.5 |
| Shatter Index (%) | 70 – 72 |
| Surface Stability Index (%) | 97.0 – 97.5 |
3.3 Backing Sand Experiments and Analysis
For the backing sand, the aim is to provide strong support to the facing sand while remaining cost-effective. The base material is the reconditioned sand. I conducted a benchmark experiment using just the reconditioned sand and water, finding that while it could meet some minimum strength requirements, its performance was inconsistent due to variations in the recycled materials and a high potential for ‘sand foundry defect’. To increase the robustness and reduce the risk of defects, I introduced small additions of fresh bentonite and starch additives.
Following a similar methodology, I performed a dedicated L9 (3³) orthogonal experiment for the backing sand, varying the added water (3.0%, 3.3%, 3.6%), added bentonite (0.1%, 0.2%, 0.3%), and added additives (0.1%, 0.2%, 0.3%). The performance metrics were the same as for the facing sand. The analysis of the experimental data, presented in the succeeding table, revealed the critical influence of water in the backing sand. The recycled sand possesses a high quantity of fines (mud), which dramatically increases the surface area for water adsorption, thereby making the sand properties highly sensitive to water additions. Too little water results in a brittle, poorly bonded sand that will easily erode; too much water drastically reduces permeability, potentially causing blowholes. This is a classic manifestation of a ‘sand foundry defect’ originating from poor sand system control.
| Property | Primary Influencing Factor (Range Value) |
|---|---|
| Compactability | Water (13) |
| Permeability | Water (8) |
| Green Compressive Strength | Water (12) |
| Hot Wet Tensile Strength | Water (0.38) |
| Shatter Index | Water (3.6) |
| Surface Stability Index | Water (0.4) |
The optimized backing sand recipe was determined to be `100% reconditioned sand + (0.2-0.3%) bentonite + (0.1-0.3%) starch additives + (3.3-3.6%) water`. The resultant performance window, as shown in the table below, indicated a higher variability compared to the facing sand, which is acceptable given its role and reinforces a critical finding: strict moisture control is the linchpin for the stability of the recycled sand system to avoid ‘sand foundry defect’.
| Property | Predicted Performance Range |
|---|---|
| Compactability (%) | 55 – 60 |
| Permeability | 200 – 210 |
| Green Compressive Strength (kPa) | 90 – 95 |
| Hot Wet Tensile Strength (kPa) | 3.8 – 4.5 |
| Shatter Index (%) | 72 – 75 |
| Surface Stability Index (%) | 97.5 – 98.0 |
3.4 Production Verification of Sand Formulations
After establishing the laboratory-based optimized formulations, the next logical step was to verify their efficacy under actual production conditions. The company’s advanced sand processing system, equipped with an on-line molding sand control system, provided an excellent platform for this validation. The system automatically adjusted the water and additive additions based on real-time measurements of moisture, compactability, and green strength. During the mass production phase in August 2011, I collected and analyzed the sand formulations and properties used on a critical production day (ambient temperature 34°C, humidity 80%). The results, shown in the table below, demonstrated an excellent agreement with my laboratory recommendations, confirming the practical applicability and reliability of the experimental data. The successful production runs served as the ultimate validation that the laboratory-derived formulations were capable of producing high-quality castings without the common pitfalls of ‘sand foundry defect’. The difference between predicted and actual values was within acceptable operational tolerance, confirming that the experimental model accurately depicted the behavior of the system in an industrial setting.
| Sand Type | Actual Production Composition (% by weight) | Actual Production Test Properties |
|---|---|---|
| Facing Sand | New Sand: 100%, Bentonite: 8.0%, Additives: 0.3%, Water: 3.3% | Compactability: 55%, Permeability: 465, Green Strength: 102 kPa, Hot Wet Tensile: 4.2 kPa, Shatter Index: 74%, Surface Stability: 98%. |
| Backing Sand | Recycled Sand: 100%, Bentonite: 0.1%, Additives: 0.2%, Water: 3.5% | Compactability: 58%, Permeability: 225, Green Strength: 90 kPa, Hot Wet Tensile: 3.8 kPa, Shatter Index: 72%, Surface Stability: 97.7%. |
4. Process Simulation and Optimization of Railway Castings
With the sand system established, the focus shifted to the casting process design itself. I utilized the advanced ProCAST simulation software to perform a physics-based analysis of the initial (original) casting process for the two critical components. The fundamental governing equations for fluid flow, heat transfer, and solidification are solved to track the evolution of various fields during mold filling and solidification.
The simulation of fluid flow is based on the Navier-Stokes equations. For an incompressible Newtonian fluid, the mass and momentum conservation equations are:
$$ \frac{\partial u_i}{\partial x_i} = 0 $$
$$ \rho \frac{\partial u_i}{\partial t} + \rho u_j \frac{\partial u_i}{\partial x_j} = -\frac{\partial p}{\partial x_i} + \mu \nabla^2 u_i + \rho g_i $$
where \( \rho \) is the density, \( u_i \) is the velocity vector, \( p \) is pressure, \( \mu \) is dynamic viscosity, and \( g_i \) is the gravitational acceleration. The heat transfer is governed by the heat conduction equation, which, in its general form, is:
$$ \rho c_p \frac{\partial T}{\partial t} = \nabla \cdot (k \nabla T) + \rho L \frac{\partial f_s}{\partial t} $$
where \( c_p \) is the specific heat, \( T \) is temperature, \( k \) is thermal conductivity, \( L \) is latent heat of fusion, and \( f_s \) is the solid fraction. These equations are discretized and solved on a finite element mesh, providing a detailed description of the flow and solidification. This detailed analysis was pivotal in predicting not only the physical quantities like velocity and temperature but also in identifying the exact locations where a ‘sand foundry defect’ is likely to initiate.
4.1 Simulation and Optimization of the 17-Type Coupler Yoke
The 17-type coupler yoke is a complex, medium-sized casting measuring 975×574×496 mm and weighing 190 kg. Its complex inner cavity and varying wall thickness make it particularly prone to defects. The original process used a two-cavity mold with multiple conventional risers. My simulation of this original process, as shown in the results for velocity and temperature fields, aimed to predict defects. The results are summarized in the table below.
| Defect Type | Predicted Locations (Original Process) | Hypothesized Root Cause |
|---|---|---|
| Sand Erosion / Inclusions | Near the pin hole at the coupler tail, lower part of the head. | High velocity of metal flow directly impinging on the core (pin hole) and the mold surface. |
| Gas Porosity / Blowholes | Large flat surfaces on the coupler body. | Poor venting; gas entrapped by the advancing metal front cannot escape from the large horizontal surfaces. |
| Shrinkage Porosity | Hot spots at inner/outer corners of the head, the base of the traction bolt, and the junction of the head and body (neck). | Isolated liquid pools in thick sections; risers are unable to feed these regions effectively due to a narrow feeding path. |
| Hot Tears | Riser roots, the area around the coupler tail pin holes, head-body junction, and complex internal corners. | High thermal stress during cooling due to sharp section changes and constrained contraction. |
Based on these simulations, I proposed a series of targeted improvements. The primary goal was to eliminate the ‘sand foundry defect’ by altering the filling pattern and improving the solidification feeding. The modifications are detailed in the following table.
| Process Change | Purpose |
|---|---|
| Moved the ingate position to the side of the coupler body. | To reduce the impact velocity of the molten steel on the core and mold surfaces, thereby minimizing erosion and the formation of sand inclusions. |
| Replaced the standard sand risers with insulating risers on the coupler head. | To extend the time the metal remains molten, thereby increasing the feeding efficiency and improving the soundness of the thick head section. |
| Added a small blind riser near the ingate. | To provide a final feed path to the hot spot created near the ingate, minimizing shrinkage porosity. |
| Added an insulating riser on the coupler body. | To control the temperature gradient in the body section, promoting a directional solidification towards the riser and mitigating porosity. |
| Relocated the vent holes of the mold to be offset from the riser vents. | To prevent loose sand from the vent area from falling into the mold cavity, reducing sand inclusions. |
| Adjusted and added steel chills in the hot spot areas. | To accelerate the cooling of hot spots, promoting a more uniform and directional solidification pattern. |
The numerical simulation of the improved design showed significant progress. The velocity field during pouring was more uniform, mitigating the risk of erosion. The liquid fraction profiles confirmed that the insulating risers created a more effective feeding channel. The predicted shrinkage porosity was significantly reduced, with the remaining porosity concentrated in the heavier sections. The stress simulation indicated a considerable decrease in overall stress levels. These improvements were validated by actual trial productions. The castings produced using the improved process exhibited a dramatically reduced occurrence of all defect types, confirming the high fidelity of the simulation in predicting and preventing ‘sand foundry defect’.
4.2 Simulation and Optimization of the 16-Type Coupler Knuckle
The 16-type coupler knuckle, weighing 40 kg, is a smaller but equally critical component with a complex geometry and smooth surfaces. The original process was an eight-cavity mold with a central blind riser serving all castings. My simulation of this original process, following the same methodology, identified several issues, which are detailed below.
| Defect Type | Predicted Locations (Original Process) | Hypothesized Root Cause |
|---|---|---|
| Sand Erosion / Inclusions | Around the knuckle pin hole, specifically on the side facing the ingate. | High-velocity metal impinging directly on the core and mold surface opposite to the ingate. |
| Shrinkage Porosity | Heavy sections of the inner wrist face, the traction block, and the radius at the outer edge of the pin guard. | Thick sections isolated from the central riser; the feeding channel is too narrow, leading to premature solidification isolation. |
| Hot Tears | Corners of the traction block, the internal corners of the knuckle wrist, and the cast-on gate. | Stress concentration at sharp corners during solidification shrinkage. |
The proposed optimization was more radical than for the yoke. It involved changing the molding layout and the feeding system. The specific changes are described in the table below.
| Process Change | Purpose |
|---|---|
| Changed the mold from an eight-cavity layout to a four-cavity layout. | To simplify the gating system, allowing for more precise control over filling and feeding. |
| Relocated the ingate to the back side, between the pin hole and the knuckle head. | To prevent the high-velocity metal stream from directly impinging on the core that forms the pin hole, thus preventing erosion and subsequent sand inclusions. |
| Replaced the single, large central riser with individual insulated top risers on each casting. | To ensure that each casting is independently fed, especially the heavy inner wrist and traction block area, improving feeding efficiency and reducing shrinkage porosity. |
The simulation of the improved process confirmed that the flow of metal was more streamlined and less erosive. The individual risers created a directional solidification pattern, feeding the hot spots effectively. The shrinkage porosity, while not eliminated, was relocated to less critical areas. The outcomes of the simulation were subsequently used to guide the design and implementation of the new process. Experimental casting with the original process indeed revealed severe defects, including significant core erosion in the pin hole area on a majority of the castings, confirming the ‘sand foundry defect’ predictions. The optimized process, as predicted by the simulation, drastically improved the quality, proving the robustness of our design changes.
5. Analysis and Prevention of Casting Defects in Production
The production journey from trial to mass production is a process of constant learning and refinement. While the initial phases, guided by simulation, successfully addressed macro-level design issues like gating and feeding, the micro-level process variability and post-casting operations introduced new challenges. This section details my analysis of defects that appeared during the trial and early mass production stages, with a specific focus on the most common and problematic type—cracks.
5.1 Overview of Production Process
The production of couplers is a complex and multi-stage process. The foundry operations encompass sand preparation, core making, molding, melting, pouring, shakeout, and initial fettling. A crucial step is the heat treatment, which includes pre-normalizing, quenching and tempering (QT), and stress-relieving. The component is subjected to rigorous inspections, including magnetic particle testing (MT) for surface defects and ultrasonic testing (UT) for internal discontinuities. Cracks, a severe type of ‘sand foundry defect’, were frequently observed, particularly in the trial production phase, even though the parts eventually passed quality checks after weld repair. The manufacturing process was closely monitored to trace these defects back to their roots.
5.2 Classification and Root Cause Analysis of Cracks
During the trial production of the 17-type coupler yoke and 16-type coupler knuckle, I systematically investigated the defects. The predominant defect was cracks. Based on their macroscopic and microscopic characteristics, I categorized them as follows:
- Hot Tears: These are characterized by their jagged, irregular, and branched appearance. The fracture surface is typically oxidized (dark, black or blue) due to its formation at high temperatures. My analysis showed they were predominantly observed at locations with sharp corners and significant changes in section thickness, such as the riser roots, the junction between the coupler head and body, and complex internal corners of the head. The root cause is the inability of the partially solidified and weak metal to withstand the tensile stresses induced by the contraction of the cooler, solidified outer shell. Factors contributing to this include:
- Insufficient fillet radii at the connections between the risers/gates and the casting.
- Excessive thermal gradients due to rapid cooling in some areas, exacerbated by the high thermal conductivity of green sand.
- Early shakeout, which exposes the still-hot and weak casting to ambient conditions, leading to stress build-up.
- High pouring temperatures, leading to a larger solidification shrinkage and a wide solidification range.
- Cold Cracks: These appear as straight, clean fracture lines, often transcrystalline, with a bright or slightly oxidized surface. They were typically found at the riser or gate roots after cutting. My investigation attributed these to the high residual stresses within the casting, combined with the thermal stresses and mechanical forces applied during post-casting operations like riser cutting (flame cutting) or abrasive blasting. The absence of proper stress-relief treatment before these operations is a key contributing factor.
- Welding Cracks: These were encountered during the weld repair of the hot tears. These are primarily hydrogen-induced cold cracks, arising due to the high hardenability of the E-grade steel, the presence of hydrogen from the welding consumables, and a high degree of restraint in the thick sections being repaired. Failure to preheat and post-heat the weld area correctly was identified as the primary cause.
5.3 Preventive Measures and Solutions
To effectively mitigate these cracking phenomena and other ‘sand foundry defect’ issues, I implemented a multi-faceted strategy targeting process design, operational control, and post-casting treatments. The following table summarizes the solutions mapped to the specific causes.
| Defect Type | Preventive / Corrective Measure |
|---|---|
| Hot Tears | – Optimize casting design by increasing fillet radii at riser/casting junctions. – Employ anti-cracking ribs at critical locations to absorb tensile stresses. – Use insulating materials or coatings on the mold surface around connections to moderate the cooling rate and minimize thermal gradients. – Control pouring temperature within a specified optimal window. – Strictly control the content of low-melting-point elements (S, P) and impurities in the steel to improve its high-temperature ductility. – Delay shakeout time to allow for a more uniform and gradual cool-down of the casting. |
| Cold Cracks | – Avoid premature shakeout and handling of castings before they have cooled below a safe temperature. – Implement a standardized procedure for riser removal. This includes preheating the riser root with a controlled thermal method or performing a step-cutting process, rather than a single, aggressive cut. – Ensure that the casting is placed in a warm, insulated location to cool slowly and uniformly after riser cutting. – Revisit the mold design to ensure uniform cooling, possibly through the strategic placement of chills to achieve a more uniform temperature distribution. |
| Welding Cracks | – Mandate a strict preheating regime before any weld repair. The preheated temperature must be high enough to ensure slow cooling of the weld zone. – Use low-hydrogen electrodes and ensure they are properly baked before use to minimize the hydrogen content in the weld metal. – Implement a post-weld heat treatment, such as immediate stress-relieving or slow cooling under insulating blankets, to allow any diffused hydrogen to escape and to temper the hardened heat-affected zone. |
6. Mass Production Verification and Quality Control
After refining the casting processes and establishing robust procedures for handling and repairing defects, the company commenced mass production in August 2011. The subsequent analysis of the products over a sustained period was crucial to confirm that the implemented changes had a lasting impact on quality and that ‘sand foundry defect’ rates were controlled.
Initial Inspection: The initial visual inspection after shakeout aimed to identify gross defects. Over a half-month period, a total of 154 yokes and 223 knuckles were inspected. The results, shown in the table, show that the output was relatively clean, with no obvious visual surface defects. The main cause of rejection was “sand inclusion/penetration”, a severe ‘sand foundry defect’ occurring in specific areas due to operational issues rather than systemic problems. This finding points towards the necessity for improved operator training and adherence to standard operating procedures.
| Component | Total Inspected | Rejected | Rejection Rate (%) | Primary Cause of Rejection |
|---|---|---|---|---|
| Coupler Yoke | 154 | 8 | 5.2 | Severe sand inclusion/penetration |
| Coupler Knuckle | 223 | 6 | 2.7 | Severe sand inclusion/penetration |
Ultrasonic and Magnetic Particle Testing: The final quality was assessed by the rigorous non-destructive testing (NDT) procedures. All castings undergo a full MT and UT examination. The initial inspection after the pre-normalizing and fettling steps showed that cracks were still present. A statistical analysis of the inspection reports, as presented in the tables, revealed that crack defects were concentrated at the riser and gate removal sites. Specifically, 63% of cracks on the yoke were found in these areas. This high concentration strongly suggested that the “cold” and “hot” cracking mechanisms, induced by the stress of cutting and insufficient pre/post heating, were the primary culprits. The reinforcing the hypothesis that post-casting operations are a major source of ‘sand foundry defect’ in steel foundries. The same logic applied to the knuckles, where 72% of the cracks were found on the “S” (inner wrist) face, which is a heavily stressed area receiving complex loading, making it sensitive to stress-related cracking after machining.
| Component | Date | Inspected Quantity | Location: “C” Face / “C1” Face / “C2” Face | Location: Knuckle Neck | Location: Coupler Tail | Location: Others (Riser/Gate) | Total Cracked Castings |
|---|---|---|---|---|---|---|---|
| Yoke | 11.10 | 4 | 1 / 1 / 0 | 0 | 0 | 9 | 11 |
| Yoke | 11.11 | 7 | 1 / 3 / 0 | 0 | 2 | 18 | 24 |
| Yoke | 11.12 | 18 | 2 / 12 / 1 | 6 | 8 | 40 | 69 |
| Yoke | 11.14 | 21 | 1 / 15 / 0 | 3 | 11 | 45 | 75 |
The inspection data indicated that while the design of the casting and the gating system was effective in preventing primary casting defects, the secondary operations were introducing defects, primarily cracks. This necessitated a dual approach. The first was to refine the casting process further (e.g., by optimizing the controlled cooling after shakeout) to minimize residual stress. The second was to enhance the strictness and quality of the post-casting procedures, providing comprehensive training to all operators on the correct preheating, welding, and stress-relieving protocols. The problem was not so much the initial casting but its post-processing. This led to a significant refinement of the operational procedures. Even with rigorous training and adherence to a strict process to prevent ‘sand foundry defect’, the intrinsic hardenability of the E-grade steel and the high restraint in the thick sections made crack formation during welding a persistent, though manageable, challenge. The final inspection after the post-weld heat treatment showed that the number of cracks was vastly reduced, and those that still appeared were within the limits allowed for weld repair, ensuring all shipped parts met the required quality standards.
| Component | Inspected Quantity | Location: Hook Ear / Hook Body / Hook Neck / Hook Tail / Other | Total Cracked Castings | % of Inspected |
|---|---|---|---|---|
| Yoke | 690 | 16 / 26 / 7 / 24 / 4 | 77 | 11.2% |
| Knuckle | 500 | Location: S-face 14, Others 1 | 15 | 3.0% |
The data from the mass production phase clearly demonstrates that the overall quality of the mass-produced castings was significantly better than that in the trial phase, indicating that the systemic improvements to the process were effective. The defect rates, especially for cracking, observed in the final inspection, were reduced by almost an order of magnitude compared to the initial trial phases. This confirms that a rigorous scientific approach to the design and control of the green sand system, combined with a disciplined execution of process parameters, is absolutely vital to mitigate and control the numerous forms of ‘sand foundry defect’. The findings from the quantitative analysis of the production data allow for continuous improvement, with the aim of further reducing if not eliminating the welding and repair load.
7. Conclusions and Future Directions
My comprehensive research, from the development of the sand system to the final quality control of the cast components, has yielded several fundamental conclusions that contribute to the knowledge and practical application of wet-type clay sand for railway steel castings:
- Optimized Sand System: Through a rigorous experimental program, I successfully developed and validated both facing and backing sand formulations. The optimized facing sand, composed of `100% new sand + 8.0-8.5% bentonite + 0.3-0.5% starch additives + 3.0-3.6% water`, and the backing sand, composed of `100% reconditioned sand + 0.2-0.3% bentonite + 0.1-0.3% starch additives + 3.3-3.6% water`, met all factory performance requirements and proved effective in large-scale production. The analysis clearly established that while facing sand properties are governed by all three components, the backing sand’s performance is overwhelmingly dependent on water content, which must be strictly controlled to prevent ‘sand foundry defect’.
- Simulation-Driven Design: The application of ProCAST proved invaluable. The simulations accurately predicted the location of defects in the original process designs. The optimized processes, which were informed by the simulation results, showed a dramatic reduction in all defect types. This validates the use of numerical simulation as a robust tool for proactive ‘sand foundry defect’ prevention. The optimized yoke process, featuring a side ingate and multiple insulating risers, resulted in improved soundness and a significant reduction in erosion and cracking. The optimized knuckle process, featuring an altered ingate location and individual risers, effectively mitigated shrinkage and sand wash defects.
- Effective Defect Management: The key challenge in mass production was the occurrence of cracks. My detailed analysis, linking crack characteristics to their process origins, allowed for the implementation of targeted corrective actions. The findings revealed that while casting process design is paramount in preventing primary defects, a holistic approach must also include stringent control over secondary processes such as riser removal, handling, and welding. In this context, controlling ‘sand foundry defect’ became a matter of controlling thermal gradients and residual stresses throughout the entire production chain. The improved procedures successfully reduced scrap rates and rework, ensuring that the final products met all higher-level quality checks. The by-product is a more robust and predictable production line for railway components.
Looking ahead, this research opens up several promising avenues for future work. There is a clear opportunity to refine the molding sand control system further. The integration of online sensors with a more comprehensive machine-learning algorithm for type of “closed-loop” control presents a major opportunity to achieve a superior consistency of molding sand properties. The technology can be adapted to predict and automatically compensate for variations in sand temperature and ambient humidity. The comprehensive dataset generated from this project can be used to create a “digital twin” of the molding sand process, enabling highly predictive and precise corrections to prevent the initiation of ‘sand foundry defect’ and minimize defects on every single casting. Moreover, the post-casting operations, particularly welding and heat treatment, were identified as significant sources of cracking. There is scope to develop more sophisticated, automated procedures for riser cutting and welding, using robots equipped with thermal imaging to control the heating and cooling cycles, thereby eliminating the variability introduced by manual operations. To that end, exploring alternate cutting techniques such as plasma or water jet cutting, which may impart less heat stress than flame cutting, deserves further investigation.
The future of green sand casting hinges on reducing its intrinsic variability, and this research provides a clear path forward by demonstrating how to leverage fundamental principles and modern simulation tools to turn a potentially problematic process into a robust platform for manufacturing high-integrity components.