1. Introduction and Research Background
As one of the most critical components in railway systems, the crossing (also known as froggy) serves the essential function of guiding train wheels from one track line to another. The quality of the crossing directly reflects the level of railway track construction and significantly influences the overall transportation efficiency. With the continuous development of China’s high manganese steel crossing manufacturing technology, the demand for monoblock casting crossings with superior performance has been steadily increasing, particularly for specialized lines and station tracks.
The first successful trial production of a monoblock casting crossing of high manganese steel occurred in the early 1960s. Since then, remarkable progress has been achieved in both structural design and manufacturing technology. According to statistical data, high manganese steel crossings account for more than 50% of all crossings in China, while in cold regions such as Northern Europe and the United States, these crossings are exclusively manufactured from high manganese steel due to their exceptional performance characteristics.

High manganese steel, also known as Hadfield steel, possesses outstanding plastic toughness and exceptional work-hardening capability. When subjected to high impact loads and contact stresses, its surface hardness can increase by 2.5 to 3.5 times, achieving excellent wear resistance. However, the casting performance of high manganese steel presents significant challenges due to its large solidification shrinkage rate (ε of up to 8.5%) and poor heat dissipation characteristics. These unfavorable properties frequently lead to casting defects such as coarse grains, shrinkage porosity, shrinkage cavities, and inclusions, which severely compromise the casting quality of the crossing.
The crossing is among the most frequently replaced components in railway systems, requiring tens of thousands of replacements annually along China’s railway network. In the context of continuously increasing train speeds and loads, the demand for higher quality crossings has become increasingly urgent. Since casting forms the first step in the production of high manganese steel crossings, reducing casting defects during this process has become paramount for improving overall crossing quality and service life. Through extensive research, it has been established that the failure of high manganese steel crossings is primarily attributed not to material fatigue but rather to casting defects introduced during the manufacturing process. The predominant failure modes include spalling, impact wear, crushing, and horizontal cracking of the heart rail and wing rail, accounting for over 50% of all failure cases.
2. Literature Review and Research Status
2.1 Crossing Materials and Casting Processes
Since the first application of high manganese steel for railway crossings in 1894, extensive research has been conducted worldwide on the material properties and performance optimization. The standard composition of the austenitic high manganese steel used for crossings, as proposed by the British engineer Hadfield, contains 1.0% to 1.4% carbon and 10% to 14% manganese (typically not less than 13%). Table 1 presents the typical chemical composition of high manganese steel used for crossings in this research.
| Main component | C | Mn | Si | P | S |
|---|---|---|---|---|---|
| Mass fraction (%) | 1.1 | 13 | 0.53 | 0.033 | 0.008 |
To further enhance the casting performance and extend the service life of crossings, many scholars have conducted in-depth theoretical research on the re-alloying treatment of high manganese steel, with several achievements successfully applied in practical production. French researchers developed a Mo-containing high manganese steel crossing that does not require heat treatment after casting. By reducing carbon content and adding elements such as molybdenum, this material demonstrates superior work-hardening capability compared to traditional high manganese steel. American researchers invented a novel high manganese steel with further reduced carbon content. The resulting material achieves a yield strength exceeding 525 MPa with approximately 30% elongation, and crossings produced from this material exhibit substantially extended service life.
Regarding the casting process for high manganese steel crossings, these components belong to the category of ultra-long castings. For sand casting foundry of such components, inclined gravity casting is typically employed to ensure complete mold filling. The casting angle significantly influences the flow field stability and casting quality. Research comparing top-pouring and bottom-pouring methods has demonstrated that the top-pouring approach (pouring with the sprue end elevated) yields superior overall performance compared to bottom-pouring.
2.2 Casting Defects in High Manganese Steel Crossings
The primary casting defects encountered in high manganese steel crossings include:
- Shrinkage porosity – Regions of insufficient density in the final solidified areas where no adequate liquid metal supply exists for feeding. This manifests as reduced load-bearing capability within the metal structure.
- Sand inclusion defects – Caused by mold sand entering the metal liquid or by slag carried into the casting. These defects present as impurities mixed within the metal.
- Shrinkage cavities – Macroscopic voids generated during the solidification process due to liquid contraction. These create stress concentration points that can lead to premature crossing failure.
After the casting model is established, the primary factors controlling defects such as shrinkage cavities and porosity include the dimensions and type of risers, the positioning of chills, and the pouring methodology. Recent studies on high manganese steel crossing casting defects have focused on various aspects. Some researchers have proposed adding exothermic compounds after pouring to elevate the temperature of the riser liquid metal, thereby improving feeding capability. Research has also been conducted on replacing insulating risers with exothermic-riser sleeves to compensate for the high solidification shrinkage of high manganese steel. Analytical approaches calculating riser modulus and feeding distance based on the material’s shrinkage rate and casting weight have also been developed. Surface defect scanning and EDS analysis of multiple high manganese steel crossings revealed that internal micropores and porosity are primarily associated with liquid iron oxidation and mold-related factors, particularly under V-method molding conditions.
2.3 Numerical Simulation of Casting Processes
Traditional foundry practice relies heavily on empirical guidelines provided in casting handbooks. However, the numerous variables in high manganese steel crossing production makes precise control challenging for such complex products. With the rapid advancement of computer-aided engineering (CAE) technology, numerical simulation has emerged as a powerful tool for pre-simulating production processes and analyzing product quality. The finite element method (FEM) approximates physical systems through discrete interconnected subdomains, with approximate solutions derived for each element to satisfy the overall system conditions.
Among the various casting simulation software packages available (including MAGMA, FtStar, and HuaZhu CAE), ProCAST was selected for this research. Developed by Universal Energy System, ProCAST offers robust finite element solvers suitable for various casting methods including sand casting, investment casting, gravity casting, high-pressure casting, and tilt pouring. The software comprises eight modules: finite element mesh generation, heat transfer analysis, flow analysis, stress analysis, thermal radiation analysis, microstructure analysis, electromagnetic induction, and inverse-solving modules.
3. Research Model and Fundamental Assumptions
3.1 Geometric Model Development
The research object of this study is the casting 75-852 high manganese steel crossing. Using SolidWorks, a complete three-dimensional geometric model of the crossing was established, as shown in Figure 1. The casting length of this crossing measures 4100 mm with a total sand mold length of 7100 mm.
For the riser design, a modulus RM of 4.33 cm was adopted. The riser dimensions were established using the perimeter quotient method proposed by Zhou. The governing equation for riser sizing is:
$$3 R^2 C M(1-\varepsilon) – \varepsilon f^2 Q = 0$$
where CM is the modulus of the casting section at the feeding location (cm), RM is the riser modulus (cm), QC is the perimeter quotient of the section to be fed, QR is the riser perimeter quotient, VC is the feeding volume (cm³), VR is the riser volume (cm³), ε is the metal volumetric shrinkage rate (%), and f is the modulus expansion coefficient defined as f = RM/CM.
| ε (%) | RM (cm) | UD (cm) | N (cm) | D0 (cm) | d0 (cm) | H (cm) | h (cm) |
|---|---|---|---|---|---|---|---|
| 8.5 | 4.33 | 132 | 88 | 222 | 175 | 222 | 90 |
Figure 2 shows the complete numerical simulation casting model including the monoblock crossing, risers, gating system, inclined sand mold, and external chills. A mesh size of 12 mm was established for the casting, while the sand mold mesh was set to 100 mm to optimize computational efficiency. For the critical heart rail and wing rail regions, localized mesh refinement was implemented with a 5 mm surface mesh size. The refined model contains 3,657,221 volume elements, with over 1.5 million elements concentrated in the heart rail and wing rail regions of highest interest.
3.2 Fundamental Assumptions for High Manganese Steel Crossing Casting
Assumption based on Niyama criterion: This research employs the classic Niyama shrinkage criterion for defect prediction. The Niyama value quantifies the tendency for micro-shrinkage formation based on the relationship between temperature gradient and cooling rate:
$$N_y = \frac{G}{\sqrt{R}}$$
where G is the temperature gradient (K/cm) and R is the cooling rate (K/min). Regions where the Niyama value Ny is less than 0.85 (K·s)1/2/mm² are considered to have significant defect formation tendency. The material properties of high manganese steel are treated as nonlinear functions of temperature. The liquidus temperature is calculated to be 1392°C and the solidus temperature is 1019°C, with a casting temperature of 1450°C.
Assumption of casting velocity attenuation: In actual production, the high manganese steel crossing employs inclined casting with the pouring end elevated at an angle θ = 2.85°. This arrangement results in substantial metal liquid overflow from the riser tops. In the actual process, a constant pouring velocity is maintained, with molten metal sequentially overflowing from the riser tops, causing an actual reduction in the metal flow rate within the mold cavity. To address the complexity of simulating this overflow phenomenon, this research innovatively proposes an equivalent pressure modeling approach with casting velocity attenuation.
The simulation establishes a geometric model where metal cannot overflow (only air can pass through riser tops), which maintains pressure consistency with actual conditions. The actual production pour rate is v₀ = 50 kg/s. Through multiple simulation calibrations, the metal overflowing from riser tops accounts for 21.1% of the total pouring volume. Therefore, negative velocity increments are introduced to compensate for the ignored overflow in the numerical model. The casting velocity attenuation curve is described by:
$$F(t) = a \cdot e^{b(x-15)} + c$$
with the constraints:
$$\int_0^{22.56} F(t)\, dt = S$$
where S represents the metal mass function, and the initial conditions are:
$$u_i = F(t_i)$$
$$v_0 = F(t_0)$$
Using MATLAB to solve the nonlinear system, the final attenuation curve for the first riser is:
$$u(t) = 1 – 2.07e^{-0.12(t-t_0)} + 52.07$$
At time t₁ when the metal liquid reaches the second riser, a second-stage overflow occurs. Since the pouring rate has already attenuated, the overflow value also attenuates. Based on multiple simulation tests, an attenuation coefficient f = 0.98 was determined:
$$u_2(t) = 0.98(1 – 2.07e^{-0.12(t-t_1)} + 52.07)$$
By iterating this process through eight risers, a complete eight-stage overflow curve is constructed, driving the numerical simulation casting process with high fidelity to actual production conditions.
4. Numerical Simulation Casting Technology and Defect Analysis
4.1 Visualization of the Sand Casting Foundry Process
Using the casting velocity attenuation curve as input and pouring temperature of 1450°C, the mold filling is completed in 22.56 seconds. For a comprehensive study of casting defect controlling factors, this research divides the casting process in five mutually interdependent stages for individual analysis:
Stage 1: Mold Filling
During this phase, the metal liquid enters through the casting end with velocity controlled by the predetermined decay curve. The risers are sequentially ignited and maintained in a continuous heating state. Due to the coupled flow-heat transfer simulation, metal liquid begins solidifying on the external chill surfaces even before mold filling is complete, forming a thin layer of fine equiaxed crystals.
Stage 2: Exothermic Reaction
This refers specifically to the exothermic-riser sleeves containing thermite, flux, oxidizing agents, and insulating materials. The risers are rapidly ignited when the hot metal enters, maintaining continuous exothermic reaction that significantly reduces the cooling rate of internal liquid metal, thereby extending the available feeding time.
Stage 3: Solidification
Following mold filling, solidification commences at the chill ends with highest cooling rates. The solidification front progresses inward while the riser metal remains at elevated temperatures with adequate fluidity. The strategic interposition of chills and risers ensures that the critical heart rail and wing rail sections solidify first with refined microstructures.
Stage 4: Feeding
As solidification proceeds, volumetric shrinkage creates pressure gradients that drive liquid metal flow from high-temperature regions (risers) toward solidification fronts. Riser metal with sufficient fluidity flows downward to compensate for the volume reduction. Regions beyond the effective feeding distance develop shrinkage porosity and cavities.
Stage 5: Defect Formation
During the mid-to-late stages of solidification, riser metal levels decrease as feeding progresses. After the exothermic reaction completes, solidification proceeds from the riser top and sleeve walls toward the center. Defects concentrate primarily in the central riser regions and in areas beyond effective feeding range.
4.2 Temperature Field and Casting Defect Analysis
To establish the relationship between temperature field and casting defect formation, the research conducted detailed thermal analysis at key locations. Temperature data from the chills show rapid initial temperature drop at the liquid metal-chill interface, approaching solidus temperature almost immediately. Interestingly, a temperature rise phenomenon occurs between 1000s and 2000s of cooling, suggesting partial remelting phenomena. The internal core region of the crossing, after mold filling, exhibits rapid cooling to near the liquidus temperature followed by slower cooling through the freezing range. The riser metal, by contrast, maintains elevated temperatures throughout the solidification period due to continuous exothermic reaction, preserving fluidity for effective feeding. The temperature curves of the exothermic-riser sleeve exhibit a convex shape, indicating the heat contribution from the combustion reaction. This thermal analysis confirms that exothermic-risers can effectively slow internal cooling, increase feeding distance, and direct shrinkage defects toward the riser in a “reverse ripple” pattern consistent with the temperature field.
The numerical simulation defect predictions correspond well with the temperature field analysis. Each riser creates a “V”-shaped feeding zone below it, with the range proportional to the spacing between adjacent risers. Between two adjacent risers, an arc-shaped un-feedable region forms, extending proportionally with increased inter-riser spacing. The edge risers exhibit severe inward contraction tendencies of their “V”-shaped feeding zones. The outer arc regions extend to the riser lower portions and concentrate severe defects.
4.3 Research on Casting Defects with Different Exothermic Riser Parameters
The key parameters distinguishing exothermic risers from conventional insulating risers are the heat generation capacity and exothermic duration. To investigate these effects systematically, a controlled variable approach was employed. The baseline comparison used insulating risers (zero exothermic capacity), which produced shrinkage porosity of 23.10% and shrinkage cavity of 1.002%.
Effect of exothermic duration: With constant heat generation of 1200 kJ/kg, exothermic durations of 120s, 240s, and 360s were tested. The results showed surprisingly minimal differences in casting defect distribution:
| Exothermic duration (s) | Shrinkage porosity ratio | Shrinkage cavity ratio |
|---|---|---|
| 120 | 17.01% | 0.530% |
| 240 | 16.51% | 0.514% |
| 360 | 16.17% | 0.504% |
This indicates that for large sand casting foundry components with extended solidification times (exceeding 6000s), variations in exothermic duration within the tested range have negligible impact on defect formation. This finding suggests a practical advantage in manufacturing flexibility.
Effect of heat generation capacity: With constant exothermic duration of 240s, varying heat capacities produced significant differences:
| Heat generation (kJ/kg) | Shrinkage porosity ratio | Shrinkage cavity ratio |
|---|---|---|
| 0 (insulating riser) | 23.10% | 1.002% |
| 400 | 19.69% | 0.670% |
| 1200 | 16.51% | 0.514% |
| 2000 | 15.84% | 0.499% |
The comparison between insulating risers and exothermic-risers clearly demonstrates the superior feeding efficiency of the latter. The exothermic-riser maintains internal metal temperature longer, extending feeding duration. However, the results also reveal a limitation: even significant increases in heat generation cannot completely eliminate defects in regions beyond the effective feeding distance. The bolt hole regions, in particular, remain problematic due to their location relative to the riser feeding zones and the insufficient fluidity of the metal at their solidification time.
Based on these findings, the 1200 kJ/kg exothermic-riser provides adequate feeding capability for this crossing. Higher heat generation shows minimal improvement while potentially causing resource waste and increasing thermal stress risks. The exothermic duration is a secondary parameter whose influence is much smaller than the total heat output for large castings.
5. Process Optimization for High Manganese Steel Crossing
5.1 Riser Addition Optimization Scheme
In the casting layout of the studied crossing, there exists an 800 mm transition zone between the pouring-end open riser and the nearest closed riser. This region lacked any feeding module in the original design, resulting in a large arc-shaped un-feedable region with defect porosity as high as 36.17%. Since increasing riser heat generation proved ineffective for this region, a new riser was proposed based on the “V”-shaped feeding zone prediction method developed in this research.
The implementation of this process optimization yielded substantial improvements:
| Defect type | Before optimization | After optimization | Reduction |
|---|---|---|---|
| Local shrinkage porosity ratio | 36.17% | 22.11% | 14.06% |
| Local shrinkage cavity ratio | 1.13% | 0.69% | 0.44% |
This optimization also revealed an interesting phenomenon: the feeding distance and defect regions matched the predicted “V”-shaped feeding zones with high accuracy, validating the prediction methodology. Additionally, the bolt hole defect regions exhibited upward migration tendencies, with defects concentrating above the bolt holes, while edge riser defects accumulated toward the crossing bottom. This upward defect migration signifies improved metal density in the train contact areas, enhancing crossing quality.
5.2 Innovative Integrated Forging-Casting Production Method
Targeting the most vulnerable heart rail and wing rail regions, this research innovatively proposes an integrated forging-casting crossing production method. Traditional approaches include monoblock casting (high manganese steel), assembly-type combined crossings (alloy steel), and integral forging crossings. The integral forging approach provides superior mechanical properties but suffers from complexity and high costs due to the large crossing dimensions. The proposed technique combines forging and casting advantages while mitigating their respective drawbacks.
The production method comprises the following steps. First, steel ingots are forged with sufficient reduction to enhance mechanical properties and toughness. Second, the forged steel ingots are machined to the shape of the critical heart rail and wing rail regions. Connection grooves with 30 mm diameter are machined on the surface contacting the molten metal to enhance bonding between forged pieces and cast metal. Third, the machined heart and wing rails are positioned within the sand mold, serving as internal chills, and the entire assembly is cast in a single operation. This method fundamentally addresses the difficulties of integral forging, reduces labor and costs, strengthens the vulnerable regions, and extends crossing service life.
To evaluate theoretical feasibility, a comparative study was conducted against external chill results. The external chills showed rapid heating at the contact surface with metal liquid temperature rising above the solidus line after pouring, with subsequent continued heating and remelting of surface layers. This confirms that the pouring metal provides sufficient heat for the chill to reach remelting temperatures.
Numerical simulation of the experimental method revealed that when employing the integrated forging piece without preheating, the surface temperature at the front and rear edges of the forging rises rapidly during pouring. The side surfaces near edges exhibit sustained temperature increases, while regions directly below riser positions reach sufficiently high temperatures for partial remelting. However, the substantial thermal mass of monolithic forging pieces prevents complete melting of the critical contact interfaces. Preheating the forging pieces to 800°C significantly improves the temperature profiles, with surface temperatures exceeding 1200°C and sustained above the solidus for extended periods, enabling proper metallurgical bonding between the forge and cast components.
5.3 Separated Forge Pieces Optimization Scheme
To more effectively address the issue of localized thin casting sections not providing sufficient heat for forging surface melting, this research proposed a separated forge pieces optimization scheme. This approach divides the integrated forging into three separate components: one heart rail and two wing rails, with the height reduced. The theoretical basis stems from the fact that smaller separated pieces present greater surface area relative to volume, allowing more thorough heating during casting.
Numerical simulation of the separated scheme without preheating shows significant improvements. The forging left edge near the pouring end achieves temperatures exceeding 1200°C with locally sustained periods above the solidus. The upper surface regions directly beneath risers rapidly heat to approximately 1000°C followed by continued gradual increases. This improved thermal profile enables more effective remelting and metallurgical bonding. Preheating to 400-800°C yields the best results, with preheating to 400°C offering optimal performance by achieving superior bonding while minimizing damage to the mechanical properties of the forging with reduced oxidation risks.
The separated forging scheme provides distinct advantages over monolithic forgings. The reduced thickness allows complete heat penetration, and the increased surface area-to-volume ratio improves thermal contact with the molten metal. The resulting bond between forged insert and cast structure is significantly superior, providing enhanced structural integrity of the critical heart rail and wing rail regions.
6. CAFE Model Simulation of Macrostructure Growth
6.1 CAFE Model Fundamentals
The cellular automaton (CA) model is a grid-based dynamical system with discrete time, space, and state variables, characterized by spatial interactions and temporal causality. The CAFE method couples finite element (FE) methods with cellular automaton probabilistic models to establish a macro-micro model. This approach is capable of simulating the formation of surface fine-grain regions, competitive growth between equiaxed and columnar dendrites, columnar zone formation, the relationship between grain orientation and thermal gradients, and columnar-to-equiaxed transition phenomena. The CAFE model employs a Gaussian distribution-based continuous nucleation model:
$$\frac{dn}{d(\Delta T)} = \frac{n_{max}}{\sqrt{2\pi} \cdot \Delta T_\sigma} \exp\left[-\frac{1}{2}\left(\frac{\Delta T – \Delta T_n}{\Delta T_\sigma}\right)^2\right]$$
where nmax is the maximum nucleation density, ΔTσ is the standard deviation of nucleation undercooling, ΔTn is the average nucleation undercooling, and ΔT is the nucleation undercooling. The grain density can be calculated from the integral:
$$n(\Delta T) = \int_0^{\Delta T} \frac{dn}{d(\Delta T’)}\, d(\Delta T’)$$
The growth kinetics are described using the KGT model, where the dendrite tip growth velocity is expressed as:
$$v = \alpha_2 \Delta T^2 + \alpha_3 \Delta T^3$$
where α₂ and α₃ are the fitting parameters representing the relationship between dendrite tip growth velocity and nucleation undercooling. The finite element and CA cell coupling employs linear interpolation factors, allowing temperature assignment from FE nodes to CA cells. The enthalpy changes are also distributed to account for latent heat evolution:
$$T_w^t = \sum_n P_{wn} \cdot T_n^t$$
$$\delta H_w = \sum_n P_{wn} \cdot \delta H_n$$
where Pwn represents the interpolation coefficients between finite element nodes and CA cell w.
6.2 Simulation Parameters and Validation
To ensure simulation fidelity, temperature-dependent nonlinear material properties were incorporated in this research. The input data include the fraction of solid and enthalpy as functions of temperature, as well as density and thermal conductivity variations. Table 6 lists the growth kinetics parameters used for the high manganese steel casting simulation.
| Component | Composition (%) | Liquidus slope | Partition coefficient | Diffusion coefficient (m²/s) |
|---|---|---|---|---|
| Mn | 13.000 | -3.81 | 0.7776 | 93×10⁶ |
| C | 1.100 | -69.53 | 0.3600 | 93×10⁶ |
| Si | 0.530 | -23.61 | 0.6129 | 93×10⁶ |
| P | 0.033 | -38.37 | 0.0579 | 93×10⁶ |
| S | 0.008 | -29.46 | 0.0088 | 93×10⁶ |
The Gaussian nucleation parameters were calibrated based on material thermodynamic properties and multiple simulation iterations: surface nucleation maximum density ns,max = 1.7×10⁵ m⁻², volume nucleation maximum density nv,max = 5.1×10⁷ m⁻³, surface nucleation average undercooling ΔTs,n = 1 K, surface nucleation undercooling standard deviation ΔTs,σ = 0.1 K, volume nucleation average undercooling ΔTv,n = 10 K, and volume nucleation undercooling standard deviation ΔTv,σ = 0.1 K. The computed growth kinetics parameters are α₂ = 5.6×10⁻⁷ and α₃ = 2.8×10⁻⁷.
6.3 Solidification Structure Growth Simulation Results
The microstructure evolution of the crossing was investigated by extracting the heart rail and wing rail sections. The simulation reveals four distinct zones forming during solidification. First, a very thin layer of fine equiaxed grains forms on the entire free surface during mold filling, resulting from the substantial temperature difference between the room-temperature sand mold and the molten metal. This chill zone is extremely thin and barely distinguishable at the surface. Second, columnar grains grow from this thin fine-grained surface layer perpendicular to the casting surface, following the direction of heat extraction toward the slower-cooling interior. Third, as the columnar growth ceases, equiaxed grains continue growing inward while new equiaxed nucleation occurs in the interior liquid. Fourth, when grains impinge upon each other, solidification is complete, resulting in a relatively fine and uniform equiaxed zone in the center.
Validation against actual production show excellent agreement. The acid-etched macrostructures of the heart rail cross-section from production display the same characteristic features: an extremely thin fine equiaxed zone at the surface, followed by a layer of coarse columnar grains with consistent orientation, and a central zone of relatively fine equiaxed grains. The simulated grain growth trends and size distributions match the actual production results closely, confirming the accuracy of the calibrated CAFE model. This comparison validates the approach for subsequent applications.
The validated CAFE model was then applied to further analyze the theoretical feasibility of the forging-casting integrated production method. Simulation results of the separated forging scheme reveal that, during the casting process, a surface layer of the forging experiences remelting. During subsequent solidification, this remelted region solidifies together with the casting liquid, bonding the two components. The microstructure demonstrates that the forging surface grains adjacent to the connection exhibit localized coarse growth, consistent with remelting and re-solidification phenomena.
At the cast surface contacting the forging, no fine chill zone forms (unlike the surfaces contacting the sand mold), but columnar grain zones develop directly, oriented perpendicular to the forging surface in the direction of heat flow. This observation confirms that the separated forging surface reaches sufficient temperatures to promote bonding-type solidification rather than merely acting as a chill. The combination of mechanical interlocking effects and metallurgical bonding in the connection zones supports the theoretical viability of the proposed production method.
7. Discussion
The integrated forging-casting approach represents a fundamental departure from both conventional sand casting foundry practice and complete forging approaches. By pre-fabricating the critical heart rail and wing rail sections through forging and incorporating them as internal chills within the sand casting foundry system, this method achieves improvements in the final product that neither approach accomplishes independently.
The numerical solutions consistently demonstrate that the casting defect distribution is primarily governed by the thermal field evolution during solidification and the effective feeding distance of the riser system. The discovery that riser exothermic duration plays only a minor role for large castings with extended solidification times provides useful practical guidance for foundry operations. The insensitivity of defect formation to exothermic duration within the tested range suggests that manufacturers can select terms based on other considerations such as cost and handling requirements.
The “V”-shaped feeding zone prediction method offers a practical tool for riser placement in similar large sand casting foundry applications. The high degree of correspondence between predicted defect locations and actual simulation results demonstrates that this method can effectively guide the design of riser systems. The predicted feeding zones directly reflect the maximum effective distance of each riser, enabling strategic placement to achieve complete coverage of critical regions.
The finite element results reveal important temperature characteristics of the forging-casting interface. Without preheating, the forging surface temperatures at the edges exceed the solidus, enabling localized remelting. However, the central regions where heat transfer depends on the thermal mass of surrounding metal cannot achieve sufficient temperatures for robust metallurgical bonding. The separated forging pieces, with their reduced cross-section, overcome this limitation through enhanced surface-to-volume ratio. Preheating further improves the thermal profile but requires optimization to balance enhanced bonding against material property degradation and oxidation risks.
In the case of the high manganese steel crossing, the shrinkage rate of up to 8.5% presents particular challenges. The integrated forging-casting method mitigates this inherent material limitation by reducing the volume of metal that must be fed in the critical regions. The pre-forged components already possess the required microstructure and mechanical properties, so the feeding distance requirements are reduced substantially. This approach may be particularly beneficial for other high-shrinkage alloys where conventional sand casting foundry techniques have reached their limits of defect reduction capability.
8. Conclusions
This comprehensive investigation into the sand casting foundry processes for the 75-852 high manganese steel crossing has yielded several significant findings and innovations:
(1) A complete geometric model and fundamental assumptions were established for the monoblock crossing castings. The novel casting velocity attenuation concept, developed to address the complexities of molten metal overflow during inclined casting, was successfully implemented. Differential mesh refinement strategies optimized computational efficiency while maintaining accuracy in the critical heart and wing rail regions (over 1.5 million elements in these areas). The casting process was visualized and analyzed in five interdependent stages, revealing the interplay between various physical phenomena during solidification.
(2) Comparative analysis of different riser configurations demonstrated that the 1200 kJ/kg exothermic-riser provides optimal performance for this crossing, reducing shrinkage porosity from 23.10% (insulating riser) to 16.51% and shrinkage cavities from 1.002% to 0.514%. Exothermic duration was found to be a secondary parameter with minimal influence for large castings. Heat generation below 1200 kJ/kg results in insufficient feeding capability, while higher values provide no additional benefits and may introduce risks of thermal stress and extended cycle times.
(3) The riser addition optimization scheme, guided by the “V”-shaped feeding zone prediction method developed in this research, achieved substantial defect reduction. The transition zone local shrinkage porosity decreased from 36.17% to 22.11%, and local shrinkage cavities from 1.13% to 0.69%. The numerical simulation results confirmed the accuracy of the prediction methodology. The innovative integrated forging-casting production method for the critical heart and wing rail regions was established and validated through thermal analysis, demonstrating theoretical viability. The separated forging scheme with 400°C preheating yielded the optimal metallurgical bonding results while minimizing mechanical property degradation.
(4) CAFE model simulation was successfully applied to visualize the solidification microstructure evolution of the high manganese steel crossing. The simulation results, calibrated through multiple Gaussian parameter adjustments, exhibited excellent agreement with actual production samples. Application of the validated model confirmed the theoretical feasibility of the integrated forging-casting approach, revealing that cast surfaces bonding with pre-forged components form columnar grain structures perpendicular to the forging surface with acceptable connectivity, distinct from the thin chill zone formed at sand mold interfaces. This research provides new insights and methods for enhancing the quality and service life of high manganese steel crossings in heavy-haul railway applications.
