In the modern foundry industry, the design of casting processes for steel castings has traditionally relied heavily on empirical experience, leading to prolonged development cycles and significant material waste. This thesis presents a comprehensive study on the numerical simulation and process optimization of steel castings using the ProCAST software, with a specific focus on the K6 bolster used in railway freight car bogies. The work is divided into two main parts: the first part involves three-dimensional modeling and simulation of the K6 bolster to validate and optimize the existing casting process, while the second part describes the design and development of a dedicated casting simulation database system to support efficient data management and retrieval for future simulations. Through detailed analysis of velocity fields, temperature fields, and solidification behavior, the study confirms the scientific rationality of the current gating system and pouring temperature range (1560°C–1580°C). Furthermore, simulation revealed a critical shrinkage defect location, which was subsequently eliminated by optimizing the process with the placement of chillers. The developed database system, built on the .NET Winform platform with a client-server architecture, provides robust functionalities for searching, adding, modifying, and deleting material and process parameters, thereby enhancing the efficiency and accuracy of future casting simulations for steel castings.
1. Introduction
The global foundry industry has witnessed significant transformations with the advent of computer-aided engineering (CAE) technologies. Although China is the largest producer of steel castings by volume, it is not yet considered a global leader in casting technology. The traditional reliance on trial-and-error methods for developing casting processes for new products often results in long development times, high costs, and inconsistent quality. For complex steel castings such as the K6 bolster, which is a critical load-bearing component in railway trucks, the internal soundness and absence of casting defects are of paramount importance. The K6 bolster experiences complex multi-axial loads during service, making it susceptible to fatigue failure. Therefore, a robust and well-designed casting process is essential to ensure the structural integrity of these steel castings.
Computer simulation of casting processes has emerged as a powerful tool to address these challenges. By numerically modeling the filling and solidification of molten metal in the mold, engineers can predict and analyze potential defects such as shrinkage porosity, gas entrapment, cold shuts, and hot tears. This allows for virtual optimization of process parameters before physical trials, significantly reducing costs and improving the quality of steel castings. Among various simulation software packages, ProCAST is widely recognized for its advanced finite-element-based capabilities in simulating casting processes. It enables detailed visualization of velocity fields, temperature distributions, and solidification sequences, providing invaluable insights into the casting behavior of steel castings.
The primary objectives of this research are to:
- Develop a three-dimensional model of the K6 bolster using Creo software and import it into ProCAST for numerical simulation.
- Verify the scientific soundness of the existing gating system design for the K6 bolster through analysis of filling and solidification phenomena.
- Evaluate the rationality of the current pouring temperature range and investigate the influence of pouring temperature on the quality of steel castings.
- Optimize the current process plan based on simulation results and confirm improvements through production trials.
- Design and develop a customized casting simulation database system to streamline data management and retrieval for future simulation tasks involving steel castings.
2. Literature Review and Numerical Foundations
Numerical simulation of casting processes has evolved over several decades, integrating disciplines such as heat transfer, fluid dynamics, solidification theory, and computational mechanics. The early work on solidification modeling laid the foundation for modern defect prediction methods. Today, simulation tools like ProCAST are capable of simulating the entire casting cycle, including mold filling, solidification, and subsequent cooling, with high accuracy. The key governing equations for fluid flow and heat transfer in casting simulation are the continuity equation, the Navier-Stokes equations, and the energy equation. For incompressible Newtonian fluids, these equations can be expressed as:
$$ \frac{\partial \rho}{\partial t} + \nabla \cdot (\rho \mathbf{u}) = 0 $$
$$ \rho \left( \frac{\partial \mathbf{u}}{\partial t} + \mathbf{u} \cdot \nabla \mathbf{u} \right) = -\nabla p + \mu \nabla^2 \mathbf{u} + \rho \mathbf{g} $$
$$ \rho C_p \left( \frac{\partial T}{\partial t} + \mathbf{u} \cdot \nabla T \right) = \nabla \cdot (k \nabla T) + Q $$
where ρ is the density, u is the velocity vector, p is the pressure, μ is the dynamic viscosity, g is the gravitational acceleration, Cp is the specific heat, T is the temperature, k is the thermal conductivity, and Q represents the latent heat source term. These equations are discretized using the finite element method (FEM) in ProCAST, allowing for complex geometries and boundary conditions typical of steel castings.
One of the critical aspects of simulation is the proper handling of the solidification process. The latent heat released during phase transformation is accounted for using methods such as the enthalpy method or the equivalent specific heat method. Additionally, the prediction of shrinkage defects is based on the criterion function, often referred to as the Niyama criterion, which can be expressed as:
$$ Niyama = \frac{G}{\sqrt{R}} $$
where G is the thermal gradient and R is the cooling rate. A low Niyama value indicates a high likelihood of microporosity formation. ProCAST uses this criterion in its post-processing module to identify regions prone to shrinkage defects in steel castings.
The accuracy of the simulation heavily depends on the quality of the finite element mesh. The mesh density and element quality directly influence both the computational time and the precision of the results. In this study, special attention was paid to mesh generation for the complex geometry of the K6 bolster. Features such as sharp corners, thin ribs, and gradual thickness transitions required careful mesh refinement to capture the thermal and flow gradients accurately. As shown in the relationship between mesh count, computational accuracy, and time, there exists an optimal mesh density beyond which additional elements yield diminishing returns while significantly increasing computational cost.
| Mesh Density | Accuracy | Computational Time |
|---|---|---|
| Coarse | Low | Short |
| Medium | Good | Moderate |
| Fine | High | Long |
For the K6 bolster simulation, a balance was struck between accuracy and computational efficiency by using a denser mesh only in critical regions such as the junction of ribs and thick sections, while a relatively coarser mesh was used in less critical areas.
3. Simulation of the K6 Bolster
3.1 Product Overview
The K6 bolster is a major structural component of the railway truck bogie. It has a complex geometry with a length of 2429 mm, a width of 470 mm, and a height of 355 mm. The casting weight is approximately 640 kg, and the material is B+ grade steel (ZG25MnCrNi). The bolster contains two rectangular openings for the brake rod, three internal ribs that connect the top and bottom walls, and spring seats on the top surface. The wall thickness varies significantly across the casting, which creates potential hot spots and non-uniform solidification, leading to shrinkage defects if not properly controlled. The current manufacturing process uses ester-cured sodium silicate sand molds, with a pouring temperature around 1570°C and a pouring rate of 40 kg/s.
To simulate the casting process of these steel castings, a detailed three-dimensional model is required. The modeling was performed using Creo 2.0 software, leveraging its powerful solid modeling capabilities. A 1/4 model was first constructed and then mirrored using symmetry features to create the full model, followed by local adjustments to match the actual product. The gating system, including the sprue, runner, ingates, and risers, was also modeled according to the existing process design. The complete assembly model is shown in the figure below.
3.2 Simulation Setup
The model was exported in IGES format and imported into ProCAST. The mesh generation was carried out using the MeshCAST module. The mesh parameters were chosen to ensure high resolution in the casting while maintaining a reasonable number of elements for the entire assembly. The initial simulation parameters are listed in the table below.
| Parameter | Value |
|---|---|
| Material | B+ grade steel (ZG25MnCrNi) |
| Mold material | Ester-cured sodium silicate sand |
| Pouring temperature | 1570°C |
| Ambient temperature | 20°C |
| Pouring rate | 40 kg/s |
| Interface heat transfer coefficient | 500 W/m²K |
The boundary conditions and initial thermal conditions were set accordingly. The simulation was run for the complete filling and solidification process, and the results were analyzed in ViewCAST.
3.3 Analysis of the Filling Process – Velocity Field
The velocity field during mold filling provides insights into the metal flow behavior and helps identify regions prone to turbulence, air entrapment, and inclusion defects. The simulation results for the K6 bolster show that during the initial stage of filling, the molten steel exhibits chaotic flow with varying velocity magnitudes and directions. This is attributed to the complex gating system and the geometry of the mold. However, as the filling progresses to about 25%, the flow becomes more stable with a dominant downward direction due to gravity. The maximum velocity occurs at this stage, but the flow direction remains relatively consistent. After the mold is half-filled, the velocity in most regions of the casting drops significantly, except within the gating channels. This indicates a smooth and progressive filling pattern, which is desirable for producing high-quality steel castings.
The velocity field analysis confirms that the existing gating system design effectively minimizes turbulent flow, reducing the risk of gas entrapment and oxide inclusions. The use of both top and bottom ingates ensures homogeneous filling and reduces the likelihood of cold shuts, which is crucial for complex steel castings like the K6 bolster.
3.4 Analysis of the Filling Process – Temperature Field
The temperature field during filling reveals the thermal history of the molten steel as it travels through the mold. Regions where the liquid cools prematurely are susceptible to misruns and cold shuts. In the simulation of the K6 bolster, it was observed that at the beginning of filling, the temperature in some extended flat areas (indicated by dark circles in the temperature contour plot) is lower than the surrounding regions. These areas correspond to the large flat surfaces of the bolster, where the molten metal spreads quickly and loses heat rapidly. The lower temperature could potentially lead to surface defects. However, subsequent simulation snapshots show that these regions are adequately filled, and no macroscopic defects are expected.
By analyzing the temperature distribution at various filling stages, it was verified that the gating system provides sufficient preheating to the mold and maintains adequate fluidity of the molten steel throughout the filling process. This is particularly important for thin-walled sections of the steel castings, where premature solidification can cause incomplete filling.
3.5 Solidification Process and Defect Prediction
The solidification sequence of the K6 bolster was examined to identify potential locations of shrinkage porosity and hot spots. The simulation results show that the casting does not solidify in an ideal progressive manner due to its complex geometry and varying wall thickness. Isolated liquid regions, or hot spots, are formed at certain junctions where ribs meet the main walls. These locations are susceptible to shrinkage defects because the surrounding material solidifies first, cutting off the liquid supply. This is a common challenge in steel castings with abrupt changes in section thickness.
ProCAST’s post-processing module uses the Niyama criterion and other algorithms to predict the probability of shrinkage defects. The simulation predicted a significant concentration of porosity at the lower part of the wedge block (斜楔) area. The predicted locations are shown in purple in the defect map. To validate these predictions, physical inspection was carried out on ten castings from different heats. Eight of these castings exhibited obvious shrinkage defects precisely at the simulated locations, confirming the accuracy of the simulation model.
4. Optimization of the Casting Process
4.1 Verification of Pouring Temperature
The pouring temperature is a critical parameter influencing the quality of steel castings. If the temperature is too high, the steel has high shrinkage, high gas absorption, and intense mold erosion, leading to defects such as shrinkage, gas porosity, and hot tears. If it is too low, the fluidity is poor, causing cold shuts and misruns. The existing process for the K6 bolster specifies a pouring temperature of 1560°C–1575°C. To verify this range and investigate the relationship between pouring temperature and defect formation, three simulations were conducted with pouring temperatures of 1560°C, 1570°C, and 1580°C.
The theoretical liquidus temperature of B+ grade steel was first calculated using the following empirical formula based on its chemical composition:
$$ T_L = 1538 – \{70w[C] + 8w[Si] + 5w[Mn] + 30w[P] + 25w[S] + 4w[Ni] + 1.5w[Cr]\} $$
Using the nominal composition of ZG25MnCrNi (C: 0.25%, Si: 0.30%, Mn: 0.80%, P: 0.02%, S: 0.02%, Ni: 0.15%, Cr: 0.40%), the liquidus temperature is calculated as:
$$ T_L = 1538 – (70 \times 0.25 + 8 \times 0.30 + 5 \times 0.80 + 30 \times 0.02 + 25 \times 0.02 + 4 \times 0.15 + 1.5 \times 0.40) = 1505.5^\circ C $$
To achieve adequate fluidity, a superheat of about 65°C is common, giving a pouring temperature of approximately 1570°C, which is consistent with the current practice.
The three simulation cases were compared with respect to the predicted shrinkage porosity. The results indicated that the total defect volume is lowest at 1570°C, while both 1560°C and 1580°C result in slightly higher defect levels. However, the differences are not significant, suggesting that the current operating range of 1560°C–1575°C is robust and well-chosen for this product. This conclusion is further supported by production records showing that actual pouring temperatures are consistently maintained around 1570°C.
| Pouring Temperature (°C) | Predicted Shrinkage Defect Volume (relative) |
|---|---|
| 1560 | 1.12 |
| 1570 | 1.00 |
| 1580 | 1.08 |
4.2 Addressing the Shrinkage Defect
As mentioned earlier, the simulation predicted a significant shrinkage defect at the lower wedge block area. In the existing process, no special measures were taken for this location. To eliminate this defect, two countermeasures were proposed and tested: placing chromite sand or applying chillers (cold iron). Chromite sand has a high thermal conductivity and acts as an effective chill, accelerating solidification locally. Chillers are external metal pieces placed in the mold to rapidly extract heat from the hot spot.
In the first trial, chromite sand was placed at the defect area. The subsequent castings showed some improvement, but the shrinkage was still visible. In the second trial, chillers were placed at the same location. The castings produced with chillers showed no shrinkage defects at the wedge block area, completely resolving the problem. This demonstrates the power of simulation-guided process optimization for steel castings: the accurate prediction of defect locations enabled a targeted solution, reducing the need for extensive trial-and-error and improving the overall quality of the product.
5. Development of a Casting Simulation Database System
5.1 Need and Motivation
One of the major obstacles in the widespread use of casting simulation software is the tedious and error-prone process of gathering and managing the required input data. For each new simulation of steel castings, engineers need to search through many product drawings, process specifications, and material datasheets to find parameters such as pouring temperature, filling rate, mold material properties, and heat transfer coefficients. This manual process is time-consuming and often leads to inconsistencies. To address this issue, a dedicated database system was developed to centralize and organize the simulation data, making it easy for process engineers to access accurate information quickly.
5.2 System Requirements and Evaluation Criteria
The database system was designed according to the following criteria:
- Content Completeness: The system should contain all materials and process data used in the production of steel castings at the company.
- Function Completeness: Users should be able to add, modify, delete, search, and view records with ease.
- Interoperability with Simulation Software: The data should be formatted in a way that can be easily used as input for ProCAST or other simulation tools.
The functional requirements were defined as: the system should support product classification (military, civil, export), provide a tree navigation structure, allow fuzzy search by product name, and display detailed information including product drawing, steel grade, casting weight, core count, pouring temperature, pouring speed, chiller count, and molding method.
5.3 Technical Architecture
The system adopts a client-server (C/S) model implemented using the .NET Winform technology. This architecture was chosen because it provides high performance, a rich user interface, and offline capabilities. The development environment consists of Microsoft Visual Studio 2010 running on Windows 10. The data storage uses a binary file serialized with JSON, which is easy to read and maintain for small to medium amounts of data. The system architecture is composed of a single application that directly manages the data file, with functions for data manipulation and visualization.
The main user interface features a left-side tree view for product type selection, a data grid for displaying records, and buttons for adding, deleting, editing, and searching. The search function performs fuzzy matching on product name and steel grade. The product drawing is stored as a byte array and can be displayed as an image when a record is selected.
5.4 Database Design
The core data structure is a Model class with properties corresponding to the fields listed in the table below.
| Field | Type | Description |
|---|---|---|
| type | string | Product category (military, civil, export) |
| no | string | Product drawing number |
| name | string | Product name |
| gType | string | Steel grade |
| zl | decimal | Casting weight (kg) |
| num | int | Number of cores |
| wendu | decimal | Pouring temperature (°C) |
| sudu | string | Pouring rate |
| ltNum | int | Number of internal/external chillers |
| zxff | string | Molding method |
| img | byte[] | Product drawing image |
This schema was designed to capture all essential parameters required for simulating steel castings, providing a one-stop reference for the engineering team.
5.5 System Implementation
The system was implemented with a clean separation between the UI and data access layers. The Common static class provides helper methods for retrieving and saving the list of models, as well as converting images to byte arrays. The main form handles the event-driven interactions. Key functionalities implemented include:
- Binding Data: The grid view displays the list of products, filtering based on search keywords and selected category. Column headers are mapped to meaningful Chinese (or English) labels.
- Adding Records: The add form allows the user to input all fields, validates the data, and appends the new record to the list.
- Editing Records: When editing, the existing data is loaded into the form, and after modifications, the record is updated and saved.
- Deleting Records: A confirmation dialog is shown, and after confirmation, the selected record is removed from the list and the data file is updated.
- Searching: The search box performs real-time filtering on the product name and steel grade columns using case-insensitive matching.
The system was tested with real process data from the company, demonstrating its usability and efficiency. It significantly reduces the time required to locate and retrieve simulation parameters, thereby accelerating the overall simulation workflow for steel castings.
6. Results and Discussion
The combined efforts of numerical simulation and database development have yielded substantial improvements in the production of K6 bolster steel castings. The simulation provided a clear understanding of the filling and solidification phenomena, confirming the robustness of the gating system and the pouring temperature range. More importantly, it identified a critical defect area that was previously unaddressed, leading to a practical solution using chillers. This improvement not only reduced the scrap rate but also enhanced the reliability of the steel castings in service.
The database system, although developed specifically for this project, is designed to be extendable. It currently contains data for the K6 bolster and other common products. Future work could expand the database to include thermal property tables for various mold materials and alloys, as well as interrelationships with simulation software to allow direct parameter import. This would further streamline the simulation process and encourage more foundries to adopt simulation-driven process design for high-quality steel castings.
7. Conclusions
This thesis has successfully demonstrated the application of ProCAST numerical simulation for the process optimization of steel castings, specifically for the K6 railway bolster. The key conclusions drawn from this research are:
- The existing gating system design is scientifically sound, as verified through the analysis of filling velocity and temperature fields during mold filling.
- The current pouring temperature range of 1560°C–1575°C is reasonable and viable. Simulation results show minimal variation in defect volume across 1560°C, 1570°C, and 1580°C, with 1570°C being optimal.
- Simulation accurately predicted the occurrence of shrinkage defects at the lower wedge block area. By implementing chillers at this location, the defects were completely eliminated, thereby optimizing the existing casting process.
- A casting simulation database system was successfully developed using C/S architecture and Winform technology. The system provides efficient data management, including search, add, modify, delete, and view functions, significantly reducing the time and effort required for parameter retrieval in future simulations of steel castings.
The integration of numerical simulation and database technologies represents a significant step forward in the modernization of casting process design for steel castings. It enables engineers to rely on scientific analysis rather than intuition, leading to higher quality products, reduced costs, and enhanced competitiveness in the global market.