Steel casting has long been recognized as one of the most fundamental manufacturing processes, enabling the production of complex metallic components with superior mechanical properties. In the field of industrial valves, the top-loaded ball valve steel casting represents a particularly challenging category due to its large dimensions, intricate geometry, and stringent quality requirements. My research focuses on the systematic investigation and optimization of the sand casting process for such steel casting components, employing advanced numerical simulation techniques through the ProCAST platform to predict and mitigate common casting defects such as shrinkage porosity and macro-porosity.
The top-loaded ball valve steel casting examined in this study has dimensions of 1920 mm × 1406 mm × 1310 mm and weighs approximately 5.76 tons. This steel casting component belongs to the medium-to-large category and is manufactured using WCB (Weldable Carbon Steel) grade material. The unique structural characteristics of this steel casting, including significant wall thickness variations and complex internal cavities, make it particularly susceptible to casting defects. The ratio between maximum and minimum wall thickness reaches 3:1, creating challenging thermal gradients during solidification. My investigation reveals that these geometric features substantially increase the difficulty of achieving proper feeding during solidification, often resulting in shrinkage defects that compromise the structural integrity of the steel casting.
1. Introduction and Research Background
My research originates from an industrial challenge encountered in a casting enterprise specializing in the production of large steel casting components. The top-loaded ball valve steel casting, unlike conventional side-mounted ball valves, features a unique two-trunnion support structure and a top-flange-reinforced design that facilitates modular rapid disassembly. While this design offers significant advantages for online maintenance and adaptability to complex service environments, it presents considerable difficulties for the casting process. The large-size cavity and differential wall thickness distribution of this steel casting result in complex molten metal flow patterns during filling, making temperature field control substantially more difficult. Additionally, the solidification shrinkage behavior of thick-walled sections creates challenging conditions for feeding channel formation, frequently leading to defects such as cold shut, shrinkage porosity, and macro-porosity.
1.1 Statistical Context of Steel Casting Industry
The steel casting industry in China has experienced remarkable growth over recent decades. According to industry statistics, China’s total casting production reached 55.16 million tons in 2024, accounting for more than 60% of global output. This positions China as the undisputed global leader in casting production, including steel casting. The continuous development of the steel casting sector has been driven by increasing demand from oil and gas, petrochemical, nuclear power, and marine engineering industries. However, the industry faces persistent challenges related to quality control, energy consumption, and defect management. Data indicates that casting defects account for approximately 12.7% of annual steel casting production, translating to substantial economic losses. My research addresses this critical issue by implementing numerical simulation techniques to optimize the steel casting process, thereby reducing defect rates and improving overall product quality.
1.2 Literature Review on Steel Casting Simulation
The numerical simulation of casting processes has evolved significantly since its inception in the 1940s. The pioneering work by Paschkis in 1949 on the “Heat and Mass Flow Analyzer” laid the foundation for computer-aided analysis of heat transfer during casting. Since then, the field has progressed through several distinct phases: initial finite difference method applications in the 1960s, finite element method integration in the 1980s, and current sophisticated multi-physics simulations incorporating fluid flow, heat transfer, and stress analysis. In the context of steel casting, simulation software has become an indispensable tool for predicting shrinkage porosity, hot tears, and other defects before actual production begins.
| Software | Country | Algorithm | Key Features | Application in Steel Casting |
|---|---|---|---|---|
| HUAZHU CAE | China | FDM | Excellent local support, rapid computation | Widely used in Chinese foundries |
| AnyCasting | Korea | FEM | High accuracy, comprehensive database | Die casting and steel casting |
| MAGMASOFT | Germany | FDM | Powerful solver, defect prediction | Automotive steel casting components |
| FLOW-3D CAST | USA | FDM | Excellent free surface tracking | Complex steel casting geometries |
| ProCAST | France | FEM/FVM | Coupled flow-thermal-stress analysis | Large steel casting optimization |
My decision to utilize ProCAST for this steel casting research was based on several critical advantages. The finite element method implemented in ProCAST provides superior handling of complex geometric boundaries, which is essential for the intricate structure of top-loaded ball valve steel castings. The software’s adaptive mesh refinement capabilities enable accurate capture of solidification front morphology in thin-thick wall transition zones. Furthermore, the hybrid FEM-FVM solver architecture demonstrates exceptional synergy in analyzing flow dynamics during mold filling and thermal stress evolution during solidification—both critical factors in steel casting quality control.
1.3 Research Objectives and Technical Route
The primary objectives formulated for this steel casting research are as follows:
1) To design and optimize the complete sand casting process system for the top-loaded ball valve steel casting, including gating system, riser system, and chill placement.
2) To develop a comprehensive multi-physics numerical simulation model using ProCAST for accurate prediction of the steel casting solidification behavior and defect formation.
3) To implement a systematic optimization methodology combining simulation analysis with orthogonal experimental design to determine the optimal process parameters.
4) To validate the optimized steel casting process through actual production trials and non-destructive testing.
The technical route of my research is summarized through the following interconnected stages, which together form a closed-loop methodology:
Stage 1: Process System Design – design of gating, riser, and chill systems based on sand casting theory.
Stage 2: Numerical Modeling – construction of the steel casting model, mesh generation, and simulation parameter definition.
Stage 3: Defect Prediction – simulation of the original process scheme and identification of potential shrinkage defects.
Stage 4: Process Optimization – modification of the feeding system and chill placement based on simulation results.
Stage 5: Experimental Validation – actual casting trials and non-destructive testing to verify simulation accuracy.
2. Sand Casting Process System Design for Steel Casting
2.1 Overview of Sand Casting Process Flow
The production of large steel casting components through sand casting involves multiple critical steps coordinated across different workshop departments. For the top-loaded ball valve steel casting, the process begins with wood pattern fabrication, followed by sand mold preparation, core assembly, metal melting, pouring, controlled cooling, shakeout, fettling, and finally, comprehensive quality inspection. Each stage requires precise control to ensure the final steel casting meets dimensional and metallurgical specifications.
I analyzed the process flow and categorized the key steps into a systematic sequence. The wood pattern is fabricated using CNC machinery with an allowance of 1.5%–2.0% shrinkage, which is essential for achieving the required final dimensions of the steel casting. The resin-bonded sand system, comprising silica sand mixed with furan resin and silane coupling agent, provides optimal strength and collapsibility characteristics. My analysis indicated that the mold preparation and core assembly stages are particularly critical, especially for the complex internal cavities of ball valve steel castings.
2.2 Gating System Design for Steel Casting
The gating system plays a fundamental role in controlling the flow of molten metal into the mold cavity. A well-designed gating system ensures smooth, controlled filling that minimizes turbulence, prevents gas entrapment, and maintains appropriate temperature distribution throughout the steel casting. My design considered five key technical dimensions: gating type selection, filling time control, flow dynamics regulation, solidification temperature field management, and process economy.
For the filling time calculation, I applied the following formulas derived from hydraulics principles:
$$t_f = \frac{m_{total}}{\dot{m}} \tag{2.1}$$
where $t_f$ is the filling time, $m_{total}$ is the total molten metal mass, and $\dot{m}$ is the mass flow rate. For the ball valve steel casting with a mass of 5.76 tons and a designed flow rate of 100 kg/s, the theoretical filling time calculates to:
$$t_f = \frac{5300 \text{ kg}}{100 \text{ kg/s}} \approx 90 \text{ s} \tag{2.2}$$
My comparative analysis of different gating system types for this steel casting is presented in Table 2.1, which evaluates their suitability based on key criteria including filling stability, temperature gradient, and defect suppression capability.
| Gating Type | Filling Speed (m/s) | Defect Rate (%) | Feeding Efficiency (%) | Suitability |
|---|---|---|---|---|
| Top gating | 1.5–2.5 | 3–5 | 50–60 | Small castings |
| Bottom gating | 0.5–1.0 | <1 | 40–50 | Large complex castings |
| Middle gating | 0.8–1.2 | 1–2 | 50–60 | Medium castings |
| Step gating | 0.6–1.0 | 0.5–1.5 | 65–75 | Tall thin-wall castings |
Based on this analysis, I selected the bottom-gating system for the top-loaded ball valve steel casting. This choice was driven by the following considerations: the large overall height of the casting makes bottom filling necessary; the bottom-gating approach provides laminar flow conditions that minimize oxidation and gas entrapment; and it allows efficient slag separation in the gating system. The multi-channel symmetric design ensures uniform filling across the casting cross-section.
2.3 Riser Design Principles
The riser system is engineered to compensate for volumetric shrinkage during solidification of the steel casting. The design process follows the modulus method, which establishes critical relationships between riser geometry, casting modulus, and solidification time. According to Chvorinov’s rule, the solidification time of a section is proportional to the square of its modulus (volume-to-surface area ratio). For effective feeding, the riser modulus ($M_r$) must exceed the casting modulus ($M_c$) at the feeding location:
$$M_r \geq M_c \cdot f \tag{2.3}$$
where $f$ is a safety factor typically ranging from 1.1 to 1.2. For side risers, the modulus relationships are established as:
$$M_c : M_n : M_r = 1 : 1.1 : 1.2 \tag{2.4}$$
where $M_n$ represents the modulus of the riser neck. The neck length is calculated as:
$$L_n = 2.4 \cdot \sqrt[3]{M_c \cdot M_r \cdot M_n} \tag{2.5}$$
Riser effectiveness is quantified by the feeding efficiency coefficient ($\eta$), which represents the volume fraction of the riser available for compensating shrinkage in the steel casting. Table 2.2 presents the typical efficiency values for different riser types.
| Riser Type | Efficiency η (%) |
|---|---|
| Cylindrical riser | 12–15 |
| Spherical riser | 15–20 |
| Insulated riser | 25–30 |
| Exothermic riser | 30–40 |
| Gas-pressurized riser | 35–40 |
| Compressed-air riser | 35–40 |
The volume requirement for the riser is established through the following equation:
$$\frac{V_r – V_{rf}}{V_r} = \eta \leq \frac{\beta (V_r + V_c)}{V_r} \tag{2.6}$$
where $V_r$ represents the initial riser volume, $V_{rf}$ the final riser volume after solidification, $V_c$ the casting volume at the hot spot, and $\beta$ the solidification shrinkage coefficient of the alloy.
For the top-loaded ball valve steel casting, I designed a total of 13 risers strategically positioned to ensure complete feeding. The arrangement includes four large waist-shaped risers (500 mm × 450 mm) on the flange-main body junctions, four cylindrical risers (325 mm diameter, 225 mm height) at complex structural locations, and five supplementary risers at intermediate wall thickness sections. This distribution ensures that all hot spots in the steel casting have access to an adequate supply of molten metal during solidification.
2.4 Insulating and Exothermic Riser Systems
To enhance feeding efficiency while minimizing riser metal consumption, I incorporated insulating and exothermic riser sleeves in the optimized steel casting process. These advanced riser systems significantly extend the solidification time of the riser by reducing heat loss to the surroundings. The modified riser modulus considering insulation effects is:
$$M_{insulated} = \frac{V}{a S_{side} + b S_{top} + S_{bottom}} \tag{2.7}$$
where $a$ and $b$ are insulation factors (typically 0.7), $S_{side}$, $S_{top}$, and $S_{bottom}$ represent the lateral, top, and bottom surface areas of the riser respectively. For a blind insulated riser with $a = b = 0.7$:
$$M_{insulated} = 1.43 \cdot M \tag{2.8}$$
For an open insulated riser with $H = 1.3D$:
$$M_{insulated} = 1.33 \cdot M \tag{2.9}$$
My design includes FT400-type exothermic sleeves (20 mm thick) on the four main risers and double-layer insulation structure on the five auxiliary risers, completed with insulation boards on top. This configuration extends the riser solidification time to 2.3 times that of conventional sand risers, significantly improving the feeding capacity for the steel casting.
2.5 Chill Design for Controlled Solidification
Chills are critical elements used to manipulate the solidification sequence of the steel casting. By accelerating local cooling, chills create favorable temperature gradients that promote directional solidification toward riser locations. My chill design considerations included: 1) location optimization to achieve controlled thermal gradients; 2) dimensional calculation to match the thermal requirements of each section; and 3) coordination with riser placement to establish effective feeding channel.
The chill thickness ($t_{chill}$) is determined by the equivalent thickness of the casting hot spot ($\delta$):
$$t_{chill} = K \cdot \delta \tag{2.10}$$
where K is a material coefficient. For steel castings, K ranges from 0.3 to 0.8. My analysis of the ball valve casting structure identified several critical zones requiring chilling: the flange sections where wall thickness changes abruptly, the central body areas with large thermal mass, and the transition zones between thick and thin sections.
The chill placement criteria are established as follows:
1) When $t \leq 2T$ (where $t$ is hot spot thickness and $T$ is connecting wall thickness), single-side chilling suffices.
2) When $2T \leq t \leq (3\sim4)T$, double-side chilling is required.
3) When $t > 4T$, a combined chill-riser system is necessary.
The effective chill surface area ($S_y$) is calculated using the modulus reduction approach:
$$S_y = \frac{V_0(1 – M_0/M_1)}{M_1(y-1)} \tag{2.11}$$
where $V_0$ is the casting volume, $M_0$ and $M_1$ are the casting modulus before and after chill application, and $y$ is the chill effectiveness coefficient. Through careful thermal calculation, I positioned chills especially at the flange sections and central cavity areas to create the desired temperature gradient field in the steel casting.
3. Casting Numerical Simulation Software and Theoretical Foundation
3.1 Overview of Commercial Simulation Software
The application of computer-aided engineering (CAE) in steel casting has become standard practice in modern foundries. My evaluation of available simulation platforms focused on their ability to handle the multi-physics nature of steel casting processes, including turbulent flow, non-linear heat transfer, phase transformation, and thermal stress. Table 3.1 provides a comprehensive comparison of major commercial casting simulation packages in the context of steel casting applications.
| Software | Method | Flow Simulation | Stress Analysis | Microstructure | Defect Criteria |
|---|---|---|---|---|---|
| HUAZHU CAE | FDM | Yes | No | No | Niyama, fs |
| AnyCasting | FEM | Yes | Partial | No | Niyama, temperature |
| MAGMASOFT | FDM | Yes | Yes | Yes | Multiple |
| FLOW-3D CAST | FDM | Yes (VOF) | No | No | Niyama |
| ProCAST | FEM/FVM | Yes (VOF) | Yes | Yes (CAFE) | Niyama, pore, etc. |
3.2 ProCAST Software Architecture and Capabilities
ProCAST, initially developed by UES Inc. for NASA in 1985 and later acquired by ESI Group, has evolved into the world’s leading finite element-based casting simulation software. The platform integrates multiple modules that collectively provide a comprehensive numerical framework for steel casting simulation. The software’s architecture comprises several key components that I utilized throughout my research:
The Visual-Mesh module serves as the pre-processing engine, enabling high-quality surface and volume mesh generation for complex casting geometries. The Visual-Cast module facilitates the definition of boundary conditions, material properties, and process parameters. The DataCast/ProCAST solver performs the core multi-physics computation, while the Visual-Viewer module presents simulation results through advanced visualization techniques. In 2024, ProCAST integrated AI-based optimization modules, further enhancing its capability for process parameter optimization in steel casting production.
3.3 Governing Equations for Casting Simulation
The numerical simulation of steel casting is fundamentally based on three conservation laws governing fluid flow, heat transfer, and solidification. For the filling process simulation, I solved the complete system of Navier-Stokes equations with volume-of-fluid (VOF) free-surface tracking.
Mass Conservation Equation:
$$\frac{\partial \rho}{\partial t} + \nabla \cdot (\rho \vec{v}) = 0 \tag{3.1}$$
For incompressible molten steel, this simplifies to:
$$\frac{\partial u}{\partial x} + \frac{\partial v}{\partial y} + \frac{\partial w}{\partial z} = 0 \tag{3.2}$$
where $u$, $v$, and $w$ are velocity components in the x, y, and z directions respectively.
Momentum Conservation (Navier-Stokes Equations):
\begin{equation}
\begin{cases}
\rho \left( u \frac{\partial u}{\partial x} + v \frac{\partial u}{\partial y} + w \frac{\partial u}{\partial z} + \frac{\partial u}{\partial t} \right) = \rho g_x – \frac{\partial P}{\partial x} + \mu \nabla^2 u \\
\rho \left( u \frac{\partial v}{\partial x} + v \frac{\partial v}{\partial y} + w \frac{\partial v}{\partial z} + \frac{\partial v}{\partial t} \right) = \rho g_y – \frac{\partial P}{\partial y} + \mu \nabla^2 v \\
\rho \left( u \frac{\partial w}{\partial x} + v \frac{\partial w}{\partial y} + w \frac{\partial w}{\partial z} + \frac{\partial w}{\partial t} \right) = \rho g_z – \frac{\partial P}{\partial z} + \mu \nabla^2 w
\end{cases}
\tag{3.3}
\end{equation}
Energy Conservation Equation:
$$\rho c_p \left( u \frac{\partial T}{\partial x} + v \frac{\partial T}{\partial y} + w \frac{\partial T}{\partial z} + \frac{\partial T}{\partial t} \right) = \frac{\partial}{\partial x} \left(k \frac{\partial T}{\partial x}\right) + \frac{\partial}{\partial y} \left(k \frac{\partial T}{\partial y}\right) + \frac{\partial}{\partial z} \left(k \frac{\partial T}{\partial z}\right) + S \tag{3.4}$$
3.4 Solidification and Latent Heat Modeling
The solidification process of steel casting involves complex heat transfer phenomena including conduction, convection, and radiation, coupled with phase transformation. The heat conduction within the solidifying steel casting follows Fourier’s law:
$$\rho c_p \frac{\partial T}{\partial t} = \frac{\partial}{\partial x} \left(\lambda \frac{\partial T}{\partial x}\right) + \frac{\partial}{\partial y} \left(\lambda \frac{\partial T}{\partial y}\right) + \frac{\partial}{\partial z} \left(\lambda \frac{\partial T}{\partial z}\right) + Q \tag{3.5}$$
Convective heat exchange at boundary surfaces is described by Newton’s law of cooling:
$$q = h_c \cdot (T_f – T_w) \tag{3.6}$$
where $h_c$ is the convective heat transfer coefficient, $T_f$ is the fluid temperature, and $T_w$ is the wall temperature. Radiative heat transfer follows the Stefan-Boltzmann law:
$$q = \varepsilon \sigma_0 T_s^4 \tag{3.7}$$
For the latent heat evolution during solidification, I implemented the enthalpy method, which transforms the energy equation into terms of enthalpy:
$$H = \int_0^T c_p \, dT + L(1 – f_s) \tag{3.8}$$
where $L$ is the latent heat of fusion and $f_s$ is the solid fraction. This formulation automatically accounts for the release of latent heat during solidification of the steel casting.
3.5 Boundary Conditions for Steel Casting Simulation
The accuracy of steel casting simulation heavily depends on proper definition of interface boundary conditions. For the ball valve casting model, I established the following interface conditions:
| Interface | Heat Transfer Coefficient (W/m²·K) |
|---|---|
| Casting – Mold | 300–1000 |
| Casting – Air | 5–10 |
| Mold – Air | 5–10 |
| Casting – Chill | 1000–5000 |
| Chill – Mold | 300–1000 |
3.6 Defect Prediction Criteria
For the prediction of shrinkage defects in steel casting, I employed the Niyama criterion, which has become the industry standard. The criterion establishes a relationship between temperature gradient and cooling rate to identify regions prone to microporosity development:
$$\frac{G}{\sqrt{R}} < C_{Niyama} \tag{3.9}$$
where $G$ is the local temperature gradient, $R$ is the cooling rate, and $C_{Niyama}$ is the critical value (1.1 for large castings, 0.8 for small castings). The temperature gradient at each simulation cell is calculated as the maximum among neighboring cells:
$$G = \max \left( \sqrt{\left(\frac{T_{i+1,j,k} – T_{i,j,k}}{\Delta x}\right)^2 + \left(\frac{T_{i,j+1,k} – T_{i,j,k}}{\Delta y}\right)^2 + \left(\frac{T_{i,j,k+1} – T_{i,j,k}}{\Delta z}\right)^2} \right) \tag{3.10}$$
Additionally, I monitored the solid fraction evolution using the critical solid fraction criterion ($f_s = f_{sc}$), where for steel casting, $f_{sc}$ is typically taken as 0.7. When the solid fraction exceeds this critical value, the remaining liquid is isolated, leading to shrinkage porosity formation.
4. ProCAST-based Optimization of Ball Valve Steel Casting
4.1 Steel Casting Structure and Initial Process Scheme
The top-loaded ball valve steel casting being investigated is a WCB-grade component with the following chemical composition requirements and mechanical properties as shown in Tables 4.1 and 4.2. The steel casting weight is approximately 5.76 tons, making it a medium-to-large steel casting that demands careful process design.
| Element | C | Si | Mn | P | S | Cr | Ni+Mo |
|---|---|---|---|---|---|---|---|
| Specification | 0.25–0.30 | ≤0.60 | ≤1.00 | ≤0.040 | ≤0.045 | ≤0.50 | ≤1.00 |
| Deviation | ±0.01 | ±0.05 | ±0.05 | ±0.005 | ±0.005 | ±0.20 | ±0.02 |
| Property | Tensile Strength (MPa) | Yield Strength (MPa) | Elongation (%) | Hardness (HB) |
|---|---|---|---|---|
| ASTM Requirement | ≥485 | ≥250 | ≥22 | 137–179 |
| Typical Value | 500–620 | 270–350 | 25–32 | 150–180 |
For this steel casting, I initially designed the casting process with a bottom-gating, four-channel symmetric filling system. The gating system consisted of a pouring cup, two symmetric sprue risers, radial runner distribution, and circumferentially positioned ingates. The riser system included four waist-shaped risers (500 mm × 450 mm) on the flange-body junctions, four cylindrical risers (325 mm × 225 mm) at complex structural locations, and four supplementary risers at intermediate wall thickness sections. Cold iron was placed at the flange and central cavity positions to promote ordered solidification of the steel casting.
4.2 Pre-processing Setup in ProCAST
My simulation workflow began with model construction using NX12.0 three-dimensional CAD software. The complete assembly of the steel casting, including the casting body, riser system, chill system, and gating system, was created in the assembly module of NX and exported in .igs format. The geometry was then imported into the Visual-Mesh module of ProCAST, where I added a virtual sand mold measuring 5000 mm × 5200 mm × 4000 mm. The mesh discretization strategy varied by region to balance computational efficiency with accuracy. The sand mold was meshed with 100 mm elements, the main risers with 50 mm elements, secondary risers with 30 mm elements, and the casting body and chills with 10 mm elements. This hierarchical meshing produced approximately 90,000 surface elements and 460,000 volume elements, totaling about 560,000 mesh cells.
For the material properties, since the WCB steel casting material was not available in the standard ProCAST database, I constructed a custom database using the thermodynamic calculation engine combined with the lever rule method. The computed liquidus temperature was 1509°C and solidus temperature was 1472°C. The thermal and mechanical property curves for WCB steel were calculated and input into the simulation model.
The volume management parameters were defined as follows: the casting, riser, and chill systems were classified as Alloy type with WCB material and initial filling fraction of 0. The sand mold was designated as Virtual Mold with resin-bonded sand properties and initial temperature of 20°C. The pouring temperature was set to 1570°C for the initial simulation.
Interface conditions were established with the casting-mold interface assigned as CONIC type with heat transfer coefficient of 500 W/(m²·K). The casting-gating system interface was treated as EQUIV since they form a continuous domain. The chill-mold interface was set with a heat transfer coefficient of 2000 W/(m²·K).
4.3 Simulation Results of the Original Process Scheme
4.3.1 Filling and Thermal Field Analysis
The simulation results for the original process showed that the filling process was completed smoothly through the bottom-gating system, with a total filling time of approximately 90 seconds. The temperature field at the completion of filling is shown to follow a natural vertical gradient, with molten metal at sections near the ingates maintaining temperatures between 1570°C and 1509°C. This temperature distribution is favorable for initiating solidification from the bottom sections of the steel casting upward. The simulation revealed that during the filling process, some localized cooling begins at thin-wall regions even before filling reaches 100%, consistent with practical observations of steel casting solidification behavior.
4.3.2 Solidification Progression
Analysis of the solid fraction evolution across the simulation time steps revealed critical information about the solidification behavior of the steel casting. At a solid fraction of 2.4% immediately after filling completion, the gating system began to solidify first, followed by the casting edges and thin-wall sections. By the time the solid fraction reached 20% (t = 1720 s), the casting edges had developed continuous solidification layers. At 50% solid fraction, the solidification front showed an inward propagation pattern, with the flange-center connection areas and riser bottoms comprising the last solidifying regions.
The final solidification regions were not confined to the interior of the risers but extended into the riser platforms and feed aids. This observation indicated that the original riser design failed to achieve the intended feeding efficacy for the steel casting process. The full solidification cycle lasted 12,513 seconds, consistent with the slow cooling expected for large steel casting components.
4.3.3 Defect Prediction Results
Using the Niyama criterion implemented in ProCAST, I obtained the defect prediction for the original process scheme. The results indicated four distinct areas of macro-porosity concentrated at the riser-casting transition zones, with shrinkage porosity distributed across several locations in the steel casting body. The total volume of shrinkage defects accounted for approximately 13% of the casting volume. The root cause analysis identified that premature closure of feeding channels due to improper temperature gradients prevented effective liquid metal supply to the solidifying sections.
To further quantify the issue, I selected five monitoring points at different locations of the steel casting. Temperature readings and solid fraction evolutions were tracked at these points. For the monitoring points located at thick-wall sections, the cooling rates were significantly lower than those at the center of the steel casting. This resulted in the solidification starting much earlier at the center points than at the thick-wall sections, demonstrating an undesirable reverse temperature gradient that deteriorates feeding conditions in the steel casting.
4.3.4 Actual Casting Verification of Original Scheme
In parallel with simulation, actual casting trials were conducted using the original process plan. The WCB steel was melted in an intermediate-frequency furnace and refined in a VOD ladle furnace. Chemical composition analysis was conducted using direct-reading spectroscopy before pouring to ensure compliance with WCB specifications. After casting and controlled cooling, the steel casting underwent multi-stage cleaning and riser removal was carried out using plasma arc thermal-mechanical cutting.
Non-destructive testing operations were subsequently performed, including visual inspection (VT), magnetic particle testing (MT), A-type pulse-echo ultrasonic testing (UT), and penetrant testing (PT). The inspection results showed clear shrinkage defects at the thick-wall sections of the steel casting, matching the simulation predictions remarkably well. This agreement validated the reliability of the ProCAST simulation framework for predicting defects in this type of steel casting.
4.4 Process Optimization and Improved Simulation Results
4.4.1 Optimization of the Feeding System
Based on the defect analysis results, I implemented optimization measures targeting the feeding system of the steel casting. The key improvements included: 1) replacing conventional sand risers with insulating-exothermic compound riser systems; 2) adding insulation boards on top of the risers; and 3) installing two additional sets of chills in the flange interior sections.
Specifically, the optimization involved applying FT400-type exothermic sleeves of 20 mm thickness to the four main risers. The five auxiliary risers in the middle sections were redesigned with a double-layer insulation structure, and insulation boards were placed on all riser tops. As established earlier, these modifications extend the riser solidification time to 2.3 times that of conventional system, providing extended molten metal supply for feeding the steel casting.
For the flange sections where defects had been observed in the lower parts, the cooling strategy was enhanced. The chill positions and sizes were adjusted on the pattern, ensuring adequate chilling to accelerate cooling and reduce the temperature gradient between thick sections and their surroundings. These changes were implemented directly on the wood pattern to accurately reflect the new chill configuration.
4.4.2 Simulation Results after Optimization
The optimized process model was constructed in ProCAST with the new riser insulation and chill configurations. Mesh generation followed the same hierarchical approach, with the insulating sleeves meshed at 30 mm. The heat transfer coefficients at the chill-casting interface were set to 2000 W/(m²·K), while the insulating sleeve-sand mold interface was set to 200 W/(m²·K).
The simulation showed that the filling time slightly increased to approximately 105 s. The solidification progression analysis revealed substantially improved behavior. At the completion of filling (t = 105 s), the solid fraction had already reached 15.6%, with early solidification initiating at edges, thin walls, and chill areas. Throughout the subsequent solidification stages, the steel casting exhibited progressively improved directional solidification from thin sections to thick sections. The final solidification locations were all within the insulating risers, confirming that the riser system could effectively provide liquid metal to the steel casting.
The temperature field analysis showed that the chills successfully created the desired temperature gradient distribution. The riser temperature profiles exhibited the characteristic bottom-to-top increasing temperature distribution, confirming proper directional solidification of the steel casting.
4.4.3 Defect Prediction after Optimization
The Niyama criterion prediction for the optimized process showed dramatic improvement. All shrinkage defects were confined to the interior of the risers, with minimal defects at the riser platform and none in the steel casting body itself. The upper surfaces of the steel casting and other critical regions displayed no defect indications. The simulation results proved that the optimized riser and chill configuration effectively eliminated shrinkage porosity in the steel casting body. However, to achieve complete process optimization, I further explored the process parameters through orthogonal experiment design.
4.5 Orthogonal Experiment for Process Parameter Optimization
4.5.1 Experiment Design
To systematically optimize the casting process parameters, I designed an L9(3³) orthogonal experiment considering three key factors, each at three levels, as the primary influences on the steel casting quality: pouring temperature (A), pouring rate (B), and initial mold temperature (C). This experimental design approach based on the orthogonal array significantly reduces the number of simulation runs required to identify the optimal parameter combination. Table 4.3 presents the factor-level settings for the experiment.
| Level | A: Pouring Temperature (°C) | B: Pouring Rate (kg/s) | C: Mold Temperature (°C) |
|---|---|---|---|
| 1 | 1580 | 100 | 20 |
| 2 | 1570 | 90 | 25 |
| 3 | 1560 | 105 | 30 |
The L9 orthogonal array allocates these factors to columns, with one column left empty for error estimation. The nine experiment combinations are shown in Table 4.4.
| Trial No. | A | Empty | B | C | Combination |
|---|---|---|---|---|---|
| L1 | 1580 | 1 | 100 | 20 | A1B1C1 |
| L2 | 1580 | 2 | 90 | 25 | A1B2C2 |
| L3 | 1580 | 3 | 105 | 30 | A1B3C3 |
| L4 | 1570 | 1 | 90 | 30 | A2B2C3 |
| L5 | 1570 | 2 | 105 | 20 | A2B3C1 |
| L6 | 1570 | 3 | 100 | 25 | A2B1C2 |
| L7 | 1560 | 1 | 105 | 25 | A3B3C2 |
| L8 | 1560 | 2 | 100 | 30 | A3B1C3 |
| L9 | 1560 | 3 | 90 | 20 | A3B2C1 |
4.5.2 Simulation Results and Range Analysis
Each of the nine experiments was simulated using ProCAST, and the shrinkage porosity percentage (porosity ratio) was calculated based on the predicted defect volumes. Table 4.5 presents the results.
| Trial No. | A | B | C | Porosity Ratio (%) |
|---|---|---|---|---|
| L1 | 1 | 1 | 1 | 12.35 |
| L2 | 1 | 2 | 2 | 11.37 |
| L3 | 1 | 3 | 3 | 13.62 |
| L4 | 2 | 2 | 3 | 10.38 |
| L5 | 2 | 3 | 1 | 7.54 |
| L6 | 2 | 1 | 2 | 9.99 |
| L7 | 3 | 3 | 2 | 14.64 |
| L8 | 3 | 1 | 3 | 11.65 |
| L9 | 3 | 2 | 1 | 15.53 |
The range analysis (Table 4.6) was performed to determine the influence ranking of the three factors and identify the optimal level combination. The range (R) value indicates the magnitude of each factor’s effect on the porosity ratio; larger R values indicate greater influence.
| Statistics | A | B | C |
|---|---|---|---|
| K1 | 37.52 | 37.37 | 33.99 |
| K2 | 27.91 | 30.56 | 37.28 |
| K3 | 41.82 | 39.14 | 35.80 |
| k1 | 12.50 | 12.46 | 11.33 |
| k2 | 9.30 | 10.19 | 12.43 |
| k3 | 13.94 | 13.05 | 11.93 |
| R | 13.91 | 8.58 | 3.29 |
From the range analysis, I determined that the ranking of influencing factors for shrinkage porosity in this steel casting is: pouring temperature (A) > pouring rate (B) > mold temperature (C). Based on the objective of minimizing the porosity ratio, the optimal level combination was identified as A2B1C1, corresponding to a pouring temperature of 1570°C, a pouring rate of 100 kg/s, and a mold temperature of 20°C.
4.5.3 Simulation Verification of Optimal Parameters
The optimal parameter combination identified from the orthogonal experiment was implemented in the ProCAST simulation to verify its efficacy. The defect prediction results showed a significant reduction in shrinkage defect volume compared to the original scheme, with the final porosity ratio reduced to approximately 6.9%. Importantly, all shrinkage defects that remained were confined to the upper portions of the risers, completely isolated from the steel casting body structure. This confirmed that the optimized process parameters combined with the enhanced feeding system produce a steel casting with excellent internal quality.
4.6 Actual Production Verification of the Optimized Steel Casting Process
With the simulation results confirming the effectiveness of the optimized process, I proceeded to actual production trials. The casting procedure followed the same steps as the initial verification: sand mold preparation, melting of WCB steel in an intermediate-frequency furnace, refining in the VOD ladle furnace, chemical composition verification, and pouring using the optimized parameters of 1570°C pouring temperature and 100 kg/s filling rate.
After the steel casting had fully solidified and cooled to ambient temperature, mold shaking out, sand removal, and riser cutting were performed. Visual examination of the separated risers clearly showed thorough feeding action—the lower parts of the risers exhibited dense, uniform metal without signs of shrinkage cavity formation. This observation confirmed that the insulating-exothermic risers successfully maintained molten metal connectivity to the steel casting throughout the solidification process.
Non-destructive testing was then conducted on the cleaned and dressed steel casting. Compared with the initial trials, the optimized casting showed complete absence of macroscopic defects. The detection results aligned exceptionally well with the simulation predictions, demonstrating the predictive capability of the ProCAST model for this type of large steel casting. The optimized steel casting satisfied all specified quality requirements, confirming the engineering feasibility and reliability of the proposed optimization methodology.
4.7 Extension and Application to Other Steel Casting Components
To validate the broader applicability of my research methodology, I applied the same “simulation-analysis-process optimization-experimental validation” framework to additional steel casting components in the production facility. One such application involved a housing-type steel casting component with complex geometry. Mesh generation and parameter setup were performed in the same systematic manner, followed by defect prediction simulation. The simulation results showed shrinkage defects located exclusively within the riser system, confirming that the optimized casting process design would produce defect-free steel casting components.
This successful extension validates that the numerical simulation-based optimization framework developed in this research can be effectively applied across different steel casting component geometries, providing significant value to the enterprise through reduced trial-and-error costs and improved first-pass yield.
5. Economic Analysis and Benefit Assessment
The economic impact of my process optimization for the top-loaded ball valve steel casting was systematically evaluated. The indicators considered included direct savings from reduced scrap rates, increased production efficiency, energy savings, and indirect benefits from quality improvements and market competitiveness. The results are summarized in Table 5.1.
| Benefit Category | Calculation Basis | Annual Value (10,000 ¥) |
|---|---|---|
| Reduced scrap cost | 70 castings saved; residual value recovery considered | 343.00 |
| Production output increase | Cycle time reduced from 18 to 13.5 days; 125 additional castings | 2000.00 |
| Energy savings | 15% reduction in energy consumption | 10.25 |
| Direct benefits subtotal | — | 2444.25 |
| Quality premium | 4.0% price increase due to improved quality | 320.00 |
| Market share growth | 10% new orders from improved delivery | 800.00 |
| Indirect benefits subtotal | — | 1120.00 |
| Total annual benefit | — | 3564.25 |
The calculations presented in Table 5.1 demonstrate substantial economic returns from the steel casting process optimization. The improved yield rate increased from 78% to 92%, representing a 70-casting reduction in annual defective products. The shortened production cycle (from 18 to 13.5 days) enabled the production of 125 additional castings annually, contributing approximately 2000 ten-thousand-yuan in new sales. Energy consumption reduction of 15% yielded additional savings. Furthermore, the enhanced quality of the steel casting resulted in price premiums and increased market share, generating significant indirect benefits.
Beyond this specific steel casting product, the technical framework developed in this research can be readily applied to other steel casting products in the enterprise’s portfolio, creating additional economic value through improved defect control and process efficiency.
6. Conclusions and Outlook
6.1 Conclusions
Based on my comprehensive research on the defect analysis and process optimization of top-loaded ball valve steel casting using ProCAST numerical simulation, the following principal conclusions were established:
1) Process Design Completeness: The complete sand casting process for the top-loaded ball valve steel casting was designed through systematic analysis, including the bottom-gating system, 13 risers with calculated dimensions, and strategically positioned chills. This process system provided the foundation for subsequent optimization and simulation studies.
2) Simulation Accuracy Validation: The original process simulation using ProCAST successfully predicted shrinkage defects matching those subsequently observed in actual casting trials. The Niyama criterion implemented in the software effectively identified defect-prone regions in the steel casting, with defects appearing at riser-casting interfaces, indicating insufficient feeding and unfavorable temperature gradients during solidification. The numerical simulation methodology demonstrated high reliability for steel casting defect prediction.
3) Feeding System Optimization Effectiveness: The optimized process incorporating insulating-exothermic riser systems and enhanced chill placement produced markedly improved solidification behavior. The simulation results showed that all shrinkage defects were confined to riser interiors, with no defects in the steel casting body. The combined insulating riser and chill configuration successfully established the desired directional solidification pattern with bottom-to-top progression, ensuring complete feeding of the steel casting.
4) Optimal Process Parameters Identified: Through L9(3³) orthogonal experiment design and range analysis, I identified both the influence ranking (pouring temperature > pouring rate > mold temperature) and the optimal parameter combination for the steel casting process: pour temperature 1570°C, filling rate 100 kg/s, initial mold temperature 20°C. These parameters combined with the optimized feeding system reduced the porosity ratio to approximately 6.9%.
5) Industrial Validation Success: The optimized process was verified through actual production, with non-destructive testing demonstrating complete absence of macroscopic defects, thereby confirming the engineering feasibility and practical value of the simulation-based optimization methodology.
6.2 Outlook and Future Work
Despite the substantial achievements of this research, several opportunities for further improvement and extension in steel casting simulation and optimization merit consideration:
1) Material Database Enhancement: The ProCAST material database contains limited entries for specialized steel casting alloys. Expanding the database with experimentally measured thermal-physical properties (liquidus/solidus temperatures, thermal conductivity, emissivity, mechanical properties) for various steel casting alloys would minimize deviations between simulation predictions and actual production outcomes.
2) Microstructure Prediction: The current simulation framework focuses primarily on macroscopic phenomena including temperature field distribution and shrinkage defect prediction. The integration of CAFE (Cellular Automaton-Finite Element) algorithms could enable the simulation of grain nucleation and growth, providing a deeper understanding of the relationship between process parameters and the microstructural evolution of steel castings.
3) Multi-objective Optimization Framework: While the orthogonal experimental design proved effective in optimizing single quality indicators such as porosity ratio, a more comprehensive multi-objective optimization framework coupling defect minimization with mechanical property enhancement and cost reduction could be developed. Recent advances in AI-based optimization algorithms present opportunities for developing intelligent steel casting process design systems.
4) Green Manufacturing Considerations: The sustainable development of the steel casting industry will increasingly emphasize energy conservation and carbon emission reduction. The digital optimization methodology developed in this research contributes to these objectives through improving first-pass yield, reducing material waste and metal consumption, and enhancing overall resource efficiency. Future extensions of this methodology could incorporate comprehensive life-cycle assessment and energy optimization objectives.