1. Introduction
During my graduate research, I focused on the optimization of the lost foam casting (LFC) process for a cast steel valve body. The foundry industry has increasingly adopted LFC due to its design flexibility, high dimensional accuracy of castings, and lower production costs, especially among small and medium-sized enterprises. However, when casting steel components using the LFC process, defects such as carburization, shrinkage cavity, and shrinkage porosity frequently occur. These defects are closely related to the mold filling and solidification processes. Therefore, employing numerical simulation techniques to analyze the filling and solidification principles, as well as to predict casting defects, is of significant guiding importance for the production practice of steel castings.
In this study, I utilized Unigraphics NX.8.0 as the three-dimensional modeling software and ProCAST as the numerical simulation software to analyze and improve the lost foam casting process of a cast steel valve body produced by a specific factory. The material used for this valve body was ZG230-250 (also known as ZG25), which is primarily used for regulating fluid flow and has strict requirements for air tightness. However, the produced cast steel valve bodies exhibited leakage phenomena, failing to meet the technical air tightness requirements, with a rejection rate exceeding 60%. To investigate this issue, I conducted a thorough examination of the products through sectioning and discovered severe shrinkage cavity and shrinkage porosity defects inside the castings. My analysis concluded that these internal shrinkage defects were the primary cause of the air tightness failure.
2. Overview of Lost Foam Casting
2.1 Principles and Characteristics of LFC
Lost foam casting is an advanced casting technique that evolved from vacuum-sealed molding combined with dry sand full mold casting. The principle involves using a foam pattern, made from expandable polystyrene (EPS) beads, which is pre-expanded, aged, and then molded into the desired shape. The foam pattern, after being coated with refractory paint and dried, is placed in a sand box. Dry sand is then filled and compacted around the pattern using vacuum and vibration. When molten metal is poured into the mold cavity, the foam pattern rapidly softens, melts, and gasifies upon contact with the high-temperature metal. The decomposition products are evacuated through the coating under the action of negative pressure within the sand box. The molten metal replaces the space occupied by the foam pattern, and after solidification and cooling, a casting with the desired shape is formed.
The lost foam casting process offers several significant advantages over traditional casting methods:
- Simplified process steps, fewer operations, and shorter production cycles
- Reduced labor intensity and improved working environment
- High dimensional accuracy of castings, with tolerances as shown in the table below
- Enhanced design freedom for component structures
- Better casting quality and lower rejection rates
- Convenient placement of risers and higher utilization of molten metal
- Easy to master process techniques and convenient production management
- Lower investment costs and quicker returns
| Casting Dimension (mm) | Tolerance (mm) |
|---|---|
| <25 | ±0.17 |
| 25～80 | ±0.25 |
| 80～125 | ±0.30 |
| 125～175 | ±0.43 |
| 175～250 | ±0.003 |
| >250 | ±0.002 |
2.2 Development History of LFC
The development of lost foam casting technology can be traced through three distinct phases. The first phase began in 1956 when American inventor H.F. Shroyer patented the full mold casting method, known as the F method. This technique involved using foam patterns with bonded sand. In 1961, Professor A. Wittmoser collaborated with several plastic companies and foundries to commercialize this technology, and by 1970, the United States had produced approximately 400,000 tons of castings using this method.
The second phase introduced dry sand full mold casting technology. In 1964, German researcher H. Nnellen and colleagues applied binder-free dry sand to casting production, though mold collapse issues frequently occurred. Subsequently, R. Hoffmann in Germany proposed the use of magnetic fields as a binding agent, leading to the patent of magnetic molding, referred to as the M method.
The third phase built upon the M method, combining its advantages with those of the vacuum-sealed molding method (V method), invented in 1969 by the Akita Corporation and the Nagano Prefectural Industrial Testing Institute in Japan. The V method used plastic film and vacuum negative pressure to compact molding sand, solving the collapse problems encountered in the M method. The combination of these three techniques, incorporating foam patterns, binderless dry sand, and vacuum compaction, formed the foundation of modern lost foam casting. During the mid-1970s, countries including the United States, Italy, the United Kingdom, Germany, and the former Soviet Union, as well as Japan, successively developed and implemented LFC production lines.
In China, research on full mold casting began in the 1960s at the Shanghai Machinery Manufacturing Technology Research Institute. Initial success was achieved with the first large cast steel component weighing 11 tons. Since the 1990s, Chinese manufacturers began introducing foreign LFC production lines and equipment, significantly advancing the overall level of the domestic LFC industry. By 2007, China had become a major producer of LFC castings, with annual output reaching approximately 648,000 tons.
3. Numerical Simulation Technology in Casting
3.1 Global Research Status
The numerical simulation of casting processes began in the 1960s with the development of computational methods for heat transfer analysis. In 1962, Danish researcher K. Forsund proposed the use of finite difference approximation for heat transfer calculations during solidification. In 1965, J.H. Henzel and K. Keverian of General Electric simulated the casting of a steel turbine housing, incorporating transient heat transfer programs. During the 1970s, the University of Michigan research group led by R.D. Pehlke and Marrone successfully simulated the solidification of various castings using alternating implicit difference and Saulyev explicit difference methods.
The numerical simulation of fluid flow during mold filling began in the early 1980s. Taiwanese scholar W.S. Hwang and R.A. Stoehr pioneered the application of computational fluid dynamics to solve casting filling problems. By 1983, researchers including Hwang were using MAC, SMAC, and SOLA-VOF methods to simulate two-dimensional and three-dimensional mold filling processes. These simulations were validated against water modeling experiments and high-speed photography, showing good agreement.
The advancement of commercial simulation software began in 1989 with the introduction of the MAGMA system by Professor Sahm and colleagues in Germany. This was followed by the development of other commercial packages including ProCAST from the United States, Cast-CAE from Finland, and various domestic Chinese software systems such as FOUNDRY STAR from Tsinghua University and HuaZhu CAE from Huazhong University of Science and Technology.
3.2 Simulation Software Comparison
| Feature | ProCAST | MAGMA Soft | HuaZhu CAE |
|---|---|---|---|
| Geometry Modeling | Poor built-in modeling, requires external CAD | Complete CAD system | Requires external CAD software |
| Analysis Method | Finite Element Method (FEM) | Finite Difference Method (FDM) | Finite Difference Method (FDM) |
| Computational Time | Longer, supports multiprocessors | Shorter | Shorter |
| Analysis Capabilities | Filling, solidification, flow, temperature, stress, electromagnetic | Filling, solidification, flow, temperature, stress | Filling, solidification, flow-temperature coupling |
| Analysis Range | Sand casting, LFC, high-pressure, gravity casting, investment casting | Sand casting, permanent mold, gravity, high-pressure die casting | Sand casting, permanent mold, shell casting, investment casting |
| Material Database | Steel, iron, aluminum, cobalt, copper, magnesium, titanium alloys | Steel, iron, aluminum, copper, magnesium | Steel, iron, aluminum, copper |
4. Numerical Simulation Methodology for Lost Foam Casting
4.1 Mold Filling Mechanism in LFC
Unlike conventional casting processes, lost foam casting retains the foam pattern inside the mold cavity. When molten metal is poured, the pattern undergoes softening, melting, gasification, and combustion upon exposure to high temperatures, involving a series of complex phase transformations. The presence of the foam pattern creates several complex phenomena during filling:
- Multi-mode heat transfer (conduction, convection, radiation) within the gas gap at the liquid metal flow front
- Physical and chemical reactions between the decomposition products and the melt, coating, and sand
- Evolution of gas pressure in the gap between the pattern and molten metal, affecting the flow front temperature and velocity
These phenomena make the mathematical modeling of LFC complex. The essential consideration is that the gas film pressure formed by pattern decomposition significantly influences the flow front and affects both process parameter selection and final casting quality.
4.2 Factors Affecting Mold Filling in LFC
Several key factors influence the mold filling process in lost foam casting:
Foam Pattern Properties: The type and density of the foam pattern determine the heat absorbed during decomposition and the quantity of decomposition products. Higher pattern density increases heat absorption and decomposition product generation, which reduces the temperature at the metal flow front and increases gas pressure resistance, thereby slowing filling speed and reducing filling capacity.
Vacuum Degree: Applying vacuum to the sealed sand box facilitates the removal of gaseous decomposition products, releasing the back pressure at the liquid metal flow front and improving filling efficiency. However, excessive vacuum can cause the metal front to preferentially fill along the pattern-mold interface, creating a concave front that may enclose the foam pattern and slow down the filling process.
Pouring Temperature: The pouring temperature affects the type, quantity, and removal rate of decomposition products. For different alloys, the optimal pouring temperature varies:(a) For aluminum and magnesium alloys at 700-800°C, decomposition products are primarily liquid; increasing temperature reduces product viscosity and improves coating wettability, facilitating product removal. (b) For cast steel and cast iron above 1500°C, decomposition products are primarily gaseous; higher temperatures increase gas volume without increasing removal rate, resulting in higher gas pressure resistance and reduced filling capacity.
Coating Properties: The permeability, wettability, and thermal insulation of the coating directly affect filling capacity. Poor coating performance impedes the escape of gaseous and liquid decomposition products, increasing gas pressure and residue in the metal.
4.3 Mathematical Models
The flow of molten metal during LFC filling is treated as viscous, incompressible Newtonian fluid flow, governed by the conservation equations of mass and momentum, coupled with the energy equation.
The continuity equation for incompressible flow:
$$\frac{\partial u}{\partial x} + \frac{\partial v}{\partial y} + \frac{\partial w}{\partial z} = 0$$
The Navier-Stokes equations for momentum conservation in the x, y, and z directions:
$$\frac{\partial u}{\partial t} + u\frac{\partial u}{\partial x} + v\frac{\partial u}{\partial y} + w\frac{\partial u}{\partial z} = -\frac{1}{\rho}\frac{\partial P}{\partial x} + g_x + \gamma\left(\frac{\partial^2 u}{\partial x^2} + \frac{\partial^2 u}{\partial y^2} + \frac{\partial^2 u}{\partial z^2}\right)$$
$$\frac{\partial v}{\partial t} + u\frac{\partial v}{\partial x} + v\frac{\partial v}{\partial y} + w\frac{\partial v}{\partial z} = -\frac{1}{\rho}\frac{\partial P}{\partial y} + g_y + \gamma\left(\frac{\partial^2 v}{\partial x^2} + \frac{\partial^2 v}{\partial y^2} + \frac{\partial^2 v}{\partial z^2}\right)$$
$$\frac{\partial w}{\partial t} + u\frac{\partial w}{\partial x} + v\frac{\partial w}{\partial y} + w\frac{\partial w}{\partial z} = -\frac{1}{\rho}\frac{\partial P}{\partial z} + g_z + \gamma\left(\frac{\partial^2 w}{\partial x^2} + \frac{\partial^2 w}{\partial y^2} + \frac{\partial^2 w}{\partial z^2}\right)$$
where u, v, w are velocity components (m·s⁻¹), P is pressure (Pa), g are gravitational acceleration components (m·s⁻²), ρ is the density of the fluid (kg·m⁻³), and γ is the kinematic viscosity (m²·s⁻¹).
The energy equation:
$$\rho_L C_{PL}\frac{\partial T}{\partial t} + \rho_L C_{PL}\mu_L\frac{\partial T}{\partial x} = \lambda_L\frac{\partial^2 T}{\partial x^2} + S$$
where ρₗ is the density of the liquid metal, Cₚₗ is the specific heat of the liquid metal (kJ/(kg·°C)), λₗ is the thermal conductivity (W/(m·°C)), T is the temperature, u is the flow velocity in the x direction, and S is the source term related to the latent heat.
For the free surface heat transfer boundary condition, the energy balance is:
$$k\frac{\partial T}{\partial n} = \rho_p L_p u$$
where k is the heat transfer coefficient, ∂T/∂n is the temperature gradient in the normal direction to the free surface, ρₚ is the pattern density, Lₚ is the latent heat of the pattern material, and u is the flow velocity normal to the free surface.
The pressure boundary condition related to gas evolution and coating permeability is represented by:
$$P_g = \frac{\rho_p Q T_g V_0 T_0}{p_0 V}$$
where Q is the gas evolution rate of the pattern (m³/kg), Tg is the gas gap temperature, V is the gasified pattern volume, p₀ is the standard atmospheric pressure, T₀ is the room temperature, k is the coating permeability, δ is the coating thickness, and the summation of contributions accounts for the gas back-pressure effect on the filling front.
Additionally, researchers like Wei Zunjie have developed gas gap pressure calculation models that account for the gas pressure and gap size as functions of time:
$$P_i = \frac{\rho_p\alpha_p V_i(T_{mi} – T_{pi})}{S L_p} – \frac{F k_g \delta_i(P_i^2 – P_{i-1}^2)}{2 R T_i \eta_g \Delta t}$$
where various parameters account for the gas pressure evolution in the gap and its effect on the filling process.
4.4 Numerical Methods for Casting Simulation
Several numerical methods have been developed for simulation of casting filling processes:
SIMPLE Method: Proposed by S.V. Patankar and Spalding, it iterates between pressure and velocity fields to solve the Navier-Stokes equations. However, it has limitations in handling free surface flows efficiently.
MAC and SMAC Methods: Developed by Harlow and Welch in 1965, these methods used marker particles to track the free surface. SMAC improved upon MAC by using a potential function to enhance computational efficiency. However, the large number of marker particles required for three-dimensional simulation limits its application.
SOLA-VOF Method: Developed at the Los Alamos National Laboratory, this method uses a volume-of-fluid function F to determine the free surface position. The value of F indicates the fraction of fluid filling each grid cell: F=1 for full cells, 0<f<1 accuracy.
5. ProCAST Simulation Software
ProCAST, developed by UES Inc., USA, since 1985, has become a comprehensive computer-aided engineering system for simulation and analysis of casting processes for various alloys. The latest versions of ProCAST, including the 2016 release, provide enhanced capabilities for improving casting quality and reducing defects. The main modules of ProCAST include:
- Visual-Mesh: The grid generation module that interfaces with CAD software and automatically generates finite element meshes. It supports various CAD formats including IGES, STEP, STL, and Parasolids.
- Basic Module (Heat Transfer): Includes pre-processing (Visual-Cast) for setting initial and boundary conditions, and post-processing (Visual-Viewer) for displaying results through contour plots, vectors, cross-sections, and animations.
- Flow Solver: Solves full three-dimensional Navier-Stokes equations with modules for turbulent flow, compressible flow, lost foam modeling, gas entrapment, and various other physical models.
- Stress Analysis Module: Incorporates thermal-stress coupling with elasto-plastic and elasto-viscoplastic material models.
- Microstructure Module: Predicts porosity, grain structure, and solidification morphology.
- Inverse Module: Determines boundary conditions and material properties from experimental data for improved simulation accuracy.
ProCAST can predict various casting defects including:
- Shrinkage cavity and porosity, using criteria such as critical solid fraction, Niyama method, and direct analysis method
- Hot cracks and residual stresses through thermal stress analysis
- Gas entrapment during mold filling
- Cold shuts and mistun due to improper pouring parameters
6. Numerical Simulation Analysis of the Original Cast Steel Process
6.1 Valve Body Characteristics and Original Process
The valve body under investigation is a butterfly valve body used in pipelines for regulating media flow including gas, natural gas, water supply, and liquefied petroleum gas. The cast steel valve body must pass a pressure test at 5 MPa to ensure internal density and prevent leakage. The three-dimensional model and physical casting are shown in the analysis.
The butterfly valve body has outer dimensions of 460 mm × 340 mm × 150 mm, with a weight of approximately 329 N. The material is ZG230-250 (ZG25) cast steel with the chemical composition listed in the table below:
| Element | C | Si | Mn | S | P | Cr | Ni | Cu | Mo | V |
|---|---|---|---|---|---|---|---|---|---|---|
| Content (%) | 0.23 | 0.50 | 0.63 | 0.04 | 0.034 | 0.35 | 0.07 | 0.30 | 0.20 | 0.05 |
The casting has a shell-like structure with a large central cavity and outer contour. Its longitudinal cross-section shows an I-shape with four uniformly distributed ribs on the cylindrical wall. The structure appears simple but contains multiple geometric hot spots. The wall thickness varies from 10 mm to 25 mm, and the casting requires internal density, smooth surfaces, and no casting defects such as cracks, cold shuts, sand holes, shrinkage cavity, or porosity.
The original process employed a bottom gating system with two ingates (15×10 mm cross-section), two runners (15×10 mm cross-section), and one sprue (15×15 mm cross-section, 250 mm high). Five identical dark risers were placed on top of the four ribs, with one additional riser on the small ring surface of the left side. The original process parameters were: pouring temperature of 1580°C, vacuum degree of 0.05 MPa, and EPS pattern density of 20 kg/m³.
6.2 Analysis of Casting Defects in the Original Process
The valve body produced by the factory exhibited poor air tightness, with a rejection rate exceeding 60%. To diagnose the problem, I employed the kerosene penetration test to check the density of the casting structure. Kerosene was applied to one side of the inner boss, and after standing for some hours, it was observed that kerosene had leaked to the other side. Sectioning analysis of the leak-prone areas revealed visible shrinkage cavity, shrinkage porosity, and gas porosity defects, particularly at the geometric hot spots where the large flange connects to the pipe body, as shown in the internal defect analysis.
6.3 Pre-processing for Simulation
I used ProCAST 2014 software for the numerical simulation. The steps included: (1) three-dimensional solid modeling using UG8.0 software; (2) importing the model into ProCAST in Parasolid format; (3) creating a sand box and assembling the model components; (4) performing geometry check and repair; (5) generating surface meshes with appropriate element sizes—8mm for the casting, 10mm for the gating system, and 20mm for the sand box; (6) generating volumetric meshes and checking mesh quality. The final finite element model for the casting contained 58,360 nodes and 2,865,437 elements.
For the lost foam casting simulation, I set the following conditions:
| Component | Material | Fill Status | Initial Temperature |
|---|---|---|---|
| Gating System | Cast Alloy | 0% | Pouring Temperature |
| Casting Cavity | Foam | 100% | Room Temperature |
| Sand Mold | Sand (Permeable) | 100% | Room Temperature |
The material properties for ZG230-250 include a solidus temperature of 1430°C, liquidus temperature of 1516°C, and latent heat of 237 kJ/kg. The EPS pattern properties used were: density 20 kg/m³, thermal conductivity 0.15 W/(m·K), specific heat 3.7 kJ/(kg·K), and latent heat 100 kJ/kg. The sand properties were: density 1520 kg/m³, thermal conductivity 0.53 W/(m·K), specific heat 1.22 kJ/(kg·K), and permeability 1×10⁻⁷ cm².
For the interface conditions, the gating system and foam pattern were connected with EQUIV interface type to allow continuous temperature and flow field transfer. For the heat transfer at other interfaces, I set a heat transfer coefficient of h=500 W/m²·K. For the pressure boundary conditions, I set the pressure at the top of the pouring cup and the outer surface of the sand box, with the difference defining the vacuum degree. The foam parameters included the heat transfer coefficient between the molten metal flow front and the foam, the burn zone distance, and the gas fraction, which describes the proportion of foam decomposition products that become gaseous.
6.4 Mold Filling Simulation Results of the Original Process
The mold filling analysis revealed that the gating system was completely filled by 5.2 seconds, with the filling progressing relatively slowly. The foam pattern created resistance to metal flow and absorbed heat, lowering the metal temperature and reducing its fluidity. Complete filling of the casting was achieved at 21.79 seconds. The time intervals to fill each 20% increment of the cavity were 5.5s, 3.3s, 2.5s, 2.67s, and 2.62s, indicating a relatively uniform filling rate. The metal flowed smoothly from the sprue through the runner to the casting bottom, filling the mold progressively without severe splashing, which is favorable for the escape of gas products and floating of inclusions.
6.5 Solidification Simulation and Defect Prediction
The solidification analysis showed that the metal generally solidified from the thin sections toward the thick sections. At 70% solidification, the upper and lower portions of the casting had begun to solidify separately. By 80% solidification, the gating system was completely solidified, and the upper half of the risers had also solidified. At this point, the casting was divided into two isolated liquid regions—one near the large rotating ring surface and one at the riser roots—which could not receive metal feeding from the gating system or the risers, leading to shrinkage cavity and porosity defects in these areas.
Using both the critical solid fraction method and the direct analysis method, I predicted the shrinkage cavity and porosity defects in the casting. The simulation predicted a total shrinkage volume of 43.61 cm³, corresponding to a shrinkage porosity ratio of 21.31%. Comparison of the simulation results with the actual defects found in the production castings showed good agreement in terms of defect location and severity. This validation confirmed the accuracy of the ProCAST simulation for this steel casting lost foam casting process.
6.6 Analysis of Defect Formation Mechanism
Analysis of the solidification-time distribution and solid fraction changes at a cross-section of the casting revealed the defect formation mechanism. The original process design had two critical flaws. First, the metal was introduced from the thick sections of the casting, artificially creating a hot spot at these thick areas that slowed heat dissipation and made them the last to solidify, preventing feed metal from reaching them. Second, the risers were placed directly above the thick sections. However, the upper portions of the risers solidified prematurely due to contact with the surrounding sand, so they could not provide effective feeding before the casting sections required it. By the time the casting was 60% solidified, the riser tops had already solidified, rendering them ineffective.
7. Process Optimization of Cast Steel Valve Body
7.1 Improvement Strategy
Based on the analysis of the original process, I developed two new gating system designs to reduce shrinkage defects. From the characteristics of cast steel—which has poor fluidity and high shrinkage—for medium and small cast steel castings, a top gating system was initially considered favorable. However, considering the LFC characteristics and defect formation mechanisms, both bottom gating and top gating approaches were evaluated.
7.2 Gating System Design
The gating system dimensions were calculated based on traditional casting design methods combined with the specific requirements of lost foam casting. The choke area (smallest cross-section) was calculated using the hydraulic formula:
$$A_{choke} = \frac{0.35 G_L}{\mu \sqrt{H_p}}$$
where GL is the weight of molten metal flowing through the choke (kg), μ is the flow coefficient (0.03-0.05 for cast steel), and Hp is the effective pouring head height (cm). The pouring time t was determined by:
$$t = C\sqrt{G_L}$$
The coefficient C was selected from the table below based on the relative density of the casting:
| Relative Density ρ_r | 0-1.0 | 1.0-2.0 | 2.0-3.0 | 3.0-4.0 | 4.0-5.0 | 5.0-6.0 | >6.0 |
|---|---|---|---|---|---|---|---|
| C | 0.8 | 0.9 | 1.0 | 1.1 | 1.2 | 1.3 | 1.4 |
For this casting, GL ≈ 53 kg, the pouring head Hp = 295 mm, and μ = 0.04, giving t ≈ 5 s and Achoke ≈ 14.8 cm². Since LFC gating systems typically require 15-20% larger dimensions than traditional sand casting, I used an area of 18 cm². The gating ratio followed the open gating system proportion:
$$F_{sprue}:F_{runner}:F_{ingate} = 1:(1.1\sim1.3):(1.2\sim1.5)$$
Three ingates were designed with cross-sections of 30 mm × 12 mm, the runner cross-section was 45 mm × 15 mm, and the sprue was a cylinder with a diameter of 32 mm and a taper of 2°.
7.3 Comparison of Two Gating System Designs
Design 1 (Bottom Gating System): In this design, metal was introduced from the side at the bottom of the casting. The mold filling simulation showed that the complete filling time was 48.96 seconds. The filling rate started slowly, then increased, and finally slowed again, consistent with the foam decomposition dynamics. The solidification analysis showed that the metal solidified progressively from the areas far from the ingates toward the ingates, improving the feeding conditions for the hot spots. The shrinkage porosity rate for this design was 13.36%, significantly lower than the original process.
Design 2 (Top Gating System): In this design, metal was introduced from the top side of the casting. The complete filling time was 68.09 seconds. Although the top gating system provided faster initial filling and better temperature distribution for some areas, it had notable disadvantages. The metal flow direction was opposite to the direction of foam decomposition product escape, making it difficult for gas and inclusions to escape. The simulation revealed gas entrapment, where foam residue was isolated in a pocket surrounded by metal at the large ring surface at 91.7% filling. This could lead to gas porosity and inclusion defects. The shrinkage porosity rate for this design was 26.08%.
Comparing the two designs, the bottom gating system (Design 1) was clearly superior. It provided smoother filling, no gas entrapment, and much lower shrinkage porosity. The bottom gating system was therefore selected as the preferred design for the cast steel valve body.
7.4 Orthogonal Experiment for Process Parameter Optimization
After selecting the appropriate gating system, I conducted a three-factor, three-level orthogonal experiment using ProCAST simulation to optimize the main process parameters affecting shrinkage defects. The three factors and their levels are shown in the table below:
| Factor | Symbol | Level | ||
|---|---|---|---|---|
| 1 | 2 | 3 | ||
| Pouring Temperature (°C) | A | 1580 | 1620 | 1660 |
| Pattern Density (kg/m³) | B | 18 | 20 | 22 |
| Vacuum Degree (MPa) | C | 0.04 | 0.05 | 0.06 |
The nine experimental runs were simulated using ProCAST, and the shrinkage porosity rate was used as the evaluation index. The results are presented in the table below:
| Run No. | A: Pouring Temperature (°C) | B: Pattern Density (kg/m³) | C: Vacuum Degree (MPa) | Shrinkage Porosity Rate (%) |
|---|---|---|---|---|
| 1 | 1580 | 18 | 0.04 | 18.352 |
| 2 | 1580 | 20 | 0.05 | 17.326 |
| 3 | 1580 | 22 | 0.06 | 17.584 |
| 4 | 1620 | 18 | 0.05 | 14.021 |
| 5 | 1620 | 20 | 0.06 | 17.562 |
| 6 | 1620 | 22 | 0.04 | 16.691 |
| 7 | 1660 | 18 | 0.06 | 16.487 |
| 8 | 1660 | 20 | 0.04 | 17.186 |
| 9 | 1660 | 22 | 0.05 | 17.265 |
To analyze the influence of each factor, I performed range analysis. The K values represent the sum of shrinkage porosity rates for each factor at each level, and k values represent the average:
| Statistical Result | A: Pouring Temperature | B: Pattern Density | C: Vacuum Degree |
|---|---|---|---|
| K1 | 53.262 | 48.860 | 52.229 |
| K2 | 48.274 | 52.074 | 48.612 |
| K3 | 50.938 | 51.540 | 51.633 |
| k1 | 17.754 | 16.286 | 17.409 |
| k2 | 16.091 | 17.358 | 16.204 |
| k3 | 16.979 | 17.180 | 17.211 |
| Range (R) | 1.663 | 1.072 | 1.205 |
From the range analysis, the order of influence on shrinkage porosity rate was: pouring temperature (A) > vacuum degree (C) > pattern density (B). Since the shrinkage porosity rate should be minimized, the optimal combination was A2B1C2, corresponding to a pouring temperature of 1620°C, a pattern density of 18 kg/m³, and a vacuum degree of 0.05 MPa.
Using this optimal parameter combination, I performed an additional simulation, which resulted in a shrinkage porosity rate of 11.80%. This was even lower than the best result from the orthogonal array (Run 4 at 14.02%), confirming the effectiveness of the optimized parameters for this steel casting application.
7.5 Riser Design Improvement
Even with the optimized gating system and process parameters, some shrinkage porosity defects remained at specific locations. To further improve the steel casting quality, I redesigned the riser placement. According to the principles of directional solidification for steel castings, risers should be placed near hot spots but not directly above them, to avoid creating additional hot spots. The redesigned riser system placed risers at appropriate positions on top of the casting to promote effective feeding during the final stage of solidification.
After adding the improved risers, the solidification simulation showed that the shrinkage-prone areas no longer formed isolated liquid regions. The shrinkage porosity distribution analysis revealed that internal shrinkage defects in the casting were completely eliminated, with shrinkage porosity only appearing in the gating system. This confirmed that the final optimized process successfully produced sound castings meeting the required quality standards.
8. Conclusions
Through my research on the lost foam casting process optimization of the cast steel valve body, I have drawn the following conclusions:
- The original process using a center bottom gating system with risers placed on the top ring surface resulted in severe shrinkage cavity and porosity defects at the large rotating ring surfaces and geometric hot spots. Analysis by the critical solid fraction method and direct analysis method consistently identified these defect locations, which matched the actual production defects.
- The comparison between simulation predictions and actual production defects validated the accuracy of ProCAST simulation settings and parameters for lost foam casting of steel castings, confirming the feasibility of using numerical simulation for LFC process optimization.
- Comparing the two new gating system designs—top gating and bottom gating—the bottom gating system showed superior performance. The top gating system resulted in gas entrapment at the large ring surface and a higher shrinkage porosity rate of 26.08%, while the bottom gating system achieved smooth filling and a reduced shrinkage porosity rate of 13.36%.
- The orthogonal experiment revealed that the order of influence on shrinkage porosity defects in the cast steel valve body is: pouring temperature > vacuum degree > pattern density. The optimal process parameters were determined as pouring temperature 1620°C, pattern density 18 kg/m³, and vacuum degree 0.05 MPa, yielding the minimal shrinkage porosity rate.
- With the optimized gating system and process parameters, the addition of properly positioned risers successfully eliminated internal shrinkage defects in the cast steel valve body. The final process produced castings meeting the quality requirements for air tightness.
The application of ProCAST numerical simulation to the lost foam casting process of the cast steel valve body demonstrated that simulation not only provides insight into the filling and solidification processes but also enables the identification and resolution of casting defects at their source. This approach saves significant time and resources compared to traditional trial-and-error methods, and provides a deeper understanding of the fundamental mechanisms of the casting process. The visualization capabilities of the simulation software offer intuitive tools for engineers and technical personnel to optimize casting processes efficiently.
</f