Abstract:
This study focuses on the high tendency of shrinkage holes and porosities in high manganese steel ball mill liners produced by lost foam casting. Utilizing ProCAST software, we simulated the filling and solidification processes for three different casting processes: top-gate pouring with eight liners per box (Process A), stepped side-gate pouring with ten liners per box (Process B), and side-gate pouring with four liners per box including feeding risers (Process C). The casting defects of these three processes were evaluated to identify the optimal casting approach.
1. Introduction
Ball mill liners are core components of grinding mills, primarily protecting the mill cylinder from the impact and wear caused by grinding media and materials. The performance of liners directly influences the service life of ball mills and the safety and reliability of mill operations [1-5]. High manganese steel (ZGMn13) is commonly used for ball mill liners due to its excellent plasticity, toughness, and work-hardening properties under high impact or compression loads [6-9]. However, compared to conventional steels, high manganese steel has a high shrinkage rate, low thermal conductivity, and a tendency to solidify in a pasty manner, which leads to a high risk of shrinkage defects [10-11].
Lost foam casting offers high design flexibility, precise casting dimensions, low production costs, and minimal environmental pollution compared to traditional casting methods [12-14]. However, determining an optimized production process for lost foam casting requires extensive experimental research or extensive experience, which is both time-consuming and labor-intensive [15]. ProCAST finite element software can effectively simulate the filling and solidification processes of lost foam casting, predict internal defects such as shrinkage holes and porosities, and assist in designing ideal casting processes [16-17].
2. Materials and Methods
2.1. Three-Dimensional Model and Simulation Parameter Settings
2.1.1. Liner Dimensions and Casting Processes Comparison
The liner castings are approximately rectangular prisms with large plane dimensions of approximately 590 mm × 340 mm. The working surface is shaped like a single large wave with a peak, i.e., thicker at the center and thinner on both sides, with the thickest part being 120 mm and the thinnest part being 80 mm. There are two through-holes at the thickest position of the liner, with an oval through-hole on the working surface and a φ150 mm through-hole on the outer surface. To ensure smooth filling of the liner mold with high manganese steel molten metal, a pouring system with few turns and low resistance should be designed, ensuring reasonable flow direction and heat distribution in all parts of the liner. Based on this, three lost foam casting processes were designed.
Table 1. Chemical Composition of High Manganese Steel Liners
| Element | Content (%) |
|---|---|
| B | 0.005 |
| C | 1.40 |
| Mn | 13.35 |
| P | 0.038 |
| Si | 0.70 |
| Cr | 2.10 |
| Ni | 0.03 |
| V | 0.025 |
| Al | 0.005 |
| Fe | Balance |
2.1.2. Thermophysical Properties
The liner material is ZGMn13 steel. The thermophysical properties of ZGMn13 were calculated using the Database module in ProCAST software. The Back-diffusion solidification module was selected with a cooling rate of 5 ℃/s. The calculated density, enthalpy, thermal conductivity, and solid fraction as a function of temperature.
The expandable polystyrene (EPS) mold has a density of 25 kg/m³, a specific heat capacity of 3.7 kJ/(kg·K), a latent heat of 100 kJ/kg, and a thermal conductivity of 0.15 W/(m·K). The foam begins to gasify within a temperature range of 330-350 ℃. Resin-bonded sand was selected as the molding sand, with a density of 1,520 kg/m³, a specific heat capacity of 1.22 kJ/(kg·K), a thermal conductivity of 0.53 W/(m·K), and an air permeability of 1e-7.
2.1.3. Mesh Generation and Process Parameter Settings
The three-dimensional CAD models were imported into ProCAST software, and the liners for the three different processes were meshed in the Visual-Mesh module. The smallest mesh was near the liner grooves, with a size of 5 mm, and the largest mesh was at the sand box, with a size of 100 mm. The resulting volume mesh, with the sand box hidden.
Process A’s entire model contains 1,352,871 tetrahedral mesh elements; Process B has 1,575,216 tetrahedral mesh elements; and Process C has 1,181,657 tetrahedral mesh elements.
2.2. Simplification of Heat Transfer at the Filling Front in Lost Foam Casting
2.3. Criteria for Predicting Shrinkage Defects
The POROS criterion and Niyama criterion embedded in the software were used to predict the distribution of internal defects in castings under different casting processes.
3. Simulation Results Analysis
3.1. Filling Process
During the filling process, gas pressure in the gas gaps significantly hindered the filling of molten metal. For Process A, the flow of molten metal became turbulent within the three liners. By 44.90 s, 90% of the filling was complete, with the last filled areas being the bottoms of four liners.
For Process B, the molten metal began to enter the liner castings from the sides in a stepped manner at 14.95 s (30% filling). By 35.81 s, 90% of the filling was complete, and due to the formation of gas gaps, the filling ability of the molten metal significantly decreased, causing severe turbulence in the filling process of two liners, which were the last to be filled.
For Process C, the molten metal began to enter the liner castings from the sides at 6.94 s (30% filling). By 21.53 s, 70% of the filling was complete, with smooth metal flow and no abnormalities observed. By 27.72 s, 90% of the filling was complete, with the last filled area being the bottom of the liners farthest from the pouring gate. Due to the fewer liners being poured simultaneously, the gas produced quickly escaped the EPS mold, resulting in a smooth filling process without turbulence.
The comparison of the filling processes showed that turbulence occurred in the flow of molten metal during Processes A and B. The reason may be the complex heat exchange between the molten metal and the EPS mold, causing changes in the flow velocity of the molten metal. Additionally, in the later stages of filling, gases produced by the decomposition of the EPS mold were not timely removed, affecting the flow of molten metal and causing turbulence. In Process C, the molten metal flowed smoothly without turbulence.
3.2. Comparison of Solidification Processes
To facilitate the observation of the solidification process of the liner castings, parts of the castings with a solid fraction greater than 70% were hidden. The solidification processes of the liners for different processes. The results indicate that the solidification of the liners in all processes follows a sequential pattern, starting from the periphery and moving towards the center. For Processes A and B, which lack feeding risers, there is no feeding function for the central area. Each liner begins to solidify from the edges and gradually progresses towards the interior, ultimately resulting in hot spots at the center of the castings. In contrast, Process C incorporates feeding risers, which can provide feeding to the central area before the risers solidify, reducing the occurrence of shrinkage cavities and porosity defects. However, as solidification progresses, hot spots also emerge in the center of the castings for all processes.
During the solidification process, the temperature distributions of the liner castings under various processes. It is evident that the temperature of all liners gradually increases from the outside to the inside, with the lowest temperatures at the edges, leading to the initiation of solidification from the periphery. For Process A, the temperature distribution across the liners is relatively uniform, with the highest temperatures at the center, and there are temperature differences between the four liners,. In Process B, the temperature distribution varies more significantly, particularly in the center of each liner, which may be attributed to the significant turbulence during the filling process. Additionally, during solidification, both Processes A and B exhibit isolated liquid regions at the center, which will result in shrinkage cavities and porosity defects after solidification [21].
For Process C, there is no significant variation in the temperature field distribution across the liners, with only small isolated liquid regions at the center. After solidification, only a small amount of shrinkage cavities and porosity defects are present in the center of the liners, and the temperature fields at the center of each liner are consistent. As solidification continues, the overall temperature of each liner gradually decreases, but the center temperature remains higher than the temperatures at the edges throughout.
All liners exhibit casting defects internally due to shrinkage during the solidification process, where the contraction of liquid metal during cooling and solidification exceeds that in the solid state, and the shrinking areas do not receive sufficient liquid metal for feeding. Therefore, shrinkage cavities and porosity defects are present within the liner castings of all three pouring processes [22]. The larger isolated liquid regions in the center of the middle liners lead to more shrinkage cavities and porosity defects during solidification.
The predictions of shrinkage cavity defects based on the POROS criterion for different casting processes. It can be observed that all liners exhibit shrinkage cavity and porosity defects under the three processes. The defects in Process A are concentrated at the center, and the defect distribution in the middle four liners is closer to the surface. In Process B, the G/R ratio is smaller on the surfaces of the middle six liners and the inner surfaces of the outer four liners, increasing the likelihood of shrinkage cavity and porosity defects. For Process C, a small number of defects within the liners are concentrated in the center, with no defects on the liner surfaces.
In summary, by utilizing ProCAST software to simulate the filling and solidification processes of three different lost foam casting processes for high manganese steel liners, this study provides insights into the solidification behavior and defect formation mechanisms. The results indicate that Process C, with its side-gating and feeding riser design, exhibits the most stable filling process and the most concentrated defect distribution, making it the optimal choice among the three processes.