Abstract
In the railway freight car industry, steel casting play a crucial role due to their stable load-bearing capacity, ease of mass production, and low manufacturing costs. Components such as bolsters, side frames, couplers, coupler yokes, and integral traction beams often employ steel casting. Both simulation and testing are essential tools for evaluating the reliability of these steel casting, and improving their conformity accuracy can significantly enhance the design quality. Several scholars have studied the conformity accuracy between simulation and testing of railway steel casting. This paper delves deeper into the factors affecting conformity accuracy, identifies key issues, and proposes improvement methods. By applying these methods to a bolster for an export vehicle, the conformity accuracy between static testing and simulation data for the steel casting bolster was increased to 90%.
1. Introduction
In the realm of railway freight car components, steel casting is widely used due to their robust load-bearing capabilities, scalability in production, and cost-effectiveness. Critical load-bearing parts such as bolsters, side frames, couplers, coupler yokes, and integral traction beams frequently utilize steel casting. Simulation and testing serve as vital means of assessing the reliability of these steel casting, and enhancing their conformity accuracy can significantly improve the overall design quality.
Previous studies have explored the conformity accuracy between simulations and tests of railway steel casting. Scholars like Shen Jingzhe, Zhang Liyuan, and Zhao Chunlei conducted simulation and experimental research on bolster components of freight car bogies. Liang Hao and Xu Wan, on the other hand, focused on the reliability of side frames in freight car bogies. These studies analyzed discrepancies between simulation and experimental results, summarized issues encountered during simulation, and offered suggestions regarding model simplification, mesh generation, and boundary condition application for steel casting simulations.
To further advance the conformity accuracy between simulations and tests of steel casting and promote mutual verification, this paper commences with identifying influencing factors. It analyzes various factors arising in models, experiments, and simulations, puts forth improvement strategies, and validates these strategies on a bolster steel casting for an export vehicle product. The findings indicate that the conformity accuracy between static testing and simulation data can be elevated to 90% by employing these refined methods.
2. Identification of Influencing Factors
To comprehensively understand the factors influencing the conformity accuracy between simulation and testing of railway freight car steel casting, this study analyzes these factors from five perspectives: people, machine, material, method, and environment.
2.1 Human Factors
Human factors primarily refer to the impact of operational processes on the conformity accuracy. These impacts can be mitigated by standardizing operational procedures.
2.2 Machine and Environmental Factors
Machine and environmental factors are external and objective, which generally involve the precision and stability of testing equipment and the environmental conditions during testing.
2.3 Material Factors
Material factors manifest as discrepancies between physical steel casting and their design models. These differences arise from the complex manufacturing process, including design, tooling, and production, which introduces deviations from the initial design.
2.4 Methodological Factors
Methodological factors concern whether the simulation method can accurately replicate the testing process. This includes the simulation model’s ability to reflect the actual physical and mechanical properties of the steel casting, as well as the appropriateness of boundary condition settings and load application methods.
3. Analysis of Influencing Factors
3.1 Differences Between Physical Steel Casting and Design Models
The manufacturing process of steel casting is intricate, involving design, tooling, mold production, and production in the workshop, all of which contribute to the evolution of the model.
3.1.1 Design to Casting Process
After completing the three-dimensional structural design, the design is converted into a two-dimensional drawing for the casting process. Based on the two-dimensional design drawing and considering factors such as product structure, steel water flow characteristics, and liquid metal cooling characteristics, the casting process engineer sets process parameters, determines parting surfaces and core surfaces, sets draft angles, adds reinforcing ribs, specifies tolerance ranges from the design drawing to specific tolerance values, according to core surfaces and draft angles. Additionally, the casting process engineer designs the casting gating and riser system, anti-deformation amounts, machining allowances, material shrinkage rates, etc., resulting in a two-dimensional casting process drawing.
3.1.2 Mold Production
The mold manufacturing unit converts the two-dimensional casting process drawing into a three-dimensional model according to the process settings. Considering the shrinkage ratio of the steel casting material, the three-dimensional model is enlarged before being used to design and manufacture the mold.
3.1.3 Casting Production
The foundry receives the manufactured metal mold and proceeds with molding, coring, mold assembly, smelting, pouring, and cleaning, resulting in the initial rough cast model. Due to the properties of the molding sand and other production factors, the rough cast model may contain errors. Therefore, a scribing inspection is conducted on the rough cast model, and the mold dimensions are adjusted according to the inspection results before further manufacturing. The qualified rough cast model then undergoes machining to produce the final product.
Throughout this process, the produced physical model deviates from the design drawing due to the evolution that occurs. This discrepancy is a comprehensive result of process control, tooling, manufacturing settings, as well as material and production conditions, reflecting the experience and expertise of technical and production personnel at each stage.
The simulation model utilized is the design three-dimensional model, which embodies nominal design dimensions. In contrast, the three-dimensional model from the mold manufacturing stage most closely resembles the physical model, differing only by the material shrinkage ratio. If this mold manufacturing stage three-dimensional model is employed in simulations, it can significantly reduce the gap between the physical object and the simulation object, thereby enhancing the conformity accuracy between simulation and testing.
3.2 Experimental Details
To improve the conformity accuracy between testing and simulation, experimental data must meet requirements for repeatability, linearity, and symmetry. A common issue with experimental data is asymmetry, which can lead to misinterpretation of test results and affect conformity accuracy.
3.2.1 Strain Gauge Placement
When placing strain gauges on steel casting specimens, attention must be paid to the position and directional accuracy of the strain gauges as well as the bonding location. Positional and directional accuracy can be aided by scribing, while the surface quality of the specimen is generally ensured by grinding to remove the casting skin. In cases of casting pores or larger surface defects, deeper grinding may be necessary, or the positions of the strain gauge and symmetrical gauge may need to be adjusted.
3.2.2 Specimen Placement
When placing the test specimen, it is crucial to ensure the accurate relative positional relationship between the testing equipment, test tooling, and the test specimen. A laser level can be used for positioning assistance, and the symmetry planes of the equipment, tooling, and workpiece should be aligned to minimize positional errors. If there are movable parts, it is recommended to scribe positioning auxiliary lines on them before placement.
3.2.3 Form and Position Tolerance of Specimens
Due to factors such as surface flatness or parallelism, steel casting specimens may exhibit local warping when placed on tooling, causing asymmetry in test data. Adjustments are necessary during testing to ensure close contact between the equipment, tooling, and workpiece.
3.3 Simulation Methods
The simulation model must accurately reflect the relationship between the equipment, tooling, and test specimen. It is essential to analyze the load transfer path in the test, correctly set the boundary conditions of the model, and utilize appropriate simulation techniques to replicate the effects of equipment and tooling. The correctness of the simulation method settings can be verified through the displacement and stress trends in the simulation results. Avoid simplifying the simulation by altering the stiffness of the test specimen, such as directly applying loads and constraints to the test specimen. Instead, model the tooling in contact with the test specimen to accurately reflect its stiffness’s impact on the local stress distribution of the test specimen, set reasonable contact friction coefficients, and solve using nonlinear methods.
Select appropriate material parameters for the steel casting. Standards such as TB 1335 specify an elastic modulus of 172 GPa for steel casting, while TB 3548’s appendix B stipulates 200 GPa, and GB 50017-2003’s steel structure design code specifies 206 GPa. Based on physical sampling and tensile tests of steel casting, the fitted elastic modulus data is approximately 200 GPa. Therefore, it is recommended to use 200 GPa as the elastic modulus for steel casting in simulation models, with discrete elements choosing either 6 mm hexahedral elements or 8-12 mm ten-node tetrahedral elements.
4. Verification of Conformity Accuracy Improvement for Steel Casting Bolsters
4.1 Case Study: Export Vehicle Bolster
This section verifies the improvements in conformity accuracy between testing and simulation using a bolster for an export vehicle product as an example.
4.1.1 Physical Testing and Impact Analysis
Industrial Computed Tomography (CT) inspection technology facilitates internal structural dimension measurements of steel casting products. Considering testing costs, a CT scanning plan targeting key cross-sections of the bolster was determined (as shown in Figure 3, with gray thin lines indicating CT scan cross-section locations).
By measuring the dimensions of various key cross-sections of the bolster and comparing them with the product drawing, differences between the actual bolster specimen and the design drawing were recorded. The following distinctions were observed:
- The measured thickness of the vertical lower wall of the bolster was greater than the nominal size on the drawing.
- The measured thickness of the vertical upper wall was smaller.
- The thickness of both side walls was slightly less than the nominal size.
- Additional process ribs were added internally between the side walls and the bottom wall.
- The internal rib shape and middle rib shape exhibited明显的拔模斜度, with rib thicknesses significantly greater than the nominal size on the drawing.
- The weight of the bolster was 10 kg heavier than indicated on the drawing.
Based on the CT test results of the physical bolster, a digital model of the sample was established. Using the same simulation method, a comparison of the heart pan vertical working condition was conducted between the sample digital model and the design digital model. The displacement values differed by 3%, with the sample digital model results being slightly smaller. For the key stresses at the vertical middle part and part A of the bolster, the average stress difference was 8.9%, with the CT model stress results being slightly smaller.
4.1.2 Adjustment of Experimental Details
During the static testing of this bolster, asymmetric strain gauge readings were observed at part A on both sides of the bolster due to warping of the bolster spring bearing platform. A scribing inspection was conducted on the bolster bearing platform:
- The flatness and post-machining relative parallelism of both spring bearing platform surfaces to the heart pan surface were measured.
- The inspection results are presented in Table 1.
| Position | Parallelism (mm) | Flatness (mm) | Flatness (Pit) (mm) |
|---|---|---|---|
| Side (Left) | 1.0 | 0.5 | 1.2 |
| Side (Right) | 1.2 | 0.8 | 1.3 |
According to the AAR M-210 standard, the allowable warping for bolster spring bearing platforms is a flatness not exceeding 3.175 mm. The TB 3012 standard specifies that the flatness of bolster spring bearing surfaces and side frame spring bearing surfaces should not exceed 3.5 mm. The parallelism between the spring bearing surfaces at both ends of the bolster relative to the heart pan mounting surface is 3 mm, and the parallelism between the spring bearing surfaces at both ends is 4 mm. The inspection results indicate that the product meets the design standards; however, as the bolster spring bearing platform serves as the primary constraint position, its contact state with the tooling directly influences the stress amplitude and trend near the bearing platform. Closely contacted areas exhibit higher stresses, while non-closely contacted areas show lower stresses. Using lead plates of different thicknesses to adjust the contact between the workpiece and tooling improved the symmetry of the results. This example demonstrates that even products meeting design standards may require local adjustments during testing to obtain better test results.
4.1.3 Simulation Settings
The static testing equipment for the bolster is a 5000 kN four-column pressure testing machine, consisting of a base worktable, bottom loading cylinder, and top fixed plate. The bolster test tooling comprises a tooling base, rotating shaft, pads, lead plates, and a loading heart pan. The simulation model is shown in Figure 4, with settings detailed in Table 2.
| Test Equipment and Tooling Function | Simulation Realization |
|---|---|
| Loading heart pan to transfer load | Establish heart pan model and apply contact |
| Lead plates to adjust tooling-specimen gap | Contact friction coefficient |
| Tooling base and pads to transfer load | Constraint elements |
| Rotating shaft and horizontal shaft for rotation | Release rotational freedom |
| Base tooling table to transfer load | – |
| Bottom cylinder of tooling table to apply load | Constraint elements, apply upward vertical force to central node |
| Top fixed plate to limit displacement | Apply vertical displacement constraint to the entire plane of the heart pan |
4.1.4 Verification Results
Using the improved testing and simulation methods, a conformity verification study was conducted on the bolster product. A total of 99 strain gauge points were arranged at positions such as the middle of the lower wall, part A of the lower wall, and the middle hole of the bolster (as shown in Figure 5).
Simulations were performed using both the design model and the CT model. The displacement conformity between the test and simulation results reached 95%. The relative error percentages of the simulation and test stresses followed a normal distribution, with an average relative error of 9.6% and a standard deviation of 2.8%, resulting in an overall accuracy of 90.4%. Considering the differences between the physical object and the design model, the conformity between simulation and test data is considered quite satisfactory, validating the reasonableness of the proposed methods.
5. Conclusion
This study on enhancing the conformity accuracy between simulation and testing of railway freight car steel casting reveals that the test specimen, experimental details, and simulation methods have the most significant impact on conformity accuracy.
Due to the complexity of the manufacturing process, the dimensional information feedback from steel casting is comprehensive result of various process and manufacturing factors, leading to discrepancies with the design model. These discrepancies, once statistically analyzed, can be referenced during design. Among the evolutions of steel casting models, the three-dimensional mold model most closely resembles the physical model, suggesting its potential use in the simulation process to achieve more ideal results.
Accurately reflecting the test setup in the simulation model is crucial for improving conformity accuracy between the two, and the impact of experimental operational details on data cannot be overlooked. Strict control over experimental details is essential to ensure data rationality.
The research methods proposed in this paper have been validated on bolster steel casting and provide a solution pathway for similar issues in other large steel casting for railway freight cars.