Casting serves as the foundational technology for modern equipment manufacturing. Traditional methods for determining casting processes, heavily reliant on trial-and-error and empirical design, are often characterized by high costs, extended production cycles, and elevated scrap rates. The application of numerical simulation to the casting process presents a significant advantage in shortening product development time, minimizing casting defects, and ultimately ensuring product quality. This study focuses on addressing the production challenges associated with a turbine rear exhaust pipe. This component is a critical connector between the turbocharger and the muffler in an automotive exhaust system. During vehicle operation, it must withstand the hot exhaust gases produced by combustion, necessitating high levels of vibration resistance, wear resistance, heat resistance, and airtightness. The initial production method for this component was lost foam casting, which unfortunately resulted in low product yield, long cycle times, material waste, and field failures such as gas leakage and fracture.
The adoption of modern numerical simulation technology allows for the virtual prototyping and analysis of castings via computer simulation. This approach drastically reduces the time and cost associated with physical trials. By simulating the process, potential defects can be identified and mitigated early in the design phase, providing a scientific basis for process optimization and guiding effective engineering decisions. This research employs the ProCAST simulation software to perform a comprehensive numerical analysis of the lost foam casting process for the turbine rear exhaust pipe. The goal is to diagnose the root causes of the observed defects and propose a validated process improvement.
Geometric Modeling and Mesh Generation
The foundation of an accurate numerical simulation is a precise digital model. The three-dimensional geometry of the turbine rear exhaust pipe was provided by the manufacturing partner. Based on the component’s complex, thin-walled, and curved structure, a detailed solid model was constructed using parametric CAD software. To maximize production efficiency, the standard practice for this part is to cast multiple units in a single mold. Therefore, the complete simulation model included not only the individual casting but also the entire gating system (pouring cup, sprue, runners, and ingates) arranged in a pattern of four castings per mold box. The assembly of the expendable polystyrene (EPS) foam patterns and the gating system was completed within the CAD environment before export.
The assembled model was then imported into a pre-processing environment capable of generating a computational mesh. For complex casting simulations, an unstructured tetrahedral mesh is typically employed due to its flexibility in conforming to intricate geometries. The mesh quality is paramount, as it directly influences the accuracy and stability of the simulation. After several iterations of mesh refinement and optimization to balance computational cost with result fidelity, a final volumetric mesh was generated. This mesh consisted of approximately 1.21 million nodes and 6.88 million tetrahedral elements, providing sufficient resolution to capture the physics of the lost foam casting process.
Key Parameters for Lost Foam Casting Simulation
Numerical simulation of lost foam casting requires the careful definition of multiple physical parameters and boundary conditions that are distinct from conventional sand casting. These parameters govern the heat transfer, fluid flow, and pattern decomposition phenomena unique to the process.
Material Properties and Initial Conditions
The casting material is gray iron, specifically grade HT250. The mold is composed of unbonded silica sand, and the expendable pattern is made of expanded polystyrene (EPS) foam. A critical parameter in lost foam casting is the pouring temperature. Because the molten metal must vaporize the foam pattern, the required temperature is generally higher than in sand casting. For铸铁件, the pouring temperature is typically elevated by 20–80 °C. For this simulation, a pouring temperature of 1400 °C was set, with an initial mold temperature of 25 °C. The material properties for density, thermal conductivity, specific heat, and latent heat were assigned from the ProCAST material database for each component.
| Parameter | Value / Specification |
|---|---|
| Casting Material | HT250 (Gray Iron) |
| Mold Material | Silica Sand (Dry, Unbonded) |
| Pattern Material | Expanded Polystyrene (EPS) Foam |
| Pouring Temperature | 1400 °C |
| Initial Mold Temperature | 25 °C |
| Interfacial Heat Transfer Coefficient (HTC) |
Mold/Casting: 500 W/m²K Mold/Pattern: 100 W/m²K |
| Back-Pressure (Vacuum Level) | 0.04 atm (Applied to mold exterior) |
| Atmospheric Pressure | 1.00 atm (Applied at top of pouring cup) |
| Cooling Condition | Air cooling at ambient temperature |
Boundary Conditions and Process Parameters
Lost foam casting involves pressure-driven flow. The decomposition of the EPS foam creates a gap filled with gaseous products, and the metal front advances under the influence of atmospheric pressure, gravity, and the vacuum applied to the mold. Two key pressure boundaries were defined: one at the top of the sprue (1.04 atm, representing atmospheric pressure) and one on the external boundaries of the sand mold (1.00 atm, representing the applied vacuum of -0.04 atm gauge pressure). The heat transfer at interfaces is modeled using heat transfer coefficients (HTC). The HTC between the sand and the metal is higher than that between the sand and the foam, reflecting the different contact conditions.
The simulation run parameters control the numerical solution scheme. The total number of time steps was set to 10,000. Adaptive time stepping was used with an initial time step (DT) of 0.0001 s and a maximum (DTMAX) of 5 s to ensure stability during rapid initial pouring and efficiency during slow solidification. To predict shrinkage porosity, a feeding length of 5 mm was specified. This criterion dictates that areas where liquid feeding cannot occur over a distance greater than this value are prone to macro-porosity. The porosity calculation level (POROS) was set to 1 for a comprehensive shrinkage analysis. Crucially, for lost foam casting, the pressure boundary conditions must also be activated as internal flow boundaries (PINLET = 1) to model the pressure-driven filling of the pattern cavity.
| Parameter Category | Setting | Purpose/Explanation |
|---|---|---|
| Total Simulation Steps | 10,000 | Defines the overall duration of the simulated process. |
| Time Step Control | DT=0.0001s, DTMAX=5s | Ensures numerical stability during fast filling and efficiency during slow cooling. |
| Feeding Length | 5 mm | Critical distance for liquid metal feeding; areas beyond this are flagged for potential shrinkage. |
| Porosity Calculation | POROS = 1 | Activates the highest level of shrinkage and porosity prediction. |
| Flow Driving Force | PINLET = 1 | Specifies that pressure boundaries drive the internal flow, essential for lost foam casting simulation. |
Numerical Simulation Results and Defect Analysis
The simulation of the lost foam casting process was executed, and the results were analyzed in two primary phases: filling (mold filling) and solidification. The post-processing module of ProCAST allows for detailed visualization of temperature fields, liquid fraction, and defect prediction.
Filling Pattern Analysis
The filling sequence in lost foam casting is generally slower than in green sand casting due to the energy consumed in vaporizing the foam pattern. The simulated filling process revealed specific characteristics for this thin-walled component. Metal entered the cavity through two ingates simultaneously. Initially, regions closest to the ingates filled rapidly. The progress was sequential overall, but the fluid dynamics within the thin, curved sections showed a more complex advancement, with metal fronts converging from different paths. The filling remained stable without indications of turbulent splashing or cold shuts, which is attributed to the dampening effect of the decomposing foam. The total filling time was significantly longer than an equivalent sand casting, a key characteristic of the lost foam casting process.
The flow during lost foam casting can be conceptually described by modifying the Bernoulli equation to account for the resisting pressure from foam decomposition:
$$P_{atm} + \rho g h – P_{vacuum} – P_{foam}(t) = \frac{1}{2}\rho v^2 + \Delta P_{loss}$$
where \(P_{foam}(t)\) is the time-varying back-pressure from the decomposing foam, \(P_{vacuum}\) is the applied vacuum, and \(\Delta P_{loss}\) accounts for viscous losses in the gating system.
Solidification and Thermal Analysis
The solidification phase is where shrinkage defects form. The analysis of the temperature field over time is crucial. The simulation showed that solidification began rapidly at the thin sections and the extremities of the casting farthest from the ingates. As time progressed, the solidification rate decreased. A longitudinal temperature gradient was observed, with the lower sections of the casting cooling faster than the upper sections. However, the thermal analysis highlighted a critical issue: the top planar section of the turbine rear exhaust pipe, along with certain junction points, became isolated liquid pools, or “hot spots,” late in the solidification process. These areas were the last to freeze and were effectively cut off from the feeding source (the gating system) by earlier solidified regions.
The solidification time for a section can be estimated using Chvorinov’s rule:
$$t_f = B \cdot \left( \frac{V}{A} \right)^n$$
where \(t_f\) is the local solidification time, \(V\) is volume, \(A\) is surface area, and \(B\) and \(n\) are constants dependent on the mold material and casting conditions. The top section, despite being relatively thin, had a less favorable \(V/A\) ratio for directional feeding compared to the walls connected to the runners, causing it to be a thermal center.
Prediction of Shrinkage Defects
Based on the solidification analysis and using the defined feeding length criterion and porosity models, the software predicted the location and severity of shrinkage porosity. The primary defect manifested as a region of macro-porosity and potential shrinkage cavity precisely in the top planar surface of the casting. This is a direct consequence of the lack of effective feeding during the final stages of solidification. The gating system, while adequate for filling, was insufficient to act as a feed source for this isolated thermal center. This simulated defect correlated directly with the reported field failure of gas leakage, as porosity in this critical sealing surface would compromise airtightness. The Niyama criterion, a common indicator for microporosity, can be expressed as:
$$N_y = \frac{G}{\sqrt{\dot{T}}}$$
where \(G\) is the temperature gradient and \(\dot{T}\) is the cooling rate. Low values of \(N_y\) indicate a high risk of shrinkage porosity. Analysis confirmed low Niyama values in the top planar region.
Process Improvement via Riser Addition and Validation
To eliminate the shrinkage defect in the lost foam casting process, the most direct and effective method is to provide a supplemental source of liquid metal to feed the shrinking region until it solidifies: a riser (or feeder). The goal of riser design is to ensure it remains molten longer than the casting section it is intended to feed, establishing a directional solidification path from the casting into the riser.
Riser Design Strategy
Based on the defect location, a riser was designed and attached to the top planar surface of the casting, which was the identified hot spot. The riser was positioned at the highest point of the casting geometry to leverage gravitational feeding. Its dimensions were calculated to provide sufficient feed metal volume and to ensure it possesses a longer solidification time than the fed region. A cylindrical riser shape was chosen for simplicity and modeling. The required riser volume \(V_r\) can be estimated as:
$$V_r = \frac{V_c \cdot \alpha}{\eta}$$
where \(V_c\) is the volume of the casting region to be fed, \(\alpha\) is the volumetric shrinkage coefficient of the molten metal (approximately 4-6% for gray iron), and \(\eta\) is the feeding efficiency of the riser (typically 10-30% for side risers in lost foam casting due to heat loss through the foam pattern).
Simulation of the Improved Lost Foam Casting Process
The modified geometry, incorporating the riser, was re-meshed and simulated under identical process parameters. The filling sequence was largely unaffected by the small addition. The critical change was observed in the solidification phase. The thermal analysis now showed a clear directional solidification pattern. The casting sections solidified first, followed by the junction with the riser, and finally the riser itself solidified last. The riser successfully acted as a thermal and feed reservoir, maintaining a liquid feed path to the top section until it was completely solidified.
The defect prediction analysis for the improved lost foam casting process confirmed the success of the modification. The region of predicted macro-porosity on the top planar surface of the casting was completely eliminated. Any residual shrinkage porosity was confined to the center of the riser body, which is an acceptable outcome as the riser will be removed during subsequent machining operations. The simulation thus validated that the proposed riser design effectively solved the shrinkage problem without introducing new defects.
Conclusion
This study successfully demonstrates the application of numerical simulation technology to diagnose and rectify a production issue in the lost foam casting of a complex turbine rear exhaust pipe. The initial simulation accurately replicated the conditions leading to shrinkage porosity in a critical sealing surface, identifying the root cause as an isolated thermal center lacking feed metal. Guided by this analysis, a targeted process improvement—the addition of a strategically designed riser—was proposed. A subsequent simulation of the modified lost foam casting process confirmed the complete elimination of the casting defect, with all predicted shrinkage being successfully redirected to the sacrificial riser.
The methodology underscores the power of virtual prototyping in foundry engineering. It enables a shift from empirical, costly trial-and-error methods to a science-based, predictive approach. For the lost foam casting process, with its unique parameters involving foam decomposition and back-pressure, numerical simulation is an indispensable tool for optimizing gating and feeding systems, reducing scrap rates, shortening development timelines, and ensuring the production of high-integrity castings. The principles and workflow detailed here are directly applicable to the optimization of other complex components manufactured via lost foam casting.