Thin-walled shell components are widely employed in high-power engines for applications such as military equipment and marine propulsion. However, their casting process stability is often poor, and the attendant shrinkage porosity and hot tearing defects severely reduce the yield. The main objective of this work is to improve the quality of a complex thin-walled shell casting made of 1Cr20Ni14Si2 heat-resistant stainless steel. In this study, a combination of theoretical calculation and numerical simulation was adopted to analyze and predict the formation of casting defects, including shrinkage porosity and hot tearing, during the investment casting process. The finite-element simulation platform ProCAST 2018 was utilized to establish a reliable material database for 1Cr20Ni14Si2. A gating system with two ingates and a runner was designed and optimized by adding feed aids. The temperature field, solid fraction evolution, shrinkage porosity, and hot tearing indicator (HTI) were used as evaluation criteria. The influence of pouring temperature, pouring speed, and shell preheating temperature on the volume and distribution of casting defects was systematically investigated. The optimized process parameters were subsequently verified by actual production trials and industrial computed tomography (CT) inspection. The results demonstrate that the addition of feed aids to the gating system enhances filling stability, prolongs solidification, and reduces the shrinkage porosity volume by about 47%. Increasing the pouring temperature to 1650 °C reduces the shrinkage porosity volume by 96.5%. An adequate preheating temperature of the ceramic shell up to 1050 °C reduces the porosity by 98%. The optimal combination of pouring temperature, pouring speed, and shell preheating temperature is 1650 °C, 3 kg/s, and 1050 °C, respectively. Under these conditions, no significant casting defects are observed in the final casting, and the effective stress at the hot-tear-sensitive region is 259 MPa, which is below the yield strength of 295 MPa. The industrial trial confirms the accuracy of the simulation and the effectiveness of the optimization.
1. Introduction
High-power diesel engines demand components with excellent high-temperature oxidation resistance, corrosion resistance, and reliable mechanical performance. The connecting shell, as an exhaust component, operates under severe thermal and mechanical loading. To achieve light weight, most of the shell wall thickness is only 3.5 mm, and the component has an eccentric three-way structure connected by thin-walled curved surfaces. Casting such a thin-walled shell is extremely challenging because the molten metal tends to solidify prematurely, leading to misruns, cold shuts, and incomplete filling. Moreover, during solidification, the complex geometry induces uneven thermal contraction, which easily causes shrinkage porosity and hot tearing. These casting defects severely deteriorate the service performance of the component and reduce the production yield.
Traditional trial-and-error methods for developing investment casting processes are time-consuming and costly. Numerical simulation offers an efficient and cost-effective route to visualize mold filling, solidification, and stress evolution, and to predict casting defects before physical trials. In the present work, ProCAST 2018, based on the finite-element method, was selected because of its robust capabilities in coupled flow-thermal-stress analysis and its well-established criteria for shrinkage porosity and hot tearing.
The main objectives of this study are as follows:
- To design and optimize the gating and feeding system for a thin-walled shell casting using the cup-capacity feeding method;
- To construct a reliable thermophysical database of 1Cr20Ni14Si2 stainless steel;
- To numerically investigate the influence of pouring temperature, pouring speed, and shell preheating temperature on casting defects;
- To determine the optimal process parameters and validate them through actual production and industrial CT inspection.
2. Experimental Material and Numerical Model
2.1 Component Characteristics
The connecting shell studied here has an overall size of 216 mm × 180 mm × 240 mm. More than 70% of its surface consists of thin walls with a thickness of 3.5 mm, with gradual transitions to 5 mm near the flanges. The middle inclined channel and the flange require minor machining, while the remaining surfaces are cast to net shape. Owing to the complex structure and variable wall thickness, the component is highly susceptible to localized hot spots and insufficient feeding during solidification, resulting in shrinkage defects and hot tearing.
2.2 Material and Properties
The selected alloy is 1Cr20Ni14Si2 heat-resistant stainless steel, which offers higher strength than the conventional 06Cr25Ni20 alloy while maintaining comparable ductility and corrosion resistance. The chemical composition of the alloy used in this study is given in Table 1.
| Element | C | Cr | Ni | Si | Mn | P | S | Fe |
|---|---|---|---|---|---|---|---|---|
| Specified | ≤0.2 | 19–22 | 12–15 | 1.5–2.5 | ≤1.5 | ≤0.035 | ≤0.03 | Bal. |
| Measured | 0.10 | 20.24 | 14.03 | 1.61 | 0.55 | 0.005 | 0.008 | Bal. |
The mechanical properties of 1Cr20Ni14Si2 at room temperature are summarized in Table 2.
| Property | Value |
|---|---|
| Tensile strength (MPa) | ≥590 |
| Yield strength (MPa) | ≥295 |
| Elongation (%) | ≥35 |
| Reduction of area (%) | ≥50 |
| Hardness (HB) | ≤187 |
The liquidus and solidus temperatures of this alloy were calculated using the Scheil solidification model in ProCAST. The liquidus temperature is approximately 1440 °C, and the solidus temperature is approximately 1268 °C. Therefore, the solidification interval is about 172 °C, which is relatively wide. This wide freezing range makes the alloy prone to micro-shrinkage and hot tearing because the interdendritic liquid film persists over a large temperature range. Consequently, the casting process must be carefully optimized to minimize casting defects.
2.3 Governing Equations for Mold Filling and Solidification
For the numerical simulation of mold filling, the incompressible Navier-Stokes equations and the continuity equation are solved. The continuity equation is expressed as:
$$ \frac{\partial u}{\partial x} + \frac{\partial v}{\partial y} + \frac{\partial w}{\partial z} = 0 \tag{1} $$
where \(u\), \(v\), and \(w\) are the velocity components in the \(x\), \(y\), and \(z\) directions, respectively. The momentum equations are:
$$ \rho \left( \frac{\partial u}{\partial t} + u \frac{\partial u}{\partial x} + v \frac{\partial u}{\partial y} + w \frac{\partial u}{\partial z} \right) = -\frac{\partial P}{\partial x} + \mu \nabla^2 u + \rho g_x \tag{2} $$
$$ \rho \left( \frac{\partial v}{\partial t} + u \frac{\partial v}{\partial x} + v \frac{\partial v}{\partial y} + w \frac{\partial v}{\partial z} \right) = -\frac{\partial P}{\partial y} + \mu \nabla^2 v + \rho g_y \tag{3} $$
$$ \rho \left( \frac{\partial w}{\partial t} + u \frac{\partial w}{\partial x} + v \frac{\partial w}{\partial y} + w \frac{\partial w}{\partial z} \right) = -\frac{\partial P}{\partial z} + \mu \nabla^2 w + \rho g_z \tag{4} $$
where \(\rho\) is the density, \(P\) is the pressure, \(\mu\) is the dynamic viscosity, and \(g_x\), \(g_y\), \(g_z\) are the gravitational acceleration components. The energy equation is:
$$ \rho c \left( \frac{\partial T}{\partial t} + u \frac{\partial T}{\partial x} + v \frac{\partial T}{\partial y} + w \frac{\partial T}{\partial z} \right) = \frac{\partial}{\partial x}\left( k \frac{\partial T}{\partial x} \right) + \frac{\partial}{\partial y}\left( k \frac{\partial T}{\partial y} \right) + \frac{\partial}{\partial z}\left( k \frac{\partial T}{\partial z} \right) + S \tag{5} $$
where \(T\) is the temperature, \(c\) is the specific heat, \(k\) is the thermal conductivity, and \(S\) is the internal heat source that includes latent heat release during solidification.
2.4 Criteria for Shrinkage and Hot Tearing
Shrinkage porosity was evaluated using the Niyama criterion. The local temperature gradient \(G\) and cooling rate \(R\) are combined in the dimensionless parameter \(G/\sqrt{R}\). The criterion states that micro-shrinkage forms when:
$$ \frac{G}{\sqrt{R}} \le M \tag{6} $$
where \(M\) is the critical value depending on the alloy composition. Lower Niyama values correspond to a higher probability of shrinkage porosity. This criterion is particularly suitable for predicting both macro-shrinkage and micro-shrinkage in steel castings.
For hot tearing prediction, the hot tearing indicator (HTI) is calculated from the stress-strain response of the mushy zone. The elastic-plastic constitutive model was adopted for the casting, while the mold was treated as a linear-elastic body. The yield stress \(\sigma_y\) and the plastic modulus \(H\) were defined as functions of temperature. The effective stress is computed from the thermal stress and mechanical constraint. A hot tear is likely to occur when the effective stress in the semisolid region exceeds the local cohesive strength of the material.
3. Design of the Gating and Feeding System
The gating system plays a vital role in minimizing casting defects. For the connecting shell, a top-gating system with two ingates and one runner was selected. The dimensions were calculated using the cup-capacity feeding method. The gate equivalent diameter was obtained from the relation:
$$ D_g = k D_c \tag{7} $$
where \(D_g\) is the gate equivalent diameter, \(k\) is a weight factor, and \(D_c\) is the equivalent thermal diameter of the hot spot. Based on the component geometry, the hot-spot diameter \(D_c\) was determined to be 14 mm, and the weight factor \(k\) was 0.92. Therefore, \(D_g\) was 13 mm. The runner diameter was set to satisfy:
$$ D_r \ge (1.15\text{–}1.25) D_g \tag{8} $$
which gives \(D_r\) in the range of 15–16 mm. A top cup with an upper diameter of 100 mm, lower diameter of 70 mm, and height of 80 mm was designed. Two different gating schemes were considered:
- Scheme A: runner dimensions 60 mm × 60 mm, runner length 260 mm, gate diameter 22 mm, without feed aids;
- Scheme B: same as Scheme A, but with feed aids (pads) added at the ingates to improve feeding and solidification temperature gradient.
Table 3 lists the dimensions of both gating schemes.
| Component | Cup upper D | Cup lower D1 | Cup height H | Runner side A | Runner length L | Gate diameter Dg | Feed aid |
|---|---|---|---|---|---|---|---|
| Scheme A | 100 | 70 | 80 | 60 | 260 | 22 | No |
| Scheme B | 100 | 70 | 80 | 60 | 260 | 22 | Yes |
3.1 Initial Casting Conditions
Because of the thin walls and the high chromium content of the alloy, the pouring temperature was initially set to 1560 °C, which is 120 °C above the liquidus temperature. The pouring speed was calculated using the empirical formula:
$$ q = K \frac{l_{\max}}{\sqrt{\delta_c}} \tag{9} $$
where \(K = 0.05\) for a top-gating system, \(l_{\max}\) is the maximum length of the thin wall, and \(\delta_c\) is the minimum wall thickness. This yielded a filling rate of approximately 3.37 kg/s; therefore, 3 kg/s was selected for convenience. The shell preheating temperature was chosen as 800 °C, typical for steel investment casting.
4. Simulation Setup
The three-dimensional model of the casting and gating system was created in UG NX and imported into ProCAST. Surface meshes were generated with edge lengths of 2 mm for the casting and 3 mm for the gating system. The ceramic shell was generated with a uniform thickness of 8 mm, as measured from actual samples. The total mesh statistics are listed in Table 4.
| Component | Element size (mm) | Surface elements | Volume elements |
|---|---|---|---|
| Casting | 2 | 140,652 | 1,384,085 |
| Gating system | 3 | 49,413 | — |
| Shell | variable | 48,952 | — |
The interface heat transfer coefficient between the casting and the mold was set to 900 W/(m²·K). The environment was modeled as ambient air at 25 °C with natural convection and radiation. The stress module was activated, and the casting was assigned an elastoplastic material model. The mold was considered rigid.
5. Results and Discussion
5.1 Mold Filling Behavior
Figure 1 shows the computed evolution of the molten metal front during filling for Scheme B. Because the connecting shell is thin-walled, the filling time is quite short: approximately 4.8 s for the complete system. The addition of feed aids reduces the initial jetting of the metal at the runner ends and promotes a more tranquil filling pattern. As a result, the risk of gas entrapment and oxide film formation is diminished.
Typical casting defects such as shrinkage, hot tearing, and gas porosity are illustrated above. These imperfections originate from complex interactions between alloy composition, mold characteristics, and process parameters. In the following sections, we quantitatively analyze the effect of the gating system and process parameters on these casting defects.
5.2 Solidification and Feeding
The solidification sequence is crucial for feeding efficiency. In Scheme A, the thin walls solidify quickly, and the ingates show a relatively high solid fraction of 0.3 when the casting is only 25% solidified. This prevents effective feeding from the runner. In contrast, Scheme B, with feed aids, keeps the ingates molten longer, with a solid fraction close to zero at the same stage. This extends the feeding time and allows the runner to compensate for the volumetric contraction of the casting. The total solidification time for Scheme A is 2436 s, while Scheme B takes 2470 s, providing a longer feeding window.
5.3 Shrinkage Porosity Prediction
The total shrinkage porosity predicted for Scheme A is 0.55 cc, whereas Scheme B reduces this volume to 0.29 cc, a reduction of about 47%. Table 5 shows the distribution of shrinkage porosity in different regions for both schemes.
| Region | Scheme A | Scheme B |
|---|---|---|
| Inclined channel | 0.32 | 0.22 |
| Thin-walled area | 0.23 | 0.07 |
| Total | 0.55 | 0.29 |
The most severe porosity appears at the inclined channel and at the connection between the flange and the thin wall. These locations correspond to local hot spots where the wall thickness changes abruptly. The temperature field shows that isolated liquid pools form in the thick section of the inclined channel while the surrounding thin walls already solidified. This prevents further liquid feeding, leading to micro-shrinkage upon final solidification. Adding feed aids improves the thermal gradient and enables the runner to feed these hot spots more effectively.
5.4 Hot Tearing Behavior
For both schemes, the highest hot tearing tendency is located near the ingates and at the transition between the flange and the thin wall, denoted as region A. The HTI values are listed in Table 6.
| Scheme | HTI | Final effective stress (MPa) | Yield strength (MPa) |
|---|---|---|---|
| A | 0.0145 | 264 | 295 |
| B | 0.0122 | 255 | 295 |
At the end of solidification, the effective stress in the mushy zone rises with time. For Scheme B, the effective stress at 800 s is about 34.3 MPa, slightly lower than the 35.7 MPa of Scheme A. The final effective stress at room temperature is 255 MPa for Scheme B, which is below the yield strength of the alloy. Therefore, the risk of hot tearing is small. The feed aids reduce the strain concentration at the junction and lower the hot tearing susceptibility.
6. Parametric Study and Optimization
Based on Scheme B, a systematic study was carried out to investigate the influence of pouring temperature, pouring speed, and shell preheating temperature on casting defects. The baseline parameters were 1560 °C, 3 kg/s, and 800 °C.
6.1 Effect of Pouring Temperature
Seven pouring temperatures were simulated: 1530, 1550, 1570, 1590, 1610, 1630, and 1650 °C. The results are shown in Table 7.
| Pouring temperature (°C) | Porosity volume (cc) |
|---|---|
| 1530 | 0.86 |
| 1550 | 0.39 |
| 1570 | 0.36 |
| 1590 | 0.25 |
| 1610 | 0.12 |
| 1630 | 0.19 |
| 1650 | 0.03 |
Increasing the pouring temperature improves the fluidity of the molten metal and delays solidification, which enhances feeding. The porosity volume decreases from 0.86 cc at 1530 °C to 0.03 cc at 1650 °C, a reduction of 96.5%. However, there is an anomaly at 1630 °C where the porosity slightly increases to 0.19 cc. This may be attributed to a change in the freezing pattern of the inclined channel, which forms a larger isolated liquid pool at this intermediate temperature. At 1650 °C, the feeding path is fully effective and the porosity almost disappears.
The effect of pouring temperature on hot tearing (HTI at region A) is shown in Table 8.
| Pouring temperature (°C) | HTI |
|---|---|
| 1590 | 0.0118 |
| 1610 | 0.0121 |
| 1630 | 0.0135 |
| 1650 | 0.0127 |
The effective stress in the mushy zone at the time when the solid fraction reaches 99% decreases with increasing pouring temperature. The final effective stress, however, increases with temperature because the higher temperature leads to a coarser grain structure and a slightly larger thermal contraction. Nevertheless, all values remain below the yield strength. The maximum final effective stress at 1630 °C is 287.9 MPa, still below 295 MPa, but the margin is small. Therefore, 1650 °C is preferable to ensure a safe margin.
6.2 Effect of Pouring Speed
The pouring speed was varied from 2.0 to 5.0 kg/s while keeping the pouring temperature at 1560 °C and the shell temperature at 800 °C. The porosity volumes are summarized in Table 9.
| Pouring speed (kg/s) | Porosity volume (cc) |
|---|---|
| 2.0 | 0.31 |
| 2.5 | 0.57 |
| 3.0 | 0.29 |
| 3.5 | 0.46 |
| 4.0 | 0.51 |
| 4.5 | 0.44 |
| 5.0 | 0.43 |
The relationship between pouring speed and porosity is non-monotonic. Too low a speed causes premature solidification and poor feeding, while too high a speed generates turbulence and gas entrapment, which increases the tendency for shrinkage and gas porosity. The minimum porosity volume occurs at 3 kg/s. Therefore, 3 kg/s is considered the optimal critical value for this casting.
Regarding hot tearing, the HTI and effective stress are nearly independent of pouring speed. The final effective stress in region A stabilizes at about 270 MPa. This suggests that the pouring speed mainly affects the mold filling and feeding behavior, but does not significantly alter the thermal stress state after complete solidification.
6.3 Effect of Shell Preheating Temperature
The shell preheating temperature was varied from 750 °C to 1050 °C while fixing the pouring temperature at 1560 °C and pouring speed at 3 kg/s. The porosity volumes are listed in Table 10.
| Shell temperature (°C) | Porosity volume (cc) |
|---|---|
| 750 | 0.71 |
| 800 | 0.29 |
| 850 | 0.24 |
| 900 | 0.22 |
| 950 | 0.05 |
| 1000 | 0.10 |
| 1050 | 0.01 |
A higher shell temperature reduces the cooling rate of the casting, allowing more time for the molten metal to feed the solidifying areas. The porosity volume decreases by 98% when the shell temperature is increased from 750 °C to 1050 °C. The distribution of porosity shifts from the inclined channel to the thin walls as the shell temperature increases. At 1050 °C, almost no shrinkage porosity remains.
The effect on hot tearing is similar to that of pouring temperature in terms of the mushy-zone stress. As the shell temperature increases, the effective stress at the end of solidification decreases. The final effective stress, however, remains constant at about 268 MPa, independent of shell temperature. Thus, a higher shell preheating temperature is beneficial for reducing both shrinkage porosity and hot tearing tendency.
6.4 Combined Optimization
Based on the single-parameter studies, four candidate sets of process parameters were selected for further evaluation. These are listed in Table 11.
| Case | Pouring temperature (°C) | Pouring speed (kg/s) | Shell temperature (°C) |
|---|---|---|---|
| 1 | 1610 | 3 | 950 |
| 2 | 1610 | 3 | 1050 |
| 3 | 1650 | 3 | 950 |
| 4 | 1650 | 3 | 1050 |
The predicted shrinkage porosity volumes for these four cases are given in Table 12.
| Case | Porosity volume (cc) |
|---|---|
| 1 | 0.13 |
| 2 | 0.02 |
| 3 | 0.06 |
| 4 | 0 |
Case 4 exhibits zero porosity. In addition, the HTI value for Case 4 is 0.012, which is identical to Cases 1 and 2 and lower than Case 3. The effective stress at the hot-tear-sensitive observation point at the end of solidification is 20.2 MPa for Case 4, the lowest among all cases. The final effective stress is 259 MPa, leaving a safety margin of 36 MPa relative to the yield strength.
Therefore, the optimal process parameters are: pouring temperature of 1650 °C, pouring speed of 3 kg/s, and shell preheating temperature of 1050 °C. These parameters were adopted for the actual production trial.
7. Industrial Trial and Verification
7.1 Wax Pattern and Shell Preparation
A trial production was carried out using the optimized process. A green recyclable wax was used for the pattern. The wax was injected into an aluminum die at a temperature of about 61 °C, with an injection pressure of 3.5 MPa, a filling time of 25 s, and a holding time of 55 s. The resulting wax pattern exhibited good surface quality and dimensional accuracy. Multiple wax patterns were assembled onto a runner to form a tree.
For the ceramic shell, a silica sol binder was used. The primary slurry consisted of 45 kg of silica sol, 160 kg of zirconium silicate powder (325 mesh), and appropriate wetting and defoaming agents. The primary stucco was zircon sand (120 mesh). The secondary and back-up coats used mullite sand with mesh sizes of 60 and 30. The shell was dried after each coat, and a final seal coat was applied. After complete drying, the wax was removed in a steam autoclave at 180 °C for 17 min. The shell was then fired at 1170 °C for 60 min and held for 30 min.
7.2 Casting and Post-Processing
The alloy was melted in a 50 kg medium-frequency induction furnace. The chemical composition was verified by optical emission spectroscopy prior to pouring. The optimized pouring parameters were 1650 °C, 3 kg/s, and 1050 °C for the shell. For comparison, an additional casting was made using the initial parameters (1560 °C, 3 kg/s, 800 °C). After solidification and cooling, the shell was removed by a vibratory hammer. The gating system was cut off, and the casting was ground, blasted, pickled, and passivated.
7.3 Surface Inspection
Visual inspection of the optimized casting revealed a clean surface without burn-on, swelling, misruns, or flash. The surface quality was more uniform compared with the casting made using the initial parameters, which exhibited slight surface blemishes and oxidation. The higher pouring temperature and shell temperature reduced the viscosity of the chromium-rich molten steel and improved the replica of the mold cavity, thereby reducing surface defects.
7.4 Industrial CT Examination
To evaluate internal casting defects, industrial computed tomography (ICT) was performed using an ICT-450 system. The scanning parameters were: voltage 430 kV, current 1.5 μA, focus size 0.4 mm, integration time 1000 ms, and voxel size 0.083 mm. The inspection covered the inclined channel and the thin-wall regions. According to the acceptance standard, no cracks are allowed, and shrinkage porosity must not exceed 2 mm in size or more than three indications per area.
For the casting produced with the initial parameters, the CT images revealed two significant shrinkage defects in the inclined channel. Their maximum diameters were 19.2 mm and 20.3 mm, respectively, far exceeding the acceptance limit. The locations of these defects agreed well with the simulation predictions, confirming the accuracy of the numerical model. No hot cracks were detected in the transition region.
For the casting produced with the optimized parameters, the CT images showed no shrinkage pores or cracks in the inclined channel or in the thin-wall sections. The internal quality fully met the specification requirements. The industrial CT inspection therefore verified that the optimized process effectively eliminates the major casting defects and that the numerical simulation is reliable.
8. Conclusions
In this work, the investment casting process of a thin-walled shell casting made of 1Cr20Ni14Si2 stainless steel was numerically simulated and experimentally validated. The main conclusions are summarized as follows:
- The addition of feed aids to the gating system significantly improves filling stability and feeding efficiency. Compared with the scheme without feed aids, the total shrinkage porosity volume is reduced by 47% (from 0.55 cc to 0.29 cc), and the hot tearing indicator at the susceptible region decreases from 0.0145 to 0.0122.
- Increasing the pouring temperature from 1530 °C to 1650 °C reduces the shrinkage porosity volume by 96.5%. At 1650 °C, the porosity volume is only 0.03 cc, and the porosity in the inclined channel disappears. The effective stress at the end of solidification decreases with increasing pouring temperature, while the final effective stress increases but remains below the yield strength.
- The pouring speed has a non-monotonic effect on shrinkage porosity. The minimal porosity is obtained at 3 kg/s. The pouring speed has negligible influence on hot tearing tendency and final effective stress.
- Raising the shell preheating temperature from 750 °C to 1050 °C reduces the shrinkage porosity volume by 98%. A higher shell temperature is beneficial for feeding and reduces the effective stress in the mushy zone. The final effective stress remains stable at about 268 MPa.
- The optimal process parameters are a pouring temperature of 1650 °C, a pouring speed of 3 kg/s, and a shell preheating temperature of 1050 °C. Under these conditions, the predicted shrinkage porosity volume is zero, and the final effective stress at the hot-tear-sensitive area is 259 MPa, which is well below the yield strength of 295 MPa.
- Industrial trials and industrial CT inspection confirmed that the optimized process yields sound castings without significant internal casting defects. The simulation predictions are in good agreement with the experimental results, demonstrating the high reliability of the numerical simulation approach for optimizing thin-walled shell castings.
The methodology presented here provides a practical and cost-effective route for the development of casting processes for complex thin-walled components, reducing the incidence of casting defects and improving production yield.