Optimizing Investment Casting for Ductile Iron Protective Parts

In my work, I focused on the investment casting process for protective parts used in oil drilling machines. These protective parts are safety-critical ductile iron castings that must operate without surface or internal defects. The main challenge was to eliminate shrinkage cavities and dispersed shrinkage while maintaining stable mold filling and avoiding gas entrapment, inclusions, and cold shuts. I combined Pro/E solid modeling with ProCAST numerical simulation to reproduce the original process, identify the causes of defects, redesign the gating system, and optimize three key process parameters through an orthogonal experiment. The optimized parameters were a pouring temperature of 1300 °C, a filling speed of 0.460 m/s, and a shell preheating temperature of 800 °C. Under these conditions, the shrinkage rate of the ductile iron castings decreased from 13.2603% to 1.321%, and actual production confirmed the improvement.

The protective part is a hollow cylindrical ductile iron casting with a relatively simple overall structure and a uniform wall thickness. It has three asymmetric slots at one end, eight through holes between the two end faces, and a serrated groove on the outer cylindrical surface. Because the part functions as a safety protector in oil drilling equipment, the quality requirements are strict. The casting must be free of visible surface defects and internal porosity. I treated the part as a 600 MPa grade ductile iron casting with a solidification range from 1129 °C to 1194 °C and a density of 7300 kg/m3. The geometry and functional requirements are summarized in Table 1.

Table 1. Geometry and functional requirements of the ductile iron protective part
Parameter Value
Application Safety protection in oil drilling machines
Material type Ductile iron castings, 600 MPa tensile strength grade
Outer envelope 148.49 mm x 148.49 mm x 71.1 mm
Mean wall thickness 20 mm
Shape Hollow cylinder with three asymmetric slots, eight through holes, and an outer serrated groove
Required internal quality No shrinkage cavities, no dispersed shrinkage, no inclusions
Required surface quality No surface defects, no cold shuts, no gas pores
Tensile strength At least 600 MPa
Hardness HB 190 to 270
Solidification range 1129 °C to 1194 °C
Density 7300 kg/m3

The chemical composition of the ductile iron castings was controlled within a narrow range to ensure graphite spheroidization, adequate fluidity, and sufficient feeding during solidification. Carbon and silicon affect the eutectic solidification behavior and the amount of graphite expansion. Manganese influences matrix hardenability, while sulfur and phosphorus must remain low to avoid hot brittleness and inclusions. Magnesium is the spheroidizing element and must be kept within a suitable range. The target composition is shown in Table 2.

Table 2. Chemical composition range of the ductile iron castings
Element C Si Mn S P Mg
Mass fraction (%) 3.0 to 3.5 2.4 to 2.8 0.3 to 0.5 0.03 to 0.035 <0.1 0.045 to 0.050

I designed the investment casting process with a one-mold-four-piece layout to improve productivity. The gating system consisted of a sprue, a runner, ingates, and vents. The objective was to achieve stable filling and directional solidification so that the ductile iron castings would solidify from the far end toward the ingate, allowing the gating system to feed shrinkage. The part was poured in the top-pouring position with gravity acting in the negative X direction. The shell had six layers, a thickness of approximately 6 mm, quartz sand as the refractory material, and silica sol as the binder. The initial process parameters are listed in Table 3.

Table 3. Initial investment casting parameters for the ductile iron castings
Parameter Initial value
Mold layout One mold with four parts
Pouring position Top pouring
Gravity direction Negative X
Pouring temperature 1300 °C
Filling speed 0.458 m/s
Shell preheating temperature 900 °C
Shell layers 6
Shell thickness About 6 mm
Refractory material Quartz sand
Binder Silica sol
Cooling condition Natural cooling in air

For the minimum filling speed, I used a Kargin-type relation. The purpose was to ensure that the metal surface inside the mold rose quickly enough to prevent premature freezing, misruns, and cold shuts. The relation can be written in the following general form:

$$ v_{\mathrm{fill}} = \frac{h}{\delta T} C_{\mathrm{K}} $$

where \( v_{\mathrm{fill}} \) is the minimum rising speed of the liquid metal in the mold cavity, \( h \) is the casting height, \( \delta \) is the wall thickness, \( T \) is the pouring temperature, and \( C_{\mathrm{K}} \) is an empirical coefficient that depends on the alloy, shell material, and pouring conditions. Using the geometric values of the protective part and the selected pouring temperature, I obtained a calculated minimum filling speed of 0.4577 m/s. Considering practical pouring control, I selected 0.458 m/s for the initial trial. This value was later used as the baseline for the filling-speed optimization.

I created the three-dimensional model in Pro/E and then imported it into ProCAST for mesh generation and numerical simulation. The mesh included the casting, gating system, shell, and surrounding air domain. The mesh size was 8 mm for the pouring cup and sprue, and 5 mm for the casting and ingates. The initial model contained 67,010 total nodes. The boundary conditions and thermal contact settings are given in Table 4. I used a casting-shell heat transfer coefficient of 1000 W/(m2·K), a casting-air heat transfer coefficient of 1000 W/(m2·K), and a shell-air heat transfer coefficient of 50 W/(m2·K). The gravity vector was set to negative X with a magnitude of 9.8 m/s2.

Table 4. Mesh details and thermal boundary conditions used in the initial simulation
Item Setting
Total nodes 67,010
Mesh length for pouring cup and sprue 8 mm
Mesh length for casting and ingates 5 mm
Casting-shell heat transfer coefficient 1000 W/(m2·K)
Casting-air heat transfer coefficient 1000 W/(m2·K)
Shell-air heat transfer coefficient 50 W/(m2·K)
Gravity direction Negative X
Gravity magnitude 9.8 m/s2
Initial shell temperature 900 °C
Initial pouring temperature 1300 °C

The initial filling simulation showed that the liquid metal entered the mold in a relatively smooth manner. The inner wall of the cavity experienced only moderate impact, and no obvious jetting was observed at the early stage. The sprue was filled at 0.7 s. The first pair of ductile iron castings was filled at 2.5 s. The second pair was more than half filled at 3.4 s. Complete filling of the entire mold occurred at 7.2 s. These results are summarized in Table 5.

Table 5. Filling timeline of the initial investment casting process
Event Time (s)
Sprue completely filled 0.7
First pair of ductile iron castings filled 2.5
Second pair more than half filled 3.4
Complete mold filling 7.2

The solidification simulation indicated that the castings began to freeze from the outer surface and then progressed inward and upward. At 944 s, the outer skin had solidified, and the solid fraction in parts of the casting reached 10% to 20%. At 1774 s, the main body of the casting was close to complete solidification. At 2704 s, the entire casting system had fully solidified. Although the general solidification sequence appeared reasonable, the final shrinkage distribution revealed severe defects. The initial process produced a shrinkage rate of 13.2603%, mainly concentrated in the sprue, runner, ingate, and the connection region between the ingate and the casting. This level of porosity would make the ductile iron castings unacceptable for service.

I analyzed the defect formation mechanism using several solidification criteria. The thermal modulus of a section can be expressed as:

$$ M = \frac{V}{A} $$

where \( M \) is the thermal modulus, \( V \) is the volume of the section, and \( A \) is the heat-dissipating surface area. A larger thermal modulus means a longer solidification time. According to Chvorinov’s rule, the solidification time can be approximated by:

$$ t_s = B M^n $$

where \( t_s \) is the solidification time, \( B \) is a mold constant, and \( n \) is an exponent typically close to 2 for many casting conditions. In the initial process, the ingate and its connection to the casting had a smaller thermal modulus than the casting body in some regions, so they solidified earlier. Once the ingate froze, the feeding channel was blocked, and liquid metal could no longer compensate for the volume contraction of the ductile iron castings. This produced shrinkage cavities and dispersed shrinkage.

I also used the Niyama criterion to evaluate the likelihood of shrinkage porosity:

$$ N_y = \frac{G}{\sqrt{\dot{T}}} $$

where \( N_y \) is the Niyama value, \( G \) is the temperature gradient, and \( \dot{T} \) is the cooling rate. Regions with low \( N_y \) values are prone to shrinkage and dispersed shrinkage. The initial simulation showed low Niyama values near the ingate connection and in the central thick sections of the gating system. The solid fraction during solidification was evaluated by:

$$ f_s = \frac{V_s}{V_s + V_l} $$

where \( f_s \) is the solid fraction, \( V_s \) is the solid volume, and \( V_l \) is the liquid volume. When \( f_s \) increased rapidly in the ingate before the casting body had finished feeding, the remaining liquid in the casting became isolated. This isolation, combined with the volume contraction of the ductile iron castings, led to the observed porosity. The main defect locations and causes are summarized in Table 6.

Table 6. Initial defect locations and cause analysis for the ductile iron castings
Location Observed defect Main cause
Sprue Shrinkage cavity Low thermal modulus and late feeding demand
Runner Dispersed shrinkage Flow friction and premature solidification
Ingate Severe shrinkage Ingate solidified before casting body
Ingate-casting connection Shrinkage cavity and porosity Blocked feeding path and isolated liquid
Casting body Scattered dispersed shrinkage Insufficient feeding and low Niyama values

Based on the initial simulation, I redesigned the gating system. The goal was to delay the solidification of the feeding channel and to shorten the feeding distance to the hot spots in the ductile iron castings. I proposed two optimization schemes. Scheme A added a side ingate on the rectangular side of the casting. Scheme B added an ingate on the pipe section of the casting. Both schemes were modeled with the same mesh settings and boundary conditions as the initial process. The simulation results showed that Scheme A reduced the shrinkage rate to 6.9025%, while Scheme B reduced it to 1.3675%. Moreover, in Scheme B, the casting body itself had no shrinkage defects. Therefore, Scheme B was selected as the improved gating design. The comparison is given in Table 7.

Table 7. Comparison of gating system schemes for the ductile iron castings
Scheme Modification Shrinkage rate (%) Casting body defect
Initial Original sprue, runner, and ingates 13.2603 Severe dispersed shrinkage
A Added side ingate on rectangular side 6.9025 Reduced but still present
B Added ingate on pipe section 1.3675 No visible shrinkage

After selecting the improved gating system, I optimized the process parameters. I selected three factors: pouring temperature, filling speed, and shell preheating temperature. The pouring temperature affects fluidity, feeding ability, and gas absorption. The filling speed affects heat loss during filling and the stability of the flow. The shell preheating temperature affects the cooling rate and the temperature gradient in the ductile iron castings. I used an orthogonal experiment to reduce the number of simulations while still evaluating the combined effects. The factors and levels are listed in Table 8.

Table 8. Orthogonal experiment factors and levels
Level A: Pouring temperature (°C) B: Filling speed (m/s) C: Shell preheating temperature (°C)
1 1250 0.450 800
2 1280 0.455 900
3 1300 0.460 1000

I evaluated each run by the shrinkage rate of the ductile iron castings. A lower shrinkage rate indicates better feeding and higher internal quality. The mean response for each factor level was calculated by:

$$ \bar{K}_{ij} = \frac{1}{n} \sum_{m=1}^{n} y_{ijm} $$

where \( \bar{K}_{ij} \) is the average response for factor \( j \) at level \( i \), \( n \) is the number of runs at that level, and \( y_{ijm} \) is the observed shrinkage rate. The range of each factor was calculated by:

$$ R_j = \max_i \bar{K}_{ij} – \min_i \bar{K}_{ij} $$

where \( R_j \) is the range for factor \( j \). A larger range means that the factor has a stronger effect on the shrinkage rate. The nine orthogonal runs and their results are shown in Table 9.

Table 9. Orthogonal experimental design and shrinkage results for the ductile iron castings
Run A B C Filling time (s) Shrinkage rate (%)
L1 1 1 1 4.881 1.451
L2 1 2 2 3.925 1.433
L3 1 3 3 3.869 1.442
L4 2 1 2 3.961 1.551
L5 2 2 3 3.918 1.452
L6 2 3 1 3.869 1.402
L7 3 1 3 3.820 1.589
L8 3 2 2 3.771 1.370
L9 3 3 1 3.858 1.321

The range analysis is shown in Table 10. The filling speed had the largest range, followed by the shell preheating temperature, and then the pouring temperature. This indicates that the filling speed was the most influential parameter for reducing shrinkage in the ductile iron castings. The filling speed controls how quickly the mold is filled and how much heat is lost before solidification begins. A higher filling speed reduced premature freezing and improved feeding. The shell preheating temperature was the second most important factor because it controls the cooling rate and thermal gradient. The pouring temperature had the smallest range in this experimental window, but its effect was still measurable.

Table 10. Range analysis of the orthogonal experiment
Factor Level 1 mean Level 2 mean Level 3 mean Range Rank
A: Pouring temperature 1.4420 1.4683 1.4267 0.0416 3
B: Filling speed 1.5303 1.4183 1.3883 0.1420 1
C: Shell preheating temperature 1.3913 1.4513 1.4943 0.1030 2

According to the principle that a smaller shrinkage rate corresponds to better quality, the optimal combination was A3B3C1. This corresponds to a pouring temperature of 1300 °C, a filling speed of 0.460 m/s, and a shell preheating temperature of 800 °C. The filling time for this optimal combination was 3.858 s, and the shrinkage rate was 1.321%. Compared with the initial process, the shrinkage rate of the ductile iron castings was reduced by approximately 90%. The optimal parameters are summarized in Table 11.

Table 11. Optimal process parameters obtained from the orthogonal experiment
Parameter Optimal value
Pouring temperature 1300 °C
Filling speed 0.460 m/s
Shell preheating temperature 800 °C
Filling time 3.858 s
Shrinkage rate 1.321%
Gating system Scheme B with added pipe-section ingate

The effect of the pouring temperature can be explained by its influence on fluidity and feeding. A higher pouring temperature increases superheat, reduces viscosity, and allows the liquid metal to flow more easily through the gating system. For ductile iron castings, sufficient superheat is important because the alloy has a relatively narrow solidification range. However, an excessively high pouring temperature can increase gas absorption and oxidation, which may introduce gas porosity and inclusions. In my orthogonal window, 1300 °C gave the lowest average shrinkage rate. This indicates that the benefit of improved feeding outweighed the risk of gas-related defects under the selected shell and filling conditions.

The filling speed had the strongest effect. A higher filling speed reduces the time available for heat loss before the mold is completely filled. This keeps the liquid metal hotter during the later stages of filling and helps maintain an open feeding channel. The mass flow rate during filling can be expressed as:

$$ \dot{m} = \rho A v $$

where \( \dot{m} \) is the mass flow rate, \( \rho \) is the density of the liquid metal, \( A \) is the cross-sectional area of the flow channel, and \( v \) is the flow velocity. The continuity equation for an incompressible melt is:

$$ A_1 v_1 = A_2 v_2 $$

where \( A_1 \) and \( A_2 \) are cross-sectional areas at two locations, and \( v_1 \) and \( v_2 \) are the corresponding velocities. When the filling speed is too low, the melt may freeze in the thinner sections before the mold is filled. When the filling speed is too high, the flow may become turbulent, causing gas entrapment and mold erosion. The Reynolds number can be used to assess the flow regime:

$$ Re = \frac{\rho v D}{\mu} $$

where \( Re \) is the Reynolds number, \( D \) is the hydraulic diameter, and \( \mu \) is the dynamic viscosity. In investment casting of ductile iron castings, a controlled increase in filling speed can improve filling without causing severe turbulence if the gating system is properly designed.

The shell preheating temperature affects the heat transfer between the casting and the shell. The heat flux can be written as:

$$ q = h A (T_1 – T_2) $$

where \( q \) is the heat flux, \( h \) is the heat transfer coefficient, \( A \) is the interfacial area, and \( T_1 \) and \( T_2 \) are the temperatures of the casting and the shell. A higher shell preheating temperature reduces the temperature difference and slows the cooling rate. In some cases, this can improve fluidity, but it can also reduce the thermal gradient needed for directional solidification. In my experiments, a lower shell preheating temperature of 800 °C produced the best result. This is because a lower shell temperature increased the thermal gradient and promoted directional solidification from the far end toward the ingate, allowing the gating system to feed the ductile iron castings more effectively.

The pressure drop in the gating system also influences filling and feeding. For a circular channel with laminar flow, the pressure drop can be approximated by:

$$ \Delta p = \frac{128 \mu L Q}{\pi D^4} $$

where \( \Delta p \) is the pressure drop, \( L \) is the channel length, \( Q \) is the volumetric flow rate, and \( D \) is the channel diameter. A smaller diameter or a longer path increases the pressure drop and reduces the effective feeding pressure. The improved gating system in Scheme B provided a shorter and more direct feeding path to the pipe section of the ductile iron castings, which reduced the pressure drop and delayed the solidification of the feeding channel.

The filling time can be estimated from the mold volume and the gating area:

$$ t_f = \frac{V_{\mathrm{mold}}}{A_g v_{\mathrm{fill}}} $$

where \( t_f \) is the filling time, \( V_{\mathrm{mold}} \) is the total mold cavity volume, \( A_g \) is the effective gating cross-sectional area, and \( v_{\mathrm{fill}} \) is the filling speed. In the optimized process, the filling time was 3.858 s, which was shorter than the initial filling time. This shorter filling time reduced heat loss and helped maintain a continuous liquid channel for feeding the ductile iron castings.

The shrinkage rate was calculated from the simulated defect volume and the casting volume:

$$ S = \frac{V_{\mathrm{defect}}}{V_{\mathrm{casting}}} \times 100\% $$

where \( S \) is the shrinkage rate, \( V_{\mathrm{defect}} \) is the volume of shrinkage cavities and dispersed shrinkage, and \( V_{\mathrm{casting}} \) is the total casting volume. This definition was used consistently for all process schemes and orthogonal runs. The initial process had \( S = 13.2603\% \). Scheme A reduced it to 6.9025%. Scheme B reduced it to 1.3675%. The orthogonal optimum achieved \( S = 1.321\% \). These results show that both the gating design and the process parameters contributed to defect reduction, but the gating design had the larger single-step effect.

I also examined the sensitivity of the process to each parameter. Table 12 summarizes the qualitative effects of the three optimized parameters on filling, solidification, and defect formation in ductile iron castings. The table is based on both the simulation results and the physical mechanisms discussed above.

Table 12. Parameter sensitivity in the investment casting of ductile iron castings
Parameter Effect on filling Effect on solidification Effect on shrinkage Optimal value
Pouring temperature Higher temperature increases fluidity and reduces viscosity Higher temperature delays solidification and increases superheat Moderate increase improves feeding but excessive increase risks gas defects 1300 °C
Filling speed Higher speed reduces heat loss and filling time Higher speed preserves superheat and delays premature freezing Strong positive effect when kept below turbulence threshold 0.460 m/s
Shell preheating temperature Higher temperature slightly improves fluidity Higher temperature reduces thermal gradient and slows cooling Lower temperature improves directional solidification and feeding 800 °C

After the numerical optimization, I validated the process through actual investment casting production. The same gating system and process parameters were used in the foundry trial. The produced ductile iron castings were inspected for surface defects, internal shrinkage, and dimensional accuracy. The optimized process significantly reduced the rejection rate. The casting body showed no visible shrinkage cavities, and the dispersed shrinkage that had been present in the initial process was largely eliminated. The verification results are summarized in Table 13.

Table 13. Verification of the optimized investment casting process
Stage Shrinkage rate (%) Casting body defect Production outcome
Initial process 13.2603 Severe shrinkage and dispersed shrinkage High rejection rate
Improved gating, Scheme B 1.3675 No visible shrinkage in the casting body Rejection rate greatly reduced
Orthogonal optimum 1.321 No visible shrinkage in the casting body Stable production with improved quality

The optimized parameters and gating design are recommended for production. Table 14 presents the recommended process window for the ductile iron castings. The values are based on the orthogonal experiment, simulation verification, and actual casting trials. The pouring temperature should be maintained near 1300 °C. The filling speed should be controlled at approximately 0.460 m/s. The shell preheating temperature should be set at 800 °C. The gating system should include the pipe-section ingate from Scheme B. These conditions produced the lowest shrinkage rate and the most stable filling behavior.

Table 14. Recommended production window for the ductile iron castings
Parameter Recommended value Acceptable range
Pouring temperature 1300 °C 1280 to 1300 °C
Filling speed 0.460 m/s 0.455 to 0.460 m/s
Shell preheating temperature 800 °C 800 to 900 °C
Gating system Scheme B with pipe-section ingate Modified gating with late-solidifying feeding channel
Shell layers 6 6 to 7
Shell thickness About 6 mm 5 to 7 mm
Pouring position Top pouring Top pouring with controlled gravity direction

I also considered the broader implications for ductile iron castings produced by investment casting. The main lesson is that gating design and process parameters must be considered together. A gating system that appears adequate in terms of filling may still fail if the ingate solidifies before the casting body has finished feeding. The thermal modulus and Niyama criteria are useful for identifying hot spots and feeding blockages. The orthogonal experiment is an efficient way to evaluate the combined effects of pouring temperature, filling speed, and shell preheating temperature. For ductile iron castings with uniform wall thickness and complex local features, the filling speed is often the dominant parameter because it controls heat loss during filling. The shell preheating temperature is also critical because it controls the thermal gradient and the solidification sequence. The pouring temperature should be high enough to ensure fluidity but not so high that gas absorption and oxidation become problematic.

The simulation results also showed that the improved gating system reduced the shrinkage rate from 13.2603% to 1.3675% before parameter optimization. This indicates that the added pipe-section ingate shortened the feeding distance and provided a late-solidifying channel. The subsequent orthogonal optimization reduced the shrinkage rate further to 1.321%. Therefore, the combination of gating redesign and parameter optimization was more effective than either approach alone. In the actual production trial, the optimized process produced ductile iron castings with no visible shrinkage in the casting body, and the rejection rate decreased significantly. This confirmed that the numerical model was reliable and that the optimized parameters were practical for foundry production.

In summary, I optimized the investment casting process for safety-critical ductile iron castings used in oil drilling machine protective parts. I used Pro/E and ProCAST to simulate filling and solidification. I identified the original defects as shrinkage cavities and dispersed shrinkage caused by premature solidification of the ingate and blocked feeding. I redesigned the gating system by adding a pipe-section ingate, which reduced the shrinkage rate to 1.3675%. I then used an orthogonal experiment to optimize the pouring temperature, filling speed, and shell preheating temperature. The optimal parameters were 1300 °C, 0.460 m/s, and 800 °C. Under these conditions, the shrinkage rate of the ductile iron castings was reduced to 1.321%, the casting body showed no visible shrinkage defects, and actual production validated the improvement. These findings provide a practical basis for producing high-quality ductile iron castings with investment casting.

Scroll to Top