I approach the manufacturing of a thin-walled shell casting through a systematic precision casting route. The component is small, yet its geometry creates a severe local thermal imbalance. The external envelope measures 28 mm × 38 mm × 14 mm, the mass is 6 g, and the wall thickness varies from 2 mm to 12 mm. Such a large thickness ratio is a classic challenge in precision casting because the thin sections cool rapidly while the thick sections remain liquid for a longer time. The thin 2 mm wall also contains a narrow groove, which makes shell building difficult and can create local shell buildup. The material is ZG35CrMnSi, and the quality requirements include magnetic particle inspection and X-ray inspection. The main defect observed in early trials was shrinkage porosity in the thin plate region. I therefore focused my process design on feeding channel continuity, shell thickness uniformity, and thermal balance during solidification. The following work describes the original process, the failure mechanism, the optimized gating and tree assembly, the shell-making parameters, dewaxing, melting and pouring, and the trial validation results.
1. Casting geometry and quality requirements
The shell casting is a small precision casting with a complex wall thickness distribution. I recorded the basic data in Table 1. The maximum thickness is six times the minimum thickness, which gives a thickness ratio of 6. In precision casting, such a ratio means that the thin wall can solidify before the thick hot spot has completed feeding. The groove on the 2 mm wall further complicates the local heat transfer because the groove can trap slurry and sand during shell making. The result is a locally thicker shell, which acts as an insulating blanket. I used the geometric data to estimate the thermal modulus of the thin wall and the hot spot.
| Parameter | Value | Unit |
|---|---|---|
| Overall length | 28 | mm |
| Overall width | 38 | mm |
| Overall height | 14 | mm |
| Mass | 6 | g |
| Minimum wall thickness | 2 | mm |
| Maximum wall thickness | 12 | mm |
| Wall thickness ratio | 6 | – |
| Material | ZG35CrMnSi | – |
| Inspection | Magnetic particle and X-ray | – |
| Main defect | Shrinkage porosity in thin plate | – |
The thickness ratio is defined as:
$$R_t = \frac{t_{\max}}{t_{\min}} = \frac{12}{2} = 6$$
A high thickness ratio promotes directional solidification only if a proper feeding path exists. The thermal modulus of a section is:
$$M = \frac{V}{A}$$
where \(V\) is the volume and \(A\) is the heat-dissipating surface area. For the thin 2 mm plate, the modulus is small, so it solidifies quickly. For the 12 mm hot spot, the modulus is larger, so it remains liquid longer. If the thin plate is isolated from the hot spot by a narrow groove or by a thick shell, the liquid metal in the hot spot cannot feed the shrinkage that forms in the plate. The result is shrinkage porosity. I used the Niyama criterion as a qualitative indicator:
$$N_y = \frac{G}{\sqrt{\dot{T}}}$$
where \(G\) is the temperature gradient and \(\dot{T}\) is the cooling rate. A low \(N_y\) value indicates a high risk of shrinkage porosity. In the original process, the groove region had a low temperature gradient and a low cooling rate because the shell was locally thick. The Niyama value was therefore below the critical threshold for this alloy. This explained the X-ray indications in the thin plate.
2. Original process and defect mechanism
The original pouring system placed inner gates at the hot spot of the casting. I used three inner gates, but the thin plate between two of them still showed shrinkage porosity. The groove was filled with shell material after the second and third slurry and sand layers. During shell building, the groove acted as a blind pocket. The slurry accumulated, and the sand could not be cleaned out completely. After firing at 1050 °C for 50 min, the groove contained a dense ceramic mass. During pouring, this ceramic mass contacted the inner side of the thin plate and reduced heat extraction. The plate therefore cooled more slowly than it should have, but it still had no direct feed path. The two adjacent inner gates did not feed the plate because the distance was too large and the thermal gradient was not favorable. The local hot spot created by the ceramic-filled groove became an artificial hot zone. The combination of slow cooling and no feeding produced shrinkage porosity.
I summarized the original process parameters in Table 2. The original tree assembly had 12 castings per group and a sprue diameter of 30 mm. The distance from each casting to the sprue was greater than 25 mm. The groove faced inward in the original tree, which made drying difficult. The shell drying air speed was not controlled tightly, and the groove region often remained wet. This caused cracks and uneven shell thickness. The original process had a yield of only 10%.
| Parameter | Original process | Optimized process |
|---|---|---|
| Number of inner gates | 3 | 4 |
| Gate at groove | No | Yes |
| Groove orientation | Inward | Outward |
| Gate cross-section | Not specified | 4 mm × 12 mm |
| Sprue diameter | 30 mm | 30 mm |
| Castings per tree | 12 | 12 |
| Distance to sprue | >25 mm | >25 mm |
| Shell layers | 5 + seal | 5 + seal |
| Face coat viscosity | 36 s | 36 s |
| Pouring temperature | 1630 ± 10 °C | 1630 ± 10 °C |
| Shell preheat | 1050 ± 10 °C | 1050 ± 10 °C |
The defect location was not random. It appeared exactly between two inner gates, in the 2 mm wall that faced the groove. The X-ray inspection showed scattered shrinkage porosity. Magnetic particle inspection did not show surface cracks, but the internal porosity was unacceptable. I measured the yield as:
$$Y = \frac{N_g}{N_t} \times 100\%$$
For the original process, \(N_g = 6\) and \(N_t = 60\), so \(Y = 10\%\). The defect rate was:
$$D = 1 – Y = 90\%$$
This high defect rate confirmed that the original gating and tree design were not robust. I needed to change both the thermal environment and the feeding path.
3. Optimized pouring system design
I redesigned the pouring system to provide a direct feed path to the thin plate. The key change was to add an inner gate at the groove location. The new gate was placed on a planar surface, not on the casting body, so that it could be cut off easily. The gate cross-section was 4 mm × 12 mm. I avoided making the gate too large because a large gate would be difficult to remove and could create a new hot spot. The gate was designed to feed the thin plate while allowing the main hot spot to feed through the original gates. The optimized system therefore had four inner gates: three at the hot spot and one at the groove. The gate at the groove provided a short feeding path to the thin wall. The other three gates maintained feeding of the thick sections.
I used the following feeding distance relationship to check the gate position:
$$L_f = k \sqrt{M}$$
where \(L_f\) is the effective feeding distance, \(k\) is an alloy-dependent coefficient, and \(M\) is the modulus. For ZG35CrMnSi, I used a conservative \(k\) value. The distance from the new gate to the center of the thin plate was less than the calculated \(L_f\). This ensured that the liquid metal could reach the last solidifying region. I also checked the thermal resistance of the shell:
$$R = \frac{t}{k A}$$
where \(t\) is the shell thickness, \(k\) is the thermal conductivity, and \(A\) is the area. By placing the groove outward, I reduced the local shell thickness because the slurry and sand could drain more easily. The thermal resistance decreased, so the cooling rate increased. A higher cooling rate increased the Niyama value and reduced the risk of shrinkage porosity.
The new gate dimensions are summarized in Table 3. The gate was 4 mm wide and 12 mm long. The contact area was small enough for easy removal. The gate was located on a flat surface so that the remaining gate root could be removed by grinding. The gate did not protrude into the groove, which avoided slurry accumulation. The gate was also oriented to avoid direct impingement of liquid metal on the thin wall. This reduced erosion and turbulence.
| Gate parameter | Value | Unit |
|---|---|---|
| Number of gates at hot spot | 3 | – |
| Number of gates at groove | 1 | – |
| Groove gate width | 4 | mm |
| Groove gate length | 12 | mm |
| Gate location | Planar surface | – |
| Removal method | Cutting and grinding | – |
| Feeding path to thin wall | Short and direct | – |
The optimized gating also changed the filling pattern. In the original design, the liquid metal entered the mold through three gates and then had to flow around the groove. The groove acted as a dead zone. In the optimized design, the fourth gate filled the thin plate directly. The filling time was slightly longer, but the velocity was lower. I estimated the Reynolds number for the new gate:
$$Re = \frac{\rho v D}{\mu}$$
where \(\rho\) is density, \(v\) is velocity, \(D\) is hydraulic diameter, and \(\mu\) is dynamic viscosity. The calculated \(Re\) was in the laminar range, so the filling was stable. This reduced the risk of gas entrapment and oxide film formation. The gate also acted as a chill because it was thinner than the hot spot. The chill effect increased the local temperature gradient and promoted directional solidification toward the gate. This further improved feeding.
4. Tree assembly and groove orientation
I kept the same basic tree layout but changed the groove orientation. The sprue diameter was 30 mm, and each group contained 12 castings. The distance from each casting to the sprue was greater than 25 mm. This distance reduced the thermal radiation from the sprue to the casting. In precision casting, the sprue is a large hot mass. If the casting is too close to the sprue, the sprue can preheat the casting and change the solidification sequence. The 25 mm minimum distance helped maintain a more uniform temperature field. I summarized the tree parameters in Table 4.
| Tree parameter | Value | Unit |
|---|---|---|
| Sprue diameter | 30 | mm |
| Castings per tree | 12 | – |
| Distance to sprue | >25 | mm |
| Groove orientation | Outward | – |
| Tree symmetry | Balanced | – |
| Sprue material | Wax | – |
| Gate connection | Direct | – |
The most important change was that the groove faced outward. In the original tree, the groove faced inward, toward the sprue and other castings. This made the groove difficult to observe during shell making. The slurry and sand accumulated in the groove, and the drying air could not reach the groove uniformly. In the optimized tree, the groove faced outward. This allowed the operator to see the groove during dipping, draining, and sanding. The slurry could drain out of the groove, and the sand could be blown out with compressed air. The drying air could flow over the groove evenly. The result was a thinner, more uniform shell at the groove. The thermal resistance decreased, and the cooling rate increased. The Niyama value increased, and the shrinkage porosity risk decreased.
I also considered the drying uniformity. The drying rate can be expressed as:
$$R_d = D \frac{\partial C}{\partial x}$$
where \(D\) is the diffusion coefficient, \(C\) is the moisture concentration, and \(x\) is the distance. When the groove faces outward, the moisture diffusion path is shorter and more uniform. When the groove faces inward, the moisture is trapped, and the drying rate is low. This leads to uneven drying, cracks, and shell thickness variation. The outward orientation solved this problem. It also made cleaning easier. The groove could be inspected visually, and any excess ceramic could be removed before firing.
5. Shell making process
I used a five-layer shell plus a seal coat. The face coat was zircon-based. The slurry viscosity was controlled at 36 s. The sand was 120 mesh zircon. For the transition and back layers, I used mullite powder and mullite sand. The details are given in Table 5. The shell-making process was designed to produce a uniform shell around the thin wall and the groove. The face coat is critical because it directly contacts the molten metal. A uniform face coat prevents metal penetration and improves surface finish. The back layers provide strength. The seal coat prevents sand fall-off and controls permeability.
| Layer | Powder and mesh | Slurry viscosity (s) | Sand and mesh |
|---|---|---|---|
| 1 | Zircon powder, 320 mesh | 36 | Zircon sand, 120 mesh |
| 2 | Mullite powder, 200 mesh | 15 | Mullite sand, 30–60 mesh |
| 3 | Mullite powder, 200 mesh | 12 | Mullite sand, 16–30 mesh |
| 4 | Mullite powder, 200 mesh | 12 | Mullite sand, 16–30 mesh |
| 5 | Mullite powder, 200 mesh | 12 | Mullite sand, 16–30 mesh |
| 6 (seal) | Mullite powder, 200 mesh | 10 | None |
Before dipping the transition and back layers, I used compressed air to clean the groove. I also used a brush to remove loose sand. This step is essential in precision casting because any loose sand in the groove can create a bridge. A bridge is a ceramic connection that spans the groove and blocks the feeding path. It also creates a thick shell region that acts as an insulator. By cleaning the groove before each dip, I ensured that the slurry could coat the surface evenly without accumulating. I also controlled the slurry viscosity. A viscosity that is too high causes thick coating and poor drainage. A viscosity that is too low causes thin coating and weak shell. The face coat viscosity of 36 s was chosen to balance these effects.
I controlled the drying process carefully. The air speed in the drying room was 3–5 m/s. The temperature and humidity were controlled according to the shell-making procedure. The drying rate is affected by air speed, temperature, and humidity. I used the following relationship to estimate the drying time:
$$t_d = \frac{\rho_s V_w}{h_m A (C_s – C_\infty)}$$
where \(t_d\) is the drying time, \(\rho_s\) is the solid density, \(V_w\) is the water volume, \(h_m\) is the mass transfer coefficient, \(A\) is the surface area, \(C_s\) is the surface moisture concentration, and \(C_\infty\) is the ambient moisture concentration. The outward groove orientation increased \(A\) and decreased the diffusion distance, which reduced \(t_d\). The uniform air speed of 3–5 m/s ensured that all parts of the tree dried at a similar rate. This reduced the risk of cracks and shell distortion.
The shell thickness is also important. The shell thickness can be estimated by:
$$t_s = k \sqrt{\eta}$$
where \(t_s\) is the shell thickness, \(k\) is a process constant, and \(\eta\) is the slurry viscosity. I kept the viscosity constant for each layer to maintain a uniform shell. The seal coat was kept thin. A thick seal coat reduces permeability and can cause gas defects. The seal coat was applied only thick enough to prevent sand fall-off. This maintained the shell permeability and allowed gases to escape during pouring.
6. Dewaxing
I used a steam autoclave for dewaxing. The shell was placed with the pouring cup facing down. The dewaxing car was not overloaded, and the shells were stable. The transfer time from the shell-making room to the autoclave was less than or equal to 60 s. This short transfer time prevented the wax from cooling and expanding, which can crack the shell. The dewaxing parameters are given in Table 6. The internal temperature was 175–185 °C. The steam boiler pressure upper limit was 0.8 ± 0.1 MPa, and the lower limit was 0.76 ± 0.1 MPa. The autoclave filling time was 1000 ± 20 s. The dewaxing time was 20 ± 5 s. The drain preheating pressure was 0.05–0.06 MPa. The wax discharge time was 100–500 s. The water discharge time was 50 ± 2 s.
| Parameter | Set value |
|---|---|
| Internal temperature | 175–185 °C |
| Steam boiler pressure upper limit | 0.8 ± 0.1 MPa |
| Steam boiler pressure lower limit | 0.76 ± 0.1 MPa |
| Autoclave filling time | 1000 ± 20 s |
| Dewaxing time | 20 ± 5 s |
| Drain preheating pressure | 0.05–0.06 MPa |
| Wax discharge time | 100–500 s |
| Water discharge time | 50 ± 2 s |
| Transfer time to autoclave | ≤60 s |
The dewaxing process must be fast and uniform. If the wax is heated too slowly, it expands and cracks the shell. If the pressure is too high, the shell can be damaged. I used the specified pressure range to ensure rapid melting and removal of the wax. The pouring cup facing down allowed the molten wax to drain out of the shell. The short dewaxing time of 20 ± 5 s was sufficient because the wax is a low-melting-point material. After dewaxing, the shell was inspected for cracks. Any cracked shell was rejected. The shell was then fired in a furnace.
7. Melting and pouring
I used a medium-frequency induction furnace to melt the ZG35CrMnSi master alloy. The pouring method was gravity pouring. Before melting, I checked the furnace body for damage, the cooling water pipes for leaks, the tilting mechanism, and the temperature measuring equipment. The master alloy bars were loaded so that the top of the bars did not exceed the induction coil height. I started with 60% power. After the current stabilized, I increased the power to the maximum. The melting process was controlled to avoid excessive overheating and gas absorption. The pouring parameters are listed in Table 7.
| Parameter | Value |
|---|---|
| Melting furnace | Medium-frequency induction furnace |
| Material | ZG35CrMnSi master alloy |
| Pouring method | Gravity pouring |
| Pouring temperature | 1630 ± 10 °C |
| Shell preheat temperature | 1050 ± 10 °C |
| Shell preheat time | 50 ± 5 min |
| Cooling method | Sand bed natural cooling |
| Insulating cover | Applied to pouring cup |
The pouring temperature of 1630 ± 10 °C was chosen to ensure good fluidity without excessive superheat. A high superheat can cause grain growth and gas porosity. A low superheat can cause cold shuts and misruns. The shell preheat temperature of 1050 ± 10 °C reduced the thermal shock on the shell and improved filling. The shell preheat time of 50 ± 5 min ensured that the shell was uniformly heated. After firing, the shell was removed from the furnace and poured directly. The shell was placed on a sand bed for natural cooling. I added an insulating cover to the pouring cup. The insulating cover reduced heat loss from the sprue and improved feeding. The feeding effect can be estimated by the thermal modulus of the sprue:
$$M_s = \frac{V_s}{A_s}$$
where \(V_s\) is the sprue volume and \(A_s\) is the sprue surface area. The insulating cover increased the effective \(M_s\) by reducing heat loss. This increased the feeding time and improved the feeding of the casting. The sand bed also reduced the cooling rate of the bottom of the shell. This helped maintain a positive temperature gradient from the thin wall to the sprue.
8. Trial results
I conducted five trial groups for each process. Each group contained 12 castings, so each process had 60 castings. The original process produced 6 good castings, a yield of 10%. The optimized process produced 52 good castings, a yield of 86.7%. The results are summarized in Table 8. The main defect in the original process was shrinkage porosity in the 2 mm wall. The X-ray inspection showed scattered porosity. The optimized process eliminated most of this defect. The few remaining defects were minor and were attributed to random process variation. After the optimized process was put into batch production, the first-pass yield reached more than 95% through refined operation. This confirmed that the optimized process was robust.
| Process | Castings invested | Good castings | Yield (%) | Main defect |
|---|---|---|---|---|
| Original | 60 | 6 | 10.0 | Shrinkage porosity |
| Optimized | 60 | 52 | 86.7 | Minor porosity |
| Batch production | – | – | >95 | – |
The yield improvement can be expressed as:
$$\Delta Y = Y_{\text{opt}} – Y_{\text{orig}} = 86.7\% – 10.0\% = 76.7\%$$
The defect reduction was:
$$\Delta D = D_{\text{orig}} – D_{\text{opt}} = 90.0\% – 13.3\% = 76.7\%$$
These values show a dramatic improvement. The optimized process changed the thermal and feeding conditions in the critical thin wall. The added gate at the groove provided a direct feed path. The outward groove orientation reduced shell thickness and improved cooling. The combination of these two changes increased the Niyama value above the critical threshold. The result was a sound thin wall with no X-ray indications of shrinkage porosity.
9. Discussion
The physics of the defect is clear. In the original process, the groove was filled with ceramic. The ceramic had a lower thermal conductivity than the metal. The local thermal resistance increased. The cooling rate decreased. The temperature gradient also decreased because the thin wall was not connected to a strong heat sink. The Niyama criterion was therefore low. The thin wall solidified last in that local region, and there was no liquid metal to feed it. Shrinkage porosity formed. The two adjacent inner gates did not help because the feeding distance was too long and the thermal gradient was not favorable. The gate at the groove solved this by providing a short feeding path. The gate also acted as a chill. The chill increased the temperature gradient toward the gate. The liquid metal in the gate and sprue could then feed the shrinkage. The outward groove orientation reduced the shell thickness. The reduced shell thickness increased the cooling rate. The increased cooling rate increased the Niyama value. The combination of feeding and cooling eliminated the defect.
I also considered the role of precision casting process control. In precision casting, the shell is the mold. The shell thickness and permeability directly affect heat transfer and gas escape. A thick shell insulates the metal and slows cooling. A thin shell cools the metal faster but may be weak. The groove is a geometric feature that can cause shell buildup. By orienting the groove outward, I made it possible to control the shell thickness. The operator could see the groove and clean it. The slurry could drain. The sand could be blown out. The drying air could reach the groove. This is a practical but critical point in precision casting. Many defects in thin-walled precision castings come from shell buildup in recesses. The solution is not always a change in gating; sometimes it is a change in orientation and handling.
I used the following heat balance equation to analyze the cooling of the thin wall:
$$q = h A (T_s – T_0)$$
where \(q\) is the heat flux, \(h\) is the heat transfer coefficient, \(A\) is the surface area, \(T_s\) is the surface temperature, and \(T_0\) is the ambient temperature. When the shell is thick, \(h\) is low. When the shell is thin, \(h\) is high. The outward groove orientation reduced the local shell thickness, which increased \(h\). The increased \(h\) increased \(q\). The increased heat flux increased the cooling rate. The increased cooling rate increased the Niyama value. This chain of effects explains why the simple change in orientation improved the yield so much.
I also considered the solidification time:
$$t_s = \frac{\rho V [L + c(T_m – T_s)]}{h A (T_s – T_0)}$$
where \(t_s\) is the solidification time, \(\rho\) is the density, \(V\) is the volume, \(L\) is the latent heat, \(c\) is the specific heat, \(T_m\) is the melting temperature, \(T_s\) is the solidus temperature, \(h\) is the heat transfer coefficient, \(A\) is the surface area, and \(T_0\) is the mold temperature. The thin wall has a small \(V/A\) ratio, so it solidifies quickly. The hot spot has a large \(V/A\) ratio, so it solidifies slowly. The feeding path must remain open until the thin wall has solidified. The added gate at the groove kept the feeding path open. The gate was not too large, so it did not become a new hot spot. The gate solidified after the thin wall but before the main hot spot. This is the ideal feeding sequence. The main hot spot then fed through the original gates. The result was a sound casting.
I also examined the effect of the insulating cover on the pouring cup. The insulating cover reduced heat loss from the sprue. The sprue acted as a reservoir. A hot sprue can feed the casting for a longer time. The thermal modulus of the sprue is:
$$M_s = \frac{V_s}{A_s}$$
The insulating cover increased the effective \(M_s\). The increased \(M_s\) increased the feeding time. The increased feeding time allowed the thin wall to receive liquid metal after it had started to contract. This reduced the shrinkage porosity. The sand bed also played a role. The sand bed reduced the cooling rate of the bottom of the shell. This helped maintain a positive temperature gradient from the thin wall to the sprue. The positive temperature gradient promoted directional solidification. The directional solidification moved the shrinkage from the thin wall to the sprue. The sprue was later removed, so the shrinkage did not affect the casting.
10. Process control and repeatability
To ensure repeatability, I standardized the following steps. First, the groove was always oriented outward in the tree. Second, the groove was cleaned with compressed air and a brush before each transition and back dip. Third, the slurry viscosity was checked before each layer. Fourth, the drying air speed was maintained at 3–5 m/s. Fifth, the dewaxing transfer time was kept below 60 s. Sixth, the pouring temperature and shell preheat temperature were controlled within the specified ranges. Seventh, the insulating cover was applied to every pouring cup. These controls reduced process variation and improved the first-pass yield to more than 95% in batch production.
I also created a defect-cause matrix, shown in Table 9. This matrix helped the team understand the relationship between process variables and defects. The matrix is a useful tool in precision casting because many defects have multiple causes. By identifying the primary cause, the team could focus on the most effective corrective action.
| Defect | Primary cause | Corrective action |
|---|---|---|
| Shrinkage porosity in thin wall | No feeding path; thick shell at groove | Add gate at groove; orient groove outward |
| Shell buildup in groove | Groove facing inward; poor drainage | Orient groove outward; blow out sand |
| Uneven drying | Poor air circulation in groove | Control air speed 3–5 m/s; outward orientation |
| Gas porosity | Thick seal coat; low permeability | Keep seal coat thin |
| Cold shut | Low pouring temperature | Control pouring at 1630 ± 10 °C |
| Shell crack | Slow dewaxing transfer | Transfer within 60 s |
The process capability can be expressed by the process capability index:
$$C_p = \frac{USL – LSL}{6\sigma}$$
where \(USL\) is the upper specification limit, \(LSL\) is the lower specification limit, and \(\sigma\) is the standard deviation. For the thin wall thickness, the specification is 2 mm. The process capability improved after the optimization because the variation in shell thickness decreased. The outward groove orientation reduced the standard deviation of the shell thickness. The added gate reduced the variation in feeding. The result was a more capable process. This is important in precision casting because small castings often have tight tolerances and strict inspection requirements.
11. Economic and production impact
The yield improvement had a direct economic impact. The original yield of 10% meant that 90% of the castings were scrapped. The optimized yield of 86.7% reduced the scrap rate to 13.3%. In batch production, the yield exceeded 95%, reducing the scrap rate to less than 5%. The reduction in scrap saved material, energy, and labor. The optimized process also reduced the need for rework and repair. The added gate was easy to remove, so the cleaning cost did not increase significantly. The outward groove orientation made shell making easier and faster. The operator could see the groove and clean it quickly. The drying time was reduced because the groove dried faster. The overall production efficiency improved.
I calculated the material utilization as:
$$U = \frac{m_{\text{casting}}}{m_{\text{pouring}}} \times 100\%$$
where \(m_{\text{casting}}\) is the casting mass and \(m_{\text{pouring}}\) is the poured mass. The optimized process reduced the scrap, so the effective utilization increased. The energy consumption per good casting also decreased. The energy consumption can be estimated by:
$$E = \frac{P t}{N_g}$$
where \(E\) is the energy per good casting, \(P\) is the power, \(t\) is the process time, and \(N_g\) is the number of good castings. As \(N_g\) increased, \(E\) decreased. This made the process more sustainable and cost-effective.
12. Conclusion
I designed and optimized a precision casting process for a thin-walled shell casting with a narrow groove. The original process had a 10% yield because the groove was filled with ceramic, which created a local hot spot and blocked feeding. The thin 2 mm wall had no direct feed path, so shrinkage porosity formed. I optimized the process by adding an inner gate at the groove and orienting the groove outward in the tree. The gate cross-section was 4 mm × 12 mm. The gate was placed on a planar surface for easy removal. The outward orientation improved shell drainage, drying, and cleaning. The shell-making process used five layers plus a seal coat, with a zircon face coat and mullite back layers. The dewaxing, melting, and pouring parameters were controlled within tight ranges. The optimized process achieved a yield of 86.7% in the trial and more than 95% in batch production. The key lessons for precision casting are: first, recesses and grooves must be oriented to allow drainage and drying; second, hot spots must have a direct feeding path; third, shell thickness must be uniform to avoid local insulation; and fourth, process control must be standardized to ensure repeatability. These principles can be applied to other thin-walled precision castings with similar geometry.
In summary, the optimized precision casting process eliminated the shrinkage porosity defect, improved the yield, and provided a robust production route for the shell casting. The combination of gating optimization and tree orientation was the critical factor. The process is now used in batch production with a first-pass yield of more than 95%. This demonstrates that careful attention to feeding and shell uniformity can solve difficult defects in precision casting.