This study focuses on the development of a novel composite casting process that combines investment casting and sand casting techniques to produce a large, complex stainless steel impeller. The impeller has an outer diameter of 700 mm, a height of 130 mm, and a net weight of 202 kg, with 54 thin, flow-shaped blades. The primary objective is to achieve high dimensional accuracy, excellent surface finish, and sound internal quality while maintaining economic feasibility under existing production conditions. The proposed method divides the impeller into two parts: the complex blade portion, produced by investment casting with a silica sol-sodium silicate composite shell, and the simple hub and spoke portions, produced by sodium silicate sand moulding. The two parts are then assembled into a composite mould. This paper details the entire process, including wax pattern fabrication, shell building, sand mould preparation, melting, pouring, post-treatment, and quality inspection. The results show that the impeller meets all technical requirements, with dimensional accuracy of CT4–CT5, surface roughness of Ra3.2–1.6, and no significant casting defects. The process is stable, economical, and suitable for producing large complex impellers, demonstrating the benefits of integrating ‘sand casting parts’ with investment casting.
1. Introduction and Background
Casting impellers are critical components in shipbuilding, power generation, pharmaceutical, pump and valve, and transportation industries. Their complex geometry and stringent quality requirements make casting the preferred or even the only viable manufacturing method. For small and intricate impellers, investment casting is widely used due to its excellent surface finish and dimensional precision. However, investment casting has limitations in producing large castings due to shell strength issues and difficulties in feeding heavy sections. On the other hand, sand casting is suitable for large and simple geometries but cannot achieve the required surface quality and dimensional accuracy for complex blades. Therefore, a hybrid approach is necessary.
Traditional production of large impellers with many complex blades often relies on sand casting with core assemblies or loose patterns. These methods suffer from poor dimensional consistency, labor-intensive operations, and difficulty in maintaining blade spacing and rim ovality. Conversely, applying full investment casting to large impellers is impractical because of the heavy hub and spoke sections that require extensive feeding and are prone to shell deformation. Hence, this research proposes a composite mould technique that combines the strengths of both processes. The impeller is partitioned at the inner diameter of the rim: the blade region is made by investment casting, while the hub and spoke region is made by sand casting. This allows the use of economical ‘sand casting parts’ for simple geometries and high-precision investment casting for complex blades.
2. Impeller Structure and Technical Requirements
The impeller under study is a key component of a large pharmaceutical machine. Figure 1 shows its schematic structure (not reproduced here). The outer diameter is Φ700 mm, height is 130 mm, and net weight is 202 kg. The hub has a thick wall, the spoke is a large flat plate, and the blades are only 5 mm thick, closely spaced, and form a complex flow path. The material is ZG1Cr18Ni9Ti stainless steel, which requires corrosion resistance and specific mechanical properties.
The technical specifications are:
- Chemical composition must conform to GB2100-80.
- Blades must be free from slag inclusions, gas pores, cracks, shrinkage cavities, and porosity. Surface roughness Ra3.2–1.6 μm, dimensional tolerance CT4–CT5.
- Rim ovality must be within 0–4 mm, and the spacing difference between adjacent blades must not exceed 2 mm.
- Defects in hub, rim, and spoke must not exceed Φ2 mm.
- After heat treatment, the impeller must be non-magnetic and meet mechanical property standards.
These requirements are difficult to satisfy with either conventional sand casting or investment casting alone, necessitating the development of the composite process.
3. Selection of Casting Method and Process Scheme
3.1 Analysis of Sand Casting and Investment Casting
Sand casting is economical and flexible for large parts, but it cannot produce the fine surface finish and tight tolerances needed for the blades. The 54 blades, each with a complex profile and thin section, would require numerous cores or loose pieces, leading to accumulation of errors and poor dimensional stability. Moreover, the ZG1Cr18Ni9Ti alloy has poor fluidity and is prone to oxidation, increasing the risk of cold shuts and misruns in thin sections. Investment casting, on the other hand, provides excellent surface quality and dimensional accuracy for complex shapes, but the large hub and spoke create heavy sections that are difficult to feed and may cause shell cracking or distortion. The shell for such a large part would need excessive thickness and reinforcement, making the process costly and impractical.
Thus, a composite approach was selected: the blade portion is produced by investment casting using a silica sol-sodium silicate shell, while the hub and spoke portion is produced by sodium silicate sand moulding. The two mould components are assembled to form a single cavity, leveraging the advantages of both techniques. This also allows the use of standard ‘sand casting parts’ (moulds, cores, runner systems) for the simple geometries.
3.2 Process Scheme Selection
Two casting schemes were evaluated:
Scheme A: Three-box moulding with bottom gating, a blind riser on the hub, and four small vent risers on the rim. This was rejected because the long runner delays filling, leading to excessive oxidation and possible cold shuts in the thin blades. The rim hot spots would not be adequately fed, risking shrinkage defects.
Scheme B: Two-box moulding with a top sprue integrated into a riser on the hub, and three open risers on the rim. This ensures fast filling, effective feeding, and simplified moulding. The top gating could cause splashing, but this can be controlled by careful pouring practice and the low height (130 mm) of the casting. Scheme B was chosen for its superior feeding and ease of mould assembly.
4. Manufacturing of the Blade Investment Shell
4.1 Wax Pattern Design and Fabrication
The blade portion consists of 54 individual blades. Each blade was formed by a metal injection die (pressure type) made of 45 steel, hardened to HB 280–350. The die cavity dimensions account for the combined shrinkage of wax, shell, and metal. The combined linear shrinkage ε was calculated as:
$$ \varepsilon = \varepsilon_1 – \varepsilon_2 + \varepsilon_3 $$
where ε₁ is the wax pattern shrinkage (0.75% for rosin-based wax), ε₂ is the shell expansion during firing (0.50% for silica sol-sodium silicate shell), and ε₃ is the metal solidification shrinkage (2.5% for ZG1Cr18Ni9Ti). Thus, ε = 0.75 – 0.50 + 2.5 = 2.75%. Based on practical experience, a corrected value of εs = 2.6% was used. The die cavity nominal dimension L is given by:
$$ L = L_p (1 + \varepsilon_s) \pm \delta $$
where Lp is the mean casting dimension and δ is the manufacturing tolerance (typically 1/3 to 1/5 of the casting tolerance). To achieve CT4–CT5, the die was designed to CT3 accuracy and then trial-machined.
The wax pattern material was a modified rosin-wax blend. The composition and properties are listed in Table 1.
| Material/Property | Value |
|---|---|
| Polymerized rosin (%) | 17 |
| Modified rosin (%) | 40 |
| Paraffin wax (%) | 30 |
| EVA (%) | 3 |
| Melting point (°C) | 74–78 |
| Softening point (°C) | >40 |
| Tensile strength (MPa) | 54 |
| Shrinkage (%) | 0.4–0.7 |
Table 1. Composition and properties of the rosin-wax mould material.
Individual blade waxes were injected at 70–85 °C, pressure 0.3–1.5 MPa, holding time 1–3 min, followed by cooling in water at 16–25 °C. The wax patterns were inspected for surface defects and dimensional accuracy.
4.2 Assembly of Blade Wax Patterns
A special assembly fixture was designed to position the 54 blades precisely into the complete blade cluster. The fixture consisted of upper and lower cover plates, a locating ring, tie rods, and spacer plates. The locating ring outer diameter was calculated as:
$$ D_{ring} = D_{inner} (1 + \varepsilon_{sand}) $$
where Dinner = 416 mm and εsand = 2.2% (metal contraction for sand casting). Thus, Dring = 425.15 ± 0.25 mm. The ring height was (130 + 3 + 5) × (1 + 0.022) ≈ 141 mm.
Assembly began by placing the inspected blades in the fixture, aligning them with the locating ring. The blades were welded together at their touching surfaces using a heated soldering iron (about 80 °C). After welding one side, the fixture was turned over to weld the other side. The assembly was then clamped between the cover plates to prevent distortion during subsequent shell coating. The whole blade cluster formed a single wax pattern with the correct geometry.
4.3 Shell Building: Silica Sol-Sodium Silicate Composite Shell
The blade cluster was coated with a composite shell: the first two layers used a silica sol (ethyl silicate) binder to ensure a smooth surface, while the subsequent layers used sodium silicate for cost reduction and strength. The raw materials and formulations are summarized in Table 2.
| Layer | Binder | Filler | Sand size (mm) | Hardening | Drying |
|---|---|---|---|---|---|
| 1 | Ethyl silicate hydrolyzed | Quartz powder | 0.425/0.212 | Ammonia (air + NH₃) | >2 h |
| 2 | Ethyl silicate hydrolyzed | Quartz powder | 0.425/0.212 | Ammonia | >2 h |
| 3 | Sodium silicate | Quartz + bauxite (50/50) | 0.850/0.425 | NH₄Cl solution | 5–10 min |
| 4–6 | Sodium silicate | Quartz + bauxite (50/50) | 1.70/0.850 | NH₄Cl solution | 5–10 min |
Table 2. Composite shell building process.
For the ethyl silicate face coats, a hydrolyzed silica sol was prepared with a SiO₂ content of 30–34%. The coating viscosity was 28–32 s (standard flow cup, Φ6 mm). After dipping or flowing the coating over the wax cluster, the assembly was drained and stuccoed with fine quartz sand. Ammonia hardening was carried out in an ammonia chamber at 18–20 °C and relative humidity 45–80% for 30–40 min. The second layer was applied similarly.
For the sodium silicate backup coats, the coating was composed of 50% quartz powder and 50% bauxite, with a powder-to-liquid ratio of 1:1 for the third layer and 1:1.12–1.2 for subsequent layers. The viscosity was adjusted to 30–45 s. After stuccoing with coarser sand, each layer was hardened in a saturated NH₄Cl solution for 5–10 min. To prevent delamination between the second and third layers, the third layer viscosity was reduced and a longer air-drying time of 1–2 h was allowed before hardening.
The complete shell thickness after six layers was about 10–12 mm. Figure 2 shows the shell after coating (not reproduced).
4.4 Shell Reinforcement with Sodium Silicate Sand Sleeve
Because the blade shell is large and fragile, a reinforcing sleeve of CO₂-hardened sodium silicate sand was formed around the outer periphery of the shell. A steel reinforcement ring with two lifting lugs was placed inside the sleeve. The sleeve outer diameter was set to Φ973 mm to match the sand mould cavity. This sleeve provided rigidity during dewaxing, firing, handling, and final mould assembly. It also served as a location feature for assembling the investment shell into the sand mould.
4.5 Dewaxing and Firing
Dewaxing was performed in a pressurized steam autoclave at 0.6–1.0 MPa with a flash time of less than 5 s and a total dewaxing time of 6–10 min. The wax was completely removed. After dewaxing, the shell was stored for more than 12 h but less than 2 days to avoid surface degradation. Firing was conducted in an electric box furnace at 850 ± 30 °C for 1.5–2 h. The shell was placed horizontally on a bed of 20/40 mesh quartz sand to prevent distortion. The fired shell exhibited a white, smooth surface. It was not allowed to be re-fired because repeated heating could cause micro-cracks or delamination.
5. Sand Mould and Core Production for Hub and Spoke (‘Sand Casting Parts’)
5.1 Process Parameters
Machining allowances were: 5 mm on the top face, 3 mm on the bottom face, and 5 mm on the internal bore. The gating/riser system was designed using the modulus method. The modulus of the hub portion was calculated as:
$$ M_{hub} = \frac{a b}{2(a + b – c)} = \frac{55 \times 138}{2(55 + 138 – 30)} = 20.7 \ \text{mm} = 2.07 \ \text{cm} $$
With a safety factor k = 1.2 for an open riser, the riser modulus was:
$$ M_{hub, riser} = 1.2 \times 2.07 = 2.48 \ \text{cm} $$
Similarly, the rim modulus was calculated as Mrim = 1.89 cm, so Mrim, riser = 2.27 cm.
The solidification shrinkage of ZG1Cr18Ni9Ti was computed from alloy composition, as shown in Table 3.
| Element | Content (%) | Contribution factor | Contribution to ε (%) |
|---|---|---|---|
| C | 0.10 | 3.2 | 3.20 |
| Mn | 1.20 | 0.0585 | 0.070 |
| Si | 1.20 | 1.03 | 1.236 |
| Cr | 18.0 | 0.10 | 1.80 |
| Ni | 10.0 | −0.0354 | −0.354 |
| Total | 5.952 ≈ 6.0 |
Table 3. Calculation of solidification shrinkage of ZG1Cr18Ni9Ti.
Based on the required modulus and shrinkage, a standard waist-shaped open riser for the hub had dimensions a = 120 mm, b = 180 mm, h = 150 mm, weight 25 kg, capable of feeding 4.3 × 10³ cm³ of solid metal. The volume to be fed by the hub riser (within a radius of 14.65 cm) was calculated as 3.94 × 10³ cm³, so one riser sufficed. For the rim, a riser of a = 110 mm, b = 165 mm, h = 150 mm, weight 21 kg, feeding 3.0 × 10³ cm³ each. The volume to be fed by the rim risers (outside a radius of 14.65 cm) was 8.89 × 10³ cm³, requiring three risers. The effective feeding distance of one rim riser was 4T = 4 × 60 = 240 mm, plus the riser length 165 mm, giving 405 mm per riser. Three risers cover 1215 mm, which is less than the rim circumference of 1431.8 mm, but the cooling effect of the 54 blades assists in feeding, so three risers were deemed sufficient.
The pouring time was set to 10–12 s, giving a mean liquid rise speed of:
$$ v = \frac{0.138}{10 \sim 12} = 0.0115 \sim 0.0138 \ \text{m/s} $$
which exceeds the minimum recommended speed of 0.01 m/s. The process yield was:
$$ \frac{202}{202 + 88} \times 100\% = 69.6\% $$
Various other parameters: mould clamping force was calculated to be 10.56 × 10³ N, using weights for holding the moulds together.
5.2 Wooden Pattern and Core Box
For single-piece production, wooden patterns were chosen. The pattern dimensions were calculated using:
$$ A_M = (A_C \pm A_t)(1 + \varepsilon) $$
where AC is the casting dimension, At is the machining allowance, and ε = 2.2% is the metal contraction. The core box for the central hole was also made of wood, using a split construction.
5.3 Moulding Sand and Core Sand
Chromite sand was used for the facing sand and cores because of its high refractoriness (>1700 °C) and resistance to metal penetration. The sand mixture (by weight) is given in Table 4.
| Component | Proportion |
|---|---|
| Chromite sand (40/70 mesh) | 100 |
| Sodium silicate | 6–7 |
| Water | 4–5 |
| Wet permeability | >200 |
| Wet compressive strength (kPa) | 17–23 |
| Hardened compressive strength (MPa) | >1.5 |
Table 4. Composition and properties of CO₂-hardened sodium silicate chromite sand.
The sand was mixed in a muller: dry materials first, then sodium silicate and water, with a dry mixing time of 2–3 min and wet mixing time of 2–4 min. The moulds and cores were compacted by hand, and then hardened by blowing CO₂ gas through Φ6–10 mm vent holes at a pressure of 0.10–0.15 MPa for 20–30 s. After hardening, the patterns were withdrawn.
5.4 Coating for Sand Mould
To improve the surface finish and prevent sand adhesion, a zirconium-based alcohol coating was brushed onto the mould and core surfaces. The coating formula is listed in Table 5.
| Ingredient | Parts by weight |
|---|---|
| Zircon powder | 70 |
| Bauxite | 30 |
| Resin | 1.2–2.0 |
| Polyvinyl butyral | 1–4 |
| Sodium bentonite | 1–5 |
| Clay | 0.5–1.5 |
| Urotropine | 8 |
| Ethanol | balance |
| Water | 3.5–4.5 |
Table 5. Zirconium alcohol-based coating formula.
The coating was brush applied to a thickness of 0.3–1 mm and ignited to dry immediately, which also strengthens the surface and improves moisture resistance.
5.5 Mould Assembly (Composite Mould)
The firing shell (with the sand sleeve) was placed into the lower sand mould cavity, with its sleeve outer diameter (Φ973 mm) indexing into the corresponding location in the sand mould. The core for the hub bore was inserted. The upper sand mould was then lowered using guide pins, closing the mould. The parting line was sealed with asbestos paste or clay strips to prevent flash. The three rim risers and the central riser/sprue were aligned. Finally, the moulds were clamped with weights to resist the ferrostatic lift.
The assembled composite mould is illustrated schematically, showing the upper and lower sand moulds, the investment shell for the blades, the core, and the risers. This integration of investment cast blades and sand cast hub/spoke forms the final ‘sand casting parts’ cavity, demonstrating the synergy of the two casting technologies.
6. Melting and Pouring of ZG1Cr18Ni9Ti Stainless Steel
6.1 Melting Equipment
A 300 kg medium-frequency induction furnace with a basic (magnesia) lining was used. The lining was made of magnesia (MgO ≥ 90%) with boracic acid (0.8–1%) and sodium silicate solution. The lining was sintered at 1500–1550 °C before use. This basic lining is essential for stainless steel melting to avoid contamination.
6.2 Charge Calculation
The charge was computed for a 300 kg melt based on 100 kg reference. The target composition and recovery rates are listed in Table 6. The final charge composition was verified with a balance table.
| Element | Target (%) | Recovery (%) | Required per 100 kg (kg) |
|---|---|---|---|
| C | 0.07 | 100 | 0.07 |
| Si | 0.80 | 90 | 0.89 |
| Mn | 1.50 | 95 | 1.58 |
| Cr | 18.0 | 98 | 18.36 |
| Ni | 10.0 | 99.9 | 10.0 |
| Ti | 0.50 | 80 | 0.625 |
| S | 0.020 | 100 | 0.020 |
| P | 0.030 | 100 | 0.030 |
Table 6. Element recovery and required amounts per 100 kg of melt.
The final charge comprised 60 kg returned scrap (from previous heats), 20.3 kg low-phosphorus steel scrap, 12.68 kg ferrochromium, 5.01 kg electrolytic nickel, 1.083 kg ferrotitanium, 0.324 kg ferrosilicon, 0.601 kg electrolytic manganese, and aluminum for deoxidation. The calculated composition satisfied all specifications.
6.3 Melting Practice
The melting process followed a non-oxidizing route. The charge was loaded tightly in the crucible, with fine material at the bottom, alloys in the middle, and large pieces on top. Power was applied at 60% for the first 6–8 min, then increased to maximum. Slag formers (lime:fluorite = 2:1) were added to cover the melt. After melting, a sample was taken for analysis, and the slag was removed. A new slag was formed, and diffusion deoxidation was performed using a mixture of lime and aluminum powder. Before tapping, the temperature was adjusted to 1630–1650 °C. Ferrotitanium was added and then aluminum (0.1%) for final deoxidation. The melt was tapped at 1620 ± 10 °C.
6.4 Pouring
The mould was poured using a 300 kg bottom-pour ladle. The pouring temperature was 1580 ± 10 °C. Argon gas was purged through the melt in the ladle to reduce oxidation. The pouring sequence was slow-fast-slow, with a fill time of 10–12 s. After filling the risers, the pouring was continued briefly to top them up, and then insulating powder was added to the risers to enhance feeding. The casting was allowed to cool in the mould for about 1 hour, then knocked out to accelerate cooling and obtain a finer structure.
7. Post-Casting Treatments
7.1 Shakeout and Cleaning
After cooling to 50–60 °C, the sand mould and sand sleeve were mechanically removed. The investment shell around the blades was removed by alkali boiling in a 15–20% NaOH solution at boiling temperature for 4–8 hours, followed by neutralization in a solution containing chromic acid (90 g), sulfuric acid (30 g), sodium chloride (1.2 g) per kg of water, and rinsing in water. This effectively cleaned the narrow inter-blade channels without damaging the surface.
7.2 Cutting of Risers and Gates
Stainless steel risers are difficult to cut by ordinary oxy-fuel methods due to the formation of refractory chromium oxide films. The vibration oxygen cutting method was adopted. Using a G01-300 torch with a No. 4 nozzle, the oxygen pressure was 0.8–1.0 MPa and acetylene pressure 0.05–0.1 MPa. The torch was vibrated with an amplitude of 10–15 mm and a frequency of about 80 cycles/min. This method separated the risers effectively, leaving a residual height of 5–8 mm, which was subsequently ground flush.
7.3 Heat Treatment
The impeller was subjected to solution treatment followed by stabilization treatment. Solution treatment: heat from below 300 °C to 1050–1100 °C at a heating rate of 200–300 °C/h, hold for 3–4 h, then quench in water (water temperature above 950 °C during quench). Stabilization treatment: heat to 850–950 °C, hold for 2–4 h, then air cool. This heat treatment dissolves chromium carbides and prevents sensitization, restoring corrosion resistance and ensuring non-magnetic behavior.
8. Results and Discussion
8.1 Chemical Composition
Spectrochemical analysis (using an XB-7503B directly reading spectrometer) was performed on test bars poured with the impeller. The results are compared with the standard in Table 7.
| Element | Analysis 1 | Analysis 2 | Target | GB2100-80 |
|---|---|---|---|---|
| C | 0.080 | 0.083 | 0.079 | ≤0.12 |
| Si | 1.0 | 1.1 | 0.887 | ≤1.5 |
| Mn | 1.5 | 1.3 | 1.585 | 0.8–2.0 |
| Cr | 18.0 | 18.5 | 18.37 | 17–20 |
| Ni | 9.0 | 10.0 | 10.10 | 8–11 |
| Ti | 0.62 | 0.58 | 0.625 | 0.5–0.7 |
| S | 0.026 | 0.025 | 0.0175 | ≤0.030 |
| P | 0.032 | 0.030 | 0.0279 | ≤0.045 |
Table 7. Chemical composition of the cast impeller.
All elements met the standard, with impurities lower than specified.
8.2 Surface and Dimensional Inspection
Visual inspection after cleaning showed no flash, sand adhesion, cold shuts, or misruns. The blade surfaces were smooth and the rim was free from defects. Dye penetrant testing on the blade roots and critical areas revealed only minor linear indications of 1.5 mm length (Grade 01) according to GB/T9443-1988, well within acceptance limits.
Surface roughness was measured against comparator blocks: Ra 1.6 μm on blades and Ra 3.2–1.6 μm on other surfaces, satisfying the specification.
Rim ovality measured with a roundness tester was less than 0.5 mm. The maximum difference between adjacent blade spacings was 0.8 mm. Flatness of the top and bottom surfaces was within 1 mm, below the machining allowances. Dimensional inspection using standard gauges confirmed that the casting achieved CT4–CT5 accuracy, and blade positions were even better (CT3).
8.3 Ultrasonic Inspection
Ultrasonic testing was performed on the hub, rim, and spoke using a CTS-26 flaw detector with a 2.5p20Z probe. The sensitivity was set to detect a Φ2 mm flat-bottom hole. The formulas used for sensitivity adjustment were:
For hub/rim (depth range 250 mm):
$$ \Delta = 20 \log \frac{2\lambda x}{\pi D^2} = 20 \log \frac{2 \times 5.9 \times 138}{2.5 \times 3.14 \times 2^2} = 34.3 \ \text{dB} $$
For spoke (depth range 50 mm):
$$ \Delta = 20 \log \frac{2 \times 5.9 \times 30}{2.5 \times 3.14 \times 2^2} = 21 \ \text{dB} $$
No defects above Φ2 mm were indicated in any of the test zones, confirming the soundness of the casting.
8.4 Microstructure
Metallographic samples were prepared from the riser contact area and observed under an optical microscope at 100× magnification. In the as-cast state, the microstructure consisted of austenite dendrites with ferrite and granular carbides. After solution treatment, the microstructure became austenite with a small amount of ferrite, and after solution plus stabilization treatment, the structure remained austenitic with minor ferrite. The absence of significant chromium carbide precipitation ensures good intergranular corrosion resistance. The impeller was confirmed to be non-magnetic after heat treatment.
8.5 Mechanical Properties
Brinell hardness measurements (HB 5/750) on the blade and rim gave values of 170, 175, and 180 HB, all below the specified maximum of 187 HB. Tensile test results from separately cast test bars (after heat treatment) are listed in Table 8.
| Property | Samples | Standard GB228-87 |
|---|---|---|
| Tensile strength σb (MPa) | 495, 510, 502 | ≥450 |
| Yield strength σs (MPa) | 240, 260, 250 | ≥196 |
| Elongation δ5 (%) | 35, 37, 39 | ≥25 |
| Reduction of area ψ (%) | 50, 50, 45 | ≥32 |
| Impact toughness Ak (×10⁵ J/m²) | 15, 17, 13 | ≥10 |
Table 8. Tensile and impact test results.
The mechanical properties comfortably exceed the requirements, with consistent results across samples. The slightly lower values are attributed to the thick section of the test bars; the thin blades are expected to exhibit even better properties.
9. Conclusions
This research successfully developed a melting modules-sand mould composite casting process for a large, complex ZG1Cr18Ni9Ti stainless steel impeller. The key conclusions are as follows:
- The division of the impeller into an investment-cast blade portion and a sand-cast hub/spoke portion (‘sand casting parts’) effectively addresses the conflicting requirements of dimensional accuracy, surface finish, and cost.
- The designed assembly fixture for the blade wax patterns enabled precise positioning and welding of the 54 blades, ensuring the required blade spacing and geometry.
- The silica sol-sodium silicate composite shell produced a smooth, accurate, and strong mould for the blades. The addition of a sodium silicate sand sleeve provided necessary rigidity and facilitated handling and alignment.
- The sand mould for the hub, spoke, and gating/riser system was economically produced using chromite sand and CO₂-hardened sodium silicate, integrating smoothly with the investment shell.
- Controlled melting in a basic lined induction furnace and optimized pouring practice (1580 °C, 10–12 s fill time) yielded clean, defect-free castings.
- Post-casting treatments, including alkali cleaning, vibration oxygen cutting, and solution plus stabilization heat treatment, ensured proper surface, dimensions, and mechanical properties.
- The produced impeller met all technical requirements: dimensions CT4–CT5, surface roughness Ra3.2–1.6, rim ovality <0.5 mm, blade spacing deviation <1 mm, no significant internal defects, and excellent mechanical properties.
- The process is stable and cost-effective, producing ten impellers with a direct economic benefit of over 100,000 RMB, and the product successfully replaced imported components.
The melting modules-sand mould composite casting process thus provides a practical solution for manufacturing large, complex impellers, combining the high precision of investment casting with the versatility and economy of sand casting parts.