As a key component in canned pumps, the self-priming pump body must withstand high impact and pressure during operation, making it a pressure-containing part that requires excellent mechanical properties. However, the casting contains many complex curved surfaces and is made of ZG07Cr19Ni10 stainless steel, which leads to poor castability. Therefore, a reasonable casting process must be designed to produce castings of qualified quality. This research focuses on the casting process optimization of the self-priming pump body using numerical simulation and experimental verification. The main objective is to reduce casting defects, improve process yield, and ensure the reliability of the final product.
1. Process Design and Numerical Simulation Fundamentals
To predict and eliminate casting defects, the filling and solidification processes were simulated using the Flow-3D casting simulation software. The theoretical basis includes the continuity equation, the Navier-Stokes equation, the energy equation, and turbulence models. For the filling stage, the standard RNG k-ε turbulence model was adopted because it is suitable for simulating low-speed turbulent flow. The governing equations are summarized below.
The continuity equation for incompressible fluid is:
$$
\frac{\partial \rho}{\partial t} + \nabla \cdot (\rho \mathbf{U}) = 0 \tag{1}
$$
The Navier-Stokes equation in Cartesian coordinates is:
$$
\begin{cases}
\dfrac{\partial u}{\partial t} + u\dfrac{\partial u}{\partial x} + v\dfrac{\partial u}{\partial y} + w\dfrac{\partial u}{\partial z} = -\dfrac{1}{\rho}\dfrac{\partial p}{\partial x} + \mu \Delta u + F_x \\[6pt]
\dfrac{\partial v}{\partial t} + u\dfrac{\partial v}{\partial x} + v\dfrac{\partial v}{\partial y} + w\dfrac{\partial v}{\partial z} = -\dfrac{1}{\rho}\dfrac{\partial p}{\partial y} + \mu \Delta v + F_y \\[6pt]
\dfrac{\partial w}{\partial t} + u\dfrac{\partial w}{\partial x} + v\dfrac{\partial w}{\partial y} + w\dfrac{\partial w}{\partial z} = -\dfrac{1}{\rho}\dfrac{\partial p}{\partial z} + \mu \Delta w + F_z
\end{cases} \tag{2}
$$
For solidification shrinkage prediction, the residual melt modulus (RMM) method and Niyama criterion were used. The Niyama criterion combines temperature gradient \(G\) and cooling rate \(R\):
$$
G = \sqrt{\left(\frac{\partial T}{\partial x}\right)^2 + \left(\frac{\partial T}{\partial y}\right)^2 + \left(\frac{\partial T}{\partial z}\right)^2}, \quad R = \frac{T_{upper} – T_{lower}}{t} \tag{3}
$$
When the value of \(G/\sqrt{R}\) is smaller than a critical threshold, shrinkage porosity is likely to form. The residual melt modulus is defined as:
$$
M_r = \frac{V_r}{A_r} \tag{4}
$$
where \(V_r\) is the residual liquid volume and \(A_r\) is the residual liquid surface area. A probability defect parameter above 0.9 indicates severe shrinkage.
2. Casting Process Design for the Pump Body
Based on the structural analysis and the actual production conditions, a sand casting process was selected. The main casting parameters are listed in Table 1.
| Parameter | Value |
|---|---|
| Material | ZG07Cr19Ni10 |
| Overall dimensions | 610 mm × 360 mm × 380 mm |
| Average wall thickness | 12 mm |
| Minimum wall thickness | 9 mm |
| Maximum wall thickness | 32 mm |
| Casting weight | 90 kg |
| Molding sand | Sodium silicate CO₂ hardened sand |
| Core sand | Alkaline phenolic resin sand |
| Pouring position | Horizontal, inlet flange at bottom |
| Molding method | Three-box molding |
| Shrinkage allowance | 2.0% (length), 3.0% (radial) |
The initial gating system adopted a stepped design with two ingates: one at the bottom inlet flange and one at the middle outlet flange. The cross-section areas are shown in Table 2.
| Component | Length (mm) | Shape | Section area (cm²) |
|---|---|---|---|
| Sprue cup | 55 | Circular Φ100 | 78.5 |
| Sprue | 440 | Circular Φ90 | 63.6 |
| Ingate 1 | 220 | Rectangular 175×38.5 | 67.4 |
| Ingate 2 | 200 | Trapezoidal 44/40×20 | 8.0 |
The riser system initially consisted of three insulating risers above the large flange, one insulating blind riser at the volute bottom, and one ordinary riser above the outlet flange. External chills were placed on the inlet and outlet flanges, and an internal chill was placed inside the exhaust boss. The initial process yield was calculated as:
$$
Y = \frac{W_c}{W_c + W_r} \times 100\% = \frac{90}{90 + 71} \times 100\% = 55.9\% \tag{5}
$$
3. Simulation of the Initial Casting Process
3.1 Filling Process
The filling process was simulated with a pouring temperature of 1560 °C, pouring time of 20 s, and initial sand temperature of 20 °C. The temperature field results showed that the metal flow was turbulent at the beginning, with a velocity difference of 1.6 m/s between the sprue and the mold cavity. This severe turbulence caused air entrainment and oxide slag formation. The pressure field analysis revealed that the maximum pressure at the bottom of the sprue was about 0.13 MPa, which is lower than the compressive strength of the sand mold (0.22 MPa), so no sand erosion was expected. The velocity field indicated significant splashing and unstable flow during the early stage, which was identified as the main cause of slag inclusion defects.
3.2 Solidification Process
The solidification simulation predicted isolated liquid zones at several locations, including the lower cavity wall, the junction between the partition and the cavity, and the exhaust boss. These regions could not be fed by the risers, resulting in shrinkage porosity. The probability defect parameter map showed that the total defect volume was \(3.0 \times 10^5 \, \text{mm}^3\), accounting for 2.6% of the total pump body volume. Figure 1 illustrates a typical automatic pouring line used in modern foundries for such castings.
4. Actual Casting Defects and Root Cause Analysis
After producing the first trial castings using the initial process, several defects were observed on the castings:
- Cold shut (cold shot) defects on the drain flange surface.
- Surface wrinkles on the upper surface.
- Cracks at the junction of the partition and cavity.
- Hole-type defects in the flange and boss areas.
Scanning electron microscopy was used to analyze the fracture surfaces. The cracking location exhibited irregular fracture morphology with coarse dendrites, which is typical of shrinkage porosity rather than hot tearing. The hole-type defects were small (about 2 mm) and located near the surface, with rough internal walls, indicating that they were caused by oxide inclusions that hindered liquid metal feeding and promoted shrinkage formation.
The root causes of these casting defects were identified as follows:
- Unreasonable gating system design leading to turbulent filling and slag entrainment.
- Unreasonable riser and chill placement causing isolated liquid zones.
- Insufficient mold rigidity and improper pouring parameters.
5. Gating System Optimization
5.1 First Optimization Scheme
The first optimization removed the upper ingate and changed the system to a closed type. The sprue diameter was reduced to Φ40 mm, and a sprue well was added. The new cross-section area relationship was \(A_{sprue}:A_{ingate} = 1.7:1\). Although the filling became smoother, a local turbulent flow still appeared when the metal impinged on the sand core surface at the inlet flange. The maximum pressure was about 0.14 MPa, still acceptable, but the velocity difference remained significant (0.8 m/s), which could cause oxide slag.
5.2 Second Optimization Scheme
In the second optimization, a runner was added between the sprue and the ingate to trap slag and gases. The gating system was changed to a semi-closed type with \(A_{sprue}:A_{runner}:A_{ingate} = 1.2:0.8:1\). The dimensions are given in Table 3.
| Component | Length (mm) | Shape | Section (mm) | Area (cm²) |
|---|---|---|---|---|
| Sprue cup | 50 | Circular Φ50 | – | 19.6 |
| Sprue | 410 | Circular Φ40 | – | 12.6 |
| Runner | 240 | Rectangular | 20×40 | 8.0 |
| Ingate | 150 | Rectangular | 40×25 | 10.0 |
| Sprue well | 80 | Circular Φ80 | – | 50.3 |
The simulation of this optimized gating system showed that:
- The temperature field remained in the range of 1440–1550 °C, with no cold shut risk.
- The maximum pressure was 0.15 MPa, below the sand strength limit.
- The velocity difference during filling was reduced to less than 0.5 m/s after the initial transient, and the metal flowed smoothly.
- The runner effectively trapped most of the oxide slag, preventing it from entering the cavity.
6. Feeding System Optimization
6.1 First Feeding System Optimization
In the first feeding system optimization, the three insulating risers above the large flange were reduced in size, the volute bottom riser was removed and replaced by chills, and the ordinary riser was downsized. The simulation showed that shrinkage defects decreased by 46% compared to the initial process, but defects still existed in the suction pipe and the large flange. The problem was insufficient feeding distance of the risers.
6.2 Second Feeding System Optimization
The second optimization restored larger insulating risers above the large flange and added an insulating blind riser at the volute bottom. The position of chills was adjusted: a chill was placed under the suction pipe, and additional chills were placed on the partition plate and bottom wall. The final feeding system layout is shown in Table 4.
| Component | Location | Size (mm) | Type |
|---|---|---|---|
| Insulating riser | Large flange | 125×65×110 | Open, waist-shaped |
| Insulating riser | Volute bottom | 70×36×60 | Blind, waist-shaped |
| Ordinary riser | Outlet flange | Φ50×120 | Open, round |
| External chill | Inlet/outlet flanges | Custom | Steel chill |
| External chill | Partition and bottom | Custom | Steel chill |
| Internal chill | Exhaust boss | Custom | Steel chill |
The solidification simulation of the final feeding system showed that:
- The temperature gradient was improved, with temperature difference over 90 °C between the bottom and top zones.
- No isolated liquid zones were formed in the critical areas.
- The final riser temperature remained about 20 °C higher than the casting until the end of solidification, ensuring continuous feeding.
- The shrinkage defect volume was reduced to \(1.0 \times 10^5 \, \text{mm}^3\), accounting for only 0.88% of the pump body volume, a 69% reduction compared to the initial process.
- The process yield increased from 55.9% to 77%, a significant improvement of 11 percentage points in process efficiency.
7. Orthogonal Experiment for Production Parameters
To further optimize the production parameters, an orthogonal experiment design was adopted. Three factors were selected: pouring temperature (A), sand mold preheating temperature (B), and pouring time (C). Each factor had three levels, as shown in Table 5.
| Level | A: Pouring temp. (°C) | B: Sand preheat temp. (°C) | C: Pouring time (s) |
|---|---|---|---|
| 1 | 1530 | 100 | 10 |
| 2 | 1560 | 200 | 15 |
| 3 | 1590 | 300 | 20 |
An \(L_9(3^4)\) orthogonal array was used, and nine simulation runs were performed. The responses included: ingate velocity, oxide slag concentration, ingate pressure, casting solidification time, ingate solidification time, and shrinkage defect volume. The results are listed in Table 6.
| Scheme | A | B | C | Ingate velocity (m/s) | Slag concentration | Ingate pressure (Pa) | Casting solid. time (s) | Defect volume (mm³) |
|---|---|---|---|---|---|---|---|---|
| L1 | 1 | 1 | 1 | 2.01 | 22.33 | 152210 | 1912 | 103809 |
| L2 | 1 | 2 | 2 | 1.54 | 39.47 | 142046 | 2022 | 102014 |
| L3 | 1 | 3 | 3 | 1.19 | 47.19 | 138700 | 1770 | 99536 |
| L4 | 2 | 1 | 3 | 1.20 | 47.94 | 137530 | 1973 | 100737 |
| L5 | 2 | 2 | 1 | 2.19 | 24.03 | 150885 | 2169 | 100289 |
| L6 | 2 | 3 | 2 | 1.44 | 36.09 | 141331 | 2378 | 98373 |
| L7 | 3 | 1 | 2 | 1.57 | 38.12 | 142228 | 2089 | 95415 |
| L8 | 3 | 2 | 3 | 1.23 | 48.21 | 137407 | 1805 | 96073 |
| L9 | 3 | 3 | 1 | 2.19 | 23.58 | 151357 | 2581 | 96169 |
Variance analysis (ANOVA) was performed to evaluate the significance of each factor. The F-values and significance levels for the filling process are summarized in Table 7.
| Indicator | Factor | F-value | Significance |
|---|---|---|---|
| Ingate velocity | Pouring temp. | 0.708 | Not significant |
| Sand preheat | 0.395 | Not significant | |
| Pouring time | 87.801 | ** | |
| Slag concentration | Pouring temp. | 0.177 | Not significant |
| Sand preheat | 1.269 | Not significant | |
| Pouring time | 281.346 | *** | |
| Ingate pressure | Pouring temp. | 10.319 | * |
| Sand preheat | 2.690 | Not significant | |
| Pouring time | 1734.410 | *** |
The ANOVA results for the solidification process are shown in Table 8.
| Indicator | Factor | F-value | Significance |
|---|---|---|---|
| Casting solidification time | Pouring temp. | 267.323 | *** |
| Sand preheat | 105.651 | *** | |
| Pouring time | 40.578 | ** | |
| Defect volume | Pouring temp. | 21.095 | ** |
| Sand preheat | 2.409 | Not significant | |
| Pouring time | 1.543 | Not significant |
The analysis indicates that for the filling process, the pouring time has the most significant effect on all indicators. A pouring time of 15 s gives a stable velocity of about 0.8 m/s, low slag concentration, and a maximum pressure of 0.17 MPa, which is safe for the sand mold. For the solidification process, the pouring temperature and sand preheat temperature have the greatest influence. At 1590 °C pouring temperature and 300 °C sand preheat temperature, the shrinkage defect volume is minimized to 0.83% of the pump body volume.
Therefore, the optimal production parameter combination is:
| Parameter | Optimal value |
|---|---|
| Pouring temperature | 1590 °C |
| Sand mold preheating temperature | 300 °C |
| Pouring time | 15 s |
8. Production Verification and Testing
Using the optimized casting process and the optimal production parameters, a verification trial was carried out. The castings were subjected to solution heat treatment and then inspected. The chemical composition of the produced pump body is shown in Table 10.
| Element | Si | Mn | Cr | Ni | Fe |
|---|---|---|---|---|---|
| Content (wt%) | 0.79 | 0.80 | 18.55 | 8.02 | Balance |
The metallographic examination showed about 7.5% α-ferrite content and a grain size of 4.5, both within the acceptable range. The mechanical properties are listed in Table 11.
| Property | Yield strength (MPa) | Tensile strength (MPa) | Elongation (%) | Impact energy (J) | Hardness (HB) |
|---|---|---|---|---|---|
| Expected | ≥175 | ≥440 | ≥30 | ≥60 | 178–230 |
| Measured | 336 | 557 | 56.5 | 247 | 221 |
Radiographic testing was performed on the flanges and other thick sections. All inspected areas were rated as Grade I according to ASTM-446, with no unacceptable indications. The pressure test results are given in Table 12.
| Test | Medium | Pressure | Result |
|---|---|---|---|
| Hydrostatic test | Water | 3 MPa | Qualified |
| Air tightness test | N₂ | 2.2 MPa | Qualified |
All test results met the acceptance criteria. The optimized casting process successfully eliminated or greatly reduced the casting defects, improved the internal quality, and ensured that the pump body could withstand the required service pressure. The combination of numerical simulation, orthogonal experiments, and production verification proved to be an effective method for optimizing the casting process of complex stainless steel pump bodies.
Conclusions
In this study, the casting process of a self-priming pump body was systematically optimized using Flow-3D simulation and experimental verification. The main conclusions are as follows:
- The initial gating system caused turbulent filling, leading to slag inclusion and gas entrainment defects. The optimized semi-closed gating system with a runner reduced the velocity difference to less than 0.5 m/s and effectively prevented casting defects.
- The feeding system was optimized by adjusting the riser sizes and chill positions. The final design achieved directional solidification, reducing shrinkage defects by 69% compared to the initial design. The process yield increased from 55.9% to 77%, a significant improvement of 11 percentage points.
- The orthogonal experiment showed that pouring time is the dominant factor affecting the filling process. A pouring time of 15 s gave stable flow and low slag concentration. For the solidification process, pouring temperature and sand preheat temperature were the most significant factors. The optimal combination was pouring temperature of 1590 °C, sand preheat temperature of 300 °C, and pouring time of 15 s.
- The verification castings met all technical requirements, including chemical composition, metallographic structure, mechanical properties, radiographic inspection, and pressure tests. The research approach combining numerical simulation with experimental validation is practical and effective for guiding the production of complex pump body castings.