Duplex Stainless Steel Pump Body Casting Research

This article presents a comprehensive investigation into the sand casting foundry of a horizontal split pump body manufactured from 00Cr25Ni6Mo2N duplex stainless steel. The study covers casting process design, numerical simulation, alloy composition optimization, process improvement, pouring parameter optimization, and a simplified metallographic phase ratio analysis method. Throughout this work, the central role of sand casting foundry techniques is emphasized, and modern CAD/CAE tools are applied to ensure high-quality castings. The findings provide practical guidance for industrial production and quality control.

1. Introduction

Horizontal split pumps are widely used in petrochemical industries for transporting corrosive and high-temperature fluids. The pump body, as a key pressure-bearing component, must withstand severe service conditions, including high pressure, erosion, and corrosion. Duplex stainless steels, with their excellent combination of mechanical strength and corrosion resistance, have become preferred materials for such applications. The material 00Cr25Ni6Mo2N belongs to the third-generation duplex stainless steel family, offering high yield strength and remarkable resistance to stress corrosion cracking. However, its castability is challenging due to wide solidification range, high shrinkage, and susceptibility to hot cracking.

In the context of sand casting foundry, the production of a sound pump body requires meticulous process design. This paper presents a systematic approach to designing and optimizing the sand casting process for the pump body, using numerical simulation and statistical experimental methods. The outcomes are verified through simulation and proven to be effective in reducing defects such as shrinkage porosity and misruns.

2. Pump Body Casting and Material Analysis

The pump body is a large, complex casting weighing approximately 1400 kg, with dimensions of 1660 mm × 1100 mm × 850 mm. It features two volute chambers, four flanges, and four support legs. The wall thickness varies from 16 mm to 89 mm, creating significant thermal gradients during solidification. The casting is required to meet stringent quality standards, including internal soundness and pressure tightness.

The material specification for the pump body is 00Cr25Ni6Mo2N, a ferritic-austenitic duplex stainless steel. Its chemical composition is listed in Table 1. The material has a solidification temperature range of approximately 100°C, which promotes a mushy solidification mode and can lead to shrinkage porosity. Furthermore, the high chromium content tends to form oxide inclusions that reduce fluidity, and the steel is sensitive to sigma phase precipitation in the temperature range of 600–800°C, which can cause embrittlement and cracking.

Table 1: Chemical composition of 00Cr25Ni6Mo2N (wt.%)
Element C Si Mn P S Cr Ni Mo N
Range ≤0.05 ≤1.00 ≤2.00 ≤0.035 ≤0.030 25–28 4.5–6.5 1.2–2.0 0.08–0.2

The required mechanical properties are an ultimate tensile strength of 800–880 MPa, a yield strength of at least 550 MPa, and an elongation of 20%. These demanding requirements necessitate precise control of both the alloy composition and the casting process.

In the sand casting foundry, the mold is made of self-hardening furan resin sand, while chromite sand is used for cores that form the flow passages. A ceramic pouring system is employed to withstand the high pouring temperature. To avoid gas defects, alcohol-based coatings with low gas evolution are applied. The casting is opened at about 1100°C and cooled rapidly to avoid sigma phase precipitation.

3. Numerical Simulation of Casting Processes

3.1 Fundamentals and Tools

To predict and minimize casting defects, the commercial software AnyCasting was used to simulate mold filling and solidification. The simulation is based on the conservation laws of mass, momentum, and energy. The governing equations are:

$$ \frac{\partial \rho}{\partial t} + \frac{\partial (\rho u)}{\partial x} + \frac{\partial (\rho v)}{\partial y} + \frac{\partial (\rho w)}{\partial z} = 0 $$

$$ \rho \left( \frac{\partial \mathbf{v}}{\partial t} + \mathbf{v} \cdot \nabla \mathbf{v} \right) = -\nabla p + \mu \nabla^2 \mathbf{v} + \rho \mathbf{g} $$

$$ \rho C_p \frac{\partial T}{\partial t} = k \nabla^2 T + Q $$

For shrinkage porosity prediction, the Niyama criterion was employed:

$$ N = \frac{G}{\sqrt{R}} < C $$

where G is the temperature gradient, R is the cooling rate, and C is a critical value (typically 0.8–1.1). Additionally, the Residual Melt Modulus (RMM) method was used to evaluate macro-shrinkage:

$$ M_r = \frac{V_r}{A_r} $$

Higher Mr values indicate a greater tendency for shrinkage porosity. The numerical simulation parameters are summarized in Table 2.

Table 2: Simulation parameters
Parameter Value
Mesh size 1,500,000 cells
Mold process Sand casting
Liquidus temperature 1453°C
Solidus temperature 1357°C
Shrinkage 2%
Initial mold temperature 25°C
Pouring temperature 1560°C
Heat transfer coefficient (alloy/mold) See material data

3.2 Initial Process Designs

Two sand casting methods were designed and modeled in UG NX. The first method used a middle injection gate located at the parting line, while the second used a stepped injection system positioned at the inlet and outlet flanges. Both methods included risers and chills but differed in gate configuration and riser placement.

Figure 1 illustrates the filling and solidification sequences for both designs. The first design showed turbulent flow at the initial stage, causing sand erosion, while the second design exhibited more stable filling but had some upper gate blockage issues. The temperature distribution at the end of filling was more uniform in the second design, promoting directional solidification.

Based on the Niyama and RMM criteria, the second design produced fewer porosity defects in the quantitative analysis. Table 3 compares the defect percentage for both designs.

Table 3: Quantitative porosity defect comparison
Criterion Design 1 defect % Design 2 defect %
Niyama 0.2738% 0.2677%
RMM 0.2915% 0.2476%

Therefore, the stepped gating system was chosen as the basis for further improvement.

4. Chemical Composition Optimization Using D-Optimal Mixture Design

The mechanical properties and corrosion resistance of duplex stainless steel are strongly influenced by its chemical composition. To optimize the composition within the specified ranges, a D-optimal mixture design was adopted. The response variables were the yield strength (σ0.2) and the pitting resistance equivalent number (PREN). The nine alloying elements (C, Si, Mn, P, S, Cr, Ni, Mo, N) were treated as mixture components, with iron as the balance.

The design matrix consisted of 55 experimental runs, including 45 model points, 5 lack-of-fit points, and 5 replicates. The yield strength was obtained using JMatPro simulation at a fixed grain size of 2 μm and solution treatment at 950°C. The PREN was calculated using the standard formula usually defined as PREN = %Cr + 3.3(%Mo + 0.5%W) + 16%N (however, in this study, JMatPro provided a more tailored calculation).

A second-order polynomial model was fitted to the data. The analysis of variance confirmed the significance of the models (p < 0.0001), with R² values of 1.000 for PREN and 0.980 for yield strength. The regression equations were used to identify the effects of individual elements and their interactions.

The optimal composition was found to be: C = 0.040%, Si = 1.0%, Mn = 1.0%, P = 0.010%, S = 0.025%, Cr = 28.0%, Ni = 4.70%, Mo = 2.0%, N = 0.20%. This composition yielded a PREN of 37.8 and a yield strength of 790 MPa, satisfying both the strength and corrosion resistance targets.

After determining the target composition, the burden calculation for the induction furnace was performed using a matrix method. The furnace charge consisted of return scrap, pure iron, low-carbon stainless steel, ferrochromium, ferromanganese, nickel plates, molybdenum wire, ferrosilicon, and chromium nitride. The absorption rates were incorporated into a linear system of mass balance equations:

$$ \sum m_i e_{ij} = M_{\text{total}} w_j^T R_j – m_{\text{return}} e_{\text{return},j} $$

The computed charge weights for a 500 kg heat are listed in Table 4.

Table 4: Calculated charge materials for 500 kg melt
Charge material Mass (kg)
Return scrap (30% of heat) 150
Pure iron 1.75
Low-carbon stainless steel 236.78
Ferrochromium 90.69
Ferromanganese 1.65
Nickel plates 6.43
Molybdenum wire 1.84
Ferrosilicon 2.55
Chromium nitride 8.30

The final phosphorus and sulfur contents were verified to be within acceptable limits (0.0127% and 0.0244%, respectively). This optimization illustrates the power of statistical mixture design in the sand casting foundry environment to achieve consistent alloy quality.

5. Casting Process Improvement and Simulation Verification

5.1 Riser and Feeder System

Based on the simulation results, the original stepped gating design lacked adequate feeding at the inlet and outlet flanges. To rectify this, two additional risers (No. 6 and No. 7) were added at these locations. The risers were designed as rectangular with dimensions 400×100×300 mm and 320×100×300 mm, respectively. In addition, the top three risers were changed to insulating riser sleeves made of aluminum silicate fiber, which prolongs the liquid metal residence time and improves feeding efficiency.

5.2 Chills

To control solidification in isolated hot spots, 24 chills were strategically placed in the mid-lower thick sections of the casting. The chills, made of cast iron or steel, accelerate cooling at these regions, thus promoting a more uniform temperature distribution and reducing the risk of shrinkage and hot tearing. The chills included contoured shapes to fit the internal geometry of the pump body. Their locations were chosen based on the RMM probability maps.

5.3 Gating System Modification

The original upper gates allowed premature entry of metal, causing turbulence. The modification involved raising the upper gate level by 140 mm to increase the height difference between the lower and upper gates to 700 mm. The upper gate cross-section was enlarged to Φ45 mm. Furthermore, a reverse sprue design was adopted, where the metal flows through a small primary sprue into a wider secondary sprue, allowing sequential filling and reducing turbulence. The improved gating system produced smooth filling without upper gate “racing” and ensured a favorable temperature gradient.

5.4 Simulation Results

The improved process was re-simulated with updated material properties (liquidus 1474°C, solidus 1333°C) and the above modifications. The filling flow at 15% and 30% stages showed calm metal advancement with no obvious turbulence. The temperature distribution at end of filling was more uniform and presented a clear upward gradient, which is ideal for directional solidification.

The solidification sequence revealed that the risers remained liquid long after the casting sections had solidified, confirming effective feeding. The overall porosity defect percentage (based on RMM) dropped to 0.2386%. The calculated casting yield was 60.8%, which is economical for a medium-sized stainless steel casting.

6. Optimization of Pouring Process Parameters Using Orthogonal Design

To further reduce shrinkage-related defects, the influence of pouring parameters was investigated using an orthogonal experimental design. Three factors were considered: pouring temperature (A), pouring time (B), and mold initial temperature (C). Each factor was set at four levels (see Table 5). An L16(4^5) orthogonal array was employed, with two columns left empty for error estimation.

Table 5: Factor levels
Level A: Pouring temp (°C) B: Pouring time (s) C: Mold temp (°C)
1 1520 20 15
2 1540 35 25
3 1560 50 35
4 1580 65 45

The response was the maximum residual melt modulus (Mr) in the casting. Sixteen simulations were carried out, and the results were analyzed by range and variance. The variance analysis indicated that pouring temperature was the most significant factor (F-ratio 2.97), followed by pouring time (F-ratio 1.00), while mold temperature was insignificant (F-ratio 0.25). The range analysis showed the optimal combination as A1 (1520°C), B4 (65 s), C3 (35°C). This combination gave the lowest Mr of 44.122 mm, which was verified by simulation. It is worth noting that the pouring temperature had a stronger influence than the pouring time, and the mold temperature had negligible effect within the tested range.

The orthogonal design proved to be an efficient and economical method for optimizing sand casting foundry parameters, especially for complex alloys like duplex stainless steel.

7. A Simplified Method for Phase Ratio Analysis of Duplex Stainless Steel

Duplex stainless steel castings require a balanced ferrite/austenite ratio (typically 30–50% ferrite) to achieve optimum mechanical properties and corrosion resistance. In many small foundries, manual metallographic counting methods are still used, which are tedious and inaccurate. To address this, a simplified digital image analysis method was developed using MATLAB. The approach converts a color micrograph into a grayscale image and then performs binary thresholding based on Otsu’s method. The threshold maximizes the inter-class variance between the two phases, effectively separating ferrite (bright) from austenite (dark). The area fraction of each phase is then calculated by counting the respective pixel populations.

The program reads an original metallographic image, converts it to grayscale, applies the Otsu threshold, and outputs the binary image. The austenite fraction is computed as the ratio of black pixels to total pixels. For a sample test image, the austenite fraction was determined as 39.23%, which closely matched the result from a commercial image analyzer (38%). This method offers an inexpensive, reliable alternative for routine quality control in sand casting foundry operations.

8. Conclusions

This study successfully developed a robust sand casting process for a duplex stainless steel pump body. The key findings are:

  1. The stepped gating system with a reverse sprue ensures smooth filling and promotes directional solidification, significantly reducing shrinkage porosity.
  2. The use of combined risers, chills, and proper pouring parameters effectively controls solidification and minimizes defects. The optimal pouring temperature was found to be 1520°C with a pouring time of 65 s.
  3. D-optimal mixture design is a powerful tool for optimizing alloy composition. The recommended composition (C=0.040%, Si=1.0%, Mn=1.0%, Cr=28%, Ni=4.7%, Mo=2%, N=0.2%) gave a good balance of strength and corrosion resistance.
  4. The MATLAB-based phase analysis method provides a simple, cost-effective approach for phase ratio measurement in duplex stainless steel castings, demonstrating practical value for small foundries.

These results demonstrate the effectiveness of integrating numerical simulation and statistical design in the sand casting foundry industry, leading to improved casting quality and production efficiency.

Future work will focus on the influence of heat treatment parameters on the microstructure and properties of these castings, as well as the exploration of energy-saving and emission-reduction measures in the foundry process.

Scroll to Top