Computing Turbomachinery Flow of Centrifugal Pump

Dynamic Meshing   |   Steady State   |   Turbulence (k-omega SST) and Turbomachinery   |   Data Analysis

Minor Project 2 - Turbomachinery,
Flowthermolab
30/10/2025

Project Overview

This project presents a comprehensive CFD investigation of centrifugal flow of a centrifugal pump to analyze:

• Flow behavior at different mass flow rate
• Head (m)
• Efficiency (%)
• Mesh Convergence
• Validation Study

The study evaluates the centrifugal pump under multiple mass flow rate conditions to map a metrics for head and efficiency values, evaluate performance and identify possible design improvements from a CFD engineering perspective.

Problem Statement

• To simulate and analyze the performance of centrifugal pump
• To understand the working of centrifugal pump, evaluate performance and provide some insight on possible design changes/optimization

Software Used

1.   Ansys Space-Claim ------- Geometry Cleanup and refinement
2.   Ansys Fluent Mesher ----- Mesh / Grid Generation
3.   Ansys Fluent ------------- Computation
4.   Ansys Post Processor ---- Post Processing
5.   MATLAB ------------------ Data Analysis & Scientific Plots

Methodology

1. Geometry Cleanup and Refinement

• Geometry was cleaned and optimized for computation.
• Dimensions of the geometry:
  • Inlet Dia (𝐷1) ---------------------------- 195 mm
  • Outlet Dia ------------------------------- 137 mm
  • Interface Dia ----------------------------- 562 mm
  • Blade Diameter -------------------------- 274 mm
  • Impeller Diameter (𝐷2) ------------------ 281 mm
  • Inlet Area (𝐴1) --------------------------- 0.0298 𝑚²
  • Impeller Area (𝐴2) ----------------------- 0.248 𝑚²
  • Backward Curved Blade angle (𝛽2) ----- 34^𝑜
  • Blade width at tip, b2 ------------------- 0.0142 m
  • Blade width at hub, b1 ------------------ 0.008 m

Centrifugal Pump Geometry and Calculations

2. Mesh Details

• Focused on mesh refinement at: Blade region.
• Mesh Details mentioned here are for Volute and Rotor together, however during meshing both volute and rotor were meshed independently.
• Following Mesh Details are for coarse mesh, medium mesh, and fine mesh:
  • Reynolds Number ------------------------------------------------------- 2.3e06
  • Y+ ------------------------------------------------------------------------ 1
  • First Layer height -------------------------------------------------------- 8e-06 m
  • Local Sizing, Face Size --------------------------------------------------- 2.7 mm
  • Surface mesh (Min Size, Curvature Normal Angle) --------------------- (2.5 mm, 12),(1 mm, 12),(0.63 mm, 12)
  • Boundary Layer (Offset Method, first height, number of layers) ------- last Ratio, 0.12 mm, 4
  • Volume Mesh (Type, Min Cell length, Max. Cell length) ---------------- (Poly-hexcore, 2.5 mm, 8 mm), (poly-hexcore, 1 mm, 8 mm), (Poly-hexcore, 0.63 mm, 2.52 mm)
  • Total mesh Elements ----------------------------------------------------- 231682, 395832, 649464
  • Average orthogonality --------------------------------------------------- 0.88, 0.92, 0.94

Centrifugal pump Meshing (Different Mesh Convergence Cases)

3. Solver Settings

Parameter	Setting
Solver Type	Pressure-Based
Time	Steady
Turbulence Model	k-ω SST
Material	Water (Liquid)
Rotational Velocity	1500 RPM and 1800 RPM
Mass Flow rate at outlet	40 kg/s to 120 kg/s
Solver Scheme	Coupled with PRESTO! Pressure Solver
Momentum Discretization	second order
Residuals & Reports	Mass flow rate at inlet and outlet, torque on impeller and velcoity at blade
Initialization	Hybrid
Iterations	2000

4. Results

1. Mesh Convergence Test

It can be concluded from the Mesh Convergence study that Medium Mesh is good to start the further study, as per the time convergence and computational power criteria.
Note:

• Convergence Criteria marked when the mass flow rate ratio of outlet to inlet got very close to 1.
• All the simulations were run for 2000 iterations
• Medium type mesh showed ~3% error for head value when compared to Fine Mesh

Error Percentage (Force and Torque), Surface (Propeller)

2. Validation Study

Validation Study shows that medium mesh showed maximum resemblance of head value with theoretical value within 81% when simulated pressure-based solver at steady time for 2000 iterations monitoring mass flow rate at inlet and outlet as convergence criteria within 5% difference for results to conclude.

Number of Elements vs Head [m] and Efficiency [%]

3. Case Results

• Optimal Head loss for mass flow rate 80 kg/s at 1500 RPM (1), ratio of theoretical to CFD head is 0.81
• Head loss reduces with increasing mass flow rate.
• Efficiency is max at 80 to 100 kg/s mass flow rate for both 1500 RPM and 1800 RPM
• From the above point we can say, for better efficiency operate centrifugal pump at above specified condition.

Head [m] and Efficiency [%] vs Mass Flow Rate [kg/s]

4. Case Results: Video Animation

pressure-animation.mp4

velocity-animation.mp4

Velocity and Pressure Conoturs Animation

5. Discussion & Future Scope

Mapping:

• Generated full characteristic curves (Head [m] & Efficiency [%] vs. mass flow rate [kg/s]) for 1500 RPM and 1800 RPM.
• Data covers a wide operational range: 40 to 120 kg/s.

Key Performance Benchmark:

• Peak Efficiency: 51.13% for 1500 RPM at 100kg/s mass flow rate.
• Best Efficiency Point (BEP): Located at 80 kg/s, 90 kg/s, 100 kg/s for 1500 RPM rotational velocity.

Head-Flow Characteristics:

• The head-Flow curve shows decreasing trend with increasing flow rate for both 1500 RPM and 1800 RPM, highlighting standard pump behavior.

Efficiency Flow Characteristics:

• The peak efficiency for 1800 RPM is lower and occurs at a shifted flow rate compared to 1500 RPM.
• This indicates that the centrifugal pump design (impeller blade angles, volute) is tuned for a specific operating point (~80-100 kg/s at 1500 RPM). When run at 1800 RPM, the increased flow velocities exacerbate losses (friction, shock, recirculation), leading to lower peak efficiency. The pump is inherently designed for its 1500 RPM duty.

Future Scope

Parameter for Improvisation	Future Scope	Design Modification Suggested
Overall Efficiency	Implement Transient studies to improvise design, maximize head and efficiency while minimizing power.	Optimize blade geometry. Vary blade number (5-7 blades)
Best Efficiency Point (BEP)	Use transient CFD with SST turbulence model to accurately capture unsteady flow phenomena and losses at off-design conditions.	Shift BEP by changing blade loading distribution. Match impeller outlet flow angle to volute tongue angle. Adjust blade leading edge angle (β₁) to match inlet flow.
Cavitation Performance	Conduct multiphase (Mixture) CFD simulations to predict vapor formation	Modify inlet blade angles (β₁) for smoother flow. Apply leading edge profile optimization (elliptical profiles).
Rotor-Dynamic Stability	Perform FSI (Fluid-Structure Interaction) analysis to predict hydraulic forces, bearing loads, and critical speeds.	Implement back vanes or pump-out vanes on the impeller rear shroud.
Manufacturing & Cost	Integrate CFD with DFM (Design for Manufacturing) for die-cast or 3D-printed impellers, balancing performance with producibility.	Simplify blade geometry for casting without performance loss. Define allowable tolerances based on sensitivity analysis. Optimize fillet radii for stress reduction and flow.

First to go with: Perform Transient simulation to get better Characteristic curves with reduced losses & study cavitation

Recommended Repository Structure

CFD-Study-Globe-Valve/
│
├── README.md
├── Fluent_Case_Files
├── animations

Author:
Ansh Vishal,
Aerospace Engineer
anshvishal215@gmail.com
LinkedIn