Difference in Results Between ANSYS Fluent and CFX for 90° Bend Pipe

[bhushan.ravindra.rode]

I am currently performing a CFD study of a 90° bend pipe using both ANSYS Fluent and CFX and have encountered a discrepancy in the computed head loss between the two solvers. The details of my simulation are as follows:

• Geometry: 90° bend pipe
• Pipe length: 100 m
• Pipe diameter: 5 m
• Mesh: Identical mesh used for both solvers
• Boundary conditions:
  • Inlet: Mass flow rate = 11,379 kg/s
  • Outlet: Static pressure = 0 Pa
• Fluids: Wate
• Turbulence model: SST k-ω
• Convergence criteria: 1 × 10⁻6 for all residuals in both cases  

Despite using the same mesh, boundary conditions, and fluid properties, the resulting head losses in Fluent and CFX are noticeably different. I would like to understand the possible reasons behind this variation and whether there are solver-specific settings or formulations that could lead to such discrepancies. Could you please advise on potential sources of this difference and suggest steps to ensure comparable results between Fluent and CFX? Thank you for your assistance.

[SRP]

Hi,

Even with an identical mesh, differences in how each solver interprets mesh topology, especially near-wall regions, can impact results. Pay particular attention to wall treatment methods (e.g., wall functions, near-wall resolution, y+ values, prism layers) and ensure they are set equivalently in both solvers

[bhushan.ravindra.rode]

Dear SRP,

Thank you for your response. The near-wall treatment methods were kept nearly similar in both solvers, with comparable prism layers and y⁺ values to ensure consistent near-wall resolution.

Regards

Bhushan