AI-based fluid mechanics optimization of a turbomachine Unleash the full power of your design workflow

Global Picture

Optimal Computing helps designers to unleash the full power of their process to increase turbomachines efficiency and related revenues.

Our keystone

Optimal Computing develops Xtreme, a versatile, agnostic and powerful optimization software based on Artificial Intelligence techniques.

Value proposition

Thanks to the AI at the core of Xtreme software, you benefit from the following advantages:

  • Complex optimization problems (e.g. eigen frequencies passing forbidden ranges) are treated easily with Xtreme.
  • A reduced number of simulations to reach the optimum. Fewer runs required with Xtreme compared to other optimisation packages.
  • Non-computable functions are treated in a unique way leading to a fast learning and solving. Learning is the foundation of Xtreme to save your time.
  • Let Xtreme adjust solvers and optimization parameters automatically while you focus on the physics of the problem to solve. Solvers are automatically adjusted behind the scene based on a learning algorithm and a large number of stored knowledge.

Xtreme helps you to master your design workflow

Example of a Fluid-Mechanical Optimization of a Radial Pump

Software workflow

  • CFTurbo for the geometry
  • TCFD for the CFD simulation
  • NetGen for the mechanical mesh
  • CalculiX for the mechanical static and dynamic calculation
  • Xtreme for the optimization
  • Xtreme for the coupling between all these software and to lead the platform

Data and Design variables

  • Impeller main diameter = 286 mm
  • Number of blades = 7
  • Number of parameters = 13

CFD analysis

  • Steady Simulation
  • 1,200,000 mesh cells
  • Computation time = 21 minutes on 6 cores

Operating conditions

  • Incompressible Fluid : water
  • Rotational speed : 1770 rpm
  • Flow rate Q = 454 m3/h
  • Head : 33 m

Mechanical analysis

  • Static calculation
  • Dynamic calculation (vibration modes)
  • 150,000 nodes

Operating conditions

  • Pump material : PVC
  • Density= 1400 kg/m3
  • Young Modulus=3500 MPa
  • Elastic limit=35 MPa

Optimization using Xtreme

  • Plug-in method: versatility and flexibility
  • Global optimization
  • Multiple objectives
  • Can deal with uncomputable simulations
  • Learning method included
  • Very fast convergence
  • Graphical user interface and meta model for 3D representation
  • Continuous / Integer / Discrete design variables
  • Continuous variable optimization and sequence optimization

Results of the radial pump optimisation

Hydraulic efficiency from 86.4 % to 88.5 % (+2.1 %)
Von Mises stresses from 2.71 MPa to 2.97 MPa < 10 MPa
Eigen frequencies (First 5 modes) increased by 7 %
Blue color tags the original blade shape

More information about Xtreme ?

Created By
Laurent D'Alvise