Shape Optimizations Using AI Agents to Accelerate Numerical Flow Field Simulation and to Control the Island Model for Massive Parallelization

Prof. Dr.-Ing. Stefan Riedelbauch (Stuttgart)


The overall aim of the project is to support the already automated design procedure for turbomachinery applications by extending individual steps within the workflow. Three main aspects are:

  • Optimization of fluid mechanically relevant systems consisting of several components by respecting multiple objective criteria
  • Reduction of the required computation time and overall number of individuals for the whole process by better initialization, supported mesh generation and improved parameterization
  • Replacement of rule-based fitness functions by data-driven functions and their flexible combinations


The typical application of a propeller turbine consisting of four runner blades, a conical hub and a S-shaped draft tube serves as a test case. The parameterization of the runner blades consists of a mean line (solid black) and a profile on suction and pressure side (solid gray). The geometry is created by a B-Spline. The control polygon and the degrees of freedom (letters and arrows) for the mean line (dashed black) and the profile on the suction side (dashed gray) allow to modify the shape of the blades.

The second test case is a kinetic river turbine. The machine consists of four runner blades and a segmented diffusor. Each diffusor segment is created from B-Splines and fully parameterized. Overall the machine has 30 degrees of freedom.

Within the project, agents will be developed to support the parameterization process, the setup of the simulation, the  meshing procedure including the block structure, the  initialization of the islands and the definition of the fitness functions. Additionally, the post-processing of the flow field will be supported by agents.

Questions to be answered in the project to achieve high quality flow field simulation results fast:

Objective function

  • Assign different fitness functions to different islands – purposeful?
  • Change fitness function in course of optimization – better overall results?

Initialization of islands

  • Omit simple randomized search of geometry parameter space
  • Generate highly diverse and at the same time valid population for starting optimization process

Computational grid

  • Automatic block-structure as well as mesh element distribution with high quality
  • Improve creation of prisms in boundary layer for hybrid meshes?


  • Evaluate flow field directly on the basis of pressure and velocity fields to rank them without evaluating the integral values, e.g. torque and efficiency?
  • Automatic setup of simulations possible with agent’s knowledge?

How can we support other projects?

  • Turbomachinery benchmark

How could other projects support our work?

  • Optimization
  • Migration
  • Fitness functions
  • AI methods


Prof. Dr.-Ing. Stefan Riedelbauch
Universität Stuttgart
Pfaffenwaldring 10
70569 Stuttgart
Room 2.47
Tel.: +49 711 685 63264

Further Information

To the top of the page