Application of a General Discrete Adjoint Method for Draft Tube Optimization

Abstract Automatic optimization is becoming increasingly important in turbomachinery design to improve the performance of machine components and Evolutionary Algorithms (EAs) play a very role this task. The main drawback EAs large number evaluations that are required obtain an “optimal” result. Consequently, order keep computational time affordable frame for purposes, either mesh size has be limited, thus reducing resolution flow phenomena, or free parameters must kept small. Adjoint does not suffer from these restrictions, i.e. affected by parameters. effort adjoint method scales only with grid usually range two times CFD simulation alone. In paper, discrete based on coupled pressure RANS solver presented applied draft tube optimization. general can therefore deal any turbulence model supported as well boundary condition, including mixing planes interfaces needed multi-stage simulations. Furthermore, there no restriction choice objective function. first baseline geometry then again its EA optimized where function was minimization losses tube. To reduce complexity proof concept but still multiple operating points optimization, peak efficiency full-load were simultaneously. significantly both cases (baseline optimization). interplay between local global seems promising strategy find optimal geometries multi-operating point/multi-objective will further investigated subsequent research.