Abstract
A multi-objective optimization system for Francis turbine runners was developed. The system considers the performances at multiple off-design points and a head at a design point as objective functions in order to obtain runner design candidates that meet required specifications. Furthermore, the number of CFD evaluations was reduced using the correlation between a head and an outlet opening area. The developed optimization system was applied to design a runner, and we obtained runner design candidates that were superior to the baseline in terms of the performances at all operating conditions with approximately half of the number of all CFD evaluations. Since an optimized runner showed runner efficiencies improved at the design point as well as both the partial load and the overload compared with the baseline by evaluating approximately half of all design candidates in the optimization, the availability of the optimization system was confirmed.
