Shape optimization of draft tubes for Agnew microhydro turbines

• The draft tube of Agnew microhydro turbine was optimized.
• Pareto optimal solutions were determined by neural networks and NSGA-II algorithm.
• The pressure recovery factor increases with height and angle over design ranges.
• The loss coefficient reaches the minimum values at angles about 2o.
• Swirl of the incoming flow has great influence on the optimization results.

In this study, the shape optimization of draft tubes utilized in Agnew type microhydro turbines has been discussed. The design parameters of the draft tube such as the cone angle and the height above the tailrace are considered in defining an optimization problem whose goal is to maximize the pressure recovery factor and minimize the energy loss coefficient of flow. The design space is determined by considering the experimental constraints and parameterized by the method of face-centered uniform ascertain distribution. The numerical simulations are performed using the boundary conditions found from laboratory tests and the obtained results are analyzed to create and validate a feed-forward neural network model, which is implemented as a surrogate model. The optimal Pareto solutions are finally determined using the NSGA-II evolutionary algorithm and compared for different inlet conditions. The results predict that the high swirl of the incoming flow drastically reduces the performance of the draft tube.