Performance optimization of a hydrokinetic hybrid Gorlov-Savonius turbine using surrogate model

The main objective of this investigation is to optimize the performance of a hydrokinetic Hybrid Gorlov-Savonius (HGS) turbine using surrogate-based modeling. A nested configuration of the HGS hydrokinetic turbine is chosen as the base design considered for optimization. The general process of the investigation starts with an experimental evaluation of an HGS reference design performance in an open channel. Then, the transient and two-phase flow around the HGS turbine in the open channel is simulated using ANSYS CFX 23.1 software. The validity of the numerical method is proven using the experimental data of the power coefficient ( \\(C_P\\) ). In the next step, using the three design parameters, including the Savonius rotor diameter and height, together with the rotational velocity, 15 different samples (or learning points) are generated by Latin Hypercube Sampling (LHS) in the design space. Then, a numerical simulation is conducted at every sample, and the power coefficient is calculated as the output variable. A Surrogate Model (or Response Surface Model) is created as an approximate function between the output ( \\(C_P\\) ) and inputs (or design parameters) by the genetic aggregation method. The constructed Surrogate Model is then applied in the optimization process using the Non-dominated Sorting Genetic Algorithm-II. The CFD-based optimization predicts that the optimum design can improve the power coefficient by 10.22% relative to the reference configuration. Also, static torque coefficient evaluation at different angular positions showed that the optimum HGS turbine has a higher \\(C_{Ts}\\) in comparison to the reference design of the turbine at most angular positions. Evaluating the power coefficients of the HGS and Gorlov turbines showed that, at the optimum design point, the Gorlov rotor inside the HGS turbine has better performance than a single Gorlov turbine.