Multi-objective Aerodynamic Design Optimization of Rotor Airfoils Using an Efficient Surrogate-Based Algorithm

To address the challenge of the aerodynamic design of rotors airfoils, this paper implements a rotor airfoils design optimization method by coupling the RANS solver and an efficient surrogate-based multi-objective algorithm (SBMO). First, the perturbed Class Function/Shape Function Transformation (CST) method is employed to parameterize airfoils and initial samples are selected by Latin Hypercube Sampling (LHS). Second, the selected solutions are evaluated by RANS solver and kriging models are established for each objective and constraint. Third, the SBMO algorithm decomposes the multi-objective problem into a set of single-objective sub-problems and obtain solutions by solving the acquisition problems for sub-problems simultaneously utilizing combined infill-sampling strategy of the excepted improvement (EI) and minimizing surrogate prediction (MSP) infill-sampling criteria. Finally, all the solutions obtained are evaluated and used to update the kriging models to share their search information. The above steps are repeated until a stop criterion is reached. A multi-objective optimization of OA309 airfoil has been done by employing the implemented method. A Pareto set contains 76 solutions and an optimal airfoil that make a great trade-off between lift coefficient in low Mach number and the drag coefficient at high Mach number are obtained. The results indicate that the implemented method has a great adaptability to design optimization of rotor airfoils and can enable designers to explore and identify the trade-offs between conflicting design objectives.