Automated Design and Evaluation of Airfoils for Rotorcraft Applications

Open Access
Author:
Stanko, Jason D
Graduate Program:
Aerospace Engineering
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
July 14, 2017
Committee Members:
  • Sven Schmitz, Thesis Advisor
Keywords:
  • parameterization
  • orthogonal
  • automation
  • airfoil performance table
  • multicollinearity
  • design
  • class shape transformation
  • genetic algorithm
  • CFD
  • Legendre Polynomial
  • CMA-ES
  • UH60
  • OVERFLOW
  • airfoil shape optimization
  • C81
  • Augmented Lagrange
Abstract:
In this work, a methodology is presented and implemented for the automated design and optimization of airfoil sections. The airfoil optimization is conducted using the CMA-ES genetic algorithm with constraints applied to the airfoil's area and pitching moment coefficient. The design variables are formed through a class shape transformation with orthogonal basis modes, allowing for decreased levels of multicollinearity in higher-order design spaces, while still maintaining the completeness of lower-order spaces. A Python framework is developed to automate the generation of airfoil performance tables using the RANS CFD solver OVERFLOW 2.2i allowing the optimization methodology to be extended to rotorcraft applications. An empirical maximum lift coefficient criteria is developed and incorporated into the table generation process to overcome inaccuracies associated with stall prediction in CFD-based methods. The methodology presented is used to perform a single point shape optimization on the tip airfoil of a UH-60A baseline rotor. The new tip section shows substantial improvements in forward-flight performance in exchange for a small reduction in the rotor's stall margin. The airfoil optimization and table generation routine shows to be effective for the single design point investigated. This holds promise that the technology developed can later be successfully extended to higher-fidelity automated routines.