Application of genetic algorithms to the design of airfoil pressure distributions

Application of genetic algorithms to the design of airfoil pressure distributions

Kuiper, H.
van der Wees, A.J.
Hendriks, C.F.W.
Labrujère, T.E.

National Aerospace Laboratory NLR

1995-09-29

NLR participates in the ECARF (European Computation Aerodynamics Research Project) "Optimum Design" project, developing an algorithm for multi-point airfoil design. Within the ECARP project a test case has been defined for two-point design with the aid of well-specified target pressure distributions. It appears that optimization of pressure distributions requires the application of efficient nonbased optimization algorithms. The investigation described in this report has been carried out in order to examine the possibilities of genetic algorithms. In this report the problem description is given, as well as how genetic algorithms can be used to optimize the problem. Next, the simulator is described which has been used to conduct some smallscale experiments. Finally, results from these simulations are presented. Summarized, the results are: - Genetic algorithms cam be used successfully to optimize pressure distributions. - Software in the public domain can be used to solve the optimalization problem. Only some problem specific operators have been added. - The current implementation of the genetic algorithm generates different minima depending on the starting point. - The results suggest that the current choice of the initial population plays an important role in the outcome of the genetic algorithm. - Although the genetic algorism uses about 100 times more function evaluations than the application of the Simplex algorithm to the present constrained minimalization problem, the efficiency of the genetic algorithm can be improved by improving the function formulation, by specifying more appropriate initial populations and by applying problem specific mutation and recombination operators.

Aerodynamic characteristics
Airfoils
Design analysis
Drag coefficients
Fitness
Genetic algorithms
Optimization
Penalty function
Pressure distribution
Software tools

http://resolver.tudelft.nl/uuid:09821146-9397-472d-909d-caded85e1d4f

Nationaal Lucht- en Ruimtevaartlaboratorium

Campus only

NLR Technical Publication TP 95342 U

Aerospace Engineering Reports

report

(c)1995 National Aerospace Laboratory NLR