Design studies of aeroelastic tailoring of forward-swept composite aircraft using hybrid genetic algorithm

The genetic/sensitivity-based hybrid algorithm was introduced and used for design studies of aeroelastic tailoring of forward-swept composite aircrafts. The program of genetic algorithm was developed while the sensitivity-based algorithm came from the modified method of feasible direction in NASTRAN. In the hybrid algorithm, the genetic algorithm was used to perform global search to escape from local optima, and the excellent individuals of every generation optimized by genetic algorithm were further optimized by the modified method of feasible direction to conduct fine-tuning to attain local optima. This paper investigates the advantages and shortcomings of the above three algorithms, and discusses the effect of initial values of lay-up thickness, ply-orientation, and forward-swept angle on optimized weight of forward-swept composite aircraft while satisfying multiple constrains on strength, displacement, lifting efficiency, aileron efficiency, divergence speed and flutter speed. The comparison of three algorithms shows that the hybrid algorithm could outperform either genetic algorithm or the modified method of feasible direction operating alone through the complementary of the properties of the two methods. The studies demonstrate that initial values of lay-up thickness have great effect on optimized weight when sensitivity-based algorithm is used, and the effect could be decreased by the use of hybrid optimization algorithm. In addition, the results also indicate that optimized weight of forward-swept composite wing with [0/90/±45]_s lay-up in skins is relatively insensitive to ply-orientation when wing is subjected to multiple constrains, and optimized weight varies significantly with variation of forward-swept angle.