Multi-Objective Experimental Optimization with Multiple Simultaneous Sampling for Flapping Wings

Chaudhuri, Anirban; Haftka, Raphael T.; Chang, Kelvin; Hall, Jordan Van; Ifju, Peter

doi:10.2514/6.2015-1586

Cited by 3 publications

(1 citation statement)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An alternative to using a computation model is to drive the optimization with experiments, using physically measured thrust forces in the place of a high-fidelity model, even possibly looking at tradeoffs between thrust and power measurements [14]. An experimental optimization of the reinforcement structure for hovering flight has been accomplished for thrust alone [15] and with thrust and power as objectives [16]. Techniques such as using different surrogates to have multiple candidate designs every optimization cycle [17] and using an adaptive sampling technique (efficient global optimization) [18] were helpful for minimizing the cost of the optimization, but the main costs pertain to the time spent fabricating and testing wings in a controlled manner.…”

Section: Introductionmentioning

confidence: 99%

Improving the Fabrication Process of Micro-Air-Vehicle Flapping Wings

Chang¹,

Chaudhuri²,

Rue³

et al. 2015

AIAA Journal

Self Cite

View full text Add to dashboard Cite

The aerodynamic performance of flapping micro air vehicles in hover conditions is dependent on many parameters, including the wing design. With the goal of optimizing the wing for hover performance, the initial focus was to reduce the uncertainty in the thrust measurements. This is because lower uncertainty in this metric enables better resolution in comparing the performance of different designs. Aerodynamic performance variability was deemed to be the fault of an imprecise manufacturing technique. Therefore, adjustments were made to the fabrication process until a permissible level of uncertainty was attained for optimization; the goal was less than 5%. This paper chronicles the progression of the wing fabrication process and details how the uncertainty was evaluated. Four fabrication methods and two different wing designs are included in this study: a carbon fiber hand layup technique, carbon fiber cured in a machined mold, and two variations of a machined plastic skeleton reinforced with a carbon fiber rod. The uncertainty in thrust production, expressed in coefficient of variation, improved from 16.8% for the hand layup method to 2.6% for the computer numerically controlled plastic skeleton adhered to the nylon membrane with transfer tape. Additionally, the coefficient of variation for wing weight also reduced (from 11.4 to 2.0%). Nomenclature C 4 = bias correction factor CV mfg = manufacturing coefficient of variation CV tot = total coefficient of variation CV test = testing coefficient of variation n = samples

show abstract