Session Track
November 15 (Friday)
비행동역학 및 제어Ⅱ
Oral,
606회의실
- 이희남
FC5-3
Intelligent Landing Gear equipped with MR damper based on Genetic Algorithm-Neural Network
Magnetorheological (MR) damper is researching to replace the oleo-strut which have a limitation under the different landing scenarios. MR damper is a smart material which can change quickly damping force by applying the electric current into coil (the response time less than 10 milliseconds). Landing gear equipped with MR dampers is a very complicated system due to nonlinearities and large uncertainties. These problems cause many challengers for the modelling and design a controller. In this paper, a genetic algorithm – neural network (GA-NN) control is adopted to improve the performance of damper without the knowledge of model. The input of NN is signal from two sensors (potential sensor and acceleration sensor), and output of NN is the electrical current. The GA is applied to create the random weight matrix and bias vector for NN. After few generations of simulation, the most importance result is that the GA-NN improve the shock absorber efficiency more than 15% compare with passive damper.
Keywords :
Genetic Algorithm, Neural Network, Landing Gear, Magnetorheological damper
Paper : FC5-3.pdf
항공우주시스템공학회, 대전광역시 중구 계백로 1719(오류동, 센트리오피스텔 제 12층 1206호)
Tel: 042-523-1978, E-mail: sase.or.kr