科研之友
微信
新浪微博
Facebook
分享链接
摘要
This paper presents an investigation on simultaneous optimization strategies for coupled aeroelastic and control systems. A representative aeroservoelastic system is considered, consisting of a two-dimensional potential flow over a deforming aerofoil, an actively controlled, but saturated compliant trailing edge, a dynamic observer that uses a series of pressure sensors on the aerofoil, and a heave/pitch linear spring model. Although computationally simple, the design allows for optimization over multiple disciplines: The structure can be designed by varying the stiffness of the springs, the control architecture through weightings in a linear quadratic regulator controller, the observer by means of the placement of pressure sensors, and the aerodynamics via the shaping of the compliant trailing edge. Optimizing the weight and a metric of performance over all disciplines simultaneously is compared with a sequential methodology of optimizing the open-loop characteristics first and subsequently adding a closed-loop controller. This paper shows that varying the parameterization and number of design variables during the optimization can lead to improvements in the final design and presents a procedure to automate this process. To accomplish this, a new basis for the design vector is created via proper orthogonal decomposition, using the trajectories of initial optimization paths as a training set. This parameterization is shown to make the optimization more robust with respect to the initial design and to facilitate an automated variable selection methodology. This variable selection allows for the dimension of the problem to be reduced temporarily, and it is shown that this makes the optimization more robust.
- 出版日期2018-3
