Genetic Spatial Optimization of Active Elements on an Aeroelastic Delta Wing

Abstract

This work outlines a cohesive approach for the design and implementation of a genetically optimized, active aeroelastic delta wing. Emphasis was placed on computational efficiency of model development and efficient means for optimizing sensor and actuator geometries. Reduced-order models of potential-flow aerodynamics were developed for facilitation of analysis and design of the aeroelastic system in the early design phase. Using these methods, models capturing “95% of the physics with 8% of the modeling effort” can be realized to evaluate various active and passive design considerations. The aeroelastic delta wing model was employed in determining the most effective locations and sizes for transducers required to provide flutter control. The basic design presented is based upon an analytical model of the structure. A comparison of optimization strategies led to the use of a genetic algorithm to determine the optimal transducer locations, sizes, and orientations required to provide effective flutter control based upon an open-loop performance metric. The genetic algorithm and performance metric essentially provided loop shaping through the adaptive structure design. An experimental model was then developed based upon the optimal transducer designs. Wind tunnel tests were performed to demonstrate closed-loop performance for flutter control. Results from this study indicate that a single sensor/actuator pair can be designed to extend the flutter boundary and selectively couple to only those modes required to control the response.