The impact of model uncertainty on spatial compensation in active structural acoustic control.

摘要

Control strategies aimed at the reduction of turbulent boundary layer noise transmission into the fuselage of an aircraft have been topics of academic and industrial research for years. There has been a great deal of knowledge culled from these research projects that has added to the understanding of the fundamental and application physics involved. However, there have been few practical realizations of these strategies due to potential limitations in application settings. This work focuses on an approach that will attempt to address one of the potential application issues, that of uncertainties in boundary conditions.; A method for optimizing a structural acoustic control system with regard to potential variations in plate boundary conditions is provided. The assumed modes method is used to build a plate model with varying levels of rotational boundary stiffness to simulate a plate with uncertain edge conditions. A transducer placement scoring process, involving Hankel singular values (HSVs), is combined with a genetic optimization routine to find locations robust to boundary condition variation. Predicted frequency response characteristics are examined, and theoretically optimized results are discussed in relation to boundary condition variations. Control system design is also impacted by uncertainty in boundary conditions, specifically in terms of the compensator required to achieve the desired control. Impact of uncertain boundary conditions on the selection of actuator and sensor locations for structural acoustic control is considered.; Results indicate it is possible to minimize the impact of uncertain boundary conditions in active structural acoustic control by optimizing the placement of transducers with respect to those uncertainties. Additionally, results indicate that optimizing with respect to a group of randomly generated boundary conditions results in a patch pair capable of performing well over the complete group of conditions. Test results were also obtained that demonstrate the importance of enforcing model coupling roll-off beyond the primary bandwidth of interest. Finally, results indicate that failure to consider possible boundary condition variations can inhibit the ability to enforce control in the presence of such variations.