Maximizing tolerable disturbances in a coupled structural system using a QFT-like method

Abstract

This thesis concentrates on developing an active control strategy that can be used to attenuate the far-field acoustic signature of a submarine by attenuating the vibration of its hull. The main goal of this thesis is to develop a controller that will maximize the magnitude of the disturbance that the system can be subjected to while satisfying magnitude constraints on the control effort and the magnitude of the vibration of the hull. In order to accomplish this, a simplified system was designed and built. An analytical model for this system was built using a combination of Lagrange's Equations and Component Mode Synthesis. These two methods, used in conjunction, allowed a single-input single-output (SISO) transfer function set to be derived to account for uncertainty in the system. The plant set was then used in a Quantitative Feedback Theory(QFT) type design method to develop a controller. Two types of magnitude-phase bounds were used during the design of the controller, stability and performance. The stability bounds were generated using standard QFT methods. The performance bounds were generated to satisfy the control effort and system output limitations. These were generated using a maximum tolerable disturbance method developed by Suhada Jayasuriya. The use of this method ensured that the resulting control system would meet the constraints set by the system constraints. The resulting control system was implemented, and its performance was tested using various disturbance signals. When the system was tested using sinusoidal disturbances within its operating frequency range, it easily met the performance objectives. At frequencies lower than its operating range, the physical system met the performance specifications; however, the analytical model did not. This is a result of more damping being present in the actual system when compared to the analytical model. When the worst possible disturbance was used, its magnitude was at little over half the magnitude of the design specified disturbance magnitude. This was due to the fact that the major frequency component of this disturbance was outside of the operating range specified for the system. This method proved to be useful in the design of a control system and the determination of what disturbance signals should be used to test the effectiveness of the system.