Metodologia de projeto ótimo-robusto de sistemas aeroeletromecânicos com harvesters piezocerâmicos e circuitos multimodais para amortecimento passivo e supressão de flutter

Abstract

The advent of increasingly lighter and more flexible structures and the possibility of higher operational speeds in the aerospace industry have required more stringent control of vibrations and instability phenomena that can lead to catastrophic failures. Consequently, efficient strategies to control these phenomena are essential given the ever-stricter safety and reliability demands, particularly in the aerospace industry. Therefore, studies focused on developing control techniques for composite structures incorporating smart materials are becoming increasingly prevalent in aerospace structural dynamics. The application of smart materials for vibration control has also opened the possibility of energy harvesting from systems subject to unwanted vibrations, given their coupling properties with the electrical domain. The flutter instability is the most complex to predict accurately and the least investigated in composite structures subjected to subsonic flow regimes. Regarding the application of techniques for flutter control, the open literature has a higher density of work focused on active techniques. Using multimodal circuits as a passive strategy for subsonic flutter control is an understudied area, especially concerning composite structures. Additionally, the simultaneous possibility of optimally harvesting energy from composite aeroelastic systems remains unexplored. Based on this, this thesis presents a passive control technique for subsonic flutter instability using a multimodal circuit applied to composite structures. The finite element modeling integrates first-order theory via equivalent layer, the Doublet Lattice Method for aerodynamic loading, and the discrete layer theory for modeling the electric potential. The configuration with partial application of piezoceramic showed a total gain of approximately 14,7% in the flutter boundary when using a multi-objective optimization technique of the circuit parameters via genetic algorithm. The robust multi-objective optimization proposed in this work was applied using an artificial neural network. Greater vulnerability was observed with regard to the harvesting power objective, given the same level of disturbance of the design variables. The system’s robustness to flutter speed was demonstrated through velocity and damping diagram envelopes, showcasing the potential of the methodology proposed here for practical industrial applications.