Contribuição ao estudo de máquinas rotativas na presença de Não-Linearidades

Abstract

This thesis focuses specifically on issues related to smart rotors, i.e., those which are able of monitoring their dynamic behavior, analyze their response and propose solutions with a certain degree of autonomy so that safe operation is ensured even in the presence of faults that may compromise either the rotors safety or efficiency. In general terms, this work encompasses rotating machines with nonlinear characteristics. More specifically, it is interested on nonlinearities associated with rotors that have time-varying parameters, such as cracked shafts and systems with electromagnetic actuators (EMA). The electromagnetic actuators introduce forces that are inversely proportional to the squared gap value. In this sense, three most significant issues are addressed in the present work, as described in the following. The first issue is related to the identification of time-varying parameters, through the expansion of the system input and output signals together with the expansion of the physical parameters of the system, in series of orthogonal functions, thus allowing the identification of the unknown parameters of the system. Both constant and varying parameters can be identified. A cracked shaft was considered as a varying parameters system and the proposed methodology was applied to determine the crack parameters. For this aim, a crack model was developed, enabling the complete simulation of the structure by using the finite element technique. The second case studied was the possible application of an electromagnetic actuator (EMA) for the purpose of maintaining the crack closed by controlling the fatigue process of a rotating shaft. The crack dynamics behavior is mainly influenced by the moments caused on the crack s surface. These moments can be determined through the variation of the angular displacements in the nodes at the ends of the cracked element. It was observed through the simulations that the use of EMA allows modifying several dynamics characteristics of the rotation system, such as critical speed, amplitude of vibration, etc. As a result of changes in the angular displacement, it was verified the possibility of using EMA to keep the crack closed along the rotation of the shaft. It follows a methodology for self-correction of the problem created by the presence of the crack (self-healing scheme). The third and last topic is related to the unbalance identification through the use of inverse problems techniques. For this, the rotating machine model must be known (or have been previously identified). Thus, a methodology based on a heuristic optimization technique was used aiming at balancing the system for the linear case (without the electromagnetic actuators) as well as for the system containing electromagnetic actuators that insert nonlinearity. It should be noted that classic techniques of balancing cannot solve such problem since they assume a linear relation between the excitation (unbalance) and the response (vibration) of the rotor.