Identificação de trincas em máquinas rotativas baseada em técnicas de segunda derivada do modo de vibrar e combinação de ressonância

Abstract

Fault detection in rotating machines is a critical topic in many industrial sectors such as aeronautics, power generation and manufacturing. Cracks can lead to catastrophic component failure, so early and accurate detection is essential to ensure safety and minimize unexpected production interruptions. A commonly used technique for Structural Integrity Monitoring (SHM) of these equipment is based on the development of mathematical models. In this context, this work proposed the comparison of three crack models available in the literature and their use in demonstrating two SHM techniques: combination resonance and second-order derivative of the mode shape. It was found that the occurrence of combination resonances can be associated with evidence of non-linear characteristics in the system, such as transverse cracks for example, however, without clearly identifying their location. The crack position in the system could be determined using operational modal shape (vibration modes) reconstructed from the rotor vibration responses, also by using a combination resonance. Concerning the evaluation of the proposed methodology, a finite element model capable of representing the dynamic behavior of a rotating machine was used, taking into account, in addition to its components, the characteristics of the breathing behavior for transverse cracks. The experimental and simulated results showed that the proposed techniques can identify the presence and location of transverse cracks in rotating machinery, contributing to improving the safety and efficiency of these equipments.