Fault detection in rotating machinery by using the modal state observer approach

Abstract

The existence of cracks in rotating shafts can lead to failures, which imply economic losses and security issues. In the last decades, attention has been devoted to this subject leading to the development of vibration-based structural health monitoring techniques. In this paper, a fault detection methodology based on the so-called modal state observer is applied for detecting transversal cracks in a horizontal rotating shaft. A rotor composed by a flexible horizontal shaft, two rigid discs, and two self-aligning ball bearings is used. The Luenberger state observer is formulated in the modal domain to estimate the rotor vibration responses described in the modal domain. Consequently, the most affected vibration modes due to the crack existence can be determined. The kernel density estimator is applied to quantify the influence of a crack over the dynamic behavior of the rotor. The time domain vibration responses measured directly on the rotating machine are used for comparison purposes. Numerical and experimental investigations are presented by considering both breathing and open cracks. The breathing behavior of the crack is simulated according to the Mayes model, in which the crack transition from fully open to fully closed is described by a cosine function. The additional flexibility introduced in the shaft by the crack is determined by using the linear fracture mechanics theory. The obtained results demonstrated that the methodology conveyed represents a good alternative technique to detect faults in rotating machines.