PSI - Issue 39

Nabam Teyi et al. / Procedia Structural Integrity 39 (2022) 333–346 Author name / Structural Integrity Procedia 00 (2019) 000–000

339

7

(2014) used MRA and an ANN to detect rotating machine faults. Three stages of filter analysis were used, as was the Jeffcott/de Laval rotor model. A supervised RBF Network was tested. MRA and RBF formed an automated monitoring system with online diagnosing capacity. Ebrahimi et al. (2014) developed a new continuous model for flexural vibration of open-edged rotors. They compared the first critical speed ratio for a fractured rotor to the exponent degradation rate. An accurate FE cracked rotor model was created to anticipate dynamic behaviour of fractured rotors to avert catastrophic failures. Haji and Oyadiji (2014) proposed a new method for locating cracks in non-rotating rotors using only orthogonal natural frequencies in both horizontal and vertical lateral bending vibration planes. An intact and cracked rotor was FE modelled. On the basis of normalised orthogonal natural frequency curves, a fracture detecting technique was devised. The cracks had sharp, notched peaks while the non-cracks had rounded peaks. These qualities aided in locating and identifying rotor cracks. In order to find cracks, Singh and Tiwari (2014) exploited slope discontinuity in the shaft elastic line. For the purpose of crack identification, an experimental set of shafts, supports, and exciters was used, as well as a shaft equipped with a laser vibrometer for comparison. The MCDLA algorithm, which uses multiple crack detection and localization techniques, was put to the test in an experiment. AL-Shudeifat (2015) studied the backward whirl of a cracked rotor using the open crack model. To solve the time periodic cracked rotor system, the harmonic balance approach used Mathieu's equation and a linear time-periodic system. The method beat Floquet's theory's time-consuming application. Open crack excited the backward whirl speeds. Peng et al. (2015) proposed a method for determining the stability of a rotating cracked rotor system by examining bifurcations at boundary points using comprehensive numerical Eigenvalues. They used a numerical transition matrix and a stability diagram. With high degree of imbalance shaft rotational motion, cracked rotor system could barely retain stable motion around first harmonic and sub harmonic resonance. It was stable for a low degree shaft imbalance. Singh and Tiwari (2015) studied the effects of AMB on a rotor-bearing system with a breathing crack. The SIMULINK model responded. The reference signal’s harmonic was plotted. A crack detection technique was developed using FFT analysis. Because the identification system used data from several spin speeds, it was proven to be resistant against signal noise and modelling mistakes. Söffker et al. (2015) compared model-based and signal based techniques for rotating machinery crack identification. The modern model-based technique was Proportional Integral, while the modern machine learning technique was a revolutionary signal-based approach based on Support Vector Machine and wavelets. In contrast, model-based solutions were more flexible to changes in system load and better able to interact with system physics and modelling parameters. Major remarks from the above mentioned literatures for the period 2013 to 2015 as possibility for future research are presented in Table 2.

Table 2. Analysis of literatures between 2013 to 2015 for work progress Author(s), year Possibility of improvement Guo et al., 2013

There is no consideration given to noise. The EMD method's decomposition procedure for super-harmonic