PSI - Issue 39

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

340

8

components can be influenced by noise if the fractures are minor. The impact of noise on fracture detection needs to be studied further in future research.

Liu and Jiang, 2013

A more robust experimentation on similar line may be performed to obtain more behavioral results.

Wang et al., 2013

Number of training samples may be reduced.

Zapico-Valle et al. 2013

Lower notch depths cause instability in the operation. Additional high modes, more precise modal testing, and a more accurate modal identification system could improve these results. There has to be greater investigation into the factors that influence vibration response, such as meshing friction and backlash, and gear manufacturing faults. The appearance of a crack causes violent excitations near the theoretical critical speed. Use of MRA ascertains that noise doesn’t affect this method. However, scope for improvement exists here. The proposed model is capable of further modifications such as implementing a breathing function to model breathing cracks or considering the gyroscopic effects. It's not a good idea to use a roving disc that's less than 8.0 percent of the shaft mass. Using a roving disc with a mass to area ratio greater than 20.4% increases the risk of the crack spreading and deepening. Possibly some investigation can be done in this area. It is possible to test the algorithm's performance for rotated cracks using a smaller fracture size to see if it detects the rotation.

Wan et al., 2014

Castejón et al., 2014

Ebrahimi et al., 2014

Haji and Oyadiji, 2014

Singh and Tiwari, 2014

AL-Shudeifat, 2015

More parameters may be tested to obtain more varied and more determinant results.

Peng et al., 2015

Fatigued rotor cracks can cause unstable behaviour and bifurcations, especially if the mass eccentricity is unbalanced at a high level while rotating horizontally. This might use some work.

Singh and Tiwari, 2015

Using the finite element method, stiffness in the multi-degree-of-freedom system can be included.

Söffker et al., 2015

More machine learning classifiers like Random Forest (RF) classifier, Logistic Regression (LR) or Naïve Bayes (NB) may be taken up for comparison.

4. Relevant literatures from 2016 to 2018 Ferjaoui et al. (2016) used a nonlinear model to investigate the effects of a transverse crack in a rotor supported by two journal bearings. Bifurcation diagrams and Poincaré maps were employed. Half rotational frequency components in the spectrum and their amplitude rise were regarded markers of rotor cracks. Gómez et al. (2016) used vibration signals to detect cracks in a rotating shaft. WPT energy used the ‘Daubechies 6’ wavelet. Individual level energies trained RBF-ANNs. The faster speed (60 Hz) produced the best outcomes in terms of success rates and cost. The false alarm rate was 1.77 percent and fracture levels exceeding 1 were reliably recognised. A rotating shaft with various crack depths and positions was researched by Gómez et al. (2016). The tests were done in a Rotokit setup at various rotation speeds. WPT was used to evaluate the signals. At the greatest speed tested, the energy of the 3× component of rotation speed was the best indicator of crack (60 Hz). The 3× energy rose for Middle Section and reduced for Side Section. Zhiwen Lu et al. (2016) proposed a proper orthogonal decomposition (POD) based method for multicrack localisation in rotors. They used FD, GSM and ODS. Cracks in the orthogonal modes caused discontinuities in POMs. The FD and GSM localisation results for the rotor with several breaks of differing depths were excellent. Zhiwen Lu et al. (2017) proposed a new crack localisation approach for stepped rotors based SVD in frequency domain. A waterfall plot of vertical response was used, as was a PSD matrix. The method worked well for single or multiple fracture localization in stepped rotors with the right rotating speed and super-harmonic component(s). To study a cracked rotor with an open edge fracture subjected to gravity and imbalance forces, Ebrahimi et al. (2017) developed the method described here. They looked at how the orbit changed with changes in rotor speed. A modified Galerkin approach was used to find each response. These results could be utilised to identify rotor cracks. A Jeffcott/de Laval rotor and an EMD based fracture detection approach were tested experimentally by Guo et al. (2017) They used FFT analysis. During the transit, the cracked rotor was whirled. The super-harmonic component variation in the 1/3