Application of dimension analysis and soft competitive tool to predict compound faults present in rotor-bearing systems

Localized and distributed faults in machinery may lead to catastrophic failures of the high-speed rotor bearings. Unbalance and misalignment present generate massive vibrations in the rotor-bearing system. This paper demonstrates duo of the matrix method of dimensional analysis (MMDA) and support vector machine (SVM) to investigate unbalance and misalignment present in the rotor-bearing system. Experimentation with different conditions is performed and compared with numerical results to reveal the effectiveness of the approach. The SVM algorithm is used to classify multiple fault classes based on vibration characteristics predicted by the MMDA model. The results comparison shows that the present duo predicts and classifies the misalignment and unbalance with acceptable error and can be useful for industrial high-speed machine diagnosis.

Automated summary

This paper introduces a combination of the matrix method of dimensional analysis (MMDA) and support vector machine (SVM) to investigate unbalance and misalignment in the rotor-bearing system. Experimental results show that this method can predict and classify these faults with acceptable error, making it useful for industrial high-speed machine diagnosis.