Weighted contribution rate of nonlinear output frequency response functions and its application to rotor system fault diagnosis

Abstract

Rotor cracking is a common fault in rotating machinery, that can often be detected based on the vibrations of such machinery. Therefore, many vibration signal based techniques have been developed for rotor fault detection. However, signal based fault diagnosis is often affected by noise and cannot be used to fully quantify the operating conditions of rotors, which are often more relevant to rotor structural characteristics compared to vibration behaviours. Nonlinear output frequency response functions (NOFRFs) represent a nonlinear system frequency analysis approach that was proposed to address challenges associated with applying signal based fault detection. NOFRFs have a wide range of applications in industry, including rotor fault diagnosis. In this study, a novel NOFRFs based criterion called the weighted contribution rate of NOFRFs is proposed to enhance the capability of the NOFRFs based rotor crack detection and quantification. The proposed criterion can amplify the impact of rotor characteristics that are sensitive to rotor cracks, which considerably improves the crack diagnosis performance. The effectiveness and advantages of the proposed rotor damage detection method are verified both experimentally and through simulations, demonstrating its potential applications in the early stages of rotor crack diagnosis.