Application of weighted contribution rate of nonlinear output frequency response functions to rotor rub-impact

The rub-impact between a rotor and a stator is a common fault of the rotor system. Various methods are available to detect the rub impact using vibration signals. However, the noise in vibration signals causes inconvenience when extracting characteristic signals. Nonlinear output frequency response functions (NOFRFs) comprise an effective method for using input and output signals to detect damage of a structure or system. A new evaluation method named the NOFRFs weighted contribution rate has been proposed in this paper. This method is an extension of traditional NOFRFs-based indexes and attempt to obtain a new index with higher recognition of faults. As a new index extracted by this method, the second-order optimal weighted contribution rate of the NOFRFs optimally amplify the extracted fault features. The characteristic signal after amplification is significantly enhanced. The new index is applied to the evaluation of a rub-impact fault. The simulation and experimental results demonstrate the sensitivity of the new index to rub-impact faults. Especially in the initial stage of the occurrence of rub-impact, the new index shows better superiority. The findings demonstrated that the second order optimal weighted contribution rate has a high application value in the fault diagnosis of rub-impact. (C) 2019 Elsevier Ltd. All rights reserved.