The Study on Robustness of CPSD-Based Estimation Method of NOFRFs

Fatigue damage of mechanical parts can be effectively revealed by the Nonlinear Output Frequency Response Functions (NOFRFs). Among various estimation methods of NOFRFs, the estimation method based on cross power spectral density (CPSD) provides a way to estimate NOFRFs using stochastic signals without identifying other models. However, the robustness of this estimation method has not been studied. This paper considers that this estimation method has the ability to reduce the impact of noise, which is demonstrated through numerical simulation and experimental study. In the numerical simulation, a multi-degree-of-freedom nonlinear system is employed, and multiple signal-to-noise ratios are set. In the experiment, a fatigued specimen is employed. The resulting error of the CPSD-based estimation method is compared with that of the general estimation method. The results show that this estimation method has better robustness than the general estimation method of NOFRFs.