Detection of nonlinear behavior using the Random Decrement Technique

This study investigates how the Random Decrement (RD) technique can be used as an output-only method of detecting nonlinear behavior. The proposed detection method is based on a grouping of the triggering points used in the RD calculation. The grouping is based on properties of the response signal at the triggering points, for instance the slope. Grouping the triggering points from the response of a system exhibiting nonlinear behavior in this way, causes the RD signatures calculated from the different groups of triggering points to have different properties. For a linear system, the grouping does not cause a change in the RD signatures, thus allowing the detection of nonlinear behavior. Linearity is assumed in operational modal analysis (OMA), but this assumption is not always valid, and the detection method proposed here could help confirm or reject the assumption.