Description

While the identification of models of linear structures from experimental data has become routine, the identifi-cation of nonlinear systems remains a much more challenging problem. To support recent theoretical and applied work reported in Refs. [1, 2, 3, 4, 5, 6, 7], we have developed a simple test structure in which the effects of nonlinear stiffness can be adjusted. This is achieved by varying one of the linear stiffness coefficients, from zero to a value large enough to reduce the influence of the nonlinear stiffness to a parasitic role.

The model structure taken as a starting point here was originally designed as a linear two-degree-of-freedom (2-DOF) model with an added essentially nonlinear spring (that is, a spring with no linear term in its force-displacement relation) between one of the structural masses and ground. That essentially nonlinear spring was formed from transversely deflected piano wire, and the essential nonlinearity was achieved only if the initial tension in the wire was zero. Several experiments were reported using this configuration and a similar arrangement in which the wire was connected between moving masses rather than between a mass and ground [4, 5, 6, 7]. In the present work, we exploit the linear stiffness terms introduced when a significant initial tension (preload) is established in the wire in its undeformed state.