Type identification for mechanical nonlinearity

For type identification of nonlinear continuum systems with stiffness nonlinearity, a classification strategy with peak detection algorithm is put forward. According to the structure of under test system, simplified models with 5 kinds of stiffness nonlinearity are constructed. The nonlinearity types are particularly interested in structural diagnostic, such as clearance nonlinearity and cubic nonlinearity. An instantaneous frequency library is then set up by studying the free oscillations and instantaneous frequencies of different models. The shape-matching algorithm is used to compare the real instantaneous frequency with the curves in the library. And the results are taken up to the index to identify the nonlinearity types. For a cantilever beam system with clearance, the matching value between the real instantaneous frequency and the curve of clearance nonlinearity measured in this strategy gets obviously smaller than the others, verifying the effectiveness.